.
diff --git a/VoucherVision_Reference.yaml b/VoucherVision_Reference.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..52ab02e4d03460a5e22faf920c7b68185aae97a2
--- /dev/null
+++ b/VoucherVision_Reference.yaml
@@ -0,0 +1,103 @@
+# To use default value, set to null
+leafmachine:
+
+ use_RGB_label_images: True
+
+ do:
+ check_for_illegal_filenames: False
+ check_for_corrupt_images_make_vertical: False
+ print:
+ verbose: True
+ optional_warnings: True
+
+ logging:
+ log_level: null
+
+
+ # Overall Project Input Settings
+ project:
+ # Image to Process
+ dir_images_local: 'D:\Dropbox\LM2_Env\VoucherVision_Datasets\2022_09_07_thru12_S3_jacortez_AllAsia' # 'D:/Dropbox/LM2_Env/VoucherVision_Datasets/Compare_Set_Easy_10imgs/imgs' #'D:\D_Desktop\Richie\Imgs' #'D:/Dropbox/LM2_Env/Image_Datasets/Acacia/Acacia_prickles_4-26-23_LANCZOS/images/short' #'D:\D_Desktop\Richie\Imgs' #'home/brlab/Dropbox/LM2_Env/Image_Datasets/Manuscript_Images' # 'D:\Dropbox\LM2_Env\Image_Datasets\SET_FieldPrism_Test\TESTING_OUTPUT\Images_Processed\REU_Field_QR-Code-Images\Cannon_Corrected\Images_Corrected' # 'F:\temp_3sppFamily' # 'D:/Dropbox/LM2_Env/Image_Datasets/GBIF_BroadSample_3SppPerFamily' # SET_Diospyros/images_short' # 'D:/Dropbox/LM2_Env/Image_Datasets/SET_Diospyros/images_short' #'D:\Dropbox\LM2_Env\Image_Datasets\GBIF_BroadSample_Herbarium' #'D:/Dropbox/LM2_Env/Image_Datasets/SET_Diospyros/images_short' # str | only for image_location:local | full path for directory containing images
+ # dir_images_local: 'D:/Dropbox/LM2_Env/VoucherVision_Datasets/Compare_Set_Easy_10imgs/imgs' #'D:\D_Desktop\Richie\Imgs' #'D:/Dropbox/LM2_Env/Image_Datasets/Acacia/Acacia_prickles_4-26-23_LANCZOS/images/short' #'D:\D_Desktop\Richie\Imgs' #'home/brlab/Dropbox/LM2_Env/Image_Datasets/Manuscript_Images' # 'D:\Dropbox\LM2_Env\Image_Datasets\SET_FieldPrism_Test\TESTING_OUTPUT\Images_Processed\REU_Field_QR-Code-Images\Cannon_Corrected\Images_Corrected' # 'F:\temp_3sppFamily' # 'D:/Dropbox/LM2_Env/Image_Datasets/GBIF_BroadSample_3SppPerFamily' # SET_Diospyros/images_short' # 'D:/Dropbox/LM2_Env/Image_Datasets/SET_Diospyros/images_short' #'D:\Dropbox\LM2_Env\Image_Datasets\GBIF_BroadSample_Herbarium' #'D:/Dropbox/LM2_Env/Image_Datasets/SET_Diospyros/images_short' # str | only for image_location:local | full path for directory containing images
+ image_location: 'local'
+
+ continue_run_from_partial_xlsx: 'D:\Dropbox\LM2_Env\VoucherVision_Datasets\POC_chatGPT__2022_09_07_thru12_S3_jacortez_AllAsia\2022_09_07_thru12_S3_jacortez_AllAsia\Transcription\transcribed.xlsx'
+ # continue_run_from_partial_xlsx: null
+
+ # Project Output Dir
+ dir_output: 'D:/Dropbox/LM2_Env/VoucherVision_Datasets/POC_chatGPT__2022_09_07_thru12_S3_jacortez_AllAsia' # 'D:/Dropbox/LM2_Env/Image_Datasets/TEST_LM2' # 'D:\D_Desktop\Richie\Richie_Out'
+ run_name: 'POC_chatGPT' #'images_short_TEST' #'images_short_landmark'
+
+ prefix_removal: 'MICH-V-'
+ suffix_removal: ''
+ catalog_numerical_only: True
+
+ # Embeddings and LLM
+ use_domain_knowledge: True
+ embeddings_database_name: 'EmbeddingsDB_all_asia_minimal_InRegion'
+ build_new_embeddings_database: False
+ path_to_domain_knowledge_xlsx: 'D:\Dropbox\LeafMachine2\leafmachine2\transcription\domain_knowledge/AllAsiaMinimalasof25May2023_2__InRegion.xlsx' #'D:/Dropbox/LeafMachine2/leafmachine2/transcription/domain_knowledge/AllAsiaMinimalasof25May2023_2__TRIMMEDtiny.xlsx'
+
+ batch_size: 500 #null # null = all
+ num_workers: 1 # int |DEFAULT| 4 # More is not always better. Most hardware loses performance after 4
+
+ modules:
+ specimen_crop: True
+
+ LLM_version: 'chatGPT' # from 'chatGPT' OR 'PaLM'
+
+ cropped_components:
+ # empty list for all, add to list to IGNORE, lowercase, comma seperated
+ # archival |FROM|
+ # ruler, barcode, colorcard, label, map, envelope, photo, attached_item, weights
+ # plant |FROM|
+ # leaf_whole, leaf_partial, leaflet, seed_fruit_one, seed_fruit_many, flower_one, flower_many, bud, specimen, roots, wood
+ do_save_cropped_annotations: True
+ save_cropped_annotations: ['label','barcode'] # 'save_all' to save all classes
+ save_per_image: False # creates a folder for each image, saves crops into class-names folders # TODO
+ save_per_annotation_class: True # saves crops into class-names folders
+ binarize_labels: False
+ binarize_labels_skeletonize: False
+
+ data:
+ save_json_rulers: False
+ save_json_measurements: False
+ save_individual_csv_files_rulers: False
+ save_individual_csv_files_measurements: False
+ include_darwin_core_data_from_combined_file: False
+ do_apply_conversion_factor: False ###########################
+
+ overlay:
+ save_overlay_to_pdf: True
+ save_overlay_to_jpgs: True
+ overlay_dpi: 300 # int |FROM| 100 to 300
+ overlay_background_color: 'black' # str |FROM| 'white' or 'black'
+
+ show_archival_detections: True
+ ignore_archival_detections_classes: []
+ show_plant_detections: True
+ ignore_plant_detections_classes: ['leaf_whole', 'specimen'] #['leaf_whole', 'leaf_partial', 'specimen']
+ show_segmentations: True
+ show_landmarks: True
+ ignore_landmark_classes: []
+
+ line_width_archival: 2 # int
+ line_width_plant: 6 # int
+ line_width_seg: 12 # int # thick = 12
+ line_width_efd: 6 # int # thick = 3
+ alpha_transparency_archival: 0.3 # float between 0 and 1
+ alpha_transparency_plant: 0
+ alpha_transparency_seg_whole_leaf: 0.4
+ alpha_transparency_seg_partial_leaf: 0.3
+
+ # Configure Archival Component Detector
+ archival_component_detector:
+ # ./leafmachine2/component_detector/runs/train/detector_type/detector_version/detector_iteration/weights/detector_weights
+ detector_type: 'Archival_Detector'
+ detector_version: 'PREP_final'
+ detector_iteration: 'PREP_final'
+ detector_weights: 'best.pt'
+ minimum_confidence_threshold: 0.5
+ do_save_prediction_overlay_images: True
+ ignore_objects_for_overlay: [] # list[str] # list of objects that can be excluded from the overlay # all = null
+
\ No newline at end of file
diff --git a/__init__.py b/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/api_cost/api_cost.yaml b/api_cost/api_cost.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d43b1bf45d2d720cf620b95f8d5e49dd9f232b05
--- /dev/null
+++ b/api_cost/api_cost.yaml
@@ -0,0 +1,9 @@
+GPT_3_5:
+ in: 0.0015
+ out: 0.002
+GPT_4:
+ in: 0.03
+ out: 0.06
+PALM2:
+ in: 0.0
+ out: 0.0
diff --git a/app.py b/app.py
new file mode 100644
index 0000000000000000000000000000000000000000..9ce0c26a54676687f07cb86724309dae694e7b92
--- /dev/null
+++ b/app.py
@@ -0,0 +1,1344 @@
+import streamlit as st
+import yaml, os, json, random, time, re
+import matplotlib.pyplot as plt
+import plotly.graph_objs as go
+import numpy as np
+from itertools import chain
+from PIL import Image
+import pandas as pd
+from typing import Union
+from streamlit_extras.let_it_rain import rain
+from vouchervision.LeafMachine2_Config_Builder import write_config_file
+from vouchervision.VoucherVision_Config_Builder import build_VV_config, run_demo_tests_GPT, run_demo_tests_Palm , TestOptionsGPT, TestOptionsPalm, check_if_usable, run_api_tests
+from vouchervision.vouchervision_main import voucher_vision, voucher_vision_OCR_test
+from vouchervision.general_utils import test_GPU, get_cfg_from_full_path, summarize_expense_report, create_google_ocr_yaml_config, validate_dir
+
+PROMPTS_THAT_NEED_DOMAIN_KNOWLEDGE = ["Version 1","Version 1 PaLM 2"]
+COLORS_EXPENSE_REPORT = {
+ 'GPT_4': '#8fff66', # Bright Green
+ 'GPT_3_5': '#006400', # Dark Green
+ 'PALM2': '#66a8ff' # blue
+ }
+
+class ProgressReport:
+ def __init__(self, overall_bar, batch_bar, text_overall, text_batch):
+ self.overall_bar = overall_bar
+ self.batch_bar = batch_bar
+ self.text_overall = text_overall
+ self.text_batch = text_batch
+ self.current_overall_step = 0
+ self.total_overall_steps = 20 # number of major steps in machine function
+ self.current_batch = 0
+ self.total_batches = 20
+
+ def update_overall(self, step_name=""):
+ self.current_overall_step += 1
+ self.overall_bar.progress(self.current_overall_step / self.total_overall_steps)
+ self.text_overall.text(step_name)
+
+ def update_batch(self, step_name=""):
+ self.current_batch += 1
+ self.batch_bar.progress(self.current_batch / self.total_batches)
+ self.text_batch.text(step_name)
+
+ def set_n_batches(self, n_batches):
+ self.total_batches = n_batches
+
+ def set_n_overall(self, total_overall_steps):
+ self.total_overall_steps = total_overall_steps
+
+ def reset_batch(self, step_name):
+ self.current_batch = 0
+ self.batch_bar.progress(0)
+ self.text_batch.text(step_name)
+ def reset_overall(self, step_name):
+ self.current_overall_step = 0
+ self.overall_bar.progress(0)
+ self.text_overall.text(step_name)
+
+ def get_n_images(self):
+ return self.n_images
+ def get_n_overall(self):
+ return self.total_overall_steps
+
+def does_private_file_exist():
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ path_cfg_private = os.path.join(dir_home, 'PRIVATE_DATA.yaml')
+ return os.path.exists(path_cfg_private)
+
+def setup_streamlit_config(dir_home):
+ # Define the directory path and filename
+ dir_path = os.path.join(dir_home, ".streamlit")
+ file_path = os.path.join(dir_path, "config.toml")
+
+ # Check if directory exists, if not create it
+ if not os.path.exists(dir_path):
+ os.makedirs(dir_path)
+
+ # Create or modify the file with the provided content
+ config_content = f"""
+ [theme]
+ base = "dark"
+ primaryColor = "#00ff00"
+
+ [server]
+ enableStaticServing = false
+ runOnSave = true
+ port = 8524
+ """
+
+ with open(file_path, "w") as f:
+ f.write(config_content.strip())
+
+def display_scrollable_results(JSON_results, test_results, OPT2, OPT3):
+ """
+ Display the results from JSON_results in a scrollable container.
+ """
+ # Initialize the container
+ con_results = st.empty()
+ with con_results.container():
+
+ # Start the custom container for all the results
+ results_html = """"""
+
+ # The CSS to make this container scrollable
+ css = """
+
+ """
+
+ # Apply the CSS and then the results
+ st.markdown(css, unsafe_allow_html=True)
+ st.markdown(results_html, unsafe_allow_html=True)
+
+def display_test_results(test_results, JSON_results, llm_version):
+ if llm_version == 'gpt':
+ OPT1, OPT2, OPT3 = TestOptionsGPT.get_options()
+ elif llm_version == 'palm':
+ OPT1, OPT2, OPT3 = TestOptionsPalm.get_options()
+ else:
+ raise
+
+ widths = [1] * (len(OPT1) + 2) + [2]
+ columns = st.columns(widths)
+
+ with columns[0]:
+ st.write("LeafMachine2")
+ with columns[1]:
+ st.write("Prompt")
+ with columns[len(OPT1) + 2]:
+ st.write("Scroll to See Last Transcription in Each Test")
+
+ already_written = set()
+
+ for test_name, result in sorted(test_results.items()):
+ _, ind_opt1, _, _ = test_name.split('__')
+ option_value = OPT1[int(ind_opt1.split('-')[1])]
+
+ if option_value not in already_written:
+ with columns[int(ind_opt1.split('-')[1]) + 2]:
+ st.write(option_value)
+ already_written.add(option_value)
+
+ printed_options = set()
+
+ with columns[-1]:
+ display_scrollable_results(JSON_results, test_results, OPT2, OPT3)
+
+ # Close the custom container
+ st.write('', unsafe_allow_html=True)
+
+
+ for idx, (test_name, result) in enumerate(sorted(test_results.items())):
+ _, ind_opt1, ind_opt2, ind_opt3 = test_name.split('__')
+ opt2_readable = "Use LeafMachine2" if OPT2[int(ind_opt2.split('-')[1])] else "Don't use LeafMachine2"
+ opt3_readable = f"{OPT3[int(ind_opt3.split('-')[1])]}"
+
+ if (opt2_readable, opt3_readable) not in printed_options:
+ with columns[0]:
+ st.info(f"{opt2_readable}")
+ st.write('---')
+ with columns[1]:
+ st.info(f"{opt3_readable}")
+ st.write('---')
+ printed_options.add((opt2_readable, opt3_readable))
+
+ with columns[int(ind_opt1.split('-')[1]) + 2]:
+ if result:
+ st.success(f"Test Passed")
+ else:
+ st.error(f"Test Failed")
+ st.write('---')
+
+ # success_count = sum(1 for result in test_results.values() if result)
+ # failure_count = len(test_results) - success_count
+ # proportional_rain("🥇", success_count, "💔", failure_count, font_size=72, falling_speed=5, animation_length="infinite")
+ rain_emojis(test_results)
+
+def add_emoji_delay():
+ time.sleep(0.3)
+
+def rain_emojis(test_results):
+ # test_results = {
+ # 'test1': True, # Test passed
+ # 'test2': True, # Test passed
+ # 'test3': True, # Test passed
+ # 'test4': False, # Test failed
+ # 'test5': False, # Test failed
+ # 'test6': False, # Test failed
+ # 'test7': False, # Test failed
+ # 'test8': False, # Test failed
+ # 'test9': False, # Test failed
+ # 'test10': False, # Test failed
+ # }
+ success_emojis = ["🥇", "🏆", "🍾", "🙌"]
+ failure_emojis = ["💔", "😭"]
+
+ success_count = sum(1 for result in test_results.values() if result)
+ failure_count = len(test_results) - success_count
+
+ chosen_emoji = random.choice(success_emojis)
+ for _ in range(success_count):
+ rain(
+ emoji=chosen_emoji,
+ font_size=72,
+ falling_speed=4,
+ animation_length=2,
+ )
+ add_emoji_delay()
+
+ chosen_emoji = random.choice(failure_emojis)
+ for _ in range(failure_count):
+ rain(
+ emoji=chosen_emoji,
+ font_size=72,
+ falling_speed=5,
+ animation_length=1,
+ )
+ add_emoji_delay()
+
+def get_prompt_versions(LLM_version):
+ yaml_files = [f for f in os.listdir(os.path.join(st.session_state.dir_home, 'custom_prompts')) if f.endswith('.yaml')]
+
+ if LLM_version in ["GPT 4", "GPT 3.5", "Azure GPT 4", "Azure GPT 3.5"]:
+ versions = ["Version 1", "Version 1 No Domain Knowledge", "Version 2"]
+ return (versions + yaml_files, "Version 2")
+ elif LLM_version in ["PaLM 2",]:
+ versions = ["Version 1 PaLM 2", "Version 1 PaLM 2 No Domain Knowledge", "Version 2 PaLM 2"]
+ return (versions + yaml_files, "Version 2 PaLM 2")
+ else:
+ # Handle other cases or raise an error
+ return (yaml_files, None)
+
+def get_private_file():
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ path_cfg_private = os.path.join(dir_home, 'PRIVATE_DATA.yaml')
+ return get_cfg_from_full_path(path_cfg_private)
+
+def create_space_saver():
+ st.subheader("Space Saving Options")
+ col_ss_1, col_ss_2 = st.columns([2,2])
+ with col_ss_1:
+ st.write("Several folders are created and populated with data during the VoucherVision transcription process.")
+ st.write("Below are several options that will allow you to automatically delete temporary files that you may not need for everyday operations.")
+ st.write("VoucherVision creates the following folders. Folders marked with a :star: are required if you want to use VoucherVisionEditor for quality control.")
+ st.write("`../[Run Name]/Archival_Components`")
+ st.write("`../[Run Name]/Config_File`")
+ st.write("`../[Run Name]/Cropped_Images` :star:")
+ st.write("`../[Run Name]/Logs`")
+ st.write("`../[Run Name]/Original_Images` :star:")
+ st.write("`../[Run Name]/Transcription` :star:")
+ with col_ss_2:
+ st.session_state.config['leafmachine']['project']['delete_temps_keep_VVE'] = st.checkbox("Delete Temporary Files (KEEP files required for VoucherVisionEditor)", st.session_state.config['leafmachine']['project'].get('delete_temps_keep_VVE', False))
+ st.session_state.config['leafmachine']['project']['delete_all_temps'] = st.checkbox("Keep only the final transcription file", st.session_state.config['leafmachine']['project'].get('delete_all_temps', False),help="*WARNING:* This limits your ability to do quality assurance. This will delete all folders created by VoucherVision, leaving only the `transcription.xlsx` file.")
+
+
+# def create_private_file():
+# st.session_state.proceed_to_main = False
+
+# if st.session_state.private_file:
+# cfg_private = get_private_file()
+# create_private_file_0(cfg_private)
+# else:
+# st.title("VoucherVision")
+# create_private_file_0()
+
+def create_private_file():
+ st.session_state.proceed_to_main = False
+ st.title("VoucherVision")
+ col_private,_= st.columns([12,2])
+
+ if st.session_state.private_file:
+ cfg_private = get_private_file()
+ else:
+ cfg_private = {}
+ cfg_private['openai'] = {}
+ cfg_private['openai']['OPENAI_API_KEY'] =''
+
+ cfg_private['openai_azure'] = {}
+ cfg_private['openai_azure']['openai_api_key'] = ''
+ cfg_private['openai_azure']['api_version'] = ''
+ cfg_private['openai_azure']['openai_api_base'] =''
+ cfg_private['openai_azure']['openai_organization'] =''
+ cfg_private['openai_azure']['openai_api_type'] =''
+
+ cfg_private['google_cloud'] = {}
+ cfg_private['google_cloud']['path_json_file'] =''
+
+ cfg_private['google_palm'] = {}
+ cfg_private['google_palm']['google_palm_api'] =''
+
+
+ with col_private:
+ st.header("Set API keys")
+ st.info("***Note:*** There is a known bug with tabs in Streamlit. If you update an input field it may take you back to the 'Project Settings' tab. Changes that you made are saved, it's just an annoying glitch. We are aware of this issue and will fix it as soon as we can.")
+ st.warning("To commit changes to API keys you must press the 'Set API Keys' button at the bottom of the page.")
+ st.write("Before using VoucherVision you must set your API keys. All keys are stored locally on your computer and are never made public.")
+ st.write("API keys are stored in `../VoucherVision/PRIVATE_DATA.yaml`.")
+ st.write("Deleting this file will allow you to reset API keys. Alternatively, you can edit the keys in the user interface.")
+ st.write("Leave keys blank if you do not intend to use that service.")
+
+ st.write("---")
+ st.subheader("Google Vision (*Required*)")
+ st.markdown("VoucherVision currently uses [Google Vision API](https://cloud.google.com/vision/docs/ocr) for OCR. Generating an API key for this is more involved than the others. [Please carefully follow the instructions outlined here to create and setup your account.](https://cloud.google.com/vision/docs/setup) ")
+ st.markdown("""
+ Once your account is created, [visit this page](https://console.cloud.google.com) and create a project. Then follow these instructions:
+
+ - **Select your Project**: If you have multiple projects, ensure you select the one where you've enabled the Vision API.
+ - **Open the Navigation Menu**: Click on the hamburger menu (three horizontal lines) in the top left corner.
+ - **Go to IAM & Admin**: In the navigation pane, hover over "IAM & Admin" and then click on "Service accounts."
+ - **Locate Your Service Account**: Find the service account for which you wish to download the JSON key. If you haven't created a service account yet, you'll need to do so by clicking the "CREATE SERVICE ACCOUNT" button at the top.
+ - **Download the JSON Key**:
+ - Click on the three dots (actions menu) on the right side of your service account name.
+ - Select "Manage keys."
+ - In the pop-up window, click on the "ADD KEY" button and select "JSON."
+ - The JSON key file will automatically be downloaded to your computer.
+ - **Store Safely**: This file contains sensitive data that can be used to authenticate and bill your Google Cloud account. Never commit it to public repositories or expose it in any way. Always keep it safe and secure.
+ """)
+ with st.container():
+ c_in_ocr, c_button_ocr = st.columns([10,2])
+ with c_in_ocr:
+ google_vision = st.text_input(label = 'Full path to Google Cloud JSON API key file', value = cfg_private['google_cloud'].get('path_json_file', ''),
+ placeholder = 'e.g. C:/Documents/Secret_Files/google_API/application_default_credentials.json',
+ help ="This API Key is in the form of a JSON file. Please save the JSON file in a safe directory. DO NOT store the JSON key inside of the VoucherVision directory.",
+ type='password',key='924857298734590283750932809238')
+ with c_button_ocr:
+ st.empty()
+
+
+ st.write("---")
+ st.subheader("OpenAI")
+ st.markdown("API key for first-party OpenAI API. Create an account with OpenAI [here](https://platform.openai.com/signup), then create an API key [here](https://platform.openai.com/account/api-keys).")
+ with st.container():
+ c_in_openai, c_button_openai = st.columns([10,2])
+ with c_in_openai:
+ openai_api_key = st.text_input("openai_api_key", cfg_private['openai'].get('OPENAI_API_KEY', ''),
+ help='The actual API key. Likely to be a string of 2 character, a dash, and then a 48-character string: sk-XXXXXXXX...',
+ placeholder = 'e.g. sk-XXXXXXXX...',
+ type='password')
+ with c_button_openai:
+ st.empty()
+
+ st.write("---")
+ st.subheader("OpenAI - Azure")
+ st.markdown("This version OpenAI relies on Azure servers directly as is intended for private enterprise instances of OpenAI's services, such as [UM-GPT](https://its.umich.edu/computing/ai). Administrators will provide you with the following information.")
+ azure_openai_api_version = st.text_input("azure_openai_api_version", cfg_private['openai_azure'].get('api_version', ''),
+ help='API Version e.g. "2023-05-15"',
+ placeholder = 'e.g. 2023-05-15',
+ type='password')
+ azure_openai_api_key = st.text_input("azure_openai_api_key", cfg_private['openai_azure'].get('openai_api_key', ''),
+ help='The actual API key. Likely to be a 32-character string',
+ placeholder = 'e.g. 12333333333333333333333333333332',
+ type='password')
+ azure_openai_api_base = st.text_input("azure_openai_api_base", cfg_private['openai_azure'].get('openai_api_base', ''),
+ help='The base url for the API e.g. "https://api.umgpt.umich.edu/azure-openai-api"',
+ placeholder = 'e.g. https://api.umgpt.umich.edu/azure-openai-api',
+ type='password')
+ azure_openai_organization = st.text_input("azure_openai_organization", cfg_private['openai_azure'].get('openai_organization', ''),
+ help='Your organization code. Likely a short string',
+ placeholder = 'e.g. 123456',
+ type='password')
+ azure_openai_api_type = st.text_input("azure_openai_api_type", cfg_private['openai_azure'].get('openai_api_type', ''),
+ help='The API type. Typically "azure"',
+ placeholder = 'e.g. azure',
+ type='password')
+ with st.container():
+ c_in_azure, c_button_azure = st.columns([10,2])
+ with c_button_azure:
+ st.empty()
+
+ st.write("---")
+ st.subheader("Google PaLM 2")
+ st.markdown('Follow these [instructions](https://developers.generativeai.google/tutorials/setup) to generate an API key for PaLM 2. You may need to also activate an account with [MakerSuite](https://makersuite.google.com/app/apikey) and enable "early access."')
+ with st.container():
+ c_in_palm, c_button_palm = st.columns([10,2])
+ with c_in_palm:
+ google_palm = st.text_input("Google PaLM 2 API Key", cfg_private['google_palm'].get('google_palm_api', ''),
+ help='The MakerSuite API key e.g. a 32-character string',
+ placeholder='e.g. SATgthsykuE64FgrrrrEervr3S4455t_geyDeGq',
+ type='password')
+
+ with st.container():
+ with c_button_ocr:
+ st.write("##")
+ st.button("Test OCR", on_click=test_API, args=['google_vision',c_in_ocr, cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key,
+ azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm])
+
+ with st.container():
+ with c_button_openai:
+ st.write("##")
+ st.button("Test OpenAI", on_click=test_API, args=['openai',c_in_openai, cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key,
+ azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm])
+
+ with st.container():
+ with c_button_azure:
+ st.write("##")
+ st.button("Test Azure OpenAI", on_click=test_API, args=['azure_openai',c_in_azure, cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key,
+ azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm])
+
+ with st.container():
+ with c_button_palm:
+ st.write("##")
+ st.button("Test PaLM 2", on_click=test_API, args=['palm',c_in_palm, cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key,
+ azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm])
+
+
+ st.button("Set API Keys",type='primary', on_click=save_changes_to_API_keys, args=[cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key,
+ azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm])
+ if st.button('Proceed to VoucherVision'):
+ st.session_state.proceed_to_private = False
+ st.session_state.proceed_to_main = True
+
+def test_API(api, message_loc, cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key, azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm):
+ # Save the API keys
+ save_changes_to_API_keys(cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key,azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm)
+
+ with st.spinner('Performing validation checks...'):
+ if api == 'google_vision':
+ print("*** Google Vision OCR API Key ***")
+ try:
+ demo_config_path = os.path.join(st.session_state.dir_home,'demo','validation_configs','google_vision_ocr_test.yaml')
+ demo_images_path = os.path.join(st.session_state.dir_home, 'demo', 'demo_images')
+ demo_out_path = os.path.join(st.session_state.dir_home, 'demo', 'demo_output','run_name')
+ create_google_ocr_yaml_config(demo_config_path, demo_images_path, demo_out_path)
+ voucher_vision_OCR_test(demo_config_path, st.session_state.dir_home, None, demo_images_path)
+ with message_loc:
+ st.success("Google Vision OCR API Key Valid :white_check_mark:")
+ return True
+ except Exception as e:
+ with message_loc:
+ st.error(f"Google Vision OCR API Key Failed! {e}")
+ return False
+
+ elif api == 'openai':
+ print("*** OpenAI API Key ***")
+ try:
+ if run_api_tests('openai'):
+ with message_loc:
+ st.success("OpenAI API Key Valid :white_check_mark:")
+ else:
+ with message_loc:
+ st.error("OpenAI API Key Failed:exclamation:")
+ return False
+ except Exception as e:
+ with message_loc:
+ st.error(f"OpenAI API Key Failed:exclamation: {e}")
+
+ elif api == 'azure_openai':
+ print("*** Azure OpenAI API Key ***")
+ try:
+ if run_api_tests('azure_openai'):
+ with message_loc:
+ st.success("Azure OpenAI API Key Valid :white_check_mark:")
+ else:
+ with message_loc:
+ st.error(f"Azure OpenAI API Key Failed:exclamation:")
+ return False
+ except Exception as e:
+ with message_loc:
+ st.error(f"Azure OpenAI API Key Failed:exclamation: {e}")
+ elif api == 'palm':
+ print("*** Google PaLM 2 API Key ***")
+ try:
+ if run_api_tests('palm'):
+ with message_loc:
+ st.success("Google PaLM 2 API Key Valid :white_check_mark:")
+ else:
+ with message_loc:
+ st.error("Google PaLM 2 API Key Failed:exclamation:")
+ return False
+ except Exception as e:
+ with message_loc:
+ st.error(f"Google PaLM 2 API Key Failed:exclamation: {e}")
+
+
+def save_changes_to_API_keys(cfg_private,openai_api_key,azure_openai_api_version,azure_openai_api_key,
+ azure_openai_api_base,azure_openai_organization,azure_openai_api_type,google_vision,google_palm):
+ # Update the configuration dictionary with the new values
+ cfg_private['openai']['OPENAI_API_KEY'] = openai_api_key
+
+ cfg_private['openai_azure']['api_version'] = azure_openai_api_version
+ cfg_private['openai_azure']['openai_api_key'] = azure_openai_api_key
+ cfg_private['openai_azure']['openai_api_base'] = azure_openai_api_base
+ cfg_private['openai_azure']['openai_organization'] = azure_openai_organization
+ cfg_private['openai_azure']['openai_api_type'] = azure_openai_api_type
+
+ cfg_private['google_cloud']['path_json_file'] = google_vision
+
+ cfg_private['google_palm']['google_palm_api'] = google_palm
+ # Call the function to write the updated configuration to the YAML file
+ write_config_file(cfg_private, st.session_state.dir_home, filename="PRIVATE_DATA.yaml")
+ st.session_state.private_file = does_private_file_exist()
+
+# Function to load a YAML file and update session_state
+def load_prompt_yaml(filename):
+ with open(filename, 'r') as file:
+ st.session_state['prompt_info'] = yaml.safe_load(file)
+ st.session_state['instructions'] = st.session_state['prompt_info'].get('instructions', st.session_state['default_instructions'])
+ st.session_state['json_formatting_instructions'] = st.session_state['prompt_info'].get('json_formatting_instructions', st.session_state['default_json_formatting_instructions'] )
+ st.session_state['rules'] = st.session_state['prompt_info'].get('rules', {})
+ st.session_state['mapping'] = st.session_state['prompt_info'].get('mapping', {})
+ st.session_state['LLM'] = st.session_state['prompt_info'].get('LLM', 'gpt')
+
+ # Placeholder:
+ st.session_state['assigned_columns'] = list(chain.from_iterable(st.session_state['mapping'].values()))
+
+def save_prompt_yaml(filename):
+ yaml_content = {
+ 'instructions': st.session_state['instructions'],
+ 'json_formatting_instructions': st.session_state['json_formatting_instructions'],
+ 'rules': st.session_state['rules'],
+ 'mapping': st.session_state['mapping'],
+ 'LLM': st.session_state['LLM']
+ }
+
+ dir_prompt = os.path.join(st.session_state.dir_home, 'custom_prompts')
+ filepath = os.path.join(dir_prompt, f"{filename}.yaml")
+
+ with open(filepath, 'w') as file:
+ yaml.safe_dump(yaml_content, file)
+
+ st.success(f"Prompt saved as '{filename}.yaml'.")
+
+def check_unique_mapping_assignments():
+ if len(st.session_state['assigned_columns']) != len(set(st.session_state['assigned_columns'])):
+ st.error("Each column name must be assigned to only one category.")
+ return False
+ else:
+ st.success("Mapping confirmed.")
+ return True
+
+def check_prompt_yaml_filename(fname):
+ # Check if the filename only contains letters, numbers, underscores, and dashes
+ pattern = r'^[\w-]+$'
+
+ # The \w matches any alphanumeric character and is equivalent to the character class [a-zA-Z0-9_].
+ # The hyphen - is literally matched.
+
+ if re.match(pattern, fname):
+ return True
+ else:
+ return False
+
+
+def btn_load_prompt(selected_yaml_file, dir_prompt):
+ if selected_yaml_file:
+ yaml_file_path = os.path.join(dir_prompt, selected_yaml_file)
+ load_prompt_yaml(yaml_file_path)
+ elif not selected_yaml_file:
+ # Directly assigning default values since no file is selected
+ st.session_state['prompt_info'] = {}
+ st.session_state['instructions'] = st.session_state['default_instructions']
+ st.session_state['json_formatting_instructions'] = st.session_state['default_json_formatting_instructions']
+ st.session_state['rules'] = {}
+ st.session_state['LLM'] = 'gpt'
+
+ st.session_state['assigned_columns'] = []
+
+ st.session_state['prompt_info'] = {
+ 'instructions': st.session_state['instructions'],
+ 'json_formatting_instructions': st.session_state['json_formatting_instructions'],
+ 'rules': st.session_state['rules'],
+ 'mapping': st.session_state['mapping'],
+ 'LLM': st.session_state['LLM']
+ }
+
+def build_LLM_prompt_config():
+ st.session_state['assigned_columns'] = []
+ st.session_state['default_instructions'] = """1. Refactor the unstructured OCR text into a dictionary based on the JSON structure outlined below.
+2. You should map the unstructured OCR text to the appropriate JSON key and then populate the field based on its rules.
+3. Some JSON key fields are permitted to remain empty if the corresponding information is not found in the unstructured OCR text.
+4. Ignore any information in the OCR text that doesn't fit into the defined JSON structure.
+5. Duplicate dictionary fields are not allowed.
+6. Ensure that all JSON keys are in lowercase.
+7. Ensure that new JSON field values follow sentence case capitalization.
+8. Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+9. Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+10. Only return a JSON dictionary represented as a string. You should not explain your answer."""
+ st.session_state['default_json_formatting_instructions'] = """The next section of instructions outlines how to format the JSON dictionary. The keys are the same as those of the final formatted JSON object.
+For each key there is a format requirement that specifies how to transcribe the information for that key.
+The possible formatting options are:
+1. "verbatim transcription" - field is populated with verbatim text from the unformatted OCR.
+2. "spell check transcription" - field is populated with spelling corrected text from the unformatted OCR.
+3. "boolean yes no" - field is populated with only yes or no.
+4. "boolean 1 0" - field is populated with only 1 or 0.
+5. "integer" - field is populated with only an integer.
+6. "[list]" - field is populated from one of the values in the list.
+7. "yyyy-mm-dd" - field is populated with a date in the format year-month-day.
+The desired null value is also given. Populate the field with the null value of the information for that key is not present in the unformatted OCR text."""
+
+ # Start building the Streamlit app
+ col_prompt_main_left, ___, col_prompt_main_right = st.columns([6,1,3])
+
+
+ with col_prompt_main_left:
+
+ st.title("Custom LLM Prompt Builder")
+ st.subheader('About')
+ st.write("This form allows you to craft a prompt for your specific task.")
+ st.subheader('How it works')
+ st.write("1. Edit this page until you are happy with your instructions. We recommend looking at the basic structure, writing down your prompt inforamtion in a Word document so that it does not randomly disappear, and then copying and pasting that info into this form once your whole prompt structure is defined.")
+ st.write("2. After you enter all of your prompt instructions, click 'Save' and give your file a name.")
+ st.write("3. This file will be saved as a yaml configuration file in the `..VoucherVision/custom_prompts` folder.")
+ st.write("4. When you go back the main VoucherVision page you will now see your custom prompt available in the 'Prompt Version' dropdown menu.")
+ st.write("5. Select your custom prompt. Note, your prompt will only be available for the LLM that you set when filling out the form below.")
+
+
+ dir_prompt = os.path.join(st.session_state.dir_home, 'custom_prompts')
+ yaml_files = [f for f in os.listdir(dir_prompt) if f.endswith('.yaml')]
+ col_load_text, col_load_btn = st.columns([8,2])
+ with col_load_text:
+ # Dropdown for selecting a YAML file
+ selected_yaml_file = st.selectbox('Select a prompt YAML file to load:', [''] + yaml_files)
+ with col_load_btn:
+ st.write('##')
+ # Button to load the selected prompt
+ st.button('Load Prompt', on_click=btn_load_prompt, args=[selected_yaml_file, dir_prompt])
+
+
+
+ # Define the options for the dropdown
+ llm_options = ['gpt', 'palm']
+ # Create the dropdown and set the value to session_state['LLM']
+ st.session_state['LLM'] = st.selectbox('Set LLM:', llm_options, index=llm_options.index(st.session_state.get('LLM', 'gpt')))
+
+
+
+ # Instructions Section
+ st.header("Instructions")
+ st.write("These are the general instructions that guide the LLM through the transcription task. We recommend using the default instructions unless you have a specific reason to change them.")
+
+ st.session_state['instructions'] = st.text_area("Enter instructions:", value=st.session_state['default_instructions'].strip(), height=350, disabled=True)
+
+ st.write('---')
+
+ # Column Instructions Section
+ st.header("JSON Formatting Instructions")
+ st.write("The following section tells the LLM how we want to structure the JSON dictionary. We do not recommend changing this section because it would likely result in unstable and inconsistent behavior.")
+ st.session_state['json_formatting_instructions'] = st.text_area("Enter column instructions:", value=st.session_state['default_json_formatting_instructions'], height=350, disabled=True)
+
+
+
+
+
+ st.write('---')
+ col_left, col_right = st.columns([6,4])
+ with col_left:
+ st.subheader('Add/Edit Columns')
+
+ # Initialize rules in session state if not already present
+ if 'rules' not in st.session_state or not st.session_state['rules']:
+ st.session_state['rules']['Dictionary'] = {
+ "catalog_number": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "The barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits."
+ }
+ }
+ st.session_state['rules']['SpeciesName'] = {
+ "taxonomy": ["Genus_species"]
+ }
+
+ # Layout for adding a new column name
+ # col_text, col_textbtn = st.columns([8, 2])
+ # with col_text:
+ new_column_name = st.text_input("Enter a new column name:")
+ # with col_textbtn:
+ # st.write('##')
+ if st.button("Add New Column") and new_column_name:
+ if new_column_name not in st.session_state['rules']['Dictionary']:
+ st.session_state['rules']['Dictionary'][new_column_name] = {"format": "", "null_value": "", "description": ""}
+ st.success(f"New column '{new_column_name}' added. Now you can edit its properties.")
+ else:
+ st.error("Column name already exists. Please enter a unique column name.")
+
+ # Get columns excluding the protected "catalog_number"
+ st.write('#')
+ editable_columns = [col for col in st.session_state['rules']['Dictionary'] if col != "catalog_number"]
+ column_name = st.selectbox("Select a column to edit:", [""] + editable_columns)
+
+ # Handle rules editing
+ current_rule = st.session_state['rules']['Dictionary'].get(column_name, {
+ "format": "",
+ "null_value": "",
+ "description": ""
+ })
+
+ if 'selected_column' not in st.session_state:
+ st.session_state['selected_column'] = column_name
+
+
+
+
+ # Form for input fields
+ with st.form(key='rule_form'):
+ format_options = ["verbatim transcription", "spell check transcription", "boolean yes no", "boolean 1 0", "integer", "[list]", "yyyy-mm-dd"]
+ current_rule["format"] = st.selectbox("Format:", format_options, index=format_options.index(current_rule["format"]) if current_rule["format"] else 0)
+ current_rule["null_value"] = st.text_input("Null value:", value=current_rule["null_value"])
+ current_rule["description"] = st.text_area("Description:", value=current_rule["description"])
+ commit_button = st.form_submit_button("Commit Column")
+
+ default_rule = {
+ "format": format_options[0], # default format
+ "null_value": "", # default null value
+ "description": "", # default description
+ }
+ if st.session_state['selected_column'] != column_name:
+ # Column has changed. Update the session_state selected column.
+ st.session_state['selected_column'] = column_name
+ # Reset the current rule to the default for this new column, or a blank rule if not set.
+ current_rule = st.session_state['rules']['Dictionary'].get(column_name, default_rule.copy())
+
+ # Handle commit action
+ if commit_button and column_name:
+ # Commit the rules to the session state.
+ st.session_state['rules']['Dictionary'][column_name] = current_rule.copy()
+ st.success(f"Column '{column_name}' added/updated in rules.")
+
+ # Force the form to reset by clearing the fields from the session state
+ st.session_state.pop('selected_column', None) # Clear the selected column to force reset
+
+ # st.session_state['rules'][column_name] = current_rule
+ # st.success(f"Column '{column_name}' added/updated in rules.")
+
+ # # Reset current_rule to default values for the next input
+ # current_rule["format"] = default_rule["format"]
+ # current_rule["null_value"] = default_rule["null_value"]
+ # current_rule["description"] = default_rule["description"]
+
+ # # To ensure that the form fields are reset, we can clear them from the session state
+ # for key in current_rule.keys():
+ # st.session_state[key] = default_rule[key]
+
+ # Layout for removing an existing column
+ # del_col, del_colbtn = st.columns([8, 2])
+ # with del_col:
+ delete_column_name = st.selectbox("Select a column to delete:", [""] + editable_columns, key='delete_column')
+ # with del_colbtn:
+ # st.write('##')
+ if st.button("Delete Column") and delete_column_name:
+ del st.session_state['rules'][delete_column_name]
+ st.success(f"Column '{delete_column_name}' removed from rules.")
+
+
+
+
+ with col_right:
+ # Display the current state of the JSON rules
+ st.subheader('Formatted Columns')
+ st.json(st.session_state['rules']['Dictionary'])
+
+ # st.subheader('All Prompt Info')
+ # st.json(st.session_state['prompt_info'])
+
+
+ st.write('---')
+
+
+ col_left_mapping, col_right_mapping = st.columns([6,4])
+ with col_left_mapping:
+ st.header("Mapping")
+ st.write("Assign each column name to a single category.")
+ st.session_state['refresh_mapping'] = False
+
+ # Dynamically create a list of all column names that can be assigned
+ # This assumes that the column names are the keys in the dictionary under 'rules'
+ all_column_names = list(st.session_state['rules']['Dictionary'].keys())
+
+ categories = ['TAXONOMY', 'GEOGRAPHY', 'LOCALITY', 'COLLECTING', 'MISCELLANEOUS']
+ if ('mapping' not in st.session_state) or (st.session_state['mapping'] == {}):
+ st.session_state['mapping'] = {category: [] for category in categories}
+ for category in categories:
+ # Filter out the already assigned columns
+ available_columns = [col for col in all_column_names if col not in st.session_state['assigned_columns'] or col in st.session_state['mapping'].get(category, [])]
+
+ # Ensure the current mapping is a subset of the available options
+ current_mapping = [col for col in st.session_state['mapping'].get(category, []) if col in available_columns]
+
+ # Provide a safe default if the current mapping is empty or contains invalid options
+ safe_default = current_mapping if all(col in available_columns for col in current_mapping) else []
+
+ # Create a multi-select widget for the category with a safe default
+ selected_columns = st.multiselect(
+ f"Select columns for {category}:",
+ available_columns,
+ default=safe_default,
+ key=f"mapping_{category}"
+ )
+ # Update the assigned_columns based on the selections
+ for col in current_mapping:
+ if col not in selected_columns and col in st.session_state['assigned_columns']:
+ st.session_state['assigned_columns'].remove(col)
+ st.session_state['refresh_mapping'] = True
+
+ for col in selected_columns:
+ if col not in st.session_state['assigned_columns']:
+ st.session_state['assigned_columns'].append(col)
+ st.session_state['refresh_mapping'] = True
+
+ # Update the mapping in session state when there's a change
+ st.session_state['mapping'][category] = selected_columns
+ if st.session_state['refresh_mapping']:
+ st.session_state['refresh_mapping'] = False
+
+ # Button to confirm and save the mapping configuration
+ if st.button('Confirm Mapping'):
+ if check_unique_mapping_assignments():
+ # Proceed with further actions since the mapping is confirmed and unique
+ pass
+
+ with col_right_mapping:
+ # Display the current state of the JSON rules
+ st.subheader('Formatted Column Maps')
+ st.json(st.session_state['mapping'])
+
+
+ col_left_save, col_right_save = st.columns([6,4])
+ with col_left_save:
+ # Input for new file name
+ new_filename = st.text_input("Enter filename to save your prompt as a configuration YAML:",placeholder='my_prompt_name')
+ # Button to save the new YAML file
+ if st.button('Save YAML', type='primary'):
+ if new_filename:
+ if check_unique_mapping_assignments():
+ if check_prompt_yaml_filename(new_filename):
+ save_prompt_yaml(new_filename)
+ else:
+ st.error("File name can only contain letters, numbers, underscores, and dashes. Cannot contain spaces.")
+ else:
+ st.error("Mapping contains an error. Make sure that each column is assigned to only ***one*** category.")
+ else:
+ st.error("Please enter a filename.")
+
+ if st.button('Exit'):
+ st.session_state.proceed_to_build_llm_prompt = False
+ st.session_state.proceed_to_main = True
+ st.rerun()
+ with col_prompt_main_right:
+ st.subheader('All Prompt Components')
+ st.session_state['prompt_info'] = {
+ 'instructions': st.session_state['instructions'],
+ 'json_formatting_instructions': st.session_state['json_formatting_instructions'],
+ 'rules': st.session_state['rules'],
+ 'mapping': st.session_state['mapping'],
+ 'LLM': st.session_state['LLM']
+ }
+ st.json(st.session_state['prompt_info'])
+
+def save_yaml(content, filename="rules_config.yaml"):
+ with open(filename, 'w') as file:
+ yaml.dump(content, file)
+
+def show_header_welcome():
+ st.session_state.logo_path = os.path.join(st.session_state.dir_home, 'img','logo.png')
+ st.session_state.logo = Image.open(st.session_state.logo_path)
+ st.image(st.session_state.logo, width=250)
+
+def content_header():
+ col_run_1, col_run_2, col_run_3 = st.columns([4,2,2])
+ col_test = st.container()
+
+ st.write("")
+ st.write("")
+ st.write("")
+ st.write("")
+ st.subheader("Overall Progress")
+ col_run_info_1 = st.columns([1])[0]
+ st.write("")
+ st.write("")
+ st.write("")
+ st.write("")
+ st.header("Configuration Settings")
+
+ with col_run_info_1:
+ # Progress
+ # Progress
+ # st.subheader('Project')
+ # bar = st.progress(0)
+ # new_text = st.empty() # Placeholder for current step name
+ # progress_report = ProgressReportVV(bar, new_text, n_images=10)
+
+ # Progress
+ overall_progress_bar = st.progress(0)
+ text_overall = st.empty() # Placeholder for current step name
+ st.subheader('Transcription Progress')
+ batch_progress_bar = st.progress(0)
+ text_batch = st.empty() # Placeholder for current step name
+ progress_report = ProgressReport(overall_progress_bar, batch_progress_bar, text_overall, text_batch)
+ st.info("***Note:*** There is a known bug with tabs in Streamlit. If you update an input field it may take you back to the 'Project Settings' tab. Changes that you made are saved, it's just an annoying glitch. We are aware of this issue and will fix it as soon as we can.")
+ st.write("If you use VoucherVision frequently, you can change the default values that are auto-populated in the form below. In a text editor or IDE, edit the first few rows in the file `../VoucherVision/vouchervision/VoucherVision_Config_Builder.py`")
+
+
+ with col_run_1:
+ show_header_welcome()
+ st.subheader('Run VoucherVision')
+ if check_if_usable():
+ if st.button("Start Processing", type='primary'):
+
+ # First, write the config file.
+ write_config_file(st.session_state.config, st.session_state.dir_home, filename="VoucherVision.yaml")
+
+ path_custom_prompts = os.path.join(st.session_state.dir_home,'custom_prompts',st.session_state.config['leafmachine']['project']['prompt_version'])
+ # Call the machine function.
+ last_JSON_response, total_cost = voucher_vision(None, st.session_state.dir_home, path_custom_prompts, None, progress_report,path_api_cost=os.path.join(st.session_state.dir_home,'api_cost','api_cost.yaml'))
+
+ if total_cost:
+ st.success(f":money_with_wings: This run cost :heavy_dollar_sign:{total_cost:.4f}")
+
+ # Format the JSON string for display.
+ if last_JSON_response is None:
+ st.markdown(f"Last JSON object in the batch: NONE")
+ else:
+ try:
+ formatted_json = json.dumps(json.loads(last_JSON_response), indent=4)
+ except:
+ formatted_json = json.dumps(last_JSON_response, indent=4)
+ st.markdown(f"Last JSON object in the batch:\n```\n{formatted_json}\n```")
+ st.balloons()
+
+ else:
+ st.button("Start Processing", type='primary', disabled=True)
+ st.error(":heavy_exclamation_mark: Required API keys not set. Please visit the 'API Keys' tab and set the Google Vision OCR API key and at least one LLM key.")
+
+ with col_run_2:
+ st.subheader('Run Tests', help="")
+ st.write('We include a single image for testing. If you want to test all of the available prompts and LLMs on a different set of images, copy your images into `../VoucherVision/demo/demo_images`.')
+ if st.button("Test GPT"):
+ progress_report.set_n_overall(TestOptionsGPT.get_length())
+ test_results, JSON_results = run_demo_tests_GPT(progress_report)
+ with col_test:
+ display_test_results(test_results, JSON_results, 'gpt')
+ st.balloons()
+
+ if st.button("Test PaLM2"):
+ progress_report.set_n_overall(TestOptionsPalm.get_length())
+ test_results, JSON_results = run_demo_tests_Palm(progress_report)
+ with col_test:
+ display_test_results(test_results, JSON_results, 'palm')
+ st.balloons()
+
+ with col_run_3:
+ st.subheader('Check GPU')
+ if st.button("GPU"):
+ success, info = test_GPU()
+
+ if success:
+ st.balloons()
+ for message in info:
+ st.success(message)
+ else:
+ for message in info:
+ st.error(message)
+
+def content_tab_settings():
+ st.header('Project')
+ col_project_1, col_project_2 = st.columns([4,2])
+
+ st.write("---")
+ st.header('Input Images')
+ col_local_1, col_local_2 = st.columns([4,2])
+
+ # st.write("---")
+ # st.header('Modules')
+ # col_m1, col_m2 = st.columns(2)
+
+ st.write("---")
+ st.header('Cropped Components')
+ col_cropped_1, col_cropped_2 = st.columns([4,4])
+
+ os.path.join(st.session_state.dir_home, )
+ ### Project
+ with col_project_1:
+ st.session_state.config['leafmachine']['project']['run_name'] = st.text_input("Run name", st.session_state.config['leafmachine']['project'].get('run_name', ''))
+ st.session_state.config['leafmachine']['project']['dir_output'] = st.text_input("Output directory", st.session_state.config['leafmachine']['project'].get('dir_output', ''))
+
+ ### Input Images Local
+ with col_local_1:
+ st.session_state.config['leafmachine']['project']['dir_images_local'] = st.text_input("Input images directory", st.session_state.config['leafmachine']['project'].get('dir_images_local', ''))
+ st.session_state.config['leafmachine']['project']['continue_run_from_partial_xlsx'] = st.text_input("Continue run from partially completed project XLSX", st.session_state.config['leafmachine']['project'].get('continue_run_from_partial_xlsx', ''), disabled=True)
+ st.write("---")
+ st.subheader('LLM Version')
+ st.markdown(
+ """
+ ***Note:*** GPT-4 is 20x more expensive than GPT-3.5
+ """
+ )
+ st.session_state.config['leafmachine']['LLM_version'] = st.selectbox("LLM version", ["GPT 4", "GPT 3.5", "Azure GPT 4", "Azure GPT 3.5", "PaLM 2"], index=["GPT 4", "GPT 3.5", "Azure GPT 4", "Azure GPT 3.5", "PaLM 2"].index(st.session_state.config['leafmachine'].get('LLM_version', 'Azure GPT 4')))
+
+ st.write("---")
+ st.subheader('Prompt Version')
+ versions, default_version = get_prompt_versions(st.session_state.config['leafmachine']['LLM_version'])
+
+ if versions:
+ selected_version = st.session_state.config['leafmachine']['project'].get('prompt_version', default_version)
+ if selected_version not in versions:
+ selected_version = default_version
+ st.session_state.config['leafmachine']['project']['prompt_version'] = st.selectbox("Prompt Version", versions, index=versions.index(selected_version))
+
+ # if st.session_state.config['leafmachine']['LLM_version'] in ["GPT 4", "GPT 3.5", "Azure GPT 4", "Azure GPT 3.5",]:
+ # st.session_state.config['leafmachine']['project']['prompt_version'] = st.selectbox("Prompt Version", ["Version 1", "Version 1 No Domain Knowledge", "Version 2"], index=["Version 1", "Version 1 No Domain Knowledge", "Version 2"].index(st.session_state.config['leafmachine']['project'].get('prompt_version', "Version 2")))
+ # elif st.session_state.config['leafmachine']['LLM_version'] in ["PaLM 2",]:
+ # st.session_state.config['leafmachine']['project']['prompt_version'] = st.selectbox("Prompt Version", ["Version 1 PaLM 2", "Version 1 PaLM 2 No Domain Knowledge", "Version 2 PaLM 2"], index=["Version 1 PaLM 2", "Version 1 PaLM 2 No Domain Knowledge", "Version 2 PaLM 2"].index(st.session_state.config['leafmachine']['project'].get('prompt_version', "Version 2 PaLM 2")))
+
+ ### Modules
+ # with col_m1:
+ # st.session_state.config['leafmachine']['modules']['specimen_crop'] = st.checkbox("Specimen Close-up", st.session_state.config['leafmachine']['modules'].get('specimen_crop', True),disabled=True)
+
+ ### cropped_components
+ # with col_cropped_1:
+ # st.session_state.config['leafmachine']['cropped_components']['do_save_cropped_annotations'] = st.checkbox("Save cropped components as images", st.session_state.config['leafmachine']['cropped_components'].get('do_save_cropped_annotations', True), disabled=True)
+ # st.session_state.config['leafmachine']['cropped_components']['save_per_image'] = st.checkbox("Save cropped components grouped by specimen", st.session_state.config['leafmachine']['cropped_components'].get('save_per_image', False), disabled=True)
+ # st.session_state.config['leafmachine']['cropped_components']['save_per_annotation_class'] = st.checkbox("Save cropped components grouped by type", st.session_state.config['leafmachine']['cropped_components'].get('save_per_annotation_class', True), disabled=True)
+ # st.session_state.config['leafmachine']['cropped_components']['binarize_labels'] = st.checkbox("Binarize labels", st.session_state.config['leafmachine']['cropped_components'].get('binarize_labels', False), disabled=True)
+ # st.session_state.config['leafmachine']['cropped_components']['binarize_labels_skeletonize'] = st.checkbox("Binarize and skeletonize labels", st.session_state.config['leafmachine']['cropped_components'].get('binarize_labels_skeletonize', False), disabled=True)
+
+ with col_cropped_1:
+ default_crops = st.session_state.config['leafmachine']['cropped_components'].get('save_cropped_annotations', ['leaf_whole'])
+ st.write("Prior to transcription, use LeafMachine2 to crop all labels from input images to create label collages for each specimen image. (Requires GPU)")
+ st.session_state.config['leafmachine']['use_RGB_label_images'] = st.checkbox("Use LeafMachine2 label collage for transcriptions", st.session_state.config['leafmachine'].get('use_RGB_label_images', False))
+
+ st.session_state.config['leafmachine']['cropped_components']['save_cropped_annotations'] = st.multiselect("Components to crop",
+ ['ruler', 'barcode','label', 'colorcard','map','envelope','photo','attached_item','weights',
+ 'leaf_whole', 'leaf_partial', 'leaflet', 'seed_fruit_one', 'seed_fruit_many', 'flower_one', 'flower_many', 'bud','specimen','roots','wood'],default=default_crops)
+ with col_cropped_2:
+ ba = os.path.join(st.session_state.dir_home,'demo', 'ba','ba2.png')
+ image = Image.open(ba)
+ st.image(image, caption='LeafMachine2 Collage', output_format = "PNG")
+
+def content_tab_component():
+ st.header('Archival Components')
+ ACD_version = st.selectbox("Archival Component Detector (ACD) Version", ["Version 2.1", "Version 2.2"])
+
+ ACD_confidence_default = int(st.session_state.config['leafmachine']['archival_component_detector']['minimum_confidence_threshold'] * 100)
+ ACD_confidence = st.number_input("ACD Confidence Threshold (%)", min_value=0, max_value=100,value=ACD_confidence_default)
+ st.session_state.config['leafmachine']['archival_component_detector']['minimum_confidence_threshold'] = float(ACD_confidence/100)
+
+ st.session_state.config['leafmachine']['archival_component_detector']['do_save_prediction_overlay_images'] = st.checkbox("Save Archival Prediction Overlay Images", st.session_state.config['leafmachine']['archival_component_detector'].get('do_save_prediction_overlay_images', True))
+
+ st.session_state.config['leafmachine']['archival_component_detector']['ignore_objects_for_overlay'] = st.multiselect("Hide Archival Components in Prediction Overlay Images",
+ ['ruler', 'barcode','label', 'colorcard','map','envelope','photo','attached_item','weights',],
+ default=[])
+
+ # Depending on the selected version, set the configuration
+ if ACD_version == "Version 2.1":
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_type'] = 'Archival_Detector'
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_version'] = 'PREP_final'
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_iteration'] = 'PREP_final'
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_weights'] = 'best.pt'
+ elif ACD_version == "Version 2.2": #TODO update this to version 2.2
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_type'] = 'Archival_Detector'
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_version'] = 'PREP_final'
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_iteration'] = 'PREP_final'
+ st.session_state.config['leafmachine']['archival_component_detector']['detector_weights'] = 'best.pt'
+
+
+def content_tab_processing():
+ st.header('Processing Options')
+ col_processing_1, col_processing_2 = st.columns([2,2,])
+ with col_processing_1:
+ st.subheader('Compute Options')
+ st.session_state.config['leafmachine']['project']['num_workers'] = st.number_input("Number of CPU workers", value=st.session_state.config['leafmachine']['project'].get('num_workers', 1), disabled=True)
+ st.session_state.config['leafmachine']['project']['batch_size'] = st.number_input("Batch size", value=st.session_state.config['leafmachine']['project'].get('batch_size', 500), help='Sets the batch size for the LeafMachine2 cropping. If computer RAM is filled, lower this value to ~100.')
+ with col_processing_2:
+ st.subheader('Misc')
+ st.session_state.config['leafmachine']['project']['prefix_removal'] = st.text_input("Remove prefix from catalog number", st.session_state.config['leafmachine']['project'].get('prefix_removal', ''))
+ st.session_state.config['leafmachine']['project']['suffix_removal'] = st.text_input("Remove suffix from catalog number", st.session_state.config['leafmachine']['project'].get('suffix_removal', ''))
+ st.session_state.config['leafmachine']['project']['catalog_numerical_only'] = st.checkbox("Require 'Catalog Number' to be numerical only", st.session_state.config['leafmachine']['project'].get('catalog_numerical_only', True))
+
+ ### Logging and Image Validation - col_v1
+ st.header('Logging and Image Validation')
+ col_v1, col_v2 = st.columns(2)
+ with col_v1:
+ st.session_state.config['leafmachine']['do']['check_for_illegal_filenames'] = st.checkbox("Check for illegal filenames", st.session_state.config['leafmachine']['do'].get('check_for_illegal_filenames', True))
+ st.session_state.config['leafmachine']['do']['check_for_corrupt_images_make_vertical'] = st.checkbox("Check for corrupt images", st.session_state.config['leafmachine']['do'].get('check_for_corrupt_images_make_vertical', True))
+
+ st.session_state.config['leafmachine']['print']['verbose'] = st.checkbox("Print verbose", st.session_state.config['leafmachine']['print'].get('verbose', True))
+ st.session_state.config['leafmachine']['print']['optional_warnings'] = st.checkbox("Show optional warnings", st.session_state.config['leafmachine']['print'].get('optional_warnings', True))
+
+ with col_v2:
+ log_level = st.session_state.config['leafmachine']['logging'].get('log_level', None)
+ log_level_display = log_level if log_level is not None else 'default'
+ selected_log_level = st.selectbox("Logging Level", ['default', 'DEBUG', 'INFO', 'WARNING', 'ERROR'], index=['default', 'DEBUG', 'INFO', 'WARNING', 'ERROR'].index(log_level_display))
+
+ if selected_log_level == 'default':
+ st.session_state.config['leafmachine']['logging']['log_level'] = None
+ else:
+ st.session_state.config['leafmachine']['logging']['log_level'] = selected_log_level
+
+def content_tab_domain():
+ st.header('Embeddings Database')
+ col_emb_1, col_emb_2 = st.columns([4,2])
+ with col_emb_1:
+ st.markdown(
+ """
+ VoucherVision includes the option of using domain knowledge inside of the dynamically generated prompts. The OCR text is queried against a database of existing label transcriptions. The most similar existing transcriptions act as an example of what the LLM should emulate and are shown to the LLM as JSON objects. VoucherVision uses cosine similarity search to return the most similar existing transcription.
+ - Note: Using domain knowledge may increase the chance that foreign text is included in the final transcription
+ - Disabling this feature will show the LLM multiple examples of an empty JSON skeleton structure instead
+ - Enabling this option requires a GPU with at least 8GB of VRAM
+ - The domain knowledge files can be located in the directory "../VoucherVision/domain_knowledge". On first run the embeddings database must be created, which takes time. If the database creation runs each time you use VoucherVision, then something is wrong.
+ """
+ )
+
+ st.write(f"Domain Knowledge is only available for the following prompts:")
+ for available_prompts in PROMPTS_THAT_NEED_DOMAIN_KNOWLEDGE:
+ st.markdown(f"- {available_prompts}")
+
+ if st.session_state.config['leafmachine']['project']['prompt_version'] in PROMPTS_THAT_NEED_DOMAIN_KNOWLEDGE:
+ st.session_state.config['leafmachine']['project']['use_domain_knowledge'] = st.checkbox("Use domain knowledge", True, disabled=True)
+ else:
+ st.session_state.config['leafmachine']['project']['use_domain_knowledge'] = st.checkbox("Use domain knowledge", False, disabled=True)
+
+ st.write("")
+ if st.session_state.config['leafmachine']['project']['use_domain_knowledge']:
+ st.session_state.config['leafmachine']['project']['embeddings_database_name'] = st.text_input("Embeddings database name (only use underscores)", st.session_state.config['leafmachine']['project'].get('embeddings_database_name', ''))
+ st.session_state.config['leafmachine']['project']['build_new_embeddings_database'] = st.checkbox("Build *new* embeddings database", st.session_state.config['leafmachine']['project'].get('build_new_embeddings_database', False))
+ st.session_state.config['leafmachine']['project']['path_to_domain_knowledge_xlsx'] = st.text_input("Path to domain knowledge CSV file (will be used to create new embeddings database)", st.session_state.config['leafmachine']['project'].get('path_to_domain_knowledge_xlsx', ''))
+ else:
+ st.session_state.config['leafmachine']['project']['embeddings_database_name'] = st.text_input("Embeddings database name (only use underscores)", st.session_state.config['leafmachine']['project'].get('embeddings_database_name', ''), disabled=True)
+ st.session_state.config['leafmachine']['project']['build_new_embeddings_database'] = st.checkbox("Build *new* embeddings database", st.session_state.config['leafmachine']['project'].get('build_new_embeddings_database', False), disabled=True)
+ st.session_state.config['leafmachine']['project']['path_to_domain_knowledge_xlsx'] = st.text_input("Path to domain knowledge CSV file (will be used to create new embeddings database)", st.session_state.config['leafmachine']['project'].get('path_to_domain_knowledge_xlsx', ''), disabled=True)
+
+def render_expense_report_summary():
+ expense_summary = st.session_state.expense_summary
+ expense_report = st.session_state.expense_report
+ st.header('Expense Report Summary')
+
+ if expense_summary:
+ st.metric(label="Total Cost", value=f"${round(expense_summary['total_cost_sum'], 4):,}")
+ col1, col2 = st.columns(2)
+
+ # Run count and total costs
+ with col1:
+ st.metric(label="Run Count", value=expense_summary['run_count'])
+ st.metric(label="Tokens In", value=f"{expense_summary['tokens_in_sum']:,}")
+
+ # Token information
+ with col2:
+ st.metric(label="Total Images", value=expense_summary['n_images_sum'])
+ st.metric(label="Tokens Out", value=f"{expense_summary['tokens_out_sum']:,}")
+
+
+ # Calculate cost proportion per image for each API version
+ st.subheader('Average Cost per Image by API Version')
+ cost_labels = []
+ cost_values = []
+ total_images = 0
+ cost_per_image_dict = {}
+ # Iterate through the expense report to accumulate costs and image counts
+ for index, row in expense_report.iterrows():
+ api_version = row['api_version']
+ total_cost = row['total_cost']
+ n_images = row['n_images']
+ total_images += n_images # Keep track of total images processed
+ if api_version not in cost_per_image_dict:
+ cost_per_image_dict[api_version] = {'total_cost': 0, 'n_images': 0}
+ cost_per_image_dict[api_version]['total_cost'] += total_cost
+ cost_per_image_dict[api_version]['n_images'] += n_images
+
+ api_versions = list(cost_per_image_dict.keys())
+ colors = [COLORS_EXPENSE_REPORT[version] if version in COLORS_EXPENSE_REPORT else '#DDDDDD' for version in api_versions]
+
+ # Calculate the cost per image for each API version
+ for version, cost_data in cost_per_image_dict.items():
+ total_cost = cost_data['total_cost']
+ n_images = cost_data['n_images']
+ # Calculate the cost per image for this version
+ cost_per_image = total_cost / n_images if n_images > 0 else 0
+ cost_labels.append(version)
+ cost_values.append(cost_per_image)
+ # Generate the pie chart
+ cost_pie_chart = go.Figure(data=[go.Pie(labels=cost_labels, values=cost_values, hole=.3)])
+ # Update traces for custom text in hoverinfo, displaying cost with a dollar sign and two decimal places
+ cost_pie_chart.update_traces(
+ marker=dict(colors=colors),
+ text=[f"${value:.2f}" for value in cost_values], # Formats the cost as a string with a dollar sign and two decimals
+ textinfo='percent+label',
+ hoverinfo='label+percent+text' # Adds custom text (formatted cost) to the hover information
+ )
+ st.plotly_chart(cost_pie_chart, use_container_width=True)
+
+
+
+ st.subheader('Proportion of Total Cost by API Version')
+ cost_labels = []
+ cost_proportions = []
+ total_cost_by_version = {}
+ # Sum the total cost for each API version
+ for index, row in expense_report.iterrows():
+ api_version = row['api_version']
+ total_cost = row['total_cost']
+ if api_version not in total_cost_by_version:
+ total_cost_by_version[api_version] = 0
+ total_cost_by_version[api_version] += total_cost
+ # Calculate the combined total cost for all versions
+ combined_total_cost = sum(total_cost_by_version.values())
+ # Calculate the proportion of total cost for each API version
+ for version, total_cost in total_cost_by_version.items():
+ proportion = (total_cost / combined_total_cost) * 100 if combined_total_cost > 0 else 0
+ cost_labels.append(version)
+ cost_proportions.append(proportion)
+ # Generate the pie chart
+ cost_pie_chart = go.Figure(data=[go.Pie(labels=cost_labels, values=cost_proportions, hole=.3)])
+ # Update traces for custom text in hoverinfo
+ cost_pie_chart.update_traces(
+ marker=dict(colors=colors),
+ text=[f"${cost:.2f}" for cost in total_cost_by_version.values()], # This will format the cost to 2 decimal places
+ textinfo='percent+label',
+ hoverinfo='label+percent+text' # This tells Plotly to show the label, percent, and custom text (cost) on hover
+ )
+ st.plotly_chart(cost_pie_chart, use_container_width=True)
+
+ # API version usage percentages pie chart
+ st.subheader('Runs by API Version')
+ api_versions = list(expense_summary['api_version_percentages'].keys())
+ percentages = [expense_summary['api_version_percentages'][version] for version in api_versions]
+ pie_chart = go.Figure(data=[go.Pie(labels=api_versions, values=percentages, hole=.3)])
+ pie_chart.update_layout(margin=dict(t=0, b=0, l=0, r=0))
+ pie_chart.update_traces(marker=dict(colors=colors),)
+ st.plotly_chart(pie_chart, use_container_width=True)
+
+ else:
+ st.error('No expense report data available.')
+
+def sidebar_content():
+ try:
+ validate_dir(os.path.join(st.session_state.dir_home,'expense_report'))
+ st.session_state.expense_summary, st.session_state.expense_report = summarize_expense_report(os.path.join(st.session_state.dir_home,'expense_report','expense_report.csv'))
+ render_expense_report_summary()
+ except:
+ st.header('Expense Report Summary')
+ st.write('Available after first run...')
+
+ # # Check if the expense summary is available in the session state
+ # if 'expense' not in st.session_state or st.session_state.expense is None:
+ # st.sidebar.write('No expense report data available.')
+ # return
+
+ # # Retrieve the expense report summary
+ # expense_summary = st.session_state.expense
+
+ # # Display the expense report summary
+ # st.sidebar.markdown('**Run Count**: ' + str(expense_summary['run_count']))
+
+ # # API version usage percentages
+ # st.sidebar.markdown('**API Version Usage**:')
+ # for version, percentage in expense_summary['api_version_percentages'].items():
+ # st.sidebar.markdown(f'- {version}: {percentage:.2f}%')
+
+ # # Summary of costs and tokens
+ # st.sidebar.markdown('**Total Cost**: $' + str(round(expense_summary['total_cost_sum'], 4)))
+ # st.sidebar.markdown('**Tokens In**: ' + str(expense_summary['tokens_in_sum']))
+ # st.sidebar.markdown('**Tokens Out**: ' + str(expense_summary['tokens_out_sum']))
+ # # st.sidebar.markdown('**Rate In**: $' + str(round(expense_summary['rate_in_sum'], 2)) + ' per 1000 tokens')
+ # # st.sidebar.markdown('**Rate Out**: $' + str(round(expense_summary['rate_out_sum'], 2)) + ' per 1000 tokens')
+ # st.sidebar.markdown('**Cost In**: $' + str(round(expense_summary['cost_in_sum'], 4)))
+ # st.sidebar.markdown('**Cost Out**: $' + str(round(expense_summary['cost_out_sum'], 4)))
+
+def main():
+ with st.sidebar:
+ sidebar_content()
+ # Main App
+ content_header()
+
+ tab_settings, tab_prompt, tab_domain, tab_component, tab_processing, tab_private, tab_delete = st.tabs(["Project Settings", "Prompt Builder", "Domain Knowledge","Component Detector", "Processing Options", "API Keys", "Space-Saver"])
+
+ with tab_settings:
+ content_tab_settings()
+
+ with tab_prompt:
+ if st.button("Build Custom LLM Prompt"):
+ st.session_state.proceed_to_build_llm_prompt = True
+ st.rerun()
+
+ with tab_component:
+ content_tab_component()
+
+ with tab_domain:
+ content_tab_domain()
+
+ with tab_processing:
+ content_tab_processing()
+
+ with tab_private:
+ if st.button("Edit API Keys"):
+ st.session_state.proceed_to_private = True
+ st.rerun()
+
+ with tab_delete:
+ create_space_saver()
+
+st.set_page_config(layout="wide", page_icon='img/icon.ico', page_title='VoucherVision')
+
+# Default YAML file path
+if 'config' not in st.session_state:
+ st.session_state.config, st.session_state.dir_home = build_VV_config()
+ setup_streamlit_config(st.session_state.dir_home)
+
+if 'proceed_to_main' not in st.session_state:
+ st.session_state.proceed_to_main = False # New state variable to control the flow
+
+if 'proceed_to_build_llm_prompt' not in st.session_state:
+ st.session_state.proceed_to_build_llm_prompt = False # New state variable to control the flow
+if 'proceed_to_private' not in st.session_state:
+ st.session_state.proceed_to_private = False # New state variable to control the flow
+
+if 'private_file' not in st.session_state:
+ st.session_state.private_file = does_private_file_exist()
+ if st.session_state.private_file:
+ st.session_state.proceed_to_main = True
+
+# Initialize session_state variables if they don't exist
+if 'prompt_info' not in st.session_state:
+ st.session_state['prompt_info'] = {}
+if 'rules' not in st.session_state:
+ st.session_state['rules'] = {}
+
+if not st.session_state.private_file:
+ create_private_file()
+elif st.session_state.proceed_to_build_llm_prompt:
+ build_LLM_prompt_config()
+elif st.session_state.proceed_to_private:
+ create_private_file()
+elif st.session_state.proceed_to_main:
+ main()
\ No newline at end of file
diff --git a/bin/version.yml b/bin/version.yml
new file mode 100644
index 0000000000000000000000000000000000000000..7c822fa9364394f2ca68c5c0a9973d102a5a842b
--- /dev/null
+++ b/bin/version.yml
@@ -0,0 +1,2 @@
+last_update: '2023-10-24'
+version: v-2-1
diff --git a/create_desktop_shortcut.py b/create_desktop_shortcut.py
new file mode 100644
index 0000000000000000000000000000000000000000..d9fbc721c5340ec913d275b5fbcadf7dc7144fa6
--- /dev/null
+++ b/create_desktop_shortcut.py
@@ -0,0 +1,69 @@
+import os, sys
+import win32com.client
+import tkinter as tk
+from tkinter import filedialog
+from PIL import Image, ImageEnhance
+
+def create_shortcut():
+ # Request user's confirmation
+ confirmation = input("Do you want to create a shortcut for the VoucherVision? (y/n): ")
+
+ if confirmation.lower() != "y":
+ print("Okay, no shortcut will be created.")
+ return
+
+ # Get the script path
+ script_path = os.path.abspath(__file__)
+ # Get the directory of the script
+ script_dir = os.path.dirname(script_path)
+
+ # Path to the icon file
+ icon_path = os.path.join(script_dir, 'img', 'icon.jpg')
+ img = Image.open(icon_path)
+ enhancer = ImageEnhance.Color(img)
+ img_enhanced = enhancer.enhance(1.5)
+ img_enhanced.save(os.path.join(script_dir, 'img', 'icon.ico'), format='ICO', sizes=[(256,256)])
+ icon_path_ico = os.path.join(script_dir, 'img', 'icon.ico')
+
+ # Construct the path to the static folder
+ static_dir = os.path.join(script_dir, "static")
+
+ # Ask for the name of the shortcut
+ shortcut_name = "Voucher Vision"
+
+ root = tk.Tk()
+ root.withdraw() # Hide the main window
+
+ root.update() # Ensures that the dialog appears on top
+ folder_path = filedialog.askdirectory(title="Choose location to save the shortcut")
+ print(f"Shortcut will be saved to {folder_path}")
+
+ venv_path = filedialog.askdirectory(title="Choose the location of your Python virtual environment")
+ print(f"Using virtual environment located at {venv_path}")
+
+ # Path to the activate script in the venv
+ activate_path = os.path.join(venv_path, "Scripts")
+
+ shortcut_path = os.path.join(folder_path, f'{shortcut_name}.lnk')
+
+ shell = win32com.client.Dispatch("WScript.Shell")
+ shortcut = shell.CreateShortCut(shortcut_path)
+ shortcut.Targetpath = "%windir%\System32\cmd.exe"
+ # The command activates the venv, navigates to the script's directory, then runs the script
+ # shortcut.Arguments = f'/K "{activate_path} & cd /D {os.path.dirname(script_path)} & streamlit run VoucherVisionEditor.py"'
+ # shortcut.Arguments = f'/K "{activate_path} & cd /D {static_dir} & start cmd /c python -m http.server & cd /D {script_dir} & streamlit run VoucherVisionEditor.py"'
+ streamlit_exe = os.path.join(venv_path, "Scripts","streamlit")
+ print(script_dir)
+ print(streamlit_exe)
+ activate_path = os.path.join(script_dir,"venv_VV","Scripts")
+ print(activate_path)
+ shortcut.Arguments = f'/K cd /D ""{activate_path}"" && activate && cd /D ""{script_dir}"" && python run_VoucherVision.py'
+ # Set the icon of the shortcut
+ shortcut.IconLocation = icon_path_ico
+
+ shortcut.save()
+
+ print(f"Shortcut created with the name '{shortcut_name}' in the chosen directory.")
+
+if __name__ == "__main__":
+ create_shortcut()
\ No newline at end of file
diff --git a/custom_prompts/required_structure.yaml b/custom_prompts/required_structure.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d73b9b286572f6fa0bd43a9b4d1669dfffc9e1b3
--- /dev/null
+++ b/custom_prompts/required_structure.yaml
@@ -0,0 +1,62 @@
+LLM: gpt
+instructions: '1. Refactor the unstructured OCR text into a dictionary based on the
+ JSON structure outlined below.
+
+ 2. You should map the unstructured OCR text to the appropriate JSON key and then
+ populate the field based on its rules.
+
+ 3. Some JSON key fields are permitted to remain empty if the corresponding information
+ is not found in the unstructured OCR text.
+
+ 4. Ignore any information in the OCR text that doesn''t fit into the defined JSON
+ structure.
+
+ 5. Duplicate dictionary fields are not allowed.
+
+ 6. Ensure that all JSON keys are in lowercase.
+
+ 7. Ensure that new JSON field values follow sentence case capitalization.
+
+ 8. Ensure all key-value pairs in the JSON dictionary strictly adhere to the format
+ and data types specified in the template.
+
+ 9. Ensure the output JSON string is valid JSON format. It should not have trailing
+ commas or unquoted keys.
+
+ 10. Only return a JSON dictionary represented as a string. You should not explain
+ your answer.'
+json_formatting_instructions: "The next section of instructions outlines how to format\
+ \ the JSON dictionary. The keys are the same as those of the final formatted JSON\
+ \ object.\nFor each key there is a format requirement that specifies how to transcribe\
+ \ the information for that key. \nThe possible formatting options are:\n1. \"verbatim\
+ \ transcription\" - field is populated with verbatim text from the unformatted OCR.\n\
+ 2. \"spell check transcription\" - field is populated with spelling corrected text\
+ \ from the unformatted OCR.\n3. \"boolean yes no\" - field is populated with only\
+ \ yes or no.\n4. \"boolean 1 0\" - field is populated with only 1 or 0.\n5. \"integer\"\
+ \ - field is populated with only an integer.\n6. \"[list]\" - field is populated\
+ \ from one of the values in the list.\n7. \"yyyy-mm-dd\" - field is populated with\
+ \ a date in the format year-month-day.\nThe desired null value is also given. Populate\
+ \ the field with the null value of the information for that key is not present in\
+ \ the unformatted OCR text."
+mapping:
+ # Add column names to the desired category. This is used to map the VV Editor.
+ COLLECTING: []
+ GEOGRAPHY: []
+ LOCALITY: []
+ MISCELLANEOUS: []
+ TAXONOMY:
+ - catalog_number
+rules:
+ Dictionary:
+ # Manually add rows here. You MUST keep 'catalog_number' unchanged. Use 'catalog_number' as a guide for adding more columns.
+ # The only values allowed in the 'format' key are those outlines above in the 'json_formatting_instructions' section.
+ # If you want an empty cell by default, use '' for the 'null_value'.
+ catalog_number:
+ description: The barcode identifier, typically a number with at least 6 digits,
+ but fewer than 30 digits.
+ format: verbatim transcription
+ null_value: ''
+ # Do not change or remove below. This is required for some LLMs
+ SpeciesName:
+ taxonomy:
+ - Genus_species
diff --git a/custom_prompts/version_2.yaml b/custom_prompts/version_2.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..254431b8aefd8cbfb0c6d558bdece3c51aabd6b3
--- /dev/null
+++ b/custom_prompts/version_2.yaml
@@ -0,0 +1,229 @@
+LLM: gpt
+instructions: '1. Refactor the unstructured OCR text into a dictionary based on the
+ JSON structure outlined below.
+
+ 2. You should map the unstructured OCR text to the appropriate JSON key and then
+ populate the field based on its rules.
+
+ 3. Some JSON key fields are permitted to remain empty if the corresponding information
+ is not found in the unstructured OCR text.
+
+ 4. Ignore any information in the OCR text that doesn''t fit into the defined JSON
+ structure.
+
+ 5. Duplicate dictionary fields are not allowed.
+
+ 6. Ensure that all JSON keys are in lowercase.
+
+ 7. Ensure that new JSON field values follow sentence case capitalization.
+
+ 8. Ensure all key-value pairs in the JSON dictionary strictly adhere to the format
+ and data types specified in the template.
+
+ 9. Ensure the output JSON string is valid JSON format. It should not have trailing
+ commas or unquoted keys.
+
+ 10. Only return a JSON dictionary represented as a string. You should not explain
+ your answer.'
+json_formatting_instructions: "The next section of instructions outlines how to format\
+ \ the JSON dictionary. The keys are the same as those of the final formatted JSON\
+ \ object.\nFor each key there is a format requirement that specifies how to transcribe\
+ \ the information for that key. \nThe possible formatting options are:\n1. \"verbatim\
+ \ transcription\" - field is populated with verbatim text from the unformatted OCR.\n\
+ 2. \"spell check transcription\" - field is populated with spelling corrected text\
+ \ from the unformatted OCR.\n3. \"boolean yes no\" - field is populated with only\
+ \ yes or no.\n4. \"boolean 1 0\" - field is populated with only 1 or 0.\n5. \"integer\"\
+ \ - field is populated with only an integer.\n6. \"[list]\" - field is populated\
+ \ from one of the values in the list.\n7. \"yyyy-mm-dd\" - field is populated with\
+ \ a date in the format year-month-day.\nThe desired null value is also given. Populate\
+ \ the field with the null value of the information for that key is not present in\
+ \ the unformatted OCR text."
+mapping:
+ COLLECTING:
+ - collectors
+ - collector_number
+ - determined_by
+ - multiple_names
+ - verbatim_date
+ - date
+ - end_date
+ GEOGRAPHY:
+ - country
+ - state
+ - county
+ - min_elevation
+ - max_elevation
+ - elevation_units
+ LOCALITY:
+ - locality_name
+ - verbatim_coordinates
+ - decimal_coordinates
+ - datum
+ - plant_description
+ - cultivated
+ - habitat
+ MISCELLANEOUS: []
+ TAXONOMY:
+ - catalog_number
+ - genus
+ - species
+ - subspecies
+ - variety
+ - forma
+rules:
+ Dictionary:
+ catalog_number:
+ description: The barcode identifier, typically a number with at least 6 digits,
+ but fewer than 30 digits.
+ format: verbatim transcription
+ null_value: ''
+ collector_number:
+ description: Unique identifier or number that denotes the specific collecting
+ event and associated with the collector.
+ format: verbatim transcription
+ null_value: s.n.
+ collectors:
+ description: Full name(s) of the individual(s) responsible for collecting the
+ specimen. When multiple collectors are involved, their names should be separated
+ by commas.
+ format: verbatim transcription
+ null_value: not present
+ country:
+ description: Country that corresponds to the current geographic location of
+ collection. Capitalize first letter of each word. If abbreviation is given
+ populate field with the full spelling of the country's name.
+ format: spell check transcription
+ null_value: ''
+ county:
+ description: Administrative division 2 that corresponds to the current geographic
+ location of collection; capitalize first letter of each word. Administrative
+ division 2 is equivalent to a U.S. county, parish, borough.
+ format: spell check transcription
+ null_value: ''
+ cultivated:
+ description: Cultivated plants are intentionally grown by humans. In text descriptions,
+ look for planting dates, garden locations, ornamental, cultivar names, garden,
+ or farm to indicate cultivated plant.
+ format: boolean yes no
+ null_value: ''
+ date:
+ description: 'Date the specimen was collected formatted as year-month-day. If
+ specific components of the date are unknown, they should be replaced with
+ zeros. Examples: ''0000-00-00'' if the entire date is unknown, ''YYYY-00-00''
+ if only the year is known, and ''YYYY-MM-00'' if year and month are known
+ but day is not.'
+ format: yyyy-mm-dd
+ null_value: ''
+ datum:
+ description: Datum of location coordinates. Possible values are include in the
+ format list. Leave field blank if unclear. [WGS84, WGS72, WGS66, WGS60, NAD83,
+ NAD27, OSGB36, ETRS89, ED50, GDA94, JGD2011, Tokyo97, KGD2002, TWD67, TWD97,
+ BJS54, XAS80, GCJ-02, BD-09, PZ-90.11, GTRF, CGCS2000, ITRF88, ITRF89, ITRF90,
+ ITRF91, ITRF92, ITRF93, ITRF94, ITRF96, ITRF97, ITRF2000, ITRF2005, ITRF2008,
+ ITRF2014, Hong Kong Principal Datum, SAD69]
+ format: '[list]'
+ null_value: ''
+ decimal_coordinates:
+ description: Correct and convert the verbatim location coordinates to conform
+ with the decimal degrees GPS coordinate format.
+ format: spell check transcription
+ null_value: ''
+ determined_by:
+ description: Full name of the individual responsible for determining the taxanomic
+ name of the specimen. Sometimes the name will be near to the characters 'det'
+ to denote determination. This name may be isolated from other names in the
+ unformatted OCR text.
+ format: verbatim transcription
+ null_value: ''
+ elevation_units:
+ description: 'Elevation units must be meters. If min_elevation field is populated,
+ then elevation_units: ''m''. Otherwise elevation_units: ''''.'
+ format: spell check transcription
+ null_value: ''
+ end_date:
+ description: 'If a date range is provided, this represents the later or ending
+ date of the collection period, formatted as year-month-day. If specific components
+ of the date are unknown, they should be replaced with zeros. Examples: ''0000-00-00''
+ if the entire end date is unknown, ''YYYY-00-00'' if only the year of the
+ end date is known, and ''YYYY-MM-00'' if year and month of the end date are
+ known but the day is not.'
+ format: yyyy-mm-dd
+ null_value: ''
+ forma:
+ description: Taxonomic determination to form (f.).
+ format: verbatim transcription
+ null_value: ''
+ genus:
+ description: Taxonomic determination to genus. Genus must be capitalized. If
+ genus is not present use the taxonomic family name followed by the word 'indet'.
+ format: verbatim transcription
+ null_value: ''
+ habitat:
+ description: Description of a plant's habitat or the location where the specimen
+ was collected. Ignore descriptions of the plant itself.
+ format: verbatim transcription
+ null_value: ''
+ locality_name:
+ description: Description of geographic location, landscape, landmarks, regional
+ features, nearby places, or any contextual information aiding in pinpointing
+ the exact origin or site of the specimen.
+ format: verbatim transcription
+ null_value: ''
+ max_elevation:
+ description: Maximum elevation or altitude in meters. If only one elevation
+ is present, then max_elevation should be set to the null_value. Only if units
+ are explicit then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m'
+ or 'm.' or 'meters'). Round to integer.
+ format: integer
+ null_value: ''
+ min_elevation:
+ description: Minimum elevation or altitude in meters. Only if units are explicit
+ then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m' or 'm.' or
+ 'meters'). Round to integer.
+ format: integer
+ null_value: ''
+ multiple_names:
+ description: Indicate whether multiple people or collector names are present
+ in the unformatted OCR text. If you see more than one person's name the value
+ is 'yes'; otherwise the value is 'no'.
+ format: boolean yes no
+ null_value: ''
+ plant_description:
+ description: Description of plant features such as leaf shape, size, color,
+ stem texture, height, flower structure, scent, fruit or seed characteristics,
+ root system type, overall growth habit and form, any notable aroma or secretions,
+ presence of hairs or bristles, and any other distinguishing morphological
+ or physiological characteristics.
+ format: verbatim transcription
+ null_value: ''
+ species:
+ description: Taxonomic determination to species, do not capitalize species.
+ format: verbatim transcription
+ null_value: ''
+ state:
+ description: Administrative division 1 that corresponds to the current geographic
+ location of collection. Capitalize first letter of each word. Administrative
+ division 1 is equivalent to a U.S. State.
+ format: spell check transcription
+ null_value: ''
+ subspecies:
+ description: Taxonomic determination to subspecies (subsp.).
+ format: verbatim transcription
+ null_value: ''
+ variety:
+ description: Taxonomic determination to variety (var).
+ format: verbatim transcription
+ null_value: ''
+ verbatim_coordinates:
+ description: Verbatim location coordinates as they appear on the label. Do not
+ convert formats. Possible coordinate types are one of [Lat, Long, UTM, TRS].
+ format: verbatim transcription
+ null_value: ''
+ verbatim_date:
+ description: Date of collection exactly as it appears on the label. Do not change
+ the format or correct typos.
+ format: verbatim transcription
+ null_value: s.d.
+ SpeciesName:
+ taxonomy:
+ - Genus_species
diff --git a/custom_prompts/version_2_OSU.yaml b/custom_prompts/version_2_OSU.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..a87999e2e6f739a4f5443e02130cf8333f0ccce2
--- /dev/null
+++ b/custom_prompts/version_2_OSU.yaml
@@ -0,0 +1,230 @@
+LLM: gpt
+instructions: '1. Refactor the unstructured OCR text into a dictionary based on the
+ JSON structure outlined below.
+
+ 2. You should map the unstructured OCR text to the appropriate JSON key and then
+ populate the field based on its rules.
+
+ 3. Some JSON key fields are permitted to remain empty if the corresponding information
+ is not found in the unstructured OCR text.
+
+ 4. Ignore any information in the OCR text that doesn''t fit into the defined JSON
+ structure.
+
+ 5. Duplicate dictionary fields are not allowed.
+
+ 6. Ensure that all JSON keys are in lowercase.
+
+ 7. Ensure that new JSON field values follow sentence case capitalization.
+
+ 8. Ensure all key-value pairs in the JSON dictionary strictly adhere to the format
+ and data types specified in the template.
+
+ 9. Ensure the output JSON string is valid JSON format. It should not have trailing
+ commas or unquoted keys.
+
+ 10. Only return a JSON dictionary represented as a string. You should not explain
+ your answer.'
+json_formatting_instructions: "The next section of instructions outlines how to format\
+ \ the JSON dictionary. The keys are the same as those of the final formatted JSON\
+ \ object.\nFor each key there is a format requirement that specifies how to transcribe\
+ \ the information for that key. \nThe possible formatting options are:\n1. \"verbatim\
+ \ transcription\" - field is populated with verbatim text from the unformatted OCR.\n\
+ 2. \"spell check transcription\" - field is populated with spelling corrected text\
+ \ from the unformatted OCR.\n3. \"boolean yes no\" - field is populated with only\
+ \ yes or no.\n4. \"boolean 1 0\" - field is populated with only 1 or 0.\n5. \"integer\"\
+ \ - field is populated with only an integer.\n6. \"[list]\" - field is populated\
+ \ from one of the values in the list.\n7. \"yyyy-mm-dd\" - field is populated with\
+ \ a date in the format year-month-day.\nThe desired null value is also given. Populate\
+ \ the field with the null value of the information for that key is not present in\
+ \ the unformatted OCR text."
+mapping:
+ COLLECTING:
+ - collectors
+ - collector_number
+ - determined_by
+ - multiple_names
+ - verbatim_date
+ - date
+ - end_date
+ GEOGRAPHY:
+ - country
+ - state
+ - county
+ - min_elevation
+ - max_elevation
+ - elevation_units
+ LOCALITY:
+ - locality_name
+ - verbatim_coordinates
+ - decimal_coordinates
+ - datum
+ - plant_description
+ - cultivated
+ - habitat
+ MISCELLANEOUS: []
+ TAXONOMY:
+ - catalog_number
+ - genus
+ - species
+ - subspecies
+ - variety
+ - forma
+rules:
+ Dictionary:
+ catalog_number:
+ description: The barcode identifier, typically a number with at least 6 digits,
+ but fewer than 30 digits.
+ format: verbatim transcription
+ null_value: ''
+ collector_number:
+ description: Unique identifier or number that denotes the specific collecting
+ event and associated with the collector.
+ format: verbatim transcription
+ null_value: s.n.
+ collectors:
+ description: Full name(s) of the individual(s) responsible for collecting the
+ specimen. When multiple collectors are involved, their names should be separated
+ by commas.
+ format: verbatim transcription
+ null_value: not present
+ country:
+ description: Country that corresponds to the current geographic location of
+ collection. Capitalize first letter of each word. If abbreviation is given
+ populate field with the full spelling of the country's name.
+ format: spell check transcription
+ null_value: ''
+ county:
+ description: Administrative division 2 that corresponds to the current geographic
+ location of collection; capitalize first letter of each word. Administrative
+ division 2 is equivalent to a U.S. county, parish, borough.
+ format: spell check transcription
+ null_value: ''
+ cultivated:
+ description: Cultivated plants are intentionally grown by humans. In text descriptions,
+ look for planting dates, garden locations, ornamental, cultivar names, garden,
+ or farm to indicate cultivated plant. The value 1 indicates that the specimen
+ was cultivated, the value zero otherwise.
+ format: boolean 1 0
+ null_value: '0'
+ date:
+ description: 'Date the specimen was collected formatted as year-month-day. If
+ specific components of the date are unknown, they should be replaced with
+ zeros. Examples: ''0000-00-00'' if the entire date is unknown, ''YYYY-00-00''
+ if only the year is known, and ''YYYY-MM-00'' if year and month are known
+ but day is not.'
+ format: yyyy-mm-dd
+ null_value: ''
+ datum:
+ description: Datum of location coordinates. Possible values are include in the
+ format list. Leave field blank if unclear. [WGS84, WGS72, WGS66, WGS60, NAD83,
+ NAD27, OSGB36, ETRS89, ED50, GDA94, JGD2011, Tokyo97, KGD2002, TWD67, TWD97,
+ BJS54, XAS80, GCJ-02, BD-09, PZ-90.11, GTRF, CGCS2000, ITRF88, ITRF89, ITRF90,
+ ITRF91, ITRF92, ITRF93, ITRF94, ITRF96, ITRF97, ITRF2000, ITRF2005, ITRF2008,
+ ITRF2014, Hong Kong Principal Datum, SAD69]
+ format: '[list]'
+ null_value: ''
+ decimal_coordinates:
+ description: Correct and convert the verbatim location coordinates to conform
+ with the decimal degrees GPS coordinate format.
+ format: spell check transcription
+ null_value: ''
+ determined_by:
+ description: Full name of the individual responsible for determining the taxanomic
+ name of the specimen. Sometimes the name will be near to the characters 'det'
+ to denote determination. This name may be isolated from other names in the
+ unformatted OCR text.
+ format: verbatim transcription
+ null_value: ''
+ elevation_units:
+ description: 'Elevation units must be meters. If min_elevation field is populated,
+ then elevation_units: ''m''. Otherwise elevation_units: ''''.'
+ format: spell check transcription
+ null_value: ''
+ end_date:
+ description: 'If a date range is provided, this represents the later or ending
+ date of the collection period, formatted as year-month-day. If specific components
+ of the date are unknown, they should be replaced with zeros. Examples: ''0000-00-00''
+ if the entire end date is unknown, ''YYYY-00-00'' if only the year of the
+ end date is known, and ''YYYY-MM-00'' if year and month of the end date are
+ known but the day is not.'
+ format: yyyy-mm-dd
+ null_value: ''
+ forma:
+ description: Taxonomic determination to form (f.).
+ format: verbatim transcription
+ null_value: ''
+ genus:
+ description: Taxonomic determination to genus. Genus must be capitalized. If
+ genus is not present use the taxonomic family name followed by the word 'indet'.
+ format: verbatim transcription
+ null_value: ''
+ habitat:
+ description: Description of a plant's habitat or the location where the specimen
+ was collected. Ignore descriptions of the plant itself.
+ format: verbatim transcription
+ null_value: ''
+ locality_name:
+ description: Description of geographic location, landscape, landmarks, regional
+ features, nearby places, or any contextual information aiding in pinpointing
+ the exact origin or site of the specimen.
+ format: verbatim transcription
+ null_value: ''
+ max_elevation:
+ description: Maximum elevation or altitude in meters. If only one elevation
+ is present, then max_elevation should be set to the null_value. Only if units
+ are explicit then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m'
+ or 'm.' or 'meters'). Round to integer.
+ format: integer
+ null_value: ''
+ min_elevation:
+ description: Minimum elevation or altitude in meters. Only if units are explicit
+ then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m' or 'm.' or
+ 'meters'). Round to integer.
+ format: integer
+ null_value: ''
+ multiple_names:
+ description: Indicate whether multiple people or collector names are present
+ in the unformatted OCR text. If you see more than one person's name the value
+ is 'yes'; otherwise the value is 'no'.
+ format: boolean yes no
+ null_value: ''
+ plant_description:
+ description: Description of plant features such as leaf shape, size, color,
+ stem texture, height, flower structure, scent, fruit or seed characteristics,
+ root system type, overall growth habit and form, any notable aroma or secretions,
+ presence of hairs or bristles, and any other distinguishing morphological
+ or physiological characteristics.
+ format: verbatim transcription
+ null_value: ''
+ species:
+ description: Taxonomic determination to species, do not capitalize species.
+ format: verbatim transcription
+ null_value: ''
+ state:
+ description: Administrative division 1 that corresponds to the current geographic
+ location of collection. Capitalize first letter of each word. Administrative
+ division 1 is equivalent to a U.S. State.
+ format: spell check transcription
+ null_value: ''
+ subspecies:
+ description: Taxonomic determination to subspecies (subsp.).
+ format: verbatim transcription
+ null_value: ''
+ variety:
+ description: Taxonomic determination to variety (var).
+ format: verbatim transcription
+ null_value: ''
+ verbatim_coordinates:
+ description: Verbatim location coordinates as they appear on the label. Do not
+ convert formats. Possible coordinate types are one of [Lat, Long, UTM, TRS].
+ format: verbatim transcription
+ null_value: ''
+ verbatim_date:
+ description: Date of collection exactly as it appears on the label. Do not change
+ the format or correct typos.
+ format: verbatim transcription
+ null_value: s.d.
+ SpeciesName:
+ taxonomy:
+ - Genus_species
diff --git a/demo/ba/ba.jpg b/demo/ba/ba.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..d7aca03f0ac413cd71c027318faa771236827ef3
--- /dev/null
+++ b/demo/ba/ba.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:852bd2d2a5c142c4aed2ccaf0e59b00b8a9a31ebfb530b2fec7b8ae741296e20
+size 780696
diff --git a/demo/ba/ba.png b/demo/ba/ba.png
new file mode 100644
index 0000000000000000000000000000000000000000..1fc6e4f7ec5fa6784c1b656070d5a1d8c2d1259e
--- /dev/null
+++ b/demo/ba/ba.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c982940062f2d4ceaa387cc58055045ea2bef69688a59262395d48943bc0b1f1
+size 5018161
diff --git a/demo/ba/ba2.png b/demo/ba/ba2.png
new file mode 100644
index 0000000000000000000000000000000000000000..fda5f26be23f0c4842600bdd21a7406c9248133a
--- /dev/null
+++ b/demo/ba/ba2.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:917ca82feb01334499b95113f9bf6e4dfb0ca8f68d33c886ca2879039dd489be
+size 4559668
diff --git a/demo/demo_gallery/NY_1928185102_Heliotropiaceae_Heliotropium_indicum.jpg b/demo/demo_gallery/NY_1928185102_Heliotropiaceae_Heliotropium_indicum.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..94ad422cb7354b3ad80c63a9b3f1df2c3d620892
--- /dev/null
+++ b/demo/demo_gallery/NY_1928185102_Heliotropiaceae_Heliotropium_indicum.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7553e2eaa8220d6cc4dd318213b4b4971509d602727dd09e7994de4f9642657b
+size 2349689
diff --git a/demo/demo_gallery/SMF_3046042583_Ebenaceae_Diospyros_mespiliformis.jpg b/demo/demo_gallery/SMF_3046042583_Ebenaceae_Diospyros_mespiliformis.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..23ad275af6aef586f0cc674d274c1cf5ef522b7d
--- /dev/null
+++ b/demo/demo_gallery/SMF_3046042583_Ebenaceae_Diospyros_mespiliformis.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb6e37ac854e96af6d5ecdee3dc2a7f3542eab2a24119148f7a8ac7d4aba397f
+size 4651022
diff --git a/demo/demo_gallery/UM_1807475718_Monimiaceae_Hedycarya_parvifolia.jpg b/demo/demo_gallery/UM_1807475718_Monimiaceae_Hedycarya_parvifolia.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..f27819db19ad5de1824cb36c7a2dcbc96f0286a3
--- /dev/null
+++ b/demo/demo_gallery/UM_1807475718_Monimiaceae_Hedycarya_parvifolia.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dcd89c2fb7f0ed5c36584647e17a9aa1bf9e80631a658a9f63d9e7f51622ec2
+size 9889840
diff --git a/demo/demo_gallery/UM_1915455196_Cardiopteridaceae_Citronella_sarmentosa.jpg b/demo/demo_gallery/UM_1915455196_Cardiopteridaceae_Citronella_sarmentosa.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..f5a011bfc47931f2557157664f6cd064b76a42e2
--- /dev/null
+++ b/demo/demo_gallery/UM_1915455196_Cardiopteridaceae_Citronella_sarmentosa.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e8dee30b9c617b746eaf6d5f9c404243ab2a80355c6ab02154fbbe27bcc5396
+size 4455062
diff --git a/demo/demo_images/UM_1807464860_Phellinaceae_Phelline_dumbeensis.jpg b/demo/demo_images/UM_1807464860_Phellinaceae_Phelline_dumbeensis.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..e7547e5e7b19412fbeec87d1a3aaba184e35aee2
--- /dev/null
+++ b/demo/demo_images/UM_1807464860_Phellinaceae_Phelline_dumbeensis.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0974bb709c651b8913f9e5d85b00f021d5d4fee683e7b9593270101be8674ad1
+size 5608440
diff --git a/demo/img/expense_report.PNG b/demo/img/expense_report.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..1e034cddf99037dfcf7bb2a7b33834902a1c25b9
--- /dev/null
+++ b/demo/img/expense_report.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc72ffe78af66b4a311882cacbf072c8f9275f35b512c00c91bd5df62778da68
+size 160658
diff --git a/demo/img/prompt_1.PNG b/demo/img/prompt_1.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..fd2f93a7506b084294de4d0500c6c225f6753aad
--- /dev/null
+++ b/demo/img/prompt_1.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee38593d57852fc310809d415a2edd1f419c207036b6d86f9fbc9309601a91db
+size 78124
diff --git a/demo/img/prompt_2.PNG b/demo/img/prompt_2.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..cb8cca9356250031137e05da63228a91e8567053
--- /dev/null
+++ b/demo/img/prompt_2.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e48061b5a244d1a5926107947f3d62c803133f848e149952e2a6335772cef91
+size 167438
diff --git a/demo/img/prompt_3.PNG b/demo/img/prompt_3.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..f121484ec60414cb033625927b04cfbda56f98b2
--- /dev/null
+++ b/demo/img/prompt_3.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:104e7d8676af9f5ed7a445d5ceac93bcb9d512b42928e1bb2abe07c6b1cb432b
+size 128105
diff --git a/demo/img/prompt_4.PNG b/demo/img/prompt_4.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..74a80c606c0b869d7daeb390f8a850160b6b9ca0
--- /dev/null
+++ b/demo/img/prompt_4.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5476d266829d37bb1601ac49299aa41e0e5bf6ec2cd3930e3fd2168b7022267
+size 277475
diff --git a/demo/img/prompt_5.PNG b/demo/img/prompt_5.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..a71a366afd4ba39b8886f7e5bc58b795223e8943
--- /dev/null
+++ b/demo/img/prompt_5.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20b5f3438c89ab352e2c3246443ccefe43e7b7e64884bdecd586b5684e958d4b
+size 101827
diff --git a/demo/img/validation_1.PNG b/demo/img/validation_1.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..0fbf9f1253c89e0e9ef1d060e5b7331738b472e8
--- /dev/null
+++ b/demo/img/validation_1.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44f29afac203b22a9422dc87e5d12e4627ea1b818bd91af48357c18bd6ca7e9a
+size 93871
diff --git a/demo/img/validation_gpt.PNG b/demo/img/validation_gpt.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..abc5b763215ff1e3c08c704c09c1e7a35ab33451
--- /dev/null
+++ b/demo/img/validation_gpt.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9e98ae5b98eb7a945ad15698508baf30327b7621f53fc62046f25a22f930a3e
+size 158816
diff --git a/demo/img/validation_gpu.PNG b/demo/img/validation_gpu.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..66345b71c138cebb352f6970db8323cde7fd7586
--- /dev/null
+++ b/demo/img/validation_gpu.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:337e2c1d5efbbc5e71b319d6d405b2aae72ae568a7176de93aa2a1dcf902af43
+size 60054
diff --git a/demo/img/validation_palm.PNG b/demo/img/validation_palm.PNG
new file mode 100644
index 0000000000000000000000000000000000000000..c47ea1973721fb5152675b33c631ff0a62cd5382
--- /dev/null
+++ b/demo/img/validation_palm.PNG
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef5c86ca77c84f2a0d399e37b41648297d1664a06329fe6d31293dfc23480d7d
+size 140458
diff --git a/domain_knowledge/SLTP_UM_AllAsiaMinimalInRegion.xlsx b/domain_knowledge/SLTP_UM_AllAsiaMinimalInRegion.xlsx
new file mode 100644
index 0000000000000000000000000000000000000000..cdee24c876d8f01de594411c54a9b46b1d3ca047
Binary files /dev/null and b/domain_knowledge/SLTP_UM_AllAsiaMinimalInRegion.xlsx differ
diff --git a/img/icon.ico b/img/icon.ico
new file mode 100644
index 0000000000000000000000000000000000000000..92e2ed05243acc6c85ba7c1e99c6117f3b255543
Binary files /dev/null and b/img/icon.ico differ
diff --git a/img/icon.jpg b/img/icon.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..b2996053e8a0d32a31371d1e6ef91c699e253285
--- /dev/null
+++ b/img/icon.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:609e0f274d52db4c0c7ac73fc9764a72a8e1c7d975bd4f98a8037db2b3c98a0c
+size 303336
diff --git a/img/icon2.ico b/img/icon2.ico
new file mode 100644
index 0000000000000000000000000000000000000000..841158445ce19faefb955b2daeeedd8283fd0c78
Binary files /dev/null and b/img/icon2.ico differ
diff --git a/img/logo.png b/img/logo.png
new file mode 100644
index 0000000000000000000000000000000000000000..4feb014c989398dc2c9681f2d0914b5d16676dac
--- /dev/null
+++ b/img/logo.png
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:760eb08b5a54ca93fcdb1b59f39f28275648855aa92a439e0f9c28dc71c70bfc
+size 169464
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..00fa4eae01e471152c09208f892e5d98d4206e5e
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,34 @@
+--extra-index-url https://download.pytorch.org/whl/cpu
+
+torch==2.0.1
+torchvision==0.15.2
+torchaudio==2.0.2
+wheel
+streamlit
+streamlit-extras
+plotly
+pyyaml
+Pillow
+pandas
+matplotlib
+matplotlib-inline
+tqdm
+openai
+langchain
+tiktoken
+openpyxl
+google-generativeai
+google-cloud-storage
+google-cloud-vision
+opencv-python
+chromadb
+chroma-migrate
+InstructorEmbedding
+transformers
+sentence-transformers
+seaborn
+dask
+psutil
+py-cpuinfo
+azureml-sdk
+azure-identity
\ No newline at end of file
diff --git a/run_VoucherVision.py b/run_VoucherVision.py
new file mode 100644
index 0000000000000000000000000000000000000000..0370ad63fb4df307d940b60c77eb2c221eabdb7f
--- /dev/null
+++ b/run_VoucherVision.py
@@ -0,0 +1,31 @@
+import streamlit.web.cli as stcli
+import os, sys
+
+# Insert a file uploader that accepts multiple files at a time:
+# import streamlit as st
+# uploaded_files = st.file_uploader("Choose a CSV file", accept_multiple_files=True)
+# for uploaded_file in uploaded_files:
+# bytes_data = uploaded_file.read()
+# st.write("filename:", uploaded_file.name)
+# st.write(bytes_data)
+
+
+def resolve_path(path):
+ resolved_path = os.path.abspath(os.path.join(os.getcwd(), path))
+ return resolved_path
+
+
+if __name__ == "__main__":
+ dir_home = os.path.dirname(__file__)
+
+ # pip install protobuf==3.20.0
+
+ sys.argv = [
+ "streamlit",
+ "run",
+ resolve_path(os.path.join(dir_home,"vouchervision", "VoucherVision_GUI.py")),
+ "--global.developmentMode=false",
+ "--server.port=8525",
+
+ ]
+ sys.exit(stcli.main())
\ No newline at end of file
diff --git a/vouchervision/LLM_Falcon.py b/vouchervision/LLM_Falcon.py
new file mode 100644
index 0000000000000000000000000000000000000000..2aea8be35ae613d7bcb759a2e7941e084f9c793a
--- /dev/null
+++ b/vouchervision/LLM_Falcon.py
@@ -0,0 +1,112 @@
+import os, sys, inspect, json, time
+
+# currentdir = os.path.dirname(os.path.abspath(
+# inspect.getfile(inspect.currentframe())))
+# parentdir = os.path.dirname(currentdir)
+# sys.path.append(parentdir)
+
+# from prompts import PROMPT_PaLM_UMICH_skeleton_all_asia, PROMPT_PaLM_OCR_Organized, PROMPT_PaLM_Redo
+# from LLM_PaLM import create_OCR_analog_for_input, num_tokens_from_string
+
+'''
+https://docs.ai21.com/docs/python-sdk-with-amazon-bedrock
+
+
+https://techcommunity.microsoft.com/t5/ai-machine-learning-blog/falcon-llms-in-azure-machine-learning/ba-p/3876847
+https://github.com/Azure/azureml-examples/blob/main/sdk/python/foundation-models/huggingface/inference/text-generation-streaming/text-generation-streaming-online-endpoint.ipynb
+https://ml.azure.com/registries/HuggingFace/models/tiiuae-falcon-40b-instruct/version/12?tid=e66e77b4-5724-44d7-8721-06df160450ce#overview
+https://azure.microsoft.com/en-us/products/machine-learning/
+'''
+
+
+
+# from azure.ai.ml import MLClient
+# from azure.identity import (
+# DefaultAzureCredential,
+# InteractiveBrowserCredential,
+# ClientSecretCredential,
+# )
+# from azure.ai.ml.entities import AmlCompute
+
+# try:
+# credential = DefaultAzureCredential()
+# credential.get_token("https://management.azure.com/.default")
+# except Exception as ex:
+# credential = InteractiveBrowserCredential()
+
+# # connect to a workspace
+# workspace_ml_client = None
+# try:
+# workspace_ml_client = MLClient.from_config(credential)
+# subscription_id = workspace_ml_client.subscription_id
+# workspace = workspace_ml_client.workspace_name
+# resource_group = workspace_ml_client.resource_group_name
+# except Exception as ex:
+# print(ex)
+# # Enter details of your workspace
+# subscription_id = ""
+# resource_group = ""
+# workspace = ""
+# workspace_ml_client = MLClient(
+# credential, subscription_id, resource_group, workspace
+# )
+# # Connect to the HuggingFaceHub registry
+# registry_ml_client = MLClient(credential, registry_name="HuggingFace")
+# print(registry_ml_client)
+
+'''
+def OCR_to_dict_Falcon(logger, OCR, VVE):
+ # Find a similar example from the domain knowledge
+ domain_knowledge_example = VVE.query_db(OCR, 4)
+ similarity = VVE.get_similarity()
+ domain_knowledge_example_string = json.dumps(domain_knowledge_example)
+
+ try:
+ logger.info(f'Length of OCR raw -- {len(OCR)}')
+ except:
+ print(f'Length of OCR raw -- {len(OCR)}')
+
+ # Create input: output: for Falcon
+ # Assuming Falcon requires a similar structure as PaLM
+ in_list, out_list = create_OCR_analog_for_input(domain_knowledge_example)
+
+ # Construct the prompt for Falcon
+ # Adjust this based on Falcon's requirements
+ # prompt = PROMPT_Falcon_skeleton(OCR, in_list, out_list)
+ prompt = PROMPT_PaLM_UMICH_skeleton_all_asia(OCR, in_list, out_list) # must provide examples to PaLM differently than for chatGPT, at least 2 examples
+
+
+ nt = num_tokens_from_string(prompt, "falcon_model_name") # Replace "falcon_model_name" with the appropriate model name for Falcon
+ try:
+ logger.info(f'Prompt token length --- {nt}')
+ except:
+ print(f'Prompt token length --- {nt}')
+
+ # Assuming Falcon has a similar API structure as PaLM
+ # Adjust the settings based on Falcon's requirements
+ Falcon_settings = {
+ 'model': 'models/falcon_model_name', # Replace with the appropriate model name for Falcon
+ 'temperature': 0,
+ 'candidate_count': 1,
+ 'top_k': 40,
+ 'top_p': 0.95,
+ 'max_output_tokens': 8000,
+ 'stop_sequences': [],
+ # Add any other required settings for Falcon
+ }
+
+ # Send the prompt to Falcon for inference
+ # Adjust the API call based on Falcon's requirements
+ response = falcon.generate_text(**Falcon_settings, prompt=prompt)
+
+ # Process the response from Falcon
+ if response and response.result:
+ if isinstance(response.result, (str, bytes)):
+ response_valid = check_and_redo_JSON(response, Falcon_settings, logger)
+ else:
+ response_valid = {}
+ else:
+ response_valid = {}
+
+ return response_valid
+'''
\ No newline at end of file
diff --git a/vouchervision/LLM_PaLM.py b/vouchervision/LLM_PaLM.py
new file mode 100644
index 0000000000000000000000000000000000000000..c2e32fcdb7feb461b080ac7d24abfe482b2c57aa
--- /dev/null
+++ b/vouchervision/LLM_PaLM.py
@@ -0,0 +1,209 @@
+import os
+import sys
+import inspect
+import json
+from json import JSONDecodeError
+import tiktoken
+import random
+import google.generativeai as palm
+
+currentdir = os.path.dirname(os.path.abspath(
+ inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+
+from prompt_catalog import PromptCatalog
+from general_utils import num_tokens_from_string
+
+"""
+DEPRECATED:
+ Safety setting regularly block a response, so set to 4 to disable
+
+ class HarmBlockThreshold(Enum):
+ HARM_BLOCK_THRESHOLD_UNSPECIFIED = 0
+ BLOCK_LOW_AND_ABOVE = 1
+ BLOCK_MEDIUM_AND_ABOVE = 2
+ BLOCK_ONLY_HIGH = 3
+ BLOCK_NONE = 4
+"""
+
+SAFETY_SETTINGS = [
+ {
+ "category": "HARM_CATEGORY_DEROGATORY",
+ "threshold": "BLOCK_NONE",
+ },
+ {
+ "category": "HARM_CATEGORY_TOXICITY",
+ "threshold": "BLOCK_NONE",
+ },
+ {
+ "category": "HARM_CATEGORY_VIOLENCE",
+ "threshold": "BLOCK_NONE",
+ },
+ {
+ "category": "HARM_CATEGORY_SEXUAL",
+ "threshold": "BLOCK_NONE",
+ },
+ {
+ "category": "HARM_CATEGORY_MEDICAL",
+ "threshold": "BLOCK_NONE",
+ },
+ {
+ "category": "HARM_CATEGORY_DANGEROUS",
+ "threshold": "BLOCK_NONE",
+ },
+]
+
+PALM_SETTINGS = {
+ 'model': 'models/text-bison-001',
+ 'temperature': 0,
+ 'candidate_count': 1,
+ 'top_k': 40,
+ 'top_p': 0.95,
+ 'max_output_tokens': 8000,
+ 'stop_sequences': [],
+ 'safety_settings': SAFETY_SETTINGS,
+}
+
+PALM_SETTINGS_REDO = {
+ 'model': 'models/text-bison-001',
+ 'temperature': 0.05,
+ 'candidate_count': 1,
+ 'top_k': 40,
+ 'top_p': 0.95,
+ 'max_output_tokens': 8000,
+ 'stop_sequences': [],
+ 'safety_settings': SAFETY_SETTINGS,
+}
+
+def OCR_to_dict_PaLM(logger, OCR, prompt_version, VVE):
+ try:
+ logger.info(f'Length of OCR raw -- {len(OCR)}')
+ except:
+ print(f'Length of OCR raw -- {len(OCR)}')
+
+ # prompt = PROMPT_PaLM_UMICH_skeleton_all_asia(OCR, in_list, out_list) # must provide examples to PaLM differently than for chatGPT, at least 2 examples
+ Prompt = PromptCatalog(OCR)
+ if prompt_version in ['prompt_v2_palm2']:
+ version = 'v2'
+ prompt = Prompt.prompt_v2_palm2(OCR)
+
+ elif prompt_version in ['prompt_v1_palm2',]:
+ version = 'v1'
+ # create input: output: for PaLM
+ # Find a similar example from the domain knowledge
+ domain_knowledge_example = VVE.query_db(OCR, 4)
+ similarity= VVE.get_similarity()
+ domain_knowledge_example_string = json.dumps(domain_knowledge_example)
+ in_list, out_list = create_OCR_analog_for_input(domain_knowledge_example)
+ prompt = Prompt.prompt_v1_palm2(in_list, out_list, OCR)
+
+ elif prompt_version in ['prompt_v1_palm2_noDomainKnowledge',]:
+ version = 'v1'
+ prompt = Prompt.prompt_v1_palm2_noDomainKnowledge(OCR)
+ else:
+ version = 'custom'
+ prompt, n_fields, xlsx_headers = Prompt.prompt_v2_custom(prompt_version, OCR=OCR, is_palm=True)
+ # raise
+
+ nt = num_tokens_from_string(prompt, "cl100k_base")
+ # try:
+ logger.info(f'Prompt token length --- {nt}')
+ # except:
+ # print(f'Prompt token length --- {nt}')
+
+ do_use_SOP = False ########
+
+ if do_use_SOP:
+ '''TODO: Check back later to see if LangChain will support PaLM'''
+ # logger.info(f'Waiting for PaLM API call --- Using StructuredOutputParser')
+ # response = structured_output_parser(OCR, prompt, logger)
+ # return response['Dictionary']
+ pass
+
+ else:
+ # try:
+ logger.info(f'Waiting for PaLM 2 API call')
+ # except:
+ # print(f'Waiting for PaLM 2 API call --- Content')
+
+ # safety_thresh = 4
+ # PaLM_settings = {'model': 'models/text-bison-001','temperature': 0,'candidate_count': 1,'top_k': 40,'top_p': 0.95,'max_output_tokens': 8000,'stop_sequences': [],
+ # 'safety_settings': [{"category":"HARM_CATEGORY_DEROGATORY","threshold":safety_thresh},{"category":"HARM_CATEGORY_TOXICITY","threshold":safety_thresh},{"category":"HARM_CATEGORY_VIOLENCE","threshold":safety_thresh},{"category":"HARM_CATEGORY_SEXUAL","threshold":safety_thresh},{"category":"HARM_CATEGORY_MEDICAL","threshold":safety_thresh},{"category":"HARM_CATEGORY_DANGEROUS","threshold":safety_thresh}],}
+ response = palm.generate_text(prompt=prompt, **PALM_SETTINGS)
+
+
+ if response and response.result:
+ if isinstance(response.result, (str, bytes)):
+ response_valid = check_and_redo_JSON(response, logger, version)
+ else:
+ response_valid = {}
+ else:
+ response_valid = {}
+
+ logger.info(f'Candidate JSON\n{response.result}')
+ return response_valid, nt
+
+def check_and_redo_JSON(response, logger, version):
+ try:
+ response_valid = json.loads(response.result)
+ logger.info(f'Response --- First call passed')
+ return response_valid
+ except JSONDecodeError:
+
+ try:
+ response_valid = json.loads(response.result.strip('```').replace('json\n', '', 1).replace('json', '', 1))
+ logger.info(f'Response --- Manual removal of ```json succeeded')
+ return response_valid
+ except:
+ logger.info(f'Response --- First call failed. Redo...')
+ Prompt = PromptCatalog()
+ if version == 'v1':
+ prompt_redo = Prompt.prompt_palm_redo_v1(response.result)
+ elif version == 'v2':
+ prompt_redo = Prompt.prompt_palm_redo_v2(response.result)
+ elif version == 'custom':
+ prompt_redo = Prompt.prompt_v2_custom_redo(response.result, is_palm=True)
+
+
+ # prompt_redo = PROMPT_PaLM_Redo(response.result)
+ try:
+ response = palm.generate_text(prompt=prompt_redo, **PALM_SETTINGS)
+ response_valid = json.loads(response.result)
+ logger.info(f'Response --- Second call passed')
+ return response_valid
+ except JSONDecodeError:
+ logger.info(f'Response --- Second call failed. Final redo. Temperature changed to 0.05')
+ try:
+ response = palm.generate_text(prompt=prompt_redo, **PALM_SETTINGS_REDO)
+ response_valid = json.loads(response.result)
+ logger.info(f'Response --- Third call passed')
+ return response_valid
+ except JSONDecodeError:
+ return None
+
+
+def create_OCR_analog_for_input(domain_knowledge_example):
+ in_list = []
+ out_list = []
+ # Iterate over the domain_knowledge_example (list of dictionaries)
+ for row_dict in domain_knowledge_example:
+ # Convert the dictionary to a JSON string and add it to the out_list
+ domain_knowledge_example_string = json.dumps(row_dict)
+ out_list.append(domain_knowledge_example_string)
+
+ # Create a single string from all values in the row_dict
+ row_text = '||'.join(str(v) for v in row_dict.values())
+
+ # Split the row text by '||', shuffle the parts, and then re-join with a single space
+ parts = row_text.split('||')
+ random.shuffle(parts)
+ shuffled_text = ' '.join(parts)
+
+ # Add the shuffled_text to the in_list
+ in_list.append(shuffled_text)
+ return in_list, out_list
+
+
+def strip_problematic_chars(s):
+ return ''.join(c for c in s if c.isprintable())
diff --git a/vouchervision/LLM_chatGPT_3_5.py b/vouchervision/LLM_chatGPT_3_5.py
new file mode 100644
index 0000000000000000000000000000000000000000..4a1402e465c6099055dad65daccaac8dc81195f7
--- /dev/null
+++ b/vouchervision/LLM_chatGPT_3_5.py
@@ -0,0 +1,420 @@
+import openai
+import os, json, sys, inspect, time, requests
+from langchain.output_parsers import StructuredOutputParser, ResponseSchema
+from langchain.prompts import PromptTemplate, ChatPromptTemplate, HumanMessagePromptTemplate
+from langchain.llms import OpenAI
+from langchain.chat_models import ChatOpenAI, AzureChatOpenAI
+from langchain.schema import HumanMessage
+from general_utils import num_tokens_from_string
+
+currentdir = os.path.dirname(os.path.abspath(
+ inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+
+from prompts import PROMPT_UMICH_skeleton_all_asia, PROMPT_OCR_Organized, PROMPT_UMICH_skeleton_all_asia_GPT4, PROMPT_OCR_Organized_GPT4, PROMPT_JSON
+from prompt_catalog import PromptCatalog
+
+RETRY_DELAY = 61 # Wait 60 seconds before retrying
+MAX_RETRIES = 5 # Maximum number of retries
+
+
+def azure_call(model, messages):
+ response = model(messages=messages)
+ return response
+
+def OCR_to_dict(is_azure, logger, MODEL, prompt, llm, prompt_version):
+ for i in range(MAX_RETRIES):
+ try:
+ do_use_SOP = True
+
+ if do_use_SOP:
+ logger.info(f'Waiting for {MODEL} API call --- Using StructuredOutputParser')
+ response = structured_output_parser(is_azure, MODEL, llm, prompt, logger, prompt_version)
+ if response is None:
+ return None
+ else:
+ return response['Dictionary']
+
+ else:
+ ### Direct GPT ###
+ logger.info(f'Waiting for {MODEL} API call')
+ if not is_azure:
+ response = openai.ChatCompletion.create(
+ model=MODEL,
+ temperature = 0,
+ messages=[
+ {"role": "system", "content": "You are a helpful assistant acting as a transcription expert and your job is to transcribe herbarium specimen labels based on OCR data and reformat it to meet Darwin Core Archive Standards into a Python dictionary based on certain rules."},
+ {"role": "user", "content": prompt},
+ ],
+ max_tokens=4096,
+ )
+ # print the model's response
+ return response.choices[0].message['content']
+ else:
+ msg = HumanMessage(
+ content=prompt
+ )
+ response = azure_call(llm, [msg])
+ return response.content
+ except Exception as e:
+ logger.error(f'{e}')
+ if i < MAX_RETRIES - 1: # No delay needed after the last try
+ time.sleep(RETRY_DELAY)
+ else:
+ raise
+
+# def OCR_to_dict(logger, MODEL, prompt, OCR, BASE_URL, HEADERS):
+# for i in range(MAX_RETRIES):
+# try:
+# do_use_SOP = False
+
+# if do_use_SOP:
+# logger.info(f'Waiting for {MODEL} API call --- Using StructuredOutputParser -- Content')
+# response = structured_output_parser(MODEL, OCR, prompt, logger)
+# if response is None:
+# return None
+# else:
+# return response['Dictionary']
+
+# else:
+# ### Direct GPT through Azure ###
+# logger.info(f'Waiting for {MODEL} API call')
+# response = azure_gpt_request(prompt, BASE_URL, HEADERS, model_name=MODEL)
+
+# # Handle the response data. Note: You might need to adjust the following line based on the exact response format of the Azure API.
+# content = response.get("choices", [{}])[0].get("message", {}).get("content", "")
+# return content
+# except requests.exceptions.RequestException as e: # Replace openai.error.APIError with requests exception.
+# # Handle HTTP exceptions. You can adjust this based on the Azure API's error responses.
+# if e.response.status_code == 502:
+# logger.info(f' *** 502 error was encountered, wait and try again ***')
+# if i < MAX_RETRIES - 1:
+# time.sleep(RETRY_DELAY)
+# else:
+# raise
+
+
+def OCR_to_dict_16k(is_azure, logger, MODEL, prompt, llm, prompt_version):
+ for i in range(MAX_RETRIES):
+ try:
+ fs = FunctionSchema()
+ response = openai.ChatCompletion.create(
+ model=MODEL,
+ temperature = 0,
+ messages=[
+ {"role": "system", "content": "You are a helpful assistant acting as a transcription expert and your job is to transcribe herbarium specimen labels based on OCR data and reformat it to meet Darwin Core Archive Standards into a Python dictionary based on certain rules."},
+ {"role": "user", "content": prompt},
+ ],
+ max_tokens=8000,
+ function_call= "none",
+ functions= fs.format_C21_AA_V1()
+
+ )
+ # Try to parse the response into JSON
+ call_failed = False
+ try:
+ response_string = response.choices[0].message['content']
+ except:
+ call_failed = True
+ response_string = prompt
+
+ if not call_failed:
+ try:
+ # Try to parse the response into JSON
+ response_dict = json.loads(response_string)
+ return response_dict['Dictionary']
+ except json.JSONDecodeError:
+ # If the response is not a valid JSON, call the structured_output_parser_for_function_calls_fail function
+ logger.info(f'Invalid JSON response, calling structured_output_parser_for_function_calls_fail function')
+ logger.info(f'Waiting for {MODEL} API call --- Using StructuredOutputParser --- JSON Fixer')
+ response_sop = structured_output_parser_for_function_calls_fail(is_azure, MODEL, response_string, logger, llm, prompt_version, is_helper=False)
+ if response_sop is None:
+ return None
+ else:
+ return response_sop['Dictionary']
+ else:
+ try:
+ logger.info(f'Call Failed. Attempting fallback JSON parse without guidance')
+ logger.info(f'Waiting for {MODEL} API call --- Using StructuredOutputParser --- JSON Fixer')
+ response_sop = structured_output_parser_for_function_calls_fail(is_azure, MODEL, response_string, logger, llm, prompt_version, is_helper=False)
+ if response_sop is None:
+ return None
+ else:
+ return response_sop['Dictionary']
+ except:
+ return None
+ except Exception as e:
+ # if e.status_code == 401: # or you can check the error message
+ logger.info(f' *** 401 error was encountered, wait and try again ***')
+ # If a 401 error was encountered, wait and try again
+ if i < MAX_RETRIES - 1: # No delay needed after the last try
+ time.sleep(RETRY_DELAY)
+ else:
+ # If it was a different error, re-raise it
+ raise
+
+def structured_output_parser(is_azure, MODEL, llm, prompt_template, logger, prompt_version, is_helper=False):
+ if not is_helper:
+ response_schemas = [
+ ResponseSchema(name="SpeciesName", description="Taxonomic determination, genus_species"),
+ ResponseSchema(name="Dictionary", description='Formatted JSON object'),]#prompt_template),]
+ elif is_helper:
+ response_schemas = [
+ ResponseSchema(name="Dictionary", description='Formatted JSON object'),#prompt_template),
+ ResponseSchema(name="Summary", description="A one sentence summary of the content"),]
+
+ output_parser = StructuredOutputParser.from_response_schemas(response_schemas)
+
+ format_instructions = output_parser.get_format_instructions()
+
+ prompt = ChatPromptTemplate(
+ messages=[
+ HumanMessagePromptTemplate.from_template("Parse the OCR text into the correct structured format.\n{format_instructions}\n{question}")
+ ],
+ input_variables=["question"],
+ partial_variables={"format_instructions": format_instructions}
+ )
+
+ # Handle Azure vs OpenAI implementation
+ if is_azure:
+ _input = prompt.format_prompt(question=prompt_template)
+ msg = HumanMessage(content=_input.to_string())
+ output = azure_call(llm, [msg])
+ else:
+ chat_model = ChatOpenAI(temperature=0, model=MODEL)
+ _input = prompt.format_prompt(question=prompt_template)
+ output = chat_model(_input.to_messages())
+
+ # Log token length if running with Gradio
+ try:
+ nt = num_tokens_from_string(_input.to_string(), "cl100k_base")
+ logger.info(f'Prompt token length --- {nt}')
+ except:
+ pass
+
+ # Parse the output
+ try:
+ # Check if output is of type 'ai' and parse accordingly
+ if output.type == 'ai':
+ parsed_content = output.content
+ logger.info(f'Formatted JSON\n{parsed_content}')
+ else:
+ # If not 'ai', log and set parsed_content to None or a default value
+ logger.error('Output type is not "ai". Unable to parse.')
+ return None
+
+ # Clean up the parsed content
+ parsed_content = parsed_content.replace('\n', "").replace('\t', "").replace('|', "")
+
+ # Attempt to parse the cleaned content
+ try:
+ refined_response = output_parser.parse(parsed_content)
+ return refined_response
+ except Exception as parse_error:
+ # Handle parsing errors specifically
+ logger.error(f'Parsing Error: {parse_error}')
+ return structured_output_parser_for_function_calls_fail(is_azure, MODEL, parsed_content, logger, llm, prompt_version, is_helper)
+
+ except Exception as e:
+ # Handle any other exceptions that might occur
+ logger.error(f'Unexpected Error: {e}')
+ return None
+
+def structured_output_parser_for_function_calls_fail(is_azure, MODEL, failed_response, logger, llm, prompt_version, is_helper=False, try_ind=0):
+ if try_ind > 5:
+ return None
+
+ # prompt_redo = PROMPT_JSON('helper' if is_helper else 'dict', failed_response)
+ Prompt = PromptCatalog()
+ if prompt_version in ['prompt_v1_verbose', 'prompt_v1_verbose_noDomainKnowledge']:
+ prompt_redo = Prompt.prompt_gpt_redo_v1(failed_response)
+ elif prompt_version in ['prompt_v2_json_rules']:
+ prompt_redo = Prompt.prompt_gpt_redo_v2(failed_response)
+ else:
+ prompt_redo = Prompt.prompt_v2_custom_redo(failed_response, is_palm=False)
+
+ response_schemas = [
+ ResponseSchema(name="Summary", description="A one sentence summary of the content"),
+ ResponseSchema(name="Dictionary", description='Formatted JSON object')
+ ]
+
+ output_parser = StructuredOutputParser.from_response_schemas(response_schemas)
+ format_instructions = output_parser.get_format_instructions()
+
+ prompt = ChatPromptTemplate(
+ messages=[
+ HumanMessagePromptTemplate.from_template("The following text contains JSON formatted text, but there is an error that you need to correct.\n{format_instructions}\n{question}")
+ ],
+ input_variables=["question"],
+ partial_variables={"format_instructions": format_instructions}
+ )
+
+ _input = prompt.format_prompt(question=prompt_redo)
+
+ # Log token length if running with Gradio
+ try:
+ nt = num_tokens_from_string(_input.to_string(), "cl100k_base")
+ logger.info(f'Prompt Redo token length --- {nt}')
+ except:
+ pass
+
+ if is_azure:
+ msg = HumanMessage(content=_input.to_string())
+ output = azure_call(llm, [msg])
+ else:
+ chat_model = ChatOpenAI(temperature=0, model=MODEL)
+ output = chat_model(_input.to_messages())
+
+ try:
+ refined_response = output_parser.parse(output.content)
+ except json.decoder.JSONDecodeError as e:
+ try_ind += 1
+ error_message = str(e)
+ redo_content = f'The error messsage is: {error_message}\nThe broken JSON object is: {output.content}'
+ logger.info(f'[Failed JSON Object]\n{output.content}')
+ refined_response = structured_output_parser_for_function_calls_fail(is_azure, MODEL, redo_content, logger, llm, prompt_version, is_helper, try_ind)
+ except:
+ try_ind += 1
+ logger.info(f'[Failed JSON Object]\n{output.content}')
+ refined_response = structured_output_parser_for_function_calls_fail(is_azure, MODEL, output.content, logger, llm, prompt_version, is_helper, try_ind)
+
+ return refined_response
+
+
+
+
+class FunctionSchema:
+ def __init__(self):
+ pass
+
+ def format_C21_AA_V1(self):
+ return [
+ {
+ "name": "format_C21_AA_V1",
+ "description": "Format the given data into a specific dictionary",
+ "parameters": {
+ "type": "object",
+ "properties": {}, # specify parameters here if your function requires any
+ "required": [] # list of required parameters
+ },
+ "output_type": "json",
+ "output_schema": {
+ "type": "object",
+ "properties": {
+ "Dictionary": {
+ "type": "object",
+ "properties": {
+ "Catalog Number": {"type": "array", "items": {"type": "string"}},
+ "Genus": {"type": "array", "items": {"type": "string"}},
+ "Species": {"type": "array", "items": {"type": "string"}},
+ "subspecies": {"type": "array", "items": {"type": "string"}},
+ "variety": {"type": "array", "items": {"type": "string"}},
+ "forma": {"type": "array", "items": {"type": "string"}},
+ "Country": {"type": "array", "items": {"type": "string"}},
+ "State": {"type": "array", "items": {"type": "string"}},
+ "County": {"type": "array", "items": {"type": "string"}},
+ "Locality Name": {"type": "array", "items": {"type": "string"}},
+ "Min Elevation": {"type": "array", "items": {"type": "string"}},
+ "Max Elevation": {"type": "array", "items": {"type": "string"}},
+ "Elevation Units": {"type": "array", "items": {"type": "string"}},
+ "Verbatim Coordinates": {"type": "array", "items": {"type": "string"}},
+ "Datum": {"type": "array", "items": {"type": "string"}},
+ "Cultivated": {"type": "array", "items": {"type": "string"}},
+ "Habitat": {"type": "array", "items": {"type": "string"}},
+ "Collectors": {"type": "array", "items": {"type": "string"}},
+ "Collector Number": {"type": "array", "items": {"type": "string"}},
+ "Verbatim Date": {"type": "array", "items": {"type": "string"}},
+ "Date": {"type": "array", "items": {"type": "string"}},
+ "End Date": {"type": "array", "items": {"type": "string"}}
+ }
+ },
+ "SpeciesName": {
+ "type": "object",
+ "properties": {
+ "taxonomy": {"type": "array", "items": {"type": "string"}}
+ }
+ }
+ }
+ }
+ }
+ ]
+
+ def format_C21_AA_V1_helper(self):
+ return [
+ {
+ "name": "format_C21_AA_V1_helper",
+ "description": "Helper function for format_C21_AA_V1 to further format the given data",
+ "parameters": {
+ "type": "object",
+ "properties": {}, # specify parameters here if your function requires any
+ "required": [] # list of required parameters
+ },
+ "output_type": "json",
+ "output_schema": {
+ "type": "object",
+ "properties": {
+ "Dictionary": {
+ "type": "object",
+ "properties": {
+ "TAXONOMY": {
+ "type": "object",
+ "properties": {
+ "Order": {"type": "array", "items": {"type": "string"}},
+ "Family": {"type": "array", "items": {"type": "string"}},
+ "Genus":{"type": "array", "items": {"type": "string"}},
+ "Species": {"type": "array", "items": {"type": "string"}},
+ "Subspecies": {"type": "array", "items": {"type": "string"}},
+ "Variety": {"type": "array", "items": {"type": "string"}},
+ "Forma": {"type": "array", "items": {"type": "string"}},
+ }
+ },
+ "GEOGRAPHY": {
+ "type": "object",
+ "properties": {
+ "Country": {"type": "array", "items": {"type": "string"}},
+ "State": {"type": "array", "items": {"type": "string"}},
+ "Prefecture": {"type": "array", "items": {"type": "string"}},
+ "Province": {"type": "array", "items": {"type": "string"}},
+ "District": {"type": "array", "items": {"type": "string"}},
+ "County": {"type": "array", "items": {"type": "string"}},
+ "City": {"type": "array", "items": {"type": "string"}},
+ "Administrative Division": {"type": "array", "items": {"type": "string"}},
+ }
+ },
+ "LOCALITY": {
+ "type": "object",
+ "properties": {
+ "Landscape": {"type": "array", "items": {"type": "string"}},
+ "Nearby Places": {"type": "array", "items": {"type": "string"}},
+ }
+ },
+ "COLLECTING": {
+ "type": "object",
+ "properties": {
+ "Collector": {"type": "array", "items": {"type": "string"}},
+ "Collector's Number": {"type": "array", "items": {"type": "string"}},
+ "Verbatim Date": {"type": "array", "items": {"type": "string"}},
+ "Formatted Date": {"type": "array", "items": {"type": "string"}},
+ "Cultivation Status": {"type": "array", "items": {"type": "string"}},
+ "Habitat Description": {"type": "array", "items": {"type": "string"}},
+ }
+ },
+ "MISCELLANEOUS": {
+ "type": "object",
+ "properties": {
+ "Additional Information": {"type": "array", "items": {"type": "string"}},
+ }
+ }
+ }
+ },
+ "Summary": {
+ "type": "object",
+ "properties": {
+ "Content Summary": {"type": "array", "items": {"type": "string"}}
+ }
+ }
+ }
+ }
+ }
+ ]
diff --git a/vouchervision/LM2_logger.py b/vouchervision/LM2_logger.py
new file mode 100644
index 0000000000000000000000000000000000000000..ffb58e6c003df356ddeb6b010d2f4d42b27f8e34
--- /dev/null
+++ b/vouchervision/LM2_logger.py
@@ -0,0 +1,117 @@
+import logging, os, psutil, torch, platform, cpuinfo, yaml #py-cpuinfo
+from vouchervision.general_utils import get_datetime, print_main_warn, print_main_info
+
+def start_logging(Dirs, cfg):
+ run_name = cfg['leafmachine']['project']['run_name']
+ path_log = os.path.join(Dirs.path_log, '__'.join(['LM2-log',str(get_datetime()), run_name])+'.log')
+
+ # Disable default StreamHandler
+ logging.getLogger().handlers = []
+
+ # create logger
+ logger = logging.getLogger('Hardware Components')
+ logger.setLevel(logging.DEBUG)
+
+ # create file handler and set level to debug
+ fh = logging.FileHandler(path_log)
+ fh.setLevel(logging.DEBUG)
+
+ # create console handler and set level to debug
+ ch = logging.StreamHandler()
+ ch.setLevel(logging.DEBUG)
+
+ # create formatter
+ formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(name)s - %(message)s')
+
+ # add formatter to handlers
+ fh.setFormatter(formatter)
+ ch.setFormatter(formatter)
+
+ # add handlers to logger
+ logger.addHandler(fh)
+ logger.addHandler(ch)
+
+ # Create a logger for the file handler
+ file_logger = logging.getLogger('file_logger')
+ file_logger.setLevel(logging.DEBUG)
+ file_formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')
+ file_handler = logging.FileHandler(path_log)
+ file_handler.setLevel(logging.DEBUG)
+ file_handler.setFormatter(file_formatter)
+ file_logger.addHandler(file_handler)
+ # Disable propagation of log messages to the root logger
+ file_logger.propagate = False
+
+ # 'application' code
+ # logger.debug('debug message')
+ # logger.info('info message')
+ # logger.warning('warn message')
+ # logger.error('error message')
+ # logger.critical('critical message')
+
+ # Get CPU information
+ logger.info(f"CPU: {find_cpu_info()}")
+
+ # Get GPU information (using PyTorch)
+ if torch.cuda.is_available():
+ num_gpus = torch.cuda.device_count()
+ if num_gpus == 1:
+ gpu = torch.cuda.get_device_properties(0)
+ logger.info(f"GPU: {gpu.name} ({gpu.total_memory // (1024 * 1024)} MB)")
+ else:
+ for i in range(num_gpus):
+ gpu = torch.cuda.get_device_properties(i)
+ logger.info(f"GPU {i}: {gpu.name} ({gpu.total_memory // (1024 * 1024)} MB)")
+ else:
+ logger.info("No GPU found")
+ logger.info("LeafMachine2 image cropping and embedding search will be extremely slow or not possible.")
+ print_main_info("No GPU found!")
+ print_main_info("LeafMachine2 image cropping and embedding search will be extremely slow or not possible.")
+
+ # Get memory information
+ mem_info = psutil.virtual_memory()
+ logger.info(f"Memory: {mem_info.total // (1024 * 1024)} MB")
+ logger.info(LM2_banner())
+ logger.info(f"Config added to log file")
+ file_logger.info('Config:\n{}'.format(yaml.dump(cfg)))
+
+
+ return logger
+
+def find_cpu_info():
+ cpu_info = []
+ cpu_info.append(platform.processor())
+ try:
+
+ with open('/proc/cpuinfo') as f:
+ for line in f:
+ if line.startswith('model name'):
+ cpu_info.append(line.split(':')[1].strip())
+ break
+ return ' / '.join(cpu_info)
+ except:
+ try:
+ info = cpuinfo.get_cpu_info()
+ cpu_info = []
+ cpu_info.append(info['brand_raw'])
+ cpu_info.append(f"{info['hz_actual_friendly']}")
+ return ' / '.join(cpu_info)
+ except:
+ return "CPU: UNKNOWN"
+
+
+def LM2_banner():
+ logo = """
+ _ __ __ __ _ _ ___
+ | | / _| \/ | | | (_) |__ \
+ | | ___ __ _| |_| \ / | __ _ ___| |__ _ _ __ ___ ) |
+ | | / _ \/ _` | _| |\/| |/ _` |/ __| '_ \| | '_ \ / _ \ / /
+ | |___| __/ (_| | | | | | | (_| | (__| | | | | | | | __// /_
+ |______\___|\__,_|_| |_| |_|\__,_|\___|_| |_|_|_| |_|\___|____|
+ __ __ _ _| |_ __ ___ _
+ \ \ / / | | |_ _| \ \ / (_) (_)
+ \ \ / /__ _ _ ___| |__ |_|_ _ _\ \ / / _ ___ _ ___ _ __
+ \ \/ / _ \| | | |/ __| '_ \ / _ \ '__\ \/ / | / __| |/ _ \| '_ \
+ \ / (_) | |_| | (__| | | | __/ | \ / | \__ \ | (_) | | | |
+ \/ \___/ \__,_|\___|_| |_|\___|_| \/ |_|___/_|\___/|_| |_|"""
+ return logo
\ No newline at end of file
diff --git a/vouchervision/LeafMachine2_Config_Builder.py b/vouchervision/LeafMachine2_Config_Builder.py
new file mode 100644
index 0000000000000000000000000000000000000000..16873de5e627254efd12b6cd0b76365cf9dfb453
--- /dev/null
+++ b/vouchervision/LeafMachine2_Config_Builder.py
@@ -0,0 +1,246 @@
+import os, yaml, platform
+
+def get_default_download_folder():
+ system_platform = platform.system() # Gets the system platform, e.g., 'Linux', 'Windows', 'Darwin'
+
+ if system_platform == "Windows":
+ # Typically, the Downloads folder for Windows is in the user's profile folder
+ default_output_folder = os.path.join(os.getenv('USERPROFILE'), 'Downloads')
+ elif system_platform == "Darwin":
+ # Typically, the Downloads folder for macOS is in the user's home directory
+ default_output_folder = os.path.join(os.path.expanduser("~"), 'Downloads')
+ elif system_platform == "Linux":
+ # Typically, the Downloads folder for Linux is in the user's home directory
+ default_output_folder = os.path.join(os.path.expanduser("~"), 'Downloads')
+ else:
+ default_output_folder = "set/path/to/downloads/folder"
+ print("Please manually set the output folder")
+ return default_output_folder
+
+def build_LM2_config():
+ dir_home = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
+
+
+ # Initialize the base structure
+ config_data = {
+ 'leafmachine': {}
+ }
+
+ # Modular sections to be added to 'leafmachine'
+ do_section = {
+ 'check_for_illegal_filenames': True,
+ 'check_for_corrupt_images_make_vertical': True,
+ 'run_leaf_processing': True
+ }
+
+ print_section = {
+ 'verbose': True,
+ 'optional_warnings': True
+ }
+
+ logging_section = {
+ 'log_level': None
+ }
+
+ default_output_folder = get_default_download_folder()
+ project_section = {
+ 'dir_output': default_output_folder,
+ # 'dir_output': 'D:/D_Desktop/LM2',
+ 'run_name': 'test',
+ 'image_location': 'local',
+ 'GBIF_mode': 'all',
+ 'batch_size': 40,
+ 'num_workers': 2,
+ 'dir_images_local': '',
+ # 'dir_images_local': 'D:\Dropbox\LM2_Env\Image_Datasets\Manuscript_Images',
+ 'path_combined_csv_local': None,
+ 'path_occurrence_csv_local': None,
+ 'path_images_csv_local': None,
+ 'use_existing_plant_component_detections': None,
+ 'use_existing_archival_component_detections': None,
+ 'process_subset_of_images': False,
+ 'dir_images_subset': '',
+ 'n_images_per_species': 10,
+ 'species_list': ''
+ }
+
+ cropped_components_section = {
+ 'do_save_cropped_annotations': False,
+ 'save_cropped_annotations': ['label'],
+ 'save_per_image': False,
+ 'save_per_annotation_class': True,
+ 'binarize_labels': False,
+ 'binarize_labels_skeletonize': False
+ }
+
+ modules_section = {
+ 'armature': False,
+ 'specimen_crop': False
+ }
+
+ data_section = {
+ 'save_json_rulers': False,
+ 'save_json_measurements': False,
+ 'save_individual_csv_files_rulers': False,
+ 'save_individual_csv_files_measurements': False,
+ 'save_individual_csv_files_landmarks': False,
+ 'save_individual_efd_files': False,
+ 'include_darwin_core_data_from_combined_file': False,
+ 'do_apply_conversion_factor': True
+ }
+
+ overlay_section = {
+ 'save_overlay_to_pdf': False,
+ 'save_overlay_to_jpgs': True,
+ 'overlay_dpi': 300, # Between 100 to 300
+ 'overlay_background_color': 'black', # Either 'white' or 'black'
+
+ 'show_archival_detections': True,
+ 'show_plant_detections': True,
+ 'show_segmentations': True,
+ 'show_landmarks': True,
+ 'ignore_archival_detections_classes': [],
+ 'ignore_plant_detections_classes': ['leaf_whole', 'specimen'], # Could also include 'leaf_partial' and others if needed
+ 'ignore_landmark_classes': [],
+
+ 'line_width_archival': 12, # Previous value given was 2
+ 'line_width_plant': 12, # Previous value given was 6
+ 'line_width_seg': 12, # 12 is specified as "thick"
+ 'line_width_efd': 12, # 3 is specified as "thick" but 12 is given here
+ 'alpha_transparency_archival': 0.3,
+ 'alpha_transparency_plant': 0,
+ 'alpha_transparency_seg_whole_leaf': 0.4,
+ 'alpha_transparency_seg_partial_leaf': 0.3
+ }
+
+ plant_component_detector_section = {
+ 'detector_type': 'Plant_Detector',
+ 'detector_version': 'PLANT_GroupAB_200',
+ 'detector_iteration': 'PLANT_GroupAB_200',
+ 'detector_weights': 'best.pt',
+ 'minimum_confidence_threshold': 0.3, # Default is 0.5
+ 'do_save_prediction_overlay_images': True,
+ 'ignore_objects_for_overlay': [] # 'leaf_partial' can be included if needed
+ }
+
+ archival_component_detector_section = {
+ 'detector_type': 'Archival_Detector',
+ 'detector_version': 'PREP_final',
+ 'detector_iteration': 'PREP_final',
+ 'detector_weights': 'best.pt',
+ 'minimum_confidence_threshold': 0.5, # Default is 0.5
+ 'do_save_prediction_overlay_images': True,
+ 'ignore_objects_for_overlay': []
+ }
+
+ armature_component_detector_section = {
+ 'detector_type': 'Armature_Detector',
+ 'detector_version': 'ARM_A_1000',
+ 'detector_iteration': 'ARM_A_1000',
+ 'detector_weights': 'best.pt',
+ 'minimum_confidence_threshold': 0.5, # Optionally: 0.2
+ 'do_save_prediction_overlay_images': True,
+ 'ignore_objects_for_overlay': []
+ }
+
+ landmark_detector_section = {
+ 'landmark_whole_leaves': True,
+ 'landmark_partial_leaves': False,
+ 'detector_type': 'Landmark_Detector_YOLO',
+ 'detector_version': 'Landmarks',
+ 'detector_iteration': 'Landmarks_V2',
+ 'detector_weights': 'best.pt',
+ 'minimum_confidence_threshold': 0.02,
+ 'do_save_prediction_overlay_images': True,
+ 'ignore_objects_for_overlay': [],
+ 'use_existing_landmark_detections': None, # Example path provided
+ 'do_show_QC_images': False,
+ 'do_save_QC_images': True,
+ 'do_show_final_images': False,
+ 'do_save_final_images': True
+ }
+
+ landmark_detector_armature_section = {
+ 'upscale_factor': 10,
+ 'detector_type': 'Landmark_Detector_YOLO',
+ 'detector_version': 'Landmarks_Arm_A_200',
+ 'detector_iteration': 'Landmarks_Arm_A_200',
+ 'detector_weights': 'last.pt',
+ 'minimum_confidence_threshold': 0.06,
+ 'do_save_prediction_overlay_images': True,
+ 'ignore_objects_for_overlay': [],
+ 'use_existing_landmark_detections': None, # Example path provided
+ 'do_show_QC_images': True,
+ 'do_save_QC_images': True,
+ 'do_show_final_images': True,
+ 'do_save_final_images': True
+ }
+
+ ruler_detection_section = {
+ 'detect_ruler_type': True,
+ 'ruler_detector': 'ruler_classifier_38classes_v-1.pt',
+ 'ruler_binary_detector': 'model_scripted_resnet_720_withCompression.pt',
+ 'minimum_confidence_threshold': 0.4,
+ 'save_ruler_validation': False,
+ 'save_ruler_validation_summary': True,
+ 'save_ruler_processed': False
+ }
+
+ leaf_segmentation_section = {
+ 'segment_whole_leaves': True,
+ 'segment_partial_leaves': False,
+
+ 'keep_only_best_one_leaf_one_petiole': True,
+
+ 'save_segmentation_overlay_images_to_pdf': True,
+ 'save_each_segmentation_overlay_image': True,
+ 'save_individual_overlay_images': True, # Not recommended due to potential file count
+ 'overlay_line_width': 1, # Default is 1
+
+ 'use_efds_for_png_masks': False, # Requires calculate_elliptic_fourier_descriptors to be True
+ 'save_masks_color': True,
+ 'save_full_image_masks_color': True,
+ 'save_rgb_cropped_images': True,
+
+ 'find_minimum_bounding_box': True,
+
+ 'calculate_elliptic_fourier_descriptors': True, # Default is True
+ 'elliptic_fourier_descriptor_order': 40, # Default is 40
+
+ 'segmentation_model': 'GroupB_Dataset_100000_Iter_1176PTS_512Batch_smooth_l1_LR00025_BGR',
+ 'minimum_confidence_threshold': 0.7, # Alternatively: 0.9
+ 'generate_overlay': True,
+ 'overlay_dpi': 300, # Range: 100 to 300
+ 'overlay_background_color': 'black' # Options: 'white' or 'black'
+ }
+
+ # Add the sections to the 'leafmachine' key
+ config_data['leafmachine']['do'] = do_section
+ config_data['leafmachine']['print'] = print_section
+ config_data['leafmachine']['logging'] = logging_section
+ config_data['leafmachine']['project'] = project_section
+ config_data['leafmachine']['cropped_components'] = cropped_components_section
+ config_data['leafmachine']['modules'] = modules_section
+ config_data['leafmachine']['data'] = data_section
+ config_data['leafmachine']['overlay'] = overlay_section
+ config_data['leafmachine']['plant_component_detector'] = plant_component_detector_section
+ config_data['leafmachine']['archival_component_detector'] = archival_component_detector_section
+ config_data['leafmachine']['armature_component_detector'] = armature_component_detector_section
+ config_data['leafmachine']['landmark_detector'] = landmark_detector_section
+ config_data['leafmachine']['landmark_detector_armature'] = landmark_detector_armature_section
+ config_data['leafmachine']['ruler_detection'] = ruler_detection_section
+ config_data['leafmachine']['leaf_segmentation'] = leaf_segmentation_section
+
+ return config_data, dir_home
+
+def write_config_file(config_data, dir_home, filename="LeafMachine2.yaml"):
+ file_path = os.path.join(dir_home, filename)
+
+ # Write the data to a YAML file
+ with open(file_path, "w") as outfile:
+ yaml.dump(config_data, outfile, default_flow_style=False)
+
+if __name__ == '__main__':
+ config_data, dir_home = build_LM2_config()
+ write_config_file(config_data, dir_home)
+
diff --git a/vouchervision/OCR_google_cloud_vision.py b/vouchervision/OCR_google_cloud_vision.py
new file mode 100644
index 0000000000000000000000000000000000000000..2bb8b87bf47a53ff0cf19695bf319e04cdcdc62f
--- /dev/null
+++ b/vouchervision/OCR_google_cloud_vision.py
@@ -0,0 +1,107 @@
+import os, io, sys, inspect
+from google.cloud import vision, storage
+from PIL import Image, ImageDraw
+
+currentdir = os.path.dirname(os.path.abspath(
+ inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+
+def draw_boxes(image, bounds, color):
+ if bounds:
+ draw = ImageDraw.Draw(image)
+ width, height = image.size
+ line_width = int((width + height) / 2 * 0.001) # This sets the line width as 0.5% of the average dimension
+
+ for bound in bounds:
+ draw.polygon(
+ [
+ bound["vertices"][0]["x"], bound["vertices"][0]["y"],
+ bound["vertices"][1]["x"], bound["vertices"][1]["y"],
+ bound["vertices"][2]["x"], bound["vertices"][2]["y"],
+ bound["vertices"][3]["x"], bound["vertices"][3]["y"],
+ ],
+ outline=color,
+ width=line_width
+ )
+ return image
+
+def detect_text(path):
+ client = vision.ImageAnnotatorClient()
+ with io.open(path, 'rb') as image_file:
+ content = image_file.read()
+ image = vision.Image(content=content)
+ response = client.document_text_detection(image=image)
+ texts = response.text_annotations
+
+ if response.error.message:
+ raise Exception(
+ '{}\nFor more info on error messages, check: '
+ 'https://cloud.google.com/apis/design/errors'.format(
+ response.error.message))
+
+ # Extract bounding boxes
+ bounds = []
+ text_to_box_mapping = {}
+ for text in texts[1:]: # Skip the first entry, as it represents the entire detected text
+ # Convert BoundingPoly to dictionary
+ bound_dict = {
+ "vertices": [
+ {"x": vertex.x, "y": vertex.y} for vertex in text.bounding_poly.vertices
+ ]
+ }
+ bounds.append(bound_dict)
+ text_to_box_mapping[str(bound_dict)] = text.description
+
+ if texts:
+ # cleaned_text = texts[0].description.replace("\n", " ").replace("\t", " ").replace("|", " ")
+ cleaned_text = texts[0].description
+ return cleaned_text, bounds, text_to_box_mapping
+ else:
+ return '', None, None
+
+def overlay_boxes_on_image(path, bounds):
+ image = Image.open(path)
+ draw_boxes(image, bounds, "green")
+ return image
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+# ''' Google Vision'''
+# def detect_text(path):
+# """Detects text in the file located in the local filesystem."""
+# client = vision.ImageAnnotatorClient()
+
+# with io.open(path, 'rb') as image_file:
+# content = image_file.read()
+
+# image = vision.Image(content=content)
+
+# response = client.document_text_detection(image=image)
+# texts = response.text_annotations
+
+# if response.error.message:
+# raise Exception(
+# '{}\nFor more info on error messages, check: '
+# 'https://cloud.google.com/apis/design/errors'.format(
+# response.error.message))
+
+# return texts[0].description if texts else ''
diff --git a/vouchervision/PaLM_example_script.py b/vouchervision/PaLM_example_script.py
new file mode 100644
index 0000000000000000000000000000000000000000..9859e7d9b19b6a769530e61f686cb01ca728f214
--- /dev/null
+++ b/vouchervision/PaLM_example_script.py
@@ -0,0 +1,70 @@
+"""
+At the command line, only need to run once to install the package via pip:
+$ pip install google-generativeai
+"""
+
+import google.generativeai as palm
+
+palm.configure(api_key="YOUR API KEY")
+
+defaults = {
+ 'model': 'models/text-bison-001',
+ 'temperature': 0,
+ 'candidate_count': 1,
+ 'top_k': 40,
+ 'top_p': 0.95,
+ 'max_output_tokens': 1024,
+ 'stop_sequences': [],
+ 'safety_settings': [{"category":"HARM_CATEGORY_DEROGATORY","threshold":1},{"category":"HARM_CATEGORY_TOXICITY","threshold":1},{"category":"HARM_CATEGORY_VIOLENCE","threshold":2},{"category":"HARM_CATEGORY_SEXUAL","threshold":2},{"category":"HARM_CATEGORY_MEDICAL","threshold":2},{"category":"HARM_CATEGORY_DANGEROUS","threshold":2}],
+}
+prompt = """1. Your job is to return a new dict based on the structure of the reference dict ref_dict and these are your rules.
+ 2. You must look at ref_dict and refactor the new text called OCR to match the same formatting.
+ 3. OCR contains unstructured text inside of [], use your knowledge to put the OCR text into the correct ref_dict column.
+ 4. If OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None" and skip all other steps.
+ 5. If there is a field that does not have a direct proxy in the OCR text, you can fill it in based on your knowledge, but you cannot generate new information.
+ 6. Never put text from the ref_dict values into the new dict, but you must use the headers from ref_dict.
+ 7. There cannot be duplicate dictionary fields.
+ 8. Only return the new dict, do not explain your answer.
+
+ "Genus" - {"format": "[Genus]" or "[Family] indet" if no genus", "null_value": "", "description": taxonomic determination to genus, do captalize genus}
+ "Species"- {"format": "[species]" or "indet" if no species, "null_value": "", "description": taxonomic determination to species, do not captalize species}
+ "subspecies" - {"format": "[subspecies]", "null_value": "", "description": taxonomic determination to subspecies (subsp.)}
+ "variety" - {"format": "[variety]", "null_value": "", "description": taxonomic determination to variety (var)}
+ "forma" - {"format": "[form]", "null_value": "", "description": taxonomic determination to form (f.)}
+
+ "Country" - {"format": "[Country]", "null_value": "no data", "description": Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abreviation is given}
+ "State" - {"format": "[Adm. Division 1]", "null_value": "no data", "description": Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "County" - {"format": "[Adm. Division 2]", "null_value": "no data", "description": Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "Locality Name" - {"format": "verbatim", if no geographic info: "no data provided on label of catalog no: [######]", or if illegible: "locality present but illegible/not translated for catalog no: #######", or if no named locality: "no named locality for catalog no: #######", "description": "Description of geographic location or landscape"}
+
+ "Min Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given}
+ "Max Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise}
+ "Elevation Units" - {format: "m", "null_value": "","description": "m" only if an elevation is present}
+
+ "Verbatim Coordinates" - {"format": "[Lat, Long | UTM | TRS]", "null_value": "", "description": Verbatim coordinates as they appear on the label, fix typos to match standardized GPS coordinate format}
+
+ "Datum" - {"format": "[WGS84, NAD23 etc.]", "null_value": "not present", "description": Datum of coordinates on label; "" is GPS coordinates are not in OCR}
+ "Cultivated" - {"format": "yes", "null_value": "", "description": Indicates if specimen was grown in cultivation}
+ "Habitat" - {"format": "verbatim", "null_value": "", "description": Description of habitat or location where specimen was collected, ignore descriptions of the plant itself}
+ "Collectors" - {"format": "[Collector]", "null_value": "not present", "description": Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas}
+ "Collector Number" - {"format": "[Collector No.]", "null_value": "s.n.", "description": Sequential number assigned to collection, associated with the collector}
+ "Verbatim Date" - {"format": "verbatim", "null_value": "s.d.", "description": Date of collection exactly as it appears on the label}
+ "Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e. 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day}
+ "End Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": If date range is listed, later date of collection range}
+input: El Kala Algeria Aegilops El Tarf 1919-05-20 locality not transcribed for catalog no: 1702723 Charles d'Alleizette ovata May 20, 1919 s.n.
+
+output: {"Genus": "Aegilops", "Species": "ovata", "subspecies": "", "variety": "", "forma": "", "Country": "Algeria", "State": "El Tarf", "County": "El Kala", "Locality Name": "locality not transcribed for catalog no: 1702723", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Charles d'Alleizette", "Collector Number": "s.n.", "Verbatim Date": "May 20, 1919", "Date": "1919-05-20", "End Date": ""}
+
+input: El Kala Algeria Agrostis El Tarf 1918-06-08 locality not transcribed for catalog no: 1702919 Charles d'Alleizette pallida 8 Juin 1918 7748
+
+output: {"Genus": "Agrostis", "Species": "pallida", "subspecies": "", "variety": "", "forma": "", "Country": "Algeria", "State": "El Tarf", "County": "El Kala", "Locality Name": "locality not transcribed for catalog no: 1702919", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Charles d'Alleizette", "Collector Number": "7748", "Verbatim Date": "8 Juin 1918", "Date": "1918-06-08", "End Date": ""}
+
+input: Gympie river nr. sawmill Australia Hydrilla Queensland 1943-12-26 locality not transcribed for catalog no: 1702580 M. S. Clemens verticillata Dec. 26/43 43329
+
+output:"""
+
+response = palm.generate_text(
+ **defaults,
+ prompt=prompt
+)
+print(response.result)
\ No newline at end of file
diff --git a/vouchervision/VoucherVision_Config_Builder.py b/vouchervision/VoucherVision_Config_Builder.py
new file mode 100644
index 0000000000000000000000000000000000000000..1446acf8a784030cb09ce467129ee93d738c1769
--- /dev/null
+++ b/vouchervision/VoucherVision_Config_Builder.py
@@ -0,0 +1,576 @@
+import os, yaml, platform, traceback
+from vouchervision.LeafMachine2_Config_Builder import get_default_download_folder, write_config_file
+from vouchervision.general_utils import validate_dir, print_main_fail
+from vouchervision.vouchervision_main import voucher_vision
+from general_utils import get_cfg_from_full_path
+
+def build_VV_config():
+ #############################################
+ ############ Set common defaults ############
+ #############################################
+ # Changing the values below will set new
+ # default values each time you open the
+ # VoucherVision user interface
+ #############################################
+ #############################################
+ #############################################
+
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ run_name = 'test'
+ # dir_images_local = 'D:/Dropbox/LM2_Env/Image_Datasets/GBIF_BroadSample_3SppPerFamily1'
+ dir_images_local = os.path.join(dir_home,'demo','demo_images')
+
+ # The default output location is the computer's "Downloads" folder
+ # You can set dir_output directly by typing the folder path,
+ # OR you can uncomment the line "dir_output = default_output_folder"
+ # to have VoucherVision save to the Downloads folder by default
+ default_output_folder = get_default_download_folder()
+ dir_output = default_output_folder
+ # dir_output = 'D:/D_Desktop/LM2'
+
+ prefix_removal = '' #'MICH-V-'
+ suffix_removal = ''
+ catalog_numerical_only = False
+
+ LLM_version_user = 'Azure GPT 4'
+ prompt_version = 'Version 2' # from ["Version 1", "Version 1 No Domain Knowledge", "Version 2"]
+ use_LeafMachine2_collage_images = False # Use LeafMachine2 collage images
+
+ batch_size = 500
+
+ path_domain_knowledge = os.path.join(dir_home,'domain_knowledge','SLTP_UM_AllAsiaMinimalInRegion.xlsx')
+ embeddings_database_name = os.path.splitext(os.path.basename(path_domain_knowledge))[0]
+
+ #############################################
+ #############################################
+ ########## DO NOT EDIT BELOW HERE ###########
+ #############################################
+ #############################################
+ return assemble_config(dir_home, run_name, dir_images_local,dir_output,
+ prefix_removal,suffix_removal,catalog_numerical_only,LLM_version_user,batch_size,
+ path_domain_knowledge,embeddings_database_name,use_LeafMachine2_collage_images,
+ prompt_version, use_domain_knowledge=False)
+
+def assemble_config(dir_home, run_name, dir_images_local,dir_output,
+ prefix_removal,suffix_removal,catalog_numerical_only,LLM_version_user,batch_size,
+ path_domain_knowledge,embeddings_database_name,use_LeafMachine2_collage_images,
+ prompt_version, use_domain_knowledge=False):
+
+
+ # Initialize the base structure
+ config_data = {
+ 'leafmachine': {}
+ }
+
+ # Modular sections to be added to 'leafmachine'
+ do_section = {
+ 'check_for_illegal_filenames': False,
+ 'check_for_corrupt_images_make_vertical': True,
+ }
+
+ print_section = {
+ 'verbose': True,
+ 'optional_warnings': True
+ }
+
+ logging_section = {
+ 'log_level': None
+ }
+
+
+ project_section = {
+ 'dir_output': dir_output,
+ 'run_name': run_name,
+ 'image_location': 'local',
+ 'batch_size': batch_size,
+ 'num_workers': 1,
+ 'dir_images_local': dir_images_local,
+ 'continue_run_from_partial_xlsx': '',
+ 'prefix_removal': prefix_removal,
+ 'suffix_removal': suffix_removal,
+ 'catalog_numerical_only': catalog_numerical_only,
+ 'use_domain_knowledge': use_domain_knowledge,
+ 'embeddings_database_name': embeddings_database_name,
+ 'build_new_embeddings_database': False,
+ 'path_to_domain_knowledge_xlsx': path_domain_knowledge,
+ 'prompt_version': prompt_version,
+ 'delete_all_temps': False,
+ 'delete_temps_keep_VVE': False,
+ }
+
+ modules_section = {
+ 'specimen_crop': True
+ }
+
+ LLM_version = LLM_version_user
+ use_RGB_label_images = use_LeafMachine2_collage_images # Use LeafMachine2 collage images
+
+ cropped_components_section = {
+ 'do_save_cropped_annotations': True,
+ 'save_cropped_annotations': ['label','barcode'],
+ 'save_per_image': False,
+ 'save_per_annotation_class': True,
+ 'binarize_labels': False,
+ 'binarize_labels_skeletonize': False
+ }
+
+ data_section = {
+ 'save_json_rulers': False,
+ 'save_json_measurements': False,
+ 'save_individual_csv_files_rulers': False,
+ 'save_individual_csv_files_measurements': False,
+ 'save_individual_csv_files_landmarks': False,
+ 'save_individual_efd_files': False,
+ 'include_darwin_core_data_from_combined_file': False,
+ 'do_apply_conversion_factor': False
+ }
+
+ overlay_section = {
+ 'save_overlay_to_pdf': False,
+ 'save_overlay_to_jpgs': True,
+ 'overlay_dpi': 300, # Between 100 to 300
+ 'overlay_background_color': 'black', # Either 'white' or 'black'
+
+ 'show_archival_detections': True,
+ 'show_plant_detections': True,
+ 'show_segmentations': True,
+ 'show_landmarks': True,
+ 'ignore_archival_detections_classes': [],
+ 'ignore_plant_detections_classes': ['leaf_whole', 'specimen'], # Could also include 'leaf_partial' and others if needed
+ 'ignore_landmark_classes': [],
+
+ 'line_width_archival': 12, # Previous value given was 2
+ 'line_width_plant': 12, # Previous value given was 6
+ 'line_width_seg': 12, # 12 is specified as "thick"
+ 'line_width_efd': 12, # 3 is specified as "thick" but 12 is given here
+ 'alpha_transparency_archival': 0.3,
+ 'alpha_transparency_plant': 0,
+ 'alpha_transparency_seg_whole_leaf': 0.4,
+ 'alpha_transparency_seg_partial_leaf': 0.3
+ }
+
+ archival_component_detector_section = {
+ 'detector_type': 'Archival_Detector',
+ 'detector_version': 'PREP_final',
+ 'detector_iteration': 'PREP_final',
+ 'detector_weights': 'best.pt',
+ 'minimum_confidence_threshold': 0.5, # Default is 0.5
+ 'do_save_prediction_overlay_images': True,
+ 'ignore_objects_for_overlay': []
+ }
+
+ # Add the sections to the 'leafmachine' key
+ config_data['leafmachine']['do'] = do_section
+ config_data['leafmachine']['print'] = print_section
+ config_data['leafmachine']['logging'] = logging_section
+ config_data['leafmachine']['project'] = project_section
+ config_data['leafmachine']['LLM_version'] = LLM_version
+ config_data['leafmachine']['use_RGB_label_images'] = use_RGB_label_images
+ config_data['leafmachine']['cropped_components'] = cropped_components_section
+ config_data['leafmachine']['modules'] = modules_section
+ config_data['leafmachine']['data'] = data_section
+ config_data['leafmachine']['overlay'] = overlay_section
+ config_data['leafmachine']['archival_component_detector'] = archival_component_detector_section
+
+ return config_data, dir_home
+
+def build_api_tests(api):
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ path_to_configs = os.path.join(dir_home,'demo','demo_configs')
+
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ dir_images_local = os.path.join(dir_home,'demo','demo_images')
+ validate_dir(os.path.join(dir_home,'demo','demo_configs'))
+ path_domain_knowledge = os.path.join(dir_home,'domain_knowledge','SLTP_UM_AllAsiaMinimalInRegion.xlsx')
+ embeddings_database_name = os.path.splitext(os.path.basename(path_domain_knowledge))[0]
+ prefix_removal = ''
+ suffix_removal = ''
+ catalog_numerical_only = False
+ batch_size = 500
+
+
+ # ### Option 1: "GPT 4" of ["GPT 4", "GPT 3.5", "Azure GPT 4", "Azure GPT 3.5", "PaLM 2"]
+ # LLM_version_user = 'Azure GPT 4'
+
+ # ### Option 2: False of [False, True]
+ # use_LeafMachine2_collage_images = False
+
+ # ### Option 3: False of [False, True]
+ # use_domain_knowledge = True
+
+ test_results = {}
+ if api == 'openai':
+ OPT1, OPT2, OPT3 = TestOptionsAPI_openai.get_options()
+ elif api == 'palm':
+ OPT1, OPT2, OPT3 = TestOptionsAPI_palm.get_options()
+ elif api == 'azure_openai':
+ OPT1, OPT2, OPT3 = TestOptionsAPI_azure_openai.get_options()
+ else:
+ raise
+
+ ind = -1
+ ind_opt1 = -1
+ ind_opt2 = -1
+ ind_opt3 = -1
+
+ for opt1 in OPT1:
+ ind_opt1+= 1
+ for opt2 in OPT2:
+ ind_opt2 += 1
+ for opt3 in OPT3:
+ ind += 1
+ ind_opt3 += 1
+
+ LLM_version_user = opt1
+ use_LeafMachine2_collage_images = opt2
+ prompt_version = opt3
+
+ filename = f"{ind}__OPT1-{ind_opt1}__OPT2-{ind_opt2}__OPT3-{ind_opt3}.yaml"
+ run_name = f"{ind}__OPT1-{ind_opt1}__OPT2-{ind_opt2}__OPT3-{ind_opt3}"
+
+ dir_output = os.path.join(dir_home,'demo','demo_output','run_name')
+ validate_dir(dir_output)
+
+ config_data, dir_home = assemble_config(dir_home, run_name, dir_images_local,dir_output,
+ prefix_removal,suffix_removal,catalog_numerical_only,LLM_version_user,batch_size,
+ path_domain_knowledge,embeddings_database_name,use_LeafMachine2_collage_images,
+ prompt_version)
+
+ write_config_file(config_data, os.path.join(dir_home,'demo','demo_configs'),filename=filename)
+
+ test_results[run_name] = False
+ ind_opt3 = -1
+ ind_opt2 = -1
+ ind_opt1 = -1
+
+ return dir_home, path_to_configs, test_results
+
+def build_demo_tests(llm_version):
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ path_to_configs = os.path.join(dir_home,'demo','demo_configs')
+
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ dir_images_local = os.path.join(dir_home,'demo','demo_images')
+ validate_dir(os.path.join(dir_home,'demo','demo_configs'))
+ path_domain_knowledge = os.path.join(dir_home,'domain_knowledge','SLTP_UM_AllAsiaMinimalInRegion.xlsx')
+ embeddings_database_name = os.path.splitext(os.path.basename(path_domain_knowledge))[0]
+ prefix_removal = ''
+ suffix_removal = ''
+ catalog_numerical_only = False
+ batch_size = 500
+
+
+ # ### Option 1: "GPT 4" of ["GPT 4", "GPT 3.5", "Azure GPT 4", "Azure GPT 3.5", "PaLM 2"]
+ # LLM_version_user = 'Azure GPT 4'
+
+ # ### Option 2: False of [False, True]
+ # use_LeafMachine2_collage_images = False
+
+ # ### Option 3: False of [False, True]
+ # use_domain_knowledge = True
+
+ test_results = {}
+ if llm_version == 'gpt':
+ OPT1, OPT2, OPT3 = TestOptionsGPT.get_options()
+ elif llm_version == 'palm':
+ OPT1, OPT2, OPT3 = TestOptionsPalm.get_options()
+ else:
+ raise
+
+ ind = -1
+ ind_opt1 = -1
+ ind_opt2 = -1
+ ind_opt3 = -1
+
+ for opt1 in OPT1:
+ ind_opt1+= 1
+ for opt2 in OPT2:
+ ind_opt2 += 1
+ for opt3 in OPT3:
+ ind += 1
+ ind_opt3 += 1
+
+ LLM_version_user = opt1
+ use_LeafMachine2_collage_images = opt2
+ prompt_version = opt3
+
+ filename = f"{ind}__OPT1-{ind_opt1}__OPT2-{ind_opt2}__OPT3-{ind_opt3}.yaml"
+ run_name = f"{ind}__OPT1-{ind_opt1}__OPT2-{ind_opt2}__OPT3-{ind_opt3}"
+
+ dir_output = os.path.join(dir_home,'demo','demo_output','run_name')
+ validate_dir(dir_output)
+
+
+ if llm_version == 'gpt':
+ if prompt_version in ['Version 1']:
+ config_data, dir_home = assemble_config(dir_home, run_name, dir_images_local,dir_output,
+ prefix_removal,suffix_removal,catalog_numerical_only,LLM_version_user,batch_size,
+ path_domain_knowledge,embeddings_database_name,use_LeafMachine2_collage_images,
+ prompt_version, use_domain_knowledge=True)
+ else:
+ config_data, dir_home = assemble_config(dir_home, run_name, dir_images_local,dir_output,
+ prefix_removal,suffix_removal,catalog_numerical_only,LLM_version_user,batch_size,
+ path_domain_knowledge,embeddings_database_name,use_LeafMachine2_collage_images,
+ prompt_version)
+ elif llm_version == 'palm':
+ if prompt_version in ['Version 1 PaLM 2']:
+ config_data, dir_home = assemble_config(dir_home, run_name, dir_images_local,dir_output,
+ prefix_removal,suffix_removal,catalog_numerical_only,LLM_version_user,batch_size,
+ path_domain_knowledge,embeddings_database_name,use_LeafMachine2_collage_images,
+ prompt_version, use_domain_knowledge=True)
+ else:
+ config_data, dir_home = assemble_config(dir_home, run_name, dir_images_local,dir_output,
+ prefix_removal,suffix_removal,catalog_numerical_only,LLM_version_user,batch_size,
+ path_domain_knowledge,embeddings_database_name,use_LeafMachine2_collage_images,
+ prompt_version)
+
+
+ write_config_file(config_data, os.path.join(dir_home,'demo','demo_configs'),filename=filename)
+
+ test_results[run_name] = False
+ ind_opt3 = -1
+ ind_opt2 = -1
+ ind_opt1 = -1
+
+ return dir_home, path_to_configs, test_results
+
+class TestOptionsGPT:
+ OPT1 = ["GPT 4", "GPT 3.5", "Azure GPT 4", "Azure GPT 3.5"]
+ OPT2 = [False, True]
+ OPT3 = ["Version 1", "Version 1 No Domain Knowledge", "Version 2"]
+
+ @classmethod
+ def get_options(cls):
+ return cls.OPT1, cls.OPT2, cls.OPT3
+ @classmethod
+ def get_length(cls):
+ return 24
+
+class TestOptionsPalm:
+ OPT1 = ["PaLM 2"]
+ OPT2 = [False, True]
+ OPT3 = ["Version 1 PaLM 2", "Version 1 PaLM 2 No Domain Knowledge", "Version 2 PaLM 2"]
+
+ @classmethod
+ def get_options(cls):
+ return cls.OPT1, cls.OPT2, cls.OPT3
+ @classmethod
+ def get_length(cls):
+ return 6
+
+class TestOptionsAPI_openai:
+ OPT1 = ["GPT 3.5"]
+ OPT2 = [False]
+ OPT3 = ["Version 2"]
+
+ @classmethod
+ def get_options(cls):
+ return cls.OPT1, cls.OPT2, cls.OPT3
+ @classmethod
+ def get_length(cls):
+ return 24
+
+class TestOptionsAPI_azure_openai:
+ OPT1 = ["Azure GPT 3.5"]
+ OPT2 = [False]
+ OPT3 = ["Version 2"]
+
+ @classmethod
+ def get_options(cls):
+ return cls.OPT1, cls.OPT2, cls.OPT3
+ @classmethod
+ def get_length(cls):
+ return 24
+
+class TestOptionsAPI_palm:
+ OPT1 = ["PaLM 2"]
+ OPT2 = [False]
+ OPT3 = ["Version 2 PaLM 2"]
+
+ @classmethod
+ def get_options(cls):
+ return cls.OPT1, cls.OPT2, cls.OPT3
+ @classmethod
+ def get_length(cls):
+ return 6
+
+def run_demo_tests_GPT(progress_report):
+ dir_home, path_to_configs, test_results = build_demo_tests('gpt')
+ progress_report.set_n_overall(len(test_results.items()))
+
+ JSON_results = {}
+
+ for ind, (cfg, result) in enumerate(test_results.items()):
+ OPT1, OPT2, OPT3 = TestOptionsGPT.get_options()
+
+ test_ind, ind_opt1, ind_opt2, ind_opt3 = cfg.split('__')
+ opt1_readable = OPT1[int(ind_opt1.split('-')[1])]
+
+ if opt1_readable in ["Azure GPT 4", "Azure GPT 3.5"]:
+ api_version = 'gpt-azure'
+ elif opt1_readable in ["GPT 4", "GPT 3.5"]:
+ api_version = 'gpt'
+ else:
+ raise
+
+ opt2_readable = "Use LeafMachine2 for Collage Images" if OPT2[int(ind_opt2.split('-')[1])] else "Don't use LeafMachine2 for Collage Images"
+ opt3_readable = f"Prompt {OPT3[int(ind_opt3.split('-')[1])]}"
+ # Construct the human-readable test name
+ human_readable_name = f"{opt1_readable}, {opt2_readable}, {opt3_readable}"
+ get_n_overall = progress_report.get_n_overall()
+ progress_report.update_overall(f"Test {int(test_ind)+1} of {get_n_overall} --- Validating {human_readable_name}")
+ print_main_fail(f"Starting validation test: {human_readable_name}")
+ cfg_file_path = os.path.join(path_to_configs,'.'.join([cfg,'yaml']))
+
+ if check_API_key(dir_home, api_version) and check_API_key(dir_home, 'google-vision-ocr'):
+ try:
+ last_JSON_response, total_cost = voucher_vision(cfg_file_path, dir_home, cfg_test=None, progress_report=progress_report, test_ind=int(test_ind))
+ test_results[cfg] = True
+ JSON_results[ind] = last_JSON_response
+ except Exception as e:
+ JSON_results[ind] = None
+ test_results[cfg] = False
+ print(f"An exception occurred: {e}")
+ traceback.print_exc() # This will print the full traceback
+ else:
+ fail_response = ''
+ if not check_API_key(dir_home, 'google-vision-ocr'):
+ fail_response += "No API key found for Google Vision OCR"
+ if not check_API_key(dir_home, api_version):
+ fail_response += f" + No API key found for {api_version}"
+ test_results[cfg] = False
+ JSON_results[ind] = fail_response
+ print(f"No API key found for {fail_response}")
+
+ return test_results, JSON_results
+
+def run_demo_tests_Palm(progress_report):
+ api_version = 'palm'
+
+ dir_home, path_to_configs, test_results = build_demo_tests('palm')
+ progress_report.set_n_overall(len(test_results.items()))
+
+ JSON_results = {}
+
+ for ind, (cfg, result) in enumerate(test_results.items()):
+ OPT1, OPT2, OPT3 = TestOptionsPalm.get_options()
+ test_ind, ind_opt1, ind_opt2, ind_opt3 = cfg.split('__')
+ opt1_readable = OPT1[int(ind_opt1.split('-')[1])]
+ opt2_readable = "Use LeafMachine2 for Collage Images" if OPT2[int(ind_opt2.split('-')[1])] else "Don't use LeafMachine2 for Collage Images"
+ opt3_readable = f"Prompt {OPT3[int(ind_opt3.split('-')[1])]}"
+ # opt3_readable = "Use Domain Knowledge" if OPT3[int(ind_opt3.split('-')[1])] else "Don't use Domain Knowledge"
+ # Construct the human-readable test name
+ human_readable_name = f"{opt1_readable}, {opt2_readable}, {opt3_readable}"
+ get_n_overall = progress_report.get_n_overall()
+ progress_report.update_overall(f"Test {int(test_ind)+1} of {get_n_overall} --- Validating {human_readable_name}")
+ print_main_fail(f"Starting validation test: {human_readable_name}")
+ cfg_file_path = os.path.join(path_to_configs,'.'.join([cfg,'yaml']))
+
+ if check_API_key(dir_home, api_version) and check_API_key(dir_home, 'google-vision-ocr') :
+ try:
+ last_JSON_response, total_cost = voucher_vision(cfg_file_path, dir_home, cfg_test=None, progress_report=progress_report, test_ind=int(test_ind))
+ test_results[cfg] = True
+ JSON_results[ind] = last_JSON_response
+ except Exception as e:
+ test_results[cfg] = False
+ JSON_results[ind] = None
+ print(f"An exception occurred: {e}")
+ traceback.print_exc() # This will print the full traceback
+ else:
+ fail_response = ''
+ if not check_API_key(dir_home, 'google-vision-ocr'):
+ fail_response += "No API key found for Google Vision OCR"
+ if not check_API_key(dir_home, api_version):
+ fail_response += f" + No API key found for {api_version}"
+ test_results[cfg] = False
+ JSON_results[ind] = fail_response
+ print(f"No API key found for {fail_response}")
+
+ return test_results, JSON_results
+
+def run_api_tests(api):
+ try:
+ dir_home, path_to_configs, test_results = build_api_tests(api)
+
+ JSON_results = {}
+
+ for ind, (cfg, result) in enumerate(test_results.items()):
+ if api == 'openai':
+ OPT1, OPT2, OPT3 = TestOptionsAPI_openai.get_options()
+ elif 'azure_openai':
+ OPT1, OPT2, OPT3 = TestOptionsAPI_azure_openai.get_options()
+ elif 'palm':
+ OPT1, OPT2, OPT3 = TestOptionsAPI_palm.get_options()
+ test_ind, ind_opt1, ind_opt2, ind_opt3 = cfg.split('__')
+ opt1_readable = OPT1[int(ind_opt1.split('-')[1])]
+ opt2_readable = "Use LeafMachine2 for Collage Images" if OPT2[int(ind_opt2.split('-')[1])] else "Don't use LeafMachine2 for Collage Images"
+ opt3_readable = f"Prompt {OPT3[int(ind_opt3.split('-')[1])]}"
+ # opt3_readable = "Use Domain Knowledge" if OPT3[int(ind_opt3.split('-')[1])] else "Don't use Domain Knowledge"
+ # Construct the human-readable test name
+ human_readable_name = f"{opt1_readable}, {opt2_readable}, {opt3_readable}"
+ print_main_fail(f"Starting validation test: {human_readable_name}")
+ cfg_file_path = os.path.join(path_to_configs,'.'.join([cfg,'yaml']))
+
+ if check_API_key(dir_home, api) and check_API_key(dir_home, 'google-vision-ocr') :
+ try:
+ last_JSON_response, total_cost = voucher_vision(cfg_file_path, dir_home, None, cfg_test=None, progress_report=None, test_ind=int(test_ind))
+ test_results[cfg] = True
+ JSON_results[ind] = last_JSON_response
+ return True
+
+ except Exception as e:
+ print(e)
+ return False
+ else:
+ return False
+ except Exception as e:
+ print(e)
+ return False
+
+def has_API_key(val):
+ if val != '':
+ return True
+ else:
+ return False
+
+def check_if_usable():
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ path_cfg_private = os.path.join(dir_home, 'PRIVATE_DATA.yaml')
+ cfg_private = get_cfg_from_full_path(path_cfg_private)
+
+ has_key_openai = has_API_key(cfg_private['openai']['OPENAI_API_KEY'])
+
+ has_key_azure_openai = has_API_key(cfg_private['openai_azure']['api_version'])
+
+ has_key_palm2 = has_API_key(cfg_private['google_palm']['google_palm_api'])
+
+ has_key_google_OCR = has_API_key(cfg_private['google_cloud']['path_json_file'])
+
+ if has_key_google_OCR and (has_key_azure_openai or has_key_openai or has_key_palm2):
+ return True
+ else:
+ return False
+
+def check_API_key(dir_home, api_version):
+ dir_home = os.path.dirname(os.path.dirname(__file__))
+ path_cfg_private = os.path.join(dir_home, 'PRIVATE_DATA.yaml')
+ cfg_private = get_cfg_from_full_path(path_cfg_private)
+
+ has_key_openai = has_API_key(cfg_private['openai']['OPENAI_API_KEY'])
+
+ has_key_azure_openai = has_API_key(cfg_private['openai_azure']['api_version'])
+
+ has_key_palm2 = has_API_key(cfg_private['google_palm']['google_palm_api'])
+
+ has_key_google_OCR = has_API_key(cfg_private['google_cloud']['path_json_file'])
+
+ if api_version == 'palm' and has_key_palm2:
+ return True
+ elif api_version in ['gpt','openai'] and has_key_openai:
+ return True
+ elif api_version in ['gpt-azure', 'azure_openai'] and has_key_azure_openai:
+ return True
+ elif api_version == 'google-vision-ocr' and has_key_google_OCR:
+ return True
+ else:
+ return False
diff --git a/vouchervision/component_detector/LICENSE b/vouchervision/component_detector/LICENSE
new file mode 100644
index 0000000000000000000000000000000000000000..92b370f0e0e1b91cf8baf5d0f78c56a9824c39f1
--- /dev/null
+++ b/vouchervision/component_detector/LICENSE
@@ -0,0 +1,674 @@
+GNU GENERAL PUBLIC LICENSE
+ Version 3, 29 June 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc.
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+ Preamble
+
+ The GNU General Public License is a free, copyleft license for
+software and other kinds of works.
+
+ The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works. By contrast,
+the GNU General Public License is intended to guarantee your freedom to
+share and change all versions of a program--to make sure it remains free
+software for all its users. We, the Free Software Foundation, use the
+GNU General Public License for most of our software; it applies also to
+any other work released this way by its authors. You can apply it to
+your programs, too.
+
+ When we speak of free software, we are referring to freedom, not
+price. Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+them if you wish), that you receive source code or can get it if you
+want it, that you can change the software or use pieces of it in new
+free programs, and that you know you can do these things.
+
+ To protect your rights, we need to prevent others from denying you
+these rights or asking you to surrender the rights. Therefore, you have
+certain responsibilities if you distribute copies of the software, or if
+you modify it: responsibilities to respect the freedom of others.
+
+ For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must pass on to the recipients the same
+freedoms that you received. You must make sure that they, too, receive
+or can get the source code. And you must show them these terms so they
+know their rights.
+
+ Developers that use the GNU GPL protect your rights with two steps:
+(1) assert copyright on the software, and (2) offer you this License
+giving you legal permission to copy, distribute and/or modify it.
+
+ For the developers' and authors' protection, the GPL clearly explains
+that there is no warranty for this free software. For both users' and
+authors' sake, the GPL requires that modified versions be marked as
+changed, so that their problems will not be attributed erroneously to
+authors of previous versions.
+
+ Some devices are designed to deny users access to install or run
+modified versions of the software inside them, although the manufacturer
+can do so. This is fundamentally incompatible with the aim of
+protecting users' freedom to change the software. The systematic
+pattern of such abuse occurs in the area of products for individuals to
+use, which is precisely where it is most unacceptable. Therefore, we
+have designed this version of the GPL to prohibit the practice for those
+products. If such problems arise substantially in other domains, we
+stand ready to extend this provision to those domains in future versions
+of the GPL, as needed to protect the freedom of users.
+
+ Finally, every program is threatened constantly by software patents.
+States should not allow patents to restrict development and use of
+software on general-purpose computers, but in those that do, we wish to
+avoid the special danger that patents applied to a free program could
+make it effectively proprietary. To prevent this, the GPL assures that
+patents cannot be used to render the program non-free.
+
+ The precise terms and conditions for copying, distribution and
+modification follow.
+
+ TERMS AND CONDITIONS
+
+ 0. Definitions.
+
+ "This License" refers to version 3 of the GNU General Public License.
+
+ "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+ "The Program" refers to any copyrightable work licensed under this
+License. Each licensee is addressed as "you". "Licensees" and
+"recipients" may be individuals or organizations.
+
+ To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy. The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+ A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+ To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy. Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+ To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies. Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+ An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License. If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+ 1. Source Code.
+
+ The "source code" for a work means the preferred form of the work
+for making modifications to it. "Object code" means any non-source
+form of a work.
+
+ A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+ The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form. A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+ The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities. However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work. For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+ The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+ The Corresponding Source for a work in source code form is that
+same work.
+
+ 2. Basic Permissions.
+
+ All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met. This License explicitly affirms your unlimited
+permission to run the unmodified Program. The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work. This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+ You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force. You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright. Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+ Conveying under any other circumstances is permitted solely under
+the conditions stated below. Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+ 3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+ No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+ When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+ 4. Conveying Verbatim Copies.
+
+ You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+ You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+ 5. Conveying Modified Source Versions.
+
+ You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+ a) The work must carry prominent notices stating that you modified
+ it, and giving a relevant date.
+
+ b) The work must carry prominent notices stating that it is
+ released under this License and any conditions added under section
+ 7. This requirement modifies the requirement in section 4 to
+ "keep intact all notices".
+
+ c) You must license the entire work, as a whole, under this
+ License to anyone who comes into possession of a copy. This
+ License will therefore apply, along with any applicable section 7
+ additional terms, to the whole of the work, and all its parts,
+ regardless of how they are packaged. This License gives no
+ permission to license the work in any other way, but it does not
+ invalidate such permission if you have separately received it.
+
+ d) If the work has interactive user interfaces, each must display
+ Appropriate Legal Notices; however, if the Program has interactive
+ interfaces that do not display Appropriate Legal Notices, your
+ work need not make them do so.
+
+ A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit. Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+ 6. Conveying Non-Source Forms.
+
+ You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+ a) Convey the object code in, or embodied in, a physical product
+ (including a physical distribution medium), accompanied by the
+ Corresponding Source fixed on a durable physical medium
+ customarily used for software interchange.
+
+ b) Convey the object code in, or embodied in, a physical product
+ (including a physical distribution medium), accompanied by a
+ written offer, valid for at least three years and valid for as
+ long as you offer spare parts or customer support for that product
+ model, to give anyone who possesses the object code either (1) a
+ copy of the Corresponding Source for all the software in the
+ product that is covered by this License, on a durable physical
+ medium customarily used for software interchange, for a price no
+ more than your reasonable cost of physically performing this
+ conveying of source, or (2) access to copy the
+ Corresponding Source from a network server at no charge.
+
+ c) Convey individual copies of the object code with a copy of the
+ written offer to provide the Corresponding Source. This
+ alternative is allowed only occasionally and noncommercially, and
+ only if you received the object code with such an offer, in accord
+ with subsection 6b.
+
+ d) Convey the object code by offering access from a designated
+ place (gratis or for a charge), and offer equivalent access to the
+ Corresponding Source in the same way through the same place at no
+ further charge. You need not require recipients to copy the
+ Corresponding Source along with the object code. If the place to
+ copy the object code is a network server, the Corresponding Source
+ may be on a different server (operated by you or a third party)
+ that supports equivalent copying facilities, provided you maintain
+ clear directions next to the object code saying where to find the
+ Corresponding Source. Regardless of what server hosts the
+ Corresponding Source, you remain obligated to ensure that it is
+ available for as long as needed to satisfy these requirements.
+
+ e) Convey the object code using peer-to-peer transmission, provided
+ you inform other peers where the object code and Corresponding
+ Source of the work are being offered to the general public at no
+ charge under subsection 6d.
+
+ A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+ A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling. In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage. For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product. A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+ "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source. The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+ If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information. But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+ The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed. Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+ Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+ 7. Additional Terms.
+
+ "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law. If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+ When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it. (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.) You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+ Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+ a) Disclaiming warranty or limiting liability differently from the
+ terms of sections 15 and 16 of this License; or
+
+ b) Requiring preservation of specified reasonable legal notices or
+ author attributions in that material or in the Appropriate Legal
+ Notices displayed by works containing it; or
+
+ c) Prohibiting misrepresentation of the origin of that material, or
+ requiring that modified versions of such material be marked in
+ reasonable ways as different from the original version; or
+
+ d) Limiting the use for publicity purposes of names of licensors or
+ authors of the material; or
+
+ e) Declining to grant rights under trademark law for use of some
+ trade names, trademarks, or service marks; or
+
+ f) Requiring indemnification of licensors and authors of that
+ material by anyone who conveys the material (or modified versions of
+ it) with contractual assumptions of liability to the recipient, for
+ any liability that these contractual assumptions directly impose on
+ those licensors and authors.
+
+ All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10. If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term. If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+ If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+ Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+ 8. Termination.
+
+ You may not propagate or modify a covered work except as expressly
+provided under this License. Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+ However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+ Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+ Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License. If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+ 9. Acceptance Not Required for Having Copies.
+
+ You are not required to accept this License in order to receive or
+run a copy of the Program. Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance. However,
+nothing other than this License grants you permission to propagate or
+modify any covered work. These actions infringe copyright if you do
+not accept this License. Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+ 10. Automatic Licensing of Downstream Recipients.
+
+ Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License. You are not responsible
+for enforcing compliance by third parties with this License.
+
+ An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations. If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+ You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License. For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+ 11. Patents.
+
+ A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based. The
+work thus licensed is called the contributor's "contributor version".
+
+ A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version. For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+ Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+ In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement). To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+ If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients. "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+ If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+ A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License. You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+ Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+ 12. No Surrender of Others' Freedom.
+
+ If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License. If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all. For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+ 13. Use with the GNU Affero General Public License.
+
+ Notwithstanding any other provision of this License, you have
+permission to link or combine any covered work with a work licensed
+under version 3 of the GNU Affero General Public License into a single
+combined work, and to convey the resulting work. The terms of this
+License will continue to apply to the part which is the covered work,
+but the special requirements of the GNU Affero General Public License,
+section 13, concerning interaction through a network will apply to the
+combination as such.
+
+ 14. Revised Versions of this License.
+
+ The Free Software Foundation may publish revised and/or new versions of
+the GNU General Public License from time to time. Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+ Each version is given a distinguishing version number. If the
+Program specifies that a certain numbered version of the GNU General
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version published by the Free Software
+Foundation. If the Program does not specify a version number of the
+GNU General Public License, you may choose any version ever published
+by the Free Software Foundation.
+
+ If the Program specifies that a proxy can decide which future
+versions of the GNU General Public License can be used, that proxy's
+public statement of acceptance of a version permanently authorizes you
+to choose that version for the Program.
+
+ Later license versions may give you additional or different
+permissions. However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+ 15. Disclaimer of Warranty.
+
+ THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+ 16. Limitation of Liability.
+
+ IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+ 17. Interpretation of Sections 15 and 16.
+
+ If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+ END OF TERMS AND CONDITIONS
+
+ How to Apply These Terms to Your New Programs
+
+ If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+ To do so, attach the following notices to the program. It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+
+ Copyright (C)
+
+ This program is free software: you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation, either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see .
+
+Also add information on how to contact you by electronic and paper mail.
+
+ If the program does terminal interaction, make it output a short
+notice like this when it starts in an interactive mode:
+
+ Copyright (C)
+ This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+ This is free software, and you are welcome to redistribute it
+ under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License. Of course, your program's commands
+might be different; for a GUI interface, you would use an "about box".
+
+ You should also get your employer (if you work as a programmer) or school,
+if any, to sign a "copyright disclaimer" for the program, if necessary.
+For more information on this, and how to apply and follow the GNU GPL, see
+.
+
+ The GNU General Public License does not permit incorporating your program
+into proprietary programs. If your program is a subroutine library, you
+may consider it more useful to permit linking proprietary applications with
+the library. If this is what you want to do, use the GNU Lesser General
+Public License instead of this License. But first, please read
+.
diff --git a/vouchervision/component_detector/__init__.py b/vouchervision/component_detector/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/vouchervision/component_detector/armature_processing.py b/vouchervision/component_detector/armature_processing.py
new file mode 100644
index 0000000000000000000000000000000000000000..9224da9bd72fe6f215475cb9105d68ba983436a4
--- /dev/null
+++ b/vouchervision/component_detector/armature_processing.py
@@ -0,0 +1,1047 @@
+import os, math, cv2, random
+import numpy as np
+from itertools import combinations
+from PIL import Image
+from dataclasses import dataclass, field
+from typing import List, Dict
+from sklearn.linear_model import LinearRegression
+from scipy.optimize import fsolve, minimize
+
+
+@dataclass()
+class ArmatureSkeleton:
+ cfg: str
+ Dirs: str
+ leaf_type: str
+ all_points: list
+ dir_temp: str
+ file_name: str
+ width: int
+ height: int
+ logger: object
+
+ is_complete: bool = False
+ keep_going: bool = False
+
+ do_show_QC_images: bool = False
+ do_save_QC_images: bool = False
+
+ classes: int = 0
+ points_list: int = 0
+
+ image: int = 0
+
+ ordered_middle: int = 0
+ midvein_fit: int = 0
+ midvein_fit_points: int = 0
+ ordered_midvein_length: float = 0.0
+ has_middle = False
+
+ has_outer = False
+ has_tip = False
+
+ is_split = False
+
+ ordered_petiole: int = 0
+ ordered_petiole_length: float = 0.0
+ has_ordered_petiole = False
+
+ has_apex: bool = False
+ apex_left: int = 0
+ apex_right: int = 0
+ apex_center: int = 0
+ apex_angle_type: str = 'NA'
+ apex_angle_degrees: float = 0.0
+
+ has_base: bool = False
+ base_left: int = 0
+ base_right: int = 0
+ base_center: int = 0
+ base_angle_type: str = 'NA'
+ base_angle_degrees: float = 0.0
+
+ has_lamina_base: bool = False
+ lamina_base: int = 0
+
+ has_lamina_length: bool = False
+ lamina_fit: int = 0
+ lamina_length: float = 0.0
+
+ has_width: bool = False
+ lamina_width: float = 0.0
+ width_left: float = 0.0
+ width_right: float = 0.0
+
+
+
+ def __init__(self, cfg, logger, Dirs, leaf_type, all_points, height, width, dir_temp, file_name) -> None:
+ # Store the necessary arguments as instance attributes
+ self.cfg = cfg
+ self.Dirs = Dirs
+ self.leaf_type = leaf_type
+ self.all_points = all_points
+ self.height = height
+ self.width = width
+ self.dir_temp = dir_temp
+ self.file_name = file_name
+
+ logger.name = f'[{leaf_type} - {file_name}]'
+ self.logger = logger
+
+ self.init_lists_dicts()
+
+ """ Setup """
+ self.set_cfg_values()
+ self.define_landmark_classes()
+
+ self.setup_QC_image()
+ self.setup_angle_image()
+ self.setup_final_image()
+
+ self.parse_all_points()
+ self.convert_YOLO_bbox_to_point()
+
+ if (len(self.points_list['outer']) > 6) and (len(self.points_list['middle']) > 3):
+ self.keep_going = True
+
+ """ Landmarks """
+ if self.keep_going:
+ # Start with ordering the midvein and petiole
+ self.order_middle()
+ # print(self.ordered_midvein)
+ if self.keep_going:
+ # Split the image using the midvein IF has_midvein == True
+ self.split_image_by_middle()
+ if self.keep_going:
+ self.group_outer_points()
+ if self.keep_going:
+ # Measure
+ self.measure_armature()
+ if self.keep_going:
+ # calc tangent angle of outer and inner polys
+ self.calc_angle_tangent()
+ if self.keep_going:
+ self.calc_angle_curl()
+ if self.keep_going:
+ # self.calc_angle_bend()
+ self.calc_curvature_radius()
+ if self.keep_going:
+ self.calc_direct_length()
+
+ # self.show_QC_image()
+ # self.show_angle_image()
+
+ self.is_complete = True # TODO add ways to set True
+
+
+ def measure_armature(self):
+ # wb = width_base = line between the last outer and inner points
+ # Define the line function
+ def line_func(x):
+ return self.wb_slope * x + self.wb_intercept
+ def middle_func(x):
+ return self.middle_poly[0]*x**2 + self.middle_poly[1]*x + self.middle_poly[2]
+ # Define the difference function
+ def line_middle_diff(x):
+ return line_func(x) - middle_func(x)
+
+ # Convert the points to numpy arrays
+ last_point_right = np.array(self.last_point_right)
+ last_point_left = np.array(self.last_point_left)
+
+ # Calculate the Euclidean distance between the points
+ self.width_base = np.linalg.norm(last_point_right - last_point_left)
+ print("The distance between the last points of the right and left segments is:", self.width_base)
+
+ # Intersection of the width and the middlepoly# Draw a line between the last points of the outer_left and outer_right segments
+ cv2.line(self.image, (int(self.last_point_left[0]), int(self.last_point_left[1])), (int(self.last_point_right[0]), int(self.last_point_right[1])), gc('white'), thickness=2)
+ cv2.line(self.image_angles, (int(self.last_point_left[0]), int(self.last_point_left[1])), (int(self.last_point_right[0]), int(self.last_point_right[1])), color=gc('white'), thickness=2)
+
+ # Calculate the slope and y-intercept of the line
+ self.wb_slope = (self.last_point_right[1] - self.last_point_left[1]) / (self.last_point_right[0] - self.last_point_left[0])
+ self.wb_intercept = self.last_point_left[1] - self.wb_slope * self.last_point_left[0]
+
+ # Find the intersection point
+ intersection_x = fsolve(line_middle_diff, 0)[0]
+ intersection_y = line_func(intersection_x)
+
+ self.width_base_inter = [(int(intersection_x), int(intersection_y))]
+ # Calculate the midpoint between the last points
+ self.width_base_mid = (last_point_right + last_point_left) / 2
+
+ cv2.circle(self.image, (int(intersection_x), int(intersection_y)), radius=2, color=gc('green'), thickness=-1)
+ cv2.circle(self.image, (int(intersection_x), int(intersection_y)), radius=4, color=gc('black'), thickness=2)
+ cv2.circle(self.image, (int(self.width_base_mid[0]), int(self.width_base_mid[1])), radius=2, color=gc('red'), thickness=-1)
+ cv2.circle(self.image, (int(self.width_base_mid[0]), int(self.width_base_mid[1])), radius=4, color=gc('black'), thickness=2)
+
+ print("The intersection point of the line and the middle polynomial is:", (intersection_x, intersection_y))
+
+
+
+ def calc_direct_length(self):
+ # Calculate the x-coordinate of the intersection point
+ x_intersection = (self.wb_intercept_perpendicular - self.wb_intercept) / (self.wb_slope - self.wb_slope_perpendicular)
+
+ # Calculate the y-coordinate of the intersection point
+ y_intersection = self.wb_slope * x_intersection + self.wb_intercept
+
+ # Store the intersection point as self.wb_origin
+ self.wb_origin = np.array([x_intersection, y_intersection])
+
+ # Calculate the distance between the intersection point and self.inter_point
+ self.length_direct = np.linalg.norm(self.wb_origin - self.inter_point)
+ # Plot a 2-pixel thick red line from self.wb_origin to self.inter_point
+ cv2.line(self.image_angles, tuple(map(int, self.wb_origin)), tuple(map(int, self.inter_point)), gc('red'), thickness=2)
+
+
+
+ def calc_curvature_radius(self):
+ def fit_circle_least_squares(points):
+ if len(points) <= 1:
+ return 0.0, (0, 0)
+
+ def calc_residuals(params, points):
+ x0, y0, r = params
+ residuals = np.sqrt((points[:, 0] - x0) ** 2 + (points[:, 1] - y0) ** 2) - r
+ return residuals
+
+ def objective(params, points):
+ return np.sum(calc_residuals(params, points) ** 2)
+
+ x_mean = np.mean(points[:, 0])
+ y_mean = np.mean(points[:, 1])
+ r_mean = np.mean(np.sqrt((points[:, 0] - x_mean) ** 2 + (points[:, 1] - y_mean) ** 2))
+ init_params = [x_mean, y_mean, r_mean]
+
+ result = minimize(objective, init_params, args=(points,), method='L-BFGS-B')
+ x0, y0, r = result.x
+
+ return r, (x0, y0)
+
+ self.radius_middle, center_middle = fit_circle_least_squares(self.ordered_middle_np)
+ self.radius_outer_left, center_outer_left = fit_circle_least_squares(self.ordered_outer_left_np)
+ self.radius_outer_right, center_outer_right = fit_circle_least_squares(self.ordered_outer_right_np)
+
+
+ # Plot the circles on self.image_angles
+ cv2.circle(self.image_angles, (int(center_middle[0]), int(center_middle[1])), int(self.radius_middle), gc('yellow'), thickness=1)
+ cv2.circle(self.image_angles, (int(center_outer_left[0]), int(center_outer_left[1])), int(self.radius_outer_left), gc('pink'), thickness=1)
+ cv2.circle(self.image_angles, (int(center_outer_right[0]), int(center_outer_right[1])), int(self.radius_outer_right), gc('cyan'), thickness=1)
+
+ print('hi')
+
+
+ def calc_angle_bend(self):
+ print('hi')
+
+
+
+ def calc_angle_curl(self):
+ # Define the perpendicular line function
+ def wb_line_perpendicular(x):
+ return self.wb_slope_perpendicular * x + self.wb_intercept_perpendicular
+
+
+ # Calculate the slope of the line perpendicular to the given line
+ self.wb_slope_perpendicular = -1 / self.wb_slope
+ # Calculate the y-intercept of the line perpendicular to the given line
+ self.wb_intercept_perpendicular = self.inter_point[1] - self.wb_slope_perpendicular * self.inter_point[0]
+
+ # Line fit to first 3 points in self.ordered_middle
+ self.middle_tip_poly = np.polyfit(self.ordered_middle_np[0:3, 0], self.ordered_middle_np[0:3, 1], 1)
+ middle_tip_slope = self.middle_tip_poly[0]
+
+ # angle between middle_tip fit the curl perpendicular
+ theta = math.atan(abs((middle_tip_slope - self.wb_slope_perpendicular) / (1 + self.wb_slope_perpendicular*middle_tip_slope)))
+
+ # Convert the angle to degrees
+ self.angle_curl = math.degrees(theta)
+
+ print("The angle between the lines is:", self.angle_curl, "degrees")
+
+ # Draw the tangents at the intersection point
+ intersection_point = np.array(self.inter_point_outer_inner, dtype=int)
+ length = 50 # Length of the tangent lines
+
+ # Calculate the points for the tangent lines
+ curl_tangent_point1 = (intersection_point[0] - length, intersection_point[1] - length * self.wb_slope_perpendicular)
+ curl_tangent_point2 = (intersection_point[0] + length, intersection_point[1] + length * self.wb_slope_perpendicular)
+ middle_tip_tangent_point1 = (intersection_point[0] - length, intersection_point[1] - length * middle_tip_slope)
+ middle_tip_tangent_point2 = (intersection_point[0] + length, intersection_point[1] + length * middle_tip_slope)
+
+ # Convert the points to integers
+ curl_tangent_point1 = tuple(map(int, curl_tangent_point1))
+ curl_tangent_point2 = tuple(map(int, curl_tangent_point2))
+ middle_tip_tangent_point1 = tuple(map(int, middle_tip_tangent_point1))
+ middle_tip_tangent_point2 = tuple(map(int, middle_tip_tangent_point2))
+
+ # Draw the tangent lines
+ cv2.line(self.image_angles, intersection_point, curl_tangent_point1, gc('teal'), 1)
+ cv2.line(self.image_angles, intersection_point, curl_tangent_point2, gc('teal'), 1)
+ cv2.line(self.image_angles, intersection_point, middle_tip_tangent_point1, gc('teal'), 1)
+ cv2.line(self.image_angles, intersection_point, middle_tip_tangent_point2, gc('teal'), 1)
+
+ # Draw the arc representing the angle
+ cv2.ellipse(self.image_angles, tuple(intersection_point), (length, length), 0, 0, self.angle_curl, gc('teal'), 2)
+ cv2.ellipse(self.image_angles, tuple(intersection_point), (length, length), 180, 0, self.angle_curl, gc('teal'), 2)
+
+ ### plot the wb_line_perpendicular
+ # Calculate the y values for the start and end points of the line
+ y_start = max(0, int(wb_line_perpendicular(0)))
+ y_end = min(self.height, int(wb_line_perpendicular(self.width)))
+
+ # Define the range of y values for the line
+ y_range = np.linspace(y_start, y_end, num=100, dtype=int) # You can adjust 'num' to control the number of points
+
+ # Draw the dotted gray line
+ for i in range(len(y_range) - 1):
+ y1, x1 = y_range[i], int((y_range[i] - self.wb_intercept_perpendicular) / self.wb_slope_perpendicular)
+ x1 = max(0, min(x1, self.width)) # Keep x1 within the bounds of the image width
+ y2, x2 = y_range[i+1], int((y_range[i+1] - self.wb_intercept_perpendicular) / self.wb_slope_perpendicular)
+ x2 = max(0, min(x2, self.width)) # Keep x2 within the bounds of the image width
+
+ if i % 2 == 0: # Change the value of 2 to adjust the spacing between the dots
+ cv2.line(self.image_angles, (x1, y1), (x2, y2), gc('white'), 1)
+
+
+
+
+ def calc_angle_tangent(self):
+ # Define the polynomial functions
+ def left_func(x):
+ return self.left_poly[0]*x**2 + self.left_poly[1]*x + self.left_poly[2]
+
+ def right_func(x):
+ return self.right_poly[0]*x**2 + self.right_poly[1]*x + self.right_poly[2]
+
+ # Define the difference function
+ def left_right_diff(x):
+ return left_func(x) - right_func(x)
+
+ # Find the x-coordinate of the intersection point
+ intersection_x = fsolve(left_right_diff, 0)[0]
+
+ # Calculate the y-coordinate of the intersection point on the left and right curves
+ intersection_y_left = left_func(intersection_x)
+ intersection_y_right = right_func(intersection_x)
+
+ # Calculate the derivatives of the polynomials at the intersection point
+ left_derivative = 2*self.left_poly[0]*intersection_x + self.left_poly[1]
+ right_derivative = 2*self.right_poly[0]*intersection_x + self.right_poly[1]
+
+ # Calculate the angle between the tangents to the polynomials at the intersection point
+ theta = math.atan(abs((right_derivative - left_derivative) / (1 + left_derivative*right_derivative)))
+
+ # Convert the angle to degrees
+ self.angle_tangent = math.degrees(theta)
+
+ print("The angle between the left and right polynomials at their point of intersection is:", theta, "degrees")
+
+ # Draw the tangents at the intersection point
+ intersection_point = np.array([int(intersection_x), int(intersection_y_left + (intersection_y_right - intersection_y_left)/2)])
+ length = 30 # Length of the tangent lines
+
+ # Calculate the points for the tangent lines
+ left_tangent_point1 = (intersection_point[0] - length, intersection_point[1] - length * left_derivative)
+ left_tangent_point2 = (intersection_point[0] + length, intersection_point[1] + length * left_derivative)
+ right_tangent_point1 = (intersection_point[0] - length, intersection_point[1] - length * right_derivative)
+ right_tangent_point2 = (intersection_point[0] + length, intersection_point[1] + length * right_derivative)
+
+ # Convert the points to integers
+ left_tangent_point1 = tuple(map(int, left_tangent_point1))
+ left_tangent_point2 = tuple(map(int, left_tangent_point2))
+ right_tangent_point1 = tuple(map(int, right_tangent_point1))
+ right_tangent_point2 = tuple(map(int, right_tangent_point2))
+
+ # # Draw the tangent lines
+ # cv2.line(self.image_angles, intersection_point, left_tangent_point1, gc('yellow'), 1)
+ # cv2.line(self.image_angles, intersection_point, left_tangent_point2, gc('yellow'), 1)
+ # cv2.line(self.image_angles, intersection_point, right_tangent_point1, gc('yellow'), 1)
+ # cv2.line(self.image_angles, intersection_point, right_tangent_point2, gc('yellow'), 1)
+
+ # Draw the arc representing the angle
+ cv2.ellipse(self.image_angles, tuple(intersection_point), (length, length), 0, 0, self.angle_tangent, gc('yellow'), 2)
+ cv2.ellipse(self.image_angles, tuple(intersection_point), (length, length), 180, 0, self.angle_tangent, gc('yellow'), 2)
+
+ # self.show_angle_image()
+ # return theta
+
+
+ def group_outer_points(self):
+ # Split the points into two groups based on their position relative to the line
+ self.outer_left = []
+ self.outer_right = []
+
+ # if 'tip' in self.points_list:
+
+ for point in self.points_list['outer']:
+ x, y = point
+ predicted_y = self.predict_y(x)
+
+ if y > predicted_y:
+ self.outer_right.append(point)
+ else:
+ self.outer_left.append(point)
+
+ self.outer_right = np.array(self.outer_right)
+ self.outer_left = np.array(self.outer_left)
+
+ if (len(self.outer_right) < 3) or (len(self.outer_left) < 3):
+ self.keep_going = False
+ else:
+ # Plot `outer_left` points in pink
+ for point in self.outer_left:
+ x, y = point
+ cv2.circle(self.image, (x, y), radius=5, color=gc('pink'), thickness=-1)
+
+ # Plot `outer_right` points in cyan
+ for point in self.outer_right:
+ x, y = point
+ cv2.circle(self.image, (x, y), radius=5, color=gc('cyan'), thickness=-1)
+
+ ### outer_left
+ self.outer_left = self.order_points(self.outer_left)
+ self.outer_left = self.remove_duplicate_points(self.outer_left)
+ # self.outer_left = self.check_momentum(self.outer_left, False)
+ self.order_points_plot(self.outer_left, 'outer_left', 'final')
+ self.order_points_plot(self.outer_left, 'outer_left', 'QC')
+ self.outer_left_length, self.outer_left = self.get_length_of_ordered_points(self.outer_left, 'outer_left')
+ self.has_outer_left = True
+
+
+ ### outer_right
+ self.outer_right = self.order_points(self.outer_right)
+ self.outer_right = self.remove_duplicate_points(self.outer_right)
+ # self.outer_right = self.check_momentum(self.outer_right, False)
+ self.order_points_plot(self.outer_right, 'outer_right', 'final')
+ self.order_points_plot(self.outer_right, 'outer_right', 'QC')
+ self.outer_right_length, self.outer_right = self.get_length_of_ordered_points(self.outer_right, 'outer_right')
+ self.has_middle = True
+
+ print(f"Length outer_left - {self.outer_left_length}")
+ print(f"Length outer_right - {self.outer_right_length}")
+
+ self.outer_right_np = np.array(self.outer_right)
+ self.outer_left_np = np.array(self.outer_left)
+ self.ordered_middle_np = np.array(self.ordered_middle)
+
+ # Fit 2nd order polynomials to the line segments
+ self.left_poly = np.polyfit(self.outer_left_np[:, 0], self.outer_left_np[:, 1], 2)
+ self.right_poly = np.polyfit(self.outer_right_np[:, 0], self.outer_right_np[:, 1], 2)
+ self.middle_poly = np.polyfit(self.ordered_middle_np[:, 0], self.ordered_middle_np[:, 1], 2)
+
+
+ # Evaluate polynomial coefficients for a range of x values
+ x_range = np.linspace(0, self.width, num=100)
+ left_line = np.polyval(self.left_poly, x_range)
+ right_line = np.polyval(self.right_poly, x_range)
+ self.middle_line = np.polyval(self.middle_poly, x_range)
+
+ # Plot lines of fit as white lines
+ for i in range(len(x_range)-1):
+ cv2.line(self.image, (int(x_range[i]), int(left_line[i])), (int(x_range[i+1]), int(left_line[i+1])), color=gc('gray'), thickness=1)
+ cv2.line(self.image, (int(x_range[i]), int(right_line[i])), (int(x_range[i+1]), int(right_line[i+1])), color=gc('white'), thickness=1)
+ cv2.line(self.image, (int(x_range[i]), int(self.middle_line[i])), (int(x_range[i+1]), int(self.middle_line[i+1])), color=gc('white'), thickness=2)
+
+ # Define the polynomial functions
+ def left_func(x):
+ return self.left_poly[0]*x**2 + self.left_poly[1]*x + self.left_poly[2]
+
+ def right_func(x):
+ return self.right_poly[0]*x**2 + self.right_poly[1]*x + self.right_poly[2]
+
+ def middle_func(x):
+ return self.middle_poly[0]*x**2 + self.middle_poly[1]*x + self.middle_poly[2]
+
+ # Define the difference functions
+ def left_middle_diff(x):
+ return left_func(x) - middle_func(x)
+
+ def right_middle_diff(x):
+ return right_func(x) - middle_func(x)
+
+ def left_right_diff(x):
+ return left_func(x) - right_func(x)
+
+ # Find the intersection points
+ left_middle_intersection_x = fsolve(left_middle_diff, 0)
+ right_middle_intersection_x = fsolve(right_middle_diff, 0)
+ left_right_intersection_x = fsolve(left_right_diff, 0)
+
+ left_middle_intersection_y = left_func(left_middle_intersection_x)[0]
+ right_middle_intersection_y = right_func(right_middle_intersection_x)[0]
+ left_right_intersection_y = left_func(left_right_intersection_x)[0]
+
+ # Keep only points within the image boundaries
+ intersection_points = np.array([[left_middle_intersection_x, left_middle_intersection_y], [right_middle_intersection_x, right_middle_intersection_y], [left_right_intersection_x, left_right_intersection_y]])
+ intersection_points = intersection_points[(intersection_points[:, 0] >= 0) & (intersection_points[:, 0] <= self.width) & (intersection_points[:, 1] >= 0) & (intersection_points[:, 1] <= self.height)]
+
+ if intersection_points.size == 0:
+ self.keep_going = False
+ else:
+ # Compute the average of the intersection points
+ intersection_x = np.mean(intersection_points[:, 0])
+ intersection_y = np.mean(intersection_points[:, 1])
+
+ self.inter_point = [int(intersection_x), int(intersection_y)]
+ self.inter_point_outer_inner = [int(left_right_intersection_x), int(left_right_intersection_y)]
+
+ # Draw intersection point on the image
+ cv2.circle(self.image, (int(intersection_x), int(intersection_y)), radius=5, color=gc('green'), thickness=-1)
+ print(f"Length outer_left - {self.outer_left_length}")
+ print(f"Length outer_right - {self.outer_right_length}")
+ print(f"Intersection point - ({int(intersection_x)}, {int(intersection_y)})")
+
+ # Make the first points be at the tip, last points far away at base
+ def reorder_segment(segment, inter):
+ # Convert to numpy arrays for easier manipulation
+ segment = np.array(segment)
+ inter = np.array(inter)
+
+ # Calculate the Euclidean distance from the INTER point to the first and last points in the segment
+ dist_first = np.linalg.norm(segment[0] - inter)
+ dist_last = np.linalg.norm(segment[-1] - inter)
+
+ # If the last point is closer to the INTER point than the first point, reverse the order of the segment
+ if dist_last < dist_first:
+ segment = segment[::-1]
+
+ return segment.tolist()
+
+ self.ordered_middle = reorder_segment(self.ordered_middle, self.inter_point)
+ self.outer_left = reorder_segment(self.outer_left, self.inter_point)
+ self.outer_right = reorder_segment(self.outer_right, self.inter_point)
+
+ self.ordered_outer_right_np = np.array(self.outer_right)
+ self.ordered_outer_left_np = np.array(self.outer_left)
+ self.ordered_middle_np = np.array(self.ordered_middle)
+
+ # Draw a black ring around the last point of the outer_left segment
+ self.last_point_left = self.outer_left[-1]
+ cv2.circle(self.image, (int(self.last_point_left[0]), int(self.last_point_left[1])), radius=4, color=gc('black'), thickness=2)
+ cv2.circle(self.image, (int(self.last_point_left[0]), int(self.last_point_left[1])), radius=6, color=gc('white'), thickness=2)
+
+ # Draw a black ring around the last point of the outer_right segment
+ self.last_point_right = self.outer_right[-1]
+ cv2.circle(self.image, (int(self.last_point_right[0]), int(self.last_point_right[1])), radius=4, color=gc('black'), thickness=2)
+ cv2.circle(self.image, (int(self.last_point_right[0]), int(self.last_point_right[1])), radius=6, color=gc('white'), thickness=2)
+
+ # self.show_QC_image()
+ # print('hi')
+
+
+
+
+ def split_image_by_middle(self):
+
+ if not self.has_middle:
+ self.keep_going = False
+ else:
+ n_fit = 2
+
+ # Convert the points to a numpy array
+ points_arr = np.array(self.ordered_middle)
+
+ # Fit a line to the points
+ self.midvein_fit = np.polyfit(points_arr[:, 0], points_arr[:, 1], n_fit)
+
+ # Plot a sample of points from along the line
+ max_dim = max(self.height, self.width)
+ if max_dim < 400:
+ num_points = 40
+ elif max_dim < 1000:
+ num_points = 80
+ else:
+ num_points = 120
+
+ # Get the endpoints of the line segment that lies within the bounds of the image
+ x1 = 0
+ y1 = int(self.midvein_fit[0] * x1**2 + self.midvein_fit[1] * x1 + self.midvein_fit[2])
+ x2 = self.width - 1
+ y2 = int(self.midvein_fit[0] * x2**2 + self.midvein_fit[1] * x2 + self.midvein_fit[2])
+
+ denom = self.midvein_fit[0]
+ if denom == 0:
+ denom = 0.0000000001
+ if y1 < 0:
+ y1 = 0
+ x1 = int((y1 - self.midvein_fit[1]) / denom)
+ if y2 >= self.height:
+ y2 = self.height - 1
+ x2 = int((y2 - self.midvein_fit[1]) / denom)
+
+ # Sample num_points points along the line segment within the bounds of the image
+ x_vals = np.linspace(x1, x2, num_points)
+ y_vals = self.midvein_fit[0] * x_vals**2 + self.midvein_fit[1] * x_vals + self.midvein_fit[2]
+
+ # Remove any points that are outside the bounds of the image
+ indices = np.where((y_vals >= 0) & (y_vals < self.height))[0]
+ x_vals = x_vals[indices]
+ y_vals = y_vals[indices]
+
+ # Recompute y-values using the line equation and updated x-values
+ y_vals = self.midvein_fit[0] * x_vals + self.midvein_fit[1]
+
+ self.midvein_fit_points = np.column_stack((x_vals, y_vals))
+ self.is_split = True
+
+ # Draw line of fit
+ # for point in self.midvein_fit_points:
+ # cv2.circle(self.image, tuple(point.astype(int)), radius=1, color=(255, 255, 255), thickness=-1)
+
+ def predict_y(self, x):
+ return self.midvein_fit[0] * x**2 + self.midvein_fit[1] * x + self.midvein_fit[2]
+
+ def order_middle(self):
+
+
+ if 'middle' not in self.points_list:
+ self.keep_going = False
+ else:
+ if len(self.points_list['middle']) >= 5:
+ self.logger.debug(f"Ordered Middle - Raw list contains {len(self.points_list['middle'])} points - using momentum")
+ self.ordered_middle = self.order_points(self.points_list['middle'])
+ self.ordered_middle = self.remove_duplicate_points(self.ordered_middle)
+
+ self.ordered_middle = self.check_momentum(self.ordered_middle, False)
+
+ self.v_tip = self.find_v_tip(self.points_list['outer'])
+ # self.ordered_middle.append(self.v_tip)
+
+
+ self.order_points_plot(self.ordered_middle, 'middle', 'QC')
+ self.ordered_middle_length, self.ordered_middle = self.get_length_of_ordered_points(self.ordered_middle, 'middle')
+
+
+ self.has_middle = True
+ else:
+ self.keep_going = False
+ self.logger.debug(f"Ordered Middle - Raw list contains {len(self.points_list['middle'])} points - SKIPPING MIDDLE")
+
+ def v_shape_template(self, tip, scale):
+ return np.array([
+ [tip[0] - scale, tip[1] + scale],
+ tip,
+ [tip[0] + scale, tip[1] + scale]
+ ])
+
+ def error_function(self, params, points):
+ tip = params[:2]
+ scale = params[2]
+ template_points = self.v_shape_template(tip, scale)
+
+ error = 0
+ for p in points:
+ dist = np.min(np.linalg.norm(template_points - p, axis=1))
+ error += dist
+
+ return error
+
+ def find_v_tip(self, points):
+ points = np.array(points)
+ initial_guess = np.mean(points, axis=0)
+ initial_scale = np.linalg.norm(np.max(points, axis=0) - np.min(points, axis=0)) / 2
+
+ result = minimize(
+ self.error_function,
+ np.hstack([initial_guess, initial_scale]),
+ args=(points,),
+ method='Nelder-Mead'
+ )
+
+ tip = result.x[:2]
+ return tuple(map(int, tip))
+
+ def show_QC_image(self):
+ if self.do_show_QC_images:
+ cv2.imshow('QC image', self.image)
+ cv2.waitKey(0)
+
+ def show_angle_image(self):
+ if self.do_show_QC_images:
+ cv2.imshow('Angles image', self.image_angles)
+ cv2.waitKey(0)
+
+ def show_final_image(self):
+ if self.do_show_final_images:
+ cv2.imshow('Final image', self.image_final)
+ cv2.waitKey(0)
+
+ def get_length_of_ordered_points(self, points, name):
+ # if self.file_name == 'B_774373631_Ebenaceae_Diospyros_buxifolia__L__438-687-578-774':
+ # print('hi')
+ total_length = 0
+ total_length_first_pass = 0
+ for i in range(len(points) - 1):
+ x1, y1 = points[i]
+ x2, y2 = points[i+1]
+ segment_length = math.sqrt((x2-x1)**2 + (y2-y1)**2)
+ total_length_first_pass += segment_length
+ cutoff = total_length_first_pass / 2
+ # print(f'Total length of {name}: {total_length_first_pass}')
+ # print(f'points length {len(points)}')
+ self.logger.debug(f"Total length of {name}: {total_length_first_pass}")
+ self.logger.debug(f"Points length {len(points)}")
+
+
+ # If there are more than 2 points, this will exclude extreme outliers, or
+ # misordered points that don't belong
+ if len(points) > 2:
+ pop_ind = []
+ for i in range(len(points) - 1):
+ x1, y1 = points[i]
+ x2, y2 = points[i+1]
+ segment_length = math.sqrt((x2-x1)**2 + (y2-y1)**2)
+ if segment_length < cutoff:
+ total_length += segment_length
+ else:
+ pop_ind.append(i)
+
+ for exclude in pop_ind:
+ points.pop(exclude)
+ # print(f'Total length of {name}: {total_length}')
+ # print(f'Excluded {len(pop_ind)} points')
+ # print(f'points length {len(points)}')
+ self.logger.debug(f"Total length of {name}: {total_length}")
+ self.logger.debug(f"Excluded {len(pop_ind)} points")
+ self.logger.debug(f"Points length {len(points)}")
+
+ else:
+ total_length = total_length_first_pass
+
+ return total_length, points
+
+ def order_points_plot(self, points, version, QC_or_final):
+ # thk_base = 0
+ thk_base = 16
+
+ if version == 'middle':
+ # color = (0, 255, 0)
+ color = gc('green') # blue
+ thick = 1 #2 + thk_base
+ elif version == 'tip':
+ color = gc('green')
+ thick = 1 #2 + thk_base
+ elif version == 'outer':
+ color = gc('red')
+ thick = 1 #2 + thk_base
+ elif version == 'outer_left':
+ color = gc('pink')
+ thick = 1 #2 + thk_base
+ elif version == 'outer_right':
+ color = gc('cyan')
+ thick = 1 #2 + thk_base
+
+
+ # elif version == 'lamina_width_alt':
+ # color = (100, 100, 255)
+ # thick = 2 + thk_base
+ # elif version == 'not_reflex':
+ # color = (200, 0, 123)
+ # thick = 3 + thk_base
+ # elif version == 'reflex':
+ # color = (0, 120, 200)
+ # thick = 3 + thk_base
+ # elif version == 'petiole_tip_alt':
+ # color = (255, 55, 100)
+ # thick = 1 + thk_base
+ # elif version == 'petiole_tip':
+ # color = (100, 255, 55)
+ # thick = 1 + thk_base
+ # elif version == 'failed_angle':
+ # color = (0, 0, 0)
+ # thick = 3 + thk_base
+ # Convert the points to a numpy array and round to integer values
+ points_arr = np.round(np.array(points)).astype(int)
+
+ # Draw a green line connecting all of the points
+ if QC_or_final == 'QC':
+ for i in range(len(points_arr) - 1):
+ cv2.line(self.image, tuple(points_arr[i]), tuple(points_arr[i+1]), color, thick)
+ else:
+ for i in range(len(points_arr) - 1):
+ cv2.line(self.image_final, tuple(points_arr[i]), tuple(points_arr[i+1]), color, thick)
+
+ def check_momentum(self, coords, info):
+ original_coords = coords
+ # find middle index of coordinates
+ mid_idx = len(coords) // 2
+
+ # set up variables for running average
+ running_avg = np.array(coords[mid_idx-1])
+ avg_count = 1
+
+ # iterate over coordinates to check momentum change
+ prev_vec = np.array(coords[mid_idx-1]) - np.array(coords[mid_idx-2])
+ cur_idx = mid_idx - 1
+ while cur_idx >= 0:
+ cur_vec = np.array(coords[cur_idx]) - np.array(coords[cur_idx-1])
+
+ # add current point to running average
+ running_avg = (running_avg * avg_count + np.array(coords[cur_idx])) / (avg_count + 1)
+ avg_count += 1
+
+ # check for momentum change
+ if self.check_momentum_change(prev_vec, cur_vec):
+ break
+
+ prev_vec = cur_vec
+ cur_idx -= 1
+
+ # use running average to check for momentum change
+ cur_vec = np.array(coords[cur_idx]) - running_avg
+ if self.check_momentum_change(prev_vec, cur_vec):
+ cur_idx += 1
+
+ prev_vec = np.array(coords[mid_idx+1]) - np.array(coords[mid_idx])
+ cur_idx2 = mid_idx + 1
+ while cur_idx2 < len(coords):
+
+ # check if current index is out of range
+ if cur_idx2 >= len(coords):
+ break
+
+ cur_vec = np.array(coords[cur_idx2]) - np.array(coords[cur_idx2-1])
+
+ # add current point to running average
+ running_avg = (running_avg * avg_count + np.array(coords[cur_idx2])) / (avg_count + 1)
+ avg_count += 1
+
+ # check for momentum change
+ if self.check_momentum_change(prev_vec, cur_vec):
+ break
+
+ prev_vec = cur_vec
+ cur_idx2 += 1
+
+ # use running average to check for momentum change
+ if cur_idx2 < len(coords):
+ cur_vec = np.array(coords[cur_idx2]) - running_avg
+ if self.check_momentum_change(prev_vec, cur_vec):
+ cur_idx2 -= 1
+
+ # remove problematic points and subsequent points from list of coordinates
+ new_coords = coords[:cur_idx2] + coords[mid_idx:cur_idx2:-1]
+ if info:
+ return new_coords, len(original_coords) != len(new_coords)
+ else:
+ return new_coords
+
+ # define function to check for momentum change
+ def check_momentum_change(self, prev_vec, cur_vec):
+ dot_product = np.dot(prev_vec, cur_vec)
+ prev_norm = np.linalg.norm(prev_vec)
+ cur_norm = np.linalg.norm(cur_vec)
+ denom = (prev_norm * cur_norm)
+ if denom == 0:
+ denom = 0.0000000001
+ cos_theta = dot_product / denom
+ theta = np.arccos(cos_theta)
+ return abs(theta) > np.pi / 2
+
+ def remove_duplicate_points(self, points):
+ unique_set = set()
+ new_list = []
+
+ for item in points:
+ if item not in unique_set:
+ unique_set.add(item)
+ new_list.append(item)
+ return new_list
+
+ def distance(self, point1, point2):
+ x1, y1 = point1
+ x2, y2 = point2
+ return math.sqrt((x2 - x1) ** 2 + (y2 - y1) ** 2)
+
+ ### Shortest distance
+ def order_points(self, points):
+ points = [tuple(point) for point in points] # Convert numpy.ndarray points to tuples
+
+ best_tour = None
+ shortest_tour_length = float('inf')
+
+ for start_point in points:
+ tour = [start_point]
+ unvisited = set(points) - {start_point}
+
+ while unvisited:
+ nearest = min(unvisited, key=lambda point: self.distance(tour[-1], point))
+ tour.append(nearest)
+ unvisited.remove(nearest)
+
+ # Calculate the length of the current tour
+ tour_length = sum(self.distance(tour[i - 1], tour[i]) for i in range(1, len(tour)))
+
+ # Update the best_tour if the current tour is shorter
+ if tour_length < shortest_tour_length:
+ shortest_tour_length = tour_length
+ best_tour = tour
+
+ return best_tour
+
+
+ ### Smoothest
+ '''
+ def angle_between_points(self, p1, p2, p3):
+ v1 = np.array([p1[0] - p2[0], p1[1] - p2[1]])
+ v2 = np.array([p3[0] - p2[0], p3[1] - p2[1]])
+ angle = np.arccos(np.dot(v1, v2) / (np.linalg.norm(v1) * np.linalg.norm(v2)))
+ return angle
+
+ def order_points(self, points):
+ points = [tuple(point) for point in points] # Convert numpy.ndarray points to tuples
+
+ best_tour = None
+ largest_sum_angles = 0
+
+ for start_point in points:
+ tour = [start_point]
+ unvisited = set(points) - {start_point}
+
+ while unvisited:
+ nearest = min(unvisited, key=lambda point: self.distance(tour[-1], point))
+ tour.append(nearest)
+ unvisited.remove(nearest)
+
+ # Calculate the sum of angles for the current tour
+ sum_angles = sum(self.angle_between_points(tour[i - 1], tour[i], tour[i + 1]) for i in range(1, len(tour) - 1))
+
+ # Update the best_tour if the current tour has a larger sum of angles
+ if sum_angles > largest_sum_angles:
+ largest_sum_angles = sum_angles
+ best_tour = tour
+
+ return best_tour
+ '''
+ ### ^^^ Smoothest
+
+
+
+
+ def convert_YOLO_bbox_to_point(self):
+ for point_type, bbox in self.points_list.items():
+ xy_points = []
+ for point in bbox:
+ x = point[0]
+ y = point[1]
+ w = point[2]
+ h = point[3]
+ x1 = int((x - w/2) * self.width)
+ y1 = int((y - h/2) * self.height)
+ x2 = int((x + w/2) * self.width)
+ y2 = int((y + h/2) * self.height)
+ xy_points.append((int((x1+x2)/2), int((y1+y2)/2)))
+ self.points_list[point_type] = xy_points
+
+ def parse_all_points(self):
+ points_list = {}
+
+ for sublist in self.all_points:
+ key = sublist[0]
+ value = sublist[1:]
+
+ key = self.swap_number_for_string(key)
+
+ if key not in points_list:
+ points_list[key] = []
+ points_list[key].append(value)
+
+ # print(points_list)
+ self.points_list = points_list
+
+ def swap_number_for_string(self, key):
+ for k, v in self.classes.items():
+ if v == key:
+ return k
+ return key
+
+ def setup_final_image(self):
+ self.image_final = cv2.imread(os.path.join(self.dir_temp, '.'.join([self.file_name, 'jpg'])))
+
+ if self.leaf_type == 'Landmarks_Armature':
+ self.path_image_final = os.path.join(self.Dirs.landmarks_armature_overlay_final, '.'.join([self.file_name, 'jpg']))
+
+ def setup_QC_image(self):
+ self.image = cv2.imread(os.path.join(self.dir_temp, '.'.join([self.file_name, 'jpg'])))
+
+ if self.leaf_type == 'Landmarks_Armature':
+ self.path_QC_image = os.path.join(self.Dirs.landmarks_armature_overlay_QC, '.'.join([self.file_name, 'jpg']))
+
+ def setup_angle_image(self):
+ self.image_angles = cv2.imread(os.path.join(self.dir_temp, '.'.join([self.file_name, 'jpg'])))
+
+ if self.leaf_type == 'Landmarks_Armature':
+ self.path_angles_image = os.path.join(self.Dirs.landmarks_armature_overlay_angles, '.'.join([self.file_name, 'jpg']))
+
+ def define_landmark_classes(self):
+ self.classes = {
+ 'tip': 0,
+ 'middle': 1,
+ 'outer': 2,
+ }
+
+ def set_cfg_values(self):
+ self.do_show_QC_images = self.cfg['leafmachine']['landmark_detector_armature']['do_show_QC_images']
+ self.do_save_QC_images = self.cfg['leafmachine']['landmark_detector_armature']['do_save_QC_images']
+ self.do_show_final_images = self.cfg['leafmachine']['landmark_detector_armature']['do_show_final_images']
+ self.do_save_final_images = self.cfg['leafmachine']['landmark_detector_armature']['do_save_final_images']
+
+ def init_lists_dicts(self):
+ # Initialize all lists and dictionaries
+ self.classes = {}
+ self.points_list = []
+ self.image = []
+
+
+ self.ordered_middle = []
+
+ self.midvein_fit = []
+ self.midvein_fit_points = []
+
+ self.outer_right = []
+ self.outer_left = []
+
+ # self.ordered_outer_left = []
+ # self.ordered_outer_right = []
+
+ self.tip = []
+
+ self.apex_left = []
+ self.apex_right = []
+ self.apex_center = []
+
+
+ self.base_left = []
+ self.base_right = []
+ self.base_center = []
+ self.lamina_base = []
+ self.width_left = []
+ self.width_right = []
+
+ def get_final(self):
+ self.image_final = np.hstack((self.image, self.image_angles))
+ return self.image_final
+
+def euclidean_distance(p1, p2):
+ return math.sqrt((p1[0] - p2[0])**2 + (p1[1] - p2[1])**2)
+
+def gc(color):
+ colors = {
+ 'red': (0, 0, 255),
+ 'green': (0, 255, 0),
+ 'blue': (255, 0, 0),
+ 'yellow': (0, 255, 255),
+ 'pink': (255, 0, 255),
+ 'cyan': (255, 255, 0),
+ 'black': (0, 0, 0),
+ 'white': (255, 255, 255),
+ 'gray': (128, 128, 128),
+ 'orange': (0, 165, 255),
+ 'purple': (128, 0, 128),
+ 'lightpink': (203, 192, 255),
+ 'brown': (42, 42, 165),
+ 'navy': (128, 0, 0),
+ 'teal': (128, 128, 0),
+ }
+ return colors.get(color.lower(), (0, 0, 0))
diff --git a/vouchervision/component_detector/color_profiles/ColorProfile__LANDMARK.csv b/vouchervision/component_detector/color_profiles/ColorProfile__LANDMARK.csv
new file mode 100644
index 0000000000000000000000000000000000000000..5230fb05996ec4cabce28e4aa6b8b2d0bd512a40
--- /dev/null
+++ b/vouchervision/component_detector/color_profiles/ColorProfile__LANDMARK.csv
@@ -0,0 +1,9 @@
+apex_angle,255,178,54
+base_angle,228,255,54
+lamina_base,52,240,233
+lamina_tip,246,33,255
+lamina_width,35,44,255
+lobe_tip,229,237,46
+midvein_trace,246,33,255
+petiole_tip,255,33,42
+petiole_trace,255,33,42
diff --git a/vouchervision/component_detector/color_profiles/ColorProfile__LANDMARK_ARM.csv b/vouchervision/component_detector/color_profiles/ColorProfile__LANDMARK_ARM.csv
new file mode 100644
index 0000000000000000000000000000000000000000..613bade2ece8190b52b7cb6280cac2f958e78a3a
--- /dev/null
+++ b/vouchervision/component_detector/color_profiles/ColorProfile__LANDMARK_ARM.csv
@@ -0,0 +1,4 @@
+tip,35,44,255
+middle,228,255,54
+outer,52,240,233
+,,,
diff --git a/vouchervision/component_detector/color_profiles/ColorProfile__PLANT.csv b/vouchervision/component_detector/color_profiles/ColorProfile__PLANT.csv
new file mode 100644
index 0000000000000000000000000000000000000000..90cae8ca1c6728a08a40fa275d086ae7024b0f7b
--- /dev/null
+++ b/vouchervision/component_detector/color_profiles/ColorProfile__PLANT.csv
@@ -0,0 +1,11 @@
+Leaf_WHOLE,0,255,55,00ff37
+Leaf_PARTIAL,0,255,250,69fffc
+Leaflet,255,203,0,ffcb00
+Seed_Fruit_ONE,252,255,0,fcff00
+Seed_Fruit_MANY,0,0,0,0
+Flower_ONE,255,52,255,ff34ff
+Flower_MANY,154,0,255,9a00ff
+Bud,255,0,9,ff0009
+Specimen,0,0,0,ceffc4
+Roots,255,134,0,ff8600
+Wood,144,22,22,901616
diff --git a/vouchervision/component_detector/color_profiles/ColorProfile__PREP.csv b/vouchervision/component_detector/color_profiles/ColorProfile__PREP.csv
new file mode 100644
index 0000000000000000000000000000000000000000..80621d2047808b76ad8afeb2922bee81c5aeb341
--- /dev/null
+++ b/vouchervision/component_detector/color_profiles/ColorProfile__PREP.csv
@@ -0,0 +1,9 @@
+Ruler, 255,0,70
+Barcode, 0,137,65
+Colorcard, 242,255,0
+Label, 0,0,255
+Map, 0,251,255
+Envelope, 163,0,89
+Photo, 255,205,220
+Attached Item, 255,172,40
+Weights, 140,140,140
diff --git a/vouchervision/component_detector/component_detector.py b/vouchervision/component_detector/component_detector.py
new file mode 100644
index 0000000000000000000000000000000000000000..f187cf0b95df15cd5723548020ce167baea35f06
--- /dev/null
+++ b/vouchervision/component_detector/component_detector.py
@@ -0,0 +1,1110 @@
+import os, sys, inspect, json, shutil, cv2, time, glob #imagesize
+import pandas as pd
+import matplotlib.pyplot as plt
+from matplotlib.backends.backend_pdf import PdfPages
+from PIL import Image
+from tqdm import tqdm
+from time import perf_counter
+import concurrent.futures
+from threading import Lock
+from collections import defaultdict
+import multiprocessing
+import torch
+
+currentdir = os.path.dirname(inspect.getfile(inspect.currentframe()))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(currentdir)
+from detect import run
+sys.path.append(parentdir)
+from landmark_processing import LeafSkeleton
+from armature_processing import ArmatureSkeleton
+
+def detect_plant_components(cfg, logger, dir_home, Project, Dirs):
+ t1_start = perf_counter()
+ logger.name = 'Locating Plant Components'
+ logger.info(f"Detecting plant components in {len(os.listdir(Project.dir_images))} images")
+
+ try:
+ dir_exisiting_labels = cfg['leafmachine']['project']['use_existing_plant_component_detections']
+ except:
+ dir_exisiting_labels = None
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 1
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ # Weights folder base
+ dir_weights = os.path.join(dir_home, 'leafmachine2', 'component_detector','runs','train')
+
+ # Detection threshold
+ threshold = cfg['leafmachine']['plant_component_detector']['minimum_confidence_threshold']
+
+ detector_version = cfg['leafmachine']['plant_component_detector']['detector_version']
+ detector_iteration = cfg['leafmachine']['plant_component_detector']['detector_iteration']
+ detector_weights = cfg['leafmachine']['plant_component_detector']['detector_weights']
+ weights = os.path.join(dir_weights,'Plant_Detector',detector_version,detector_iteration,'weights',detector_weights)
+
+ do_save_prediction_overlay_images = not cfg['leafmachine']['plant_component_detector']['do_save_prediction_overlay_images']
+ ignore_objects = cfg['leafmachine']['plant_component_detector']['ignore_objects_for_overlay']
+ ignore_objects = ignore_objects or []
+
+ if dir_exisiting_labels != None:
+ logger.info("Loading existing plant labels")
+ fetch_labels(dir_exisiting_labels, os.path.join(Dirs.path_plant_components, 'labels'))
+ if len(Project.dir_images) <= 4000:
+ logger.debug("Single-threaded create_dictionary_from_txt() len(Project.dir_images) <= 4000")
+ A = create_dictionary_from_txt(logger, dir_exisiting_labels, 'Detections_Plant_Components', Project)
+ else:
+ logger.debug(f"Multi-threaded with ({str(cfg['leafmachine']['project']['num_workers'])}) threads create_dictionary_from_txt() len(Project.dir_images) > 4000")
+ A = create_dictionary_from_txt_parallel(logger, cfg, dir_exisiting_labels, 'Detections_Plant_Components', Project)
+
+ else:
+ logger.info("Running YOLOv5 to generate plant labels")
+ # run(weights = weights,
+ # source = Project.dir_images,
+ # project = Dirs.path_plant_components,
+ # name = Dirs.run_name,
+ # imgsz = (1280, 1280),
+ # nosave = do_save_prediction_overlay_images,
+ # anno_type = 'Plant_Detector',
+ # conf_thres = threshold,
+ # ignore_objects_for_overlay = ignore_objects,
+ # mode = 'LM2',
+ # LOGGER=logger,)
+ source = Project.dir_images
+ project = Dirs.path_plant_components
+ name = Dirs.run_name
+ imgsz = (1280, 1280)
+ nosave = do_save_prediction_overlay_images
+ anno_type = 'Plant_Detector'
+ conf_thres = threshold
+ ignore_objects_for_overlay = ignore_objects
+ mode = 'LM2'
+ LOGGER = logger
+
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = [executor.submit(run_in_parallel, weights, source, project, name, imgsz, nosave, anno_type,
+ conf_thres, 10, ignore_objects_for_overlay, mode, LOGGER, i, num_workers) for i in
+ range(num_workers)]
+ for future in concurrent.futures.as_completed(futures):
+ try:
+ _ = future.result()
+ except Exception as e:
+ logger.error(f'Error in thread: {e}')
+ continue
+
+ t2_stop = perf_counter()
+ logger.info(f"[Plant components detection elapsed time] {round(t2_stop - t1_start)} seconds")
+ logger.info(f"Threads [{num_workers}]")
+
+ if len(Project.dir_images) <= 4000:
+ logger.debug("Single-threaded create_dictionary_from_txt() len(Project.dir_images) <= 4000")
+ A = create_dictionary_from_txt(logger, os.path.join(Dirs.path_plant_components, 'labels'), 'Detections_Plant_Components', Project)
+ else:
+ logger.debug(f"Multi-threaded with ({str(cfg['leafmachine']['project']['num_workers'])}) threads create_dictionary_from_txt() len(Project.dir_images) > 4000")
+ A = create_dictionary_from_txt_parallel(logger, cfg, os.path.join(Dirs.path_plant_components, 'labels'), 'Detections_Plant_Components', Project)
+
+ dict_to_json(Project.project_data, Dirs.path_plant_components, 'Detections_Plant_Components.json')
+
+ t1_stop = perf_counter()
+ logger.info(f"[Processing plant components elapsed time] {round(t1_stop - t1_start)} seconds")
+ torch.cuda.empty_cache()
+ return Project
+
+
+def detect_archival_components(cfg, logger, dir_home, Project, Dirs):
+ if not cfg['leafmachine']['use_RGB_label_images']:
+ logger.name = 'Skipping LeafMachine2 Label Detection'
+ logger.info(f"Full image will be used instead of the label collage")
+ else:
+ t1_start = perf_counter()
+ logger.name = 'Locating Archival Components'
+ logger.info(f"Detecting archival components in {len(os.listdir(Project.dir_images))} images")
+
+
+ try:
+ dir_exisiting_labels = cfg['leafmachine']['project']['use_existing_archival_component_detections']
+ except:
+ dir_exisiting_labels = None
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 1
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ # Weights folder base
+ dir_weights = os.path.join(dir_home, 'leafmachine2', 'component_detector','runs','train')
+
+ # Detection threshold
+ threshold = cfg['leafmachine']['archival_component_detector']['minimum_confidence_threshold']
+
+ detector_version = cfg['leafmachine']['archival_component_detector']['detector_version']
+ detector_iteration = cfg['leafmachine']['archival_component_detector']['detector_iteration']
+ detector_weights = cfg['leafmachine']['archival_component_detector']['detector_weights']
+ weights = os.path.join(dir_weights,'Archival_Detector',detector_version,detector_iteration,'weights',detector_weights)
+
+ do_save_prediction_overlay_images = not cfg['leafmachine']['archival_component_detector']['do_save_prediction_overlay_images']
+ ignore_objects = cfg['leafmachine']['archival_component_detector']['ignore_objects_for_overlay']
+ ignore_objects = ignore_objects or []
+
+
+ if dir_exisiting_labels != None:
+ logger.info("Loading existing archival labels")
+ fetch_labels(dir_exisiting_labels, os.path.join(Dirs.path_archival_components, 'labels'))
+ if len(Project.dir_images) <= 4000:
+ logger.debug("Single-threaded create_dictionary_from_txt() len(Project.dir_images) <= 4000")
+ A = create_dictionary_from_txt(logger, dir_exisiting_labels, 'Detections_Archival_Components', Project)
+ else:
+ logger.debug(f"Multi-threaded with ({str(cfg['leafmachine']['project']['num_workers'])}) threads create_dictionary_from_txt() len(Project.dir_images) > 4000")
+ A = create_dictionary_from_txt_parallel(logger, cfg, dir_exisiting_labels, 'Detections_Archival_Components', Project)
+
+ else:
+ logger.info("Running YOLOv5 to generate archival labels")
+ # run(weights = weights,
+ # source = Project.dir_images,
+ # project = Dirs.path_archival_components,
+ # name = Dirs.run_name,
+ # imgsz = (1280, 1280),
+ # nosave = do_save_prediction_overlay_images,
+ # anno_type = 'Archival_Detector',
+ # conf_thres = threshold,
+ # ignore_objects_for_overlay = ignore_objects,
+ # mode = 'LM2',
+ # LOGGER=logger)
+ # split the image paths into 4 chunks
+ source = Project.dir_images
+ project = Dirs.path_archival_components
+ name = Dirs.run_name
+ imgsz = (1280, 1280)
+ nosave = do_save_prediction_overlay_images
+ anno_type = 'Archival_Detector'
+ conf_thres = threshold
+ ignore_objects_for_overlay = ignore_objects
+ mode = 'LM2'
+ LOGGER = logger
+
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = [executor.submit(run_in_parallel, weights, source, project, name, imgsz, nosave, anno_type,
+ conf_thres, 10, ignore_objects_for_overlay, mode, LOGGER, i, num_workers) for i in
+ range(num_workers)]
+ for future in concurrent.futures.as_completed(futures):
+ try:
+ _ = future.result()
+ except Exception as e:
+ logger.error(f'Error in thread: {e}')
+ continue
+
+ t2_stop = perf_counter()
+ logger.info(f"[Archival components detection elapsed time] {round(t2_stop - t1_start)} seconds")
+ logger.info(f"Threads [{num_workers}]")
+
+ if len(Project.dir_images) <= 4000:
+ logger.debug("Single-threaded create_dictionary_from_txt() len(Project.dir_images) <= 4000")
+ A = create_dictionary_from_txt(logger, os.path.join(Dirs.path_archival_components, 'labels'), 'Detections_Archival_Components', Project)
+ else:
+ logger.debug(f"Multi-threaded with ({str(cfg['leafmachine']['project']['num_workers'])}) threads create_dictionary_from_txt() len(Project.dir_images) > 4000")
+ A = create_dictionary_from_txt_parallel(logger, cfg, os.path.join(Dirs.path_archival_components, 'labels'), 'Detections_Archival_Components', Project)
+
+ dict_to_json(Project.project_data, Dirs.path_archival_components, 'Detections_Archival_Components.json')
+
+ t1_stop = perf_counter()
+ logger.info(f"[Processing archival components elapsed time] {round(t1_stop - t1_start)} seconds")
+ torch.cuda.empty_cache()
+ return Project
+
+
+def detect_armature_components(cfg, logger, dir_home, Project, Dirs):
+ t1_start = perf_counter()
+ logger.name = 'Locating Armature Components'
+ logger.info(f"Detecting armature components in {len(os.listdir(Project.dir_images))} images")
+
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 1
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ # Weights folder base
+ dir_weights = os.path.join(dir_home, 'leafmachine2', 'component_detector','runs','train')
+
+ # Detection threshold
+ threshold = cfg['leafmachine']['armature_component_detector']['minimum_confidence_threshold']
+
+ detector_version = cfg['leafmachine']['armature_component_detector']['detector_version']
+ detector_iteration = cfg['leafmachine']['armature_component_detector']['detector_iteration']
+ detector_weights = cfg['leafmachine']['armature_component_detector']['detector_weights']
+ weights = os.path.join(dir_weights,'Armature_Detector',detector_version,detector_iteration,'weights',detector_weights)
+
+ do_save_prediction_overlay_images = not cfg['leafmachine']['armature_component_detector']['do_save_prediction_overlay_images']
+ ignore_objects = cfg['leafmachine']['armature_component_detector']['ignore_objects_for_overlay']
+ ignore_objects = ignore_objects or []
+
+ logger.info("Running YOLOv5 to generate armature labels")
+
+ source = Project.dir_images
+ project = Dirs.path_armature_components
+ name = Dirs.run_name
+ imgsz = (1280, 1280)
+ nosave = do_save_prediction_overlay_images
+ anno_type = 'Armature_Detector'
+ conf_thres = threshold
+ ignore_objects_for_overlay = ignore_objects
+ mode = 'LM2'
+ LOGGER = logger
+
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = [executor.submit(run_in_parallel, weights, source, project, name, imgsz, nosave, anno_type,
+ conf_thres, 10, ignore_objects_for_overlay, mode, LOGGER, i, num_workers) for i in
+ range(num_workers)]
+ for future in concurrent.futures.as_completed(futures):
+ try:
+ _ = future.result()
+ except Exception as e:
+ logger.error(f'Error in thread: {e}')
+ continue
+
+ t2_stop = perf_counter()
+ logger.info(f"[Plant components detection elapsed time] {round(t2_stop - t1_start)} seconds")
+ logger.info(f"Threads [{num_workers}]")
+
+ if len(Project.dir_images) <= 4000:
+ logger.debug("Single-threaded create_dictionary_from_txt() len(Project.dir_images) <= 4000")
+ A = create_dictionary_from_txt(logger, os.path.join(Dirs.path_armature_components, 'labels'), 'Detections_Armature_Components', Project)
+ else:
+ logger.debug(f"Multi-threaded with ({str(cfg['leafmachine']['project']['num_workers'])}) threads create_dictionary_from_txt() len(Project.dir_images) > 4000")
+ A = create_dictionary_from_txt_parallel(logger, cfg, os.path.join(Dirs.path_armature_components, 'labels'), 'Detections_Armature_Components', Project)
+
+ dict_to_json(Project.project_data, Dirs.path_armature_components, 'Detections_Armature_Components.json')
+
+ t1_stop = perf_counter()
+ logger.info(f"[Processing armature components elapsed time] {round(t1_stop - t1_start)} seconds")
+ torch.cuda.empty_cache()
+ return Project
+
+
+''' RUN IN PARALLEL'''
+def run_in_parallel(weights, source, project, name, imgsz, nosave, anno_type, conf_thres, line_thickness, ignore_objects_for_overlay, mode, LOGGER, chunk, n_workers):
+ num_files = len(os.listdir(source))
+ LOGGER.info(f"The number of worker threads: ({n_workers}), number of files ({num_files}).")
+
+ chunk_size = len(os.listdir(source)) // n_workers
+ start = chunk * chunk_size
+ end = start + chunk_size if chunk < (n_workers-1) else len(os.listdir(source))
+
+ sub_source = [os.path.join(source, f) for f in os.listdir(source)[start:end] if f.lower().endswith('.jpg')]
+
+ run(weights=weights,
+ source=sub_source,
+ project=project,
+ name=name,
+ imgsz=imgsz,
+ nosave=nosave,
+ anno_type=anno_type,
+ conf_thres=conf_thres,
+ ignore_objects_for_overlay=ignore_objects_for_overlay,
+ mode=mode,
+ LOGGER=LOGGER)
+
+''' RUN IN PARALLEL'''
+
+
+###### Multi-thread NOTE this works, but unless there are several thousand images, it will be slower
+def process_file(logger, file, dir_components, component, Project, lock):
+ file_name = str(file.split('.')[0])
+ with open(os.path.join(dir_components, file), "r") as f:
+ with lock:
+ Project.project_data[file_name][component] = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ try:
+ image_path = glob.glob(os.path.join(Project.dir_images, file_name + '.*'))[0]
+ name_ext = os.path.basename(image_path)
+ with Image.open(image_path) as im:
+ _, ext = os.path.splitext(name_ext)
+ if ext not in ['.jpg']:
+ im = im.convert('RGB')
+ im.save(os.path.join(Project.dir_images, file_name) + '.jpg', quality=100)
+ # file_name += '.jpg'
+ width, height = im.size
+ except Exception as e:
+ print(f"Unable to get image dimensions. Error: {e}")
+ logger.info(f"Unable to get image dimensions. Error: {e}")
+ width, height = None, None
+ if width and height:
+ Project.project_data[file_name]['height'] = int(height)
+ Project.project_data[file_name]['width'] = int(width)
+
+
+def create_dictionary_from_txt_parallel(logger, cfg, dir_components, component, Project):
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 4
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ files = [file for file in os.listdir(dir_components) if file.endswith(".txt")]
+ lock = Lock()
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = []
+ for file in files:
+ futures.append(executor.submit(process_file, logger, file, dir_components, component, Project, lock))
+ for future in concurrent.futures.as_completed(futures):
+ pass
+ return Project.project_data
+
+######
+
+
+
+
+
+# Single threaded
+def create_dictionary_from_txt(logger, dir_components, component, Project):
+ # dict_labels = {}
+ for file in tqdm(os.listdir(dir_components), desc="Loading Annotations", colour='green'):
+ if file.endswith(".txt"):
+ file_name = str(file.split('.')[0])
+ with open(os.path.join(dir_components, file), "r") as f:
+ # dict_labels[file] = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ Project.project_data[file_name][component] = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ try:
+ image_path = glob.glob(os.path.join(Project.dir_images, file_name + '.*'))[0]
+ name_ext = os.path.basename(image_path)
+ with Image.open(image_path) as im:
+ _, ext = os.path.splitext(name_ext)
+ if ext not in ['.jpg']:
+ im = im.convert('RGB')
+ im.save(os.path.join(Project.dir_images, file_name) + '.jpg', quality=100)
+ # file_name += '.jpg'
+ width, height = im.size
+ except Exception as e:
+ # print(f"Unable to get image dimensions. Error: {e}")
+ logger.info(f"Unable to get image dimensions. Error: {e}")
+ width, height = None, None
+ if width and height:
+ Project.project_data[file_name]['height'] = int(height)
+ Project.project_data[file_name]['width'] = int(width)
+ # for key, value in dict_labels.items():
+ # print(f'{key} --> {value}')
+ return Project.project_data
+
+
+# old below
+'''def create_dictionary_from_txt(dir_components, component, Project):
+ # dict_labels = {}
+ for file in os.listdir(dir_components):
+ if file.endswith(".txt"):
+ file_name = str(file.split('.')[0])
+ with open(os.path.join(dir_components, file), "r") as f:
+ # dict_labels[file] = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ Project.project_data[file_name][component] = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ try:
+ width, height = imagesize.get(os.path.join(Project.dir_images, '.'.join([file_name,'jpg'])))
+ except Exception as e:
+ print(f"Image not in 'jpg' format. Trying 'jpeg'. Note that other formats are not supported.{e}")
+ width, height = imagesize.get(os.path.join(Project.dir_images, '.'.join([file_name,'jpeg'])))
+ Project.project_data[file_name]['height'] = int(height)
+ Project.project_data[file_name]['width'] = int(width)
+ # for key, value in dict_labels.items():
+ # print(f'{key} --> {value}')
+ return Project.project_data'''
+
+
+
+def dict_to_json(dict_labels, dir_components, name_json):
+ dir_components = os.path.join(dir_components, 'JSON')
+ with open(os.path.join(dir_components, name_json), "w") as outfile:
+ json.dump(dict_labels, outfile)
+
+def fetch_labels(dir_exisiting_labels, new_dir):
+ shutil.copytree(dir_exisiting_labels, new_dir)
+
+
+'''Landmarks - uses YOLO, but works differently than above. A hybrid between segmentation and component detector'''
+def detect_landmarks(cfg, logger, dir_home, Project, batch, n_batches, Dirs, segmentation_complete):
+ start_t = perf_counter()
+ logger.name = f'[BATCH {batch+1} Detect Landmarks]'
+ logger.info(f'Detecting landmarks for batch {batch+1} of {n_batches}')
+
+ landmark_whole_leaves = cfg['leafmachine']['landmark_detector']['landmark_whole_leaves']
+ landmark_partial_leaves = cfg['leafmachine']['landmark_detector']['landmark_partial_leaves']
+
+ landmarks_whole_leaves_props = {}
+ landmarks_whole_leaves_overlay = {}
+ landmarks_partial_leaves_props = {}
+ landmarks_partial_leaves_overlay = {}
+
+ if landmark_whole_leaves:
+ run_landmarks(cfg, logger, dir_home, Project, batch, n_batches, Dirs, 'Landmarks_Whole_Leaves', segmentation_complete)
+ if landmark_partial_leaves:
+ run_landmarks(cfg, logger, dir_home, Project, batch, n_batches, Dirs, 'Landmarks_Partial_Leaves', segmentation_complete)
+
+ # if cfg['leafmachine']['leaf_segmentation']['segment_whole_leaves']:
+ # landmarks_whole_leaves_props_batch, landmarks_whole_leaves_overlay_batch = run_landmarks(Instance_Detector_Whole, Project.project_data_list[batch], 0,
+ # "Segmentation_Whole_Leaf", "Whole_Leaf_Cropped", cfg, Project, Dirs, batch, n_batches)#, start+1, end)
+ # landmarks_whole_leaves_props.update(landmarks_whole_leaves_props_batch)
+ # landmarks_whole_leaves_overlay.update(landmarks_whole_leaves_overlay_batch)
+ # if cfg['leafmachine']['leaf_segmentation']['segment_partial_leaves']:
+ # landmarks_partial_leaves_props_batch, landmarks_partial_leaves_overlay_batch = run_landmarks(Instance_Detector_Partial, Project.project_data_list[batch], 1,
+ # "Segmentation_Partial_Leaf", "Partial_Leaf_Cropped", cfg, Project, Dirs, batch, n_batches)#, start+1, end)
+ # landmarks_partial_leaves_props.update(landmarks_partial_leaves_props_batch)
+ # landmarks_partial_leaves_overlay.update(landmarks_partial_leaves_overlay_batch)
+
+ end_t = perf_counter()
+ logger.info(f'Batch {batch+1}/{n_batches}: Landmark Detection Duration --> {round((end_t - start_t)/60)} minutes')
+ return Project
+
+
+def detect_armature(cfg, logger, dir_home, Project, batch, n_batches, Dirs, segmentation_complete):
+ start_t = perf_counter()
+ logger.name = f'[BATCH {batch+1} Detect Armature]'
+ logger.info(f'Detecting armature for batch {batch+1} of {n_batches}')
+
+ landmark_armature = cfg['leafmachine']['modules']['armature']
+
+ landmarks_armature_props = {}
+ landmarks_armature_overlay = {}
+
+ if landmark_armature:
+ run_armature(cfg, logger, dir_home, Project, batch, n_batches, Dirs, 'Landmarks_Armature', segmentation_complete)
+
+ end_t = perf_counter()
+ logger.info(f'Batch {batch+1}/{n_batches}: Armature Detection Duration --> {round((end_t - start_t)/60)} minutes')
+ return Project
+
+
+def run_armature(cfg, logger, dir_home, Project, batch, n_batches, Dirs, leaf_type, segmentation_complete):
+
+ logger.info('Detecting armature landmarks from scratch')
+ if leaf_type == 'Landmarks_Armature':
+ dir_overlay = os.path.join(Dirs.landmarks_armature_overlay, ''.join(['batch_',str(batch+1)]))
+
+ # if not segmentation_complete: # If segmentation was run, then don't redo the unpack, just do the crop into the temp folder
+ if leaf_type == 'Landmarks_Armature': # TODO THE 0 is for prickles. For spines I'll need to add a 1 like with partial_leaves or just do it for all
+ Project.project_data_list[batch] = unpack_class_from_components_armature(Project.project_data_list[batch], 0, 'Armature_YOLO', 'Armature_BBoxes', Project)
+ Project.project_data_list[batch], dir_temp = crop_images_to_bbox_armature(Project.project_data_list[batch], 0, 'Armature_Cropped', "Armature_BBoxes", Project, Dirs, True, cfg)
+
+ # Weights folder base
+ dir_weights = os.path.join(dir_home, 'leafmachine2', 'component_detector','runs','train')
+
+ # Detection threshold
+ threshold = cfg['leafmachine']['landmark_detector_armature']['minimum_confidence_threshold']
+
+ detector_version = cfg['leafmachine']['landmark_detector_armature']['detector_version']
+ detector_iteration = cfg['leafmachine']['landmark_detector_armature']['detector_iteration']
+ detector_weights = cfg['leafmachine']['landmark_detector_armature']['detector_weights']
+ weights = os.path.join(dir_weights,'Landmark_Detector_YOLO',detector_version,detector_iteration,'weights',detector_weights)
+
+ do_save_prediction_overlay_images = not cfg['leafmachine']['landmark_detector_armature']['do_save_prediction_overlay_images']
+ ignore_objects = cfg['leafmachine']['landmark_detector_armature']['ignore_objects_for_overlay']
+ ignore_objects = ignore_objects or []
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 1
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ has_images = False
+ if len(os.listdir(dir_temp)) > 0:
+ has_images = True
+ source = dir_temp
+ project = dir_overlay
+ name = Dirs.run_name
+ imgsz = (1280, 1280)
+ nosave = do_save_prediction_overlay_images
+ anno_type = 'Armature_Detector'
+ conf_thres = threshold
+ line_thickness = 2
+ ignore_objects_for_overlay = ignore_objects
+ mode = 'Landmark'
+ LOGGER = logger
+
+ # Initialize a Lock object to ensure thread safety
+ lock = Lock()
+
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = [executor.submit(run_in_parallel, weights, source, project, name, imgsz, nosave, anno_type,
+ conf_thres, line_thickness, ignore_objects_for_overlay, mode, LOGGER, i, num_workers) for i in
+ range(num_workers)]
+ for future in concurrent.futures.as_completed(futures):
+ try:
+ _ = future.result()
+ except Exception as e:
+ logger.error(f'Error in thread: {e}')
+ continue
+
+ with lock:
+ if has_images:
+ dimensions_dict = get_cropped_dimensions(dir_temp)
+ A = add_to_dictionary_from_txt_armature(cfg, logger, Dirs, leaf_type, os.path.join(dir_overlay, 'labels'), leaf_type, Project, dimensions_dict, dir_temp, batch, n_batches)
+ else:
+ # TODO add empty placeholder to the image data
+ pass
+
+ # delete the temp dir
+ try:
+ shutil.rmtree(dir_temp)
+ except:
+ try:
+ time.sleep(5)
+ shutil.rmtree(dir_temp)
+ except:
+ try:
+ time.sleep(5)
+ shutil.rmtree(dir_temp)
+ except:
+ pass
+
+ torch.cuda.empty_cache()
+
+ return Project
+
+
+def run_landmarks(cfg, logger, dir_home, Project, batch, n_batches, Dirs, leaf_type, segmentation_complete):
+ use_existing_landmark_detections = cfg['leafmachine']['landmark_detector']['use_existing_landmark_detections']
+
+ if use_existing_landmark_detections is None:
+ logger.info('Detecting landmarks from scratch')
+ if leaf_type == 'Landmarks_Whole_Leaves':
+ dir_overlay = os.path.join(Dirs.landmarks_whole_leaves_overlay, ''.join(['batch_',str(batch+1)]))
+ elif leaf_type == 'Landmarks_Partial_Leaves':
+ dir_overlay = os.path.join(Dirs.landmarks_partial_leaves_overlay, ''.join(['batch_',str(batch+1)]))
+
+ # if not segmentation_complete: # If segmentation was run, then don't redo the unpack, just do the crop into the temp folder
+ if leaf_type == 'Landmarks_Whole_Leaves':
+ Project.project_data_list[batch] = unpack_class_from_components(Project.project_data_list[batch], 0, 'Whole_Leaf_BBoxes_YOLO', 'Whole_Leaf_BBoxes', Project)
+ Project.project_data_list[batch], dir_temp = crop_images_to_bbox(Project.project_data_list[batch], 0, 'Whole_Leaf_Cropped', "Whole_Leaf_BBoxes", Project, Dirs)
+
+ elif leaf_type == 'Landmarks_Partial_Leaves':
+ Project.project_data_list[batch] = unpack_class_from_components(Project.project_data_list[batch], 1, 'Partial_Leaf_BBoxes_YOLO', 'Partial_Leaf_BBoxes', Project)
+ Project.project_data_list[batch], dir_temp = crop_images_to_bbox(Project.project_data_list[batch], 1, 'Partial_Leaf_Cropped', "Partial_Leaf_BBoxes", Project, Dirs)
+ # else:
+ # if leaf_type == 'Landmarks_Whole_Leaves':
+ # Project.project_data_list[batch], dir_temp = crop_images_to_bbox(Project.project_data_list[batch], 0, 'Whole_Leaf_Cropped', "Whole_Leaf_BBoxes", Project, Dirs)
+ # elif leaf_type == 'Landmarks_Partial_Leaves':
+ # Project.project_data_list[batch], dir_temp = crop_images_to_bbox(Project.project_data_list[batch], 1, 'Partial_Leaf_Cropped', "Partial_Leaf_BBoxes", Project, Dirs)
+
+ # Weights folder base
+ dir_weights = os.path.join(dir_home, 'leafmachine2', 'component_detector','runs','train')
+
+ # Detection threshold
+ threshold = cfg['leafmachine']['landmark_detector']['minimum_confidence_threshold']
+
+ detector_version = cfg['leafmachine']['landmark_detector']['detector_version']
+ detector_iteration = cfg['leafmachine']['landmark_detector']['detector_iteration']
+ detector_weights = cfg['leafmachine']['landmark_detector']['detector_weights']
+ weights = os.path.join(dir_weights,'Landmark_Detector_YOLO',detector_version,detector_iteration,'weights',detector_weights)
+
+ do_save_prediction_overlay_images = not cfg['leafmachine']['landmark_detector']['do_save_prediction_overlay_images']
+ ignore_objects = cfg['leafmachine']['landmark_detector']['ignore_objects_for_overlay']
+ ignore_objects = ignore_objects or []
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 1
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ has_images = False
+ if len(os.listdir(dir_temp)) > 0:
+ has_images = True
+ # run(weights = weights,
+ # source = dir_temp,
+ # project = dir_overlay,
+ # name = Dirs.run_name,
+ # imgsz = (1280, 1280),
+ # nosave = do_save_prediction_overlay_images,
+ # anno_type = 'Landmark_Detector_YOLO',
+ # conf_thres = threshold,
+ # line_thickness = 2,
+ # ignore_objects_for_overlay = ignore_objects,
+ # mode = 'Landmark')
+ source = dir_temp
+ project = dir_overlay
+ name = Dirs.run_name
+ imgsz = (1280, 1280)
+ nosave = do_save_prediction_overlay_images
+ anno_type = 'Landmark_Detector'
+ conf_thres = threshold
+ line_thickness = 2
+ ignore_objects_for_overlay = ignore_objects
+ mode = 'Landmark'
+ LOGGER = logger
+
+ # Initialize a Lock object to ensure thread safety
+ lock = Lock()
+
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = [executor.submit(run_in_parallel, weights, source, project, name, imgsz, nosave, anno_type,
+ conf_thres, line_thickness, ignore_objects_for_overlay, mode, LOGGER, i, num_workers) for i in
+ range(num_workers)]
+ for future in concurrent.futures.as_completed(futures):
+ try:
+ _ = future.result()
+ except Exception as e:
+ logger.error(f'Error in thread: {e}')
+ continue
+
+ with lock:
+ if has_images:
+ dimensions_dict = get_cropped_dimensions(dir_temp)
+ A = add_to_dictionary_from_txt(cfg, logger, Dirs, leaf_type, os.path.join(dir_overlay, 'labels'), leaf_type, Project, dimensions_dict, dir_temp, batch, n_batches)
+ else:
+ # TODO add empty placeholder to the image data
+ pass
+ else:
+ logger.info('Loading existing landmark annotations')
+ dir_temp = os.path.join(use_existing_landmark_detections, f'batch_{str(batch+1)}', 'labels')
+ dimensions_dict = get_cropped_dimensions(dir_temp)
+ A = add_to_dictionary_from_txt(cfg, logger, Dirs, leaf_type, use_existing_landmark_detections, leaf_type, Project, dimensions_dict, dir_temp, batch, n_batches)
+
+
+ # delete the temp dir
+ try:
+ shutil.rmtree(dir_temp)
+ except:
+ try:
+ time.sleep(5)
+ shutil.rmtree(dir_temp)
+ except:
+ try:
+ time.sleep(5)
+ shutil.rmtree(dir_temp)
+ except:
+ pass
+
+ torch.cuda.empty_cache()
+
+ return Project
+ '''def add_to_dictionary_from_txt(cfg, Dirs, leaf_type, dir_components, component, Project, dimensions_dict, dir_temp):
+ # dict_labels = {}
+ for file in os.listdir(dir_components):
+ file_name = str(file.split('.')[0])
+ file_name_parent = file_name.split('__')[0]
+ Project.project_data[file_name_parent][component] = {}
+
+ if file.endswith(".txt"):
+ with open(os.path.join(dir_components, file), "r") as f:
+ all_points = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ Project.project_data[file_name_parent][component][file_name] = all_points
+
+ height = dimensions_dict[file_name][0]
+ width = dimensions_dict[file_name][1]
+
+ Leaf_Skeleton = LeafSkeleton(cfg, Dirs, leaf_type, all_points, height, width, dir_temp, file_name)
+ QC_add = Leaf_Skeleton.get_QC()'''
+
+
+ return Project.project_data
+
+def add_to_dictionary_from_txt_armature(cfg, logger, Dirs, leaf_type, dir_components, component, Project, dimensions_dict, dir_temp, batch, n_batches):
+ dpi = cfg['leafmachine']['overlay']['overlay_dpi']
+ if leaf_type == 'Landmarks_Armature':
+ logger.info(f'Detecting landmarks armature')
+ pdf_path = os.path.join(Dirs.landmarks_armature_overlay_QC, ''.join(['landmarks_armature_overlay_QC__',str(batch+1), 'of', str(n_batches), '.pdf']))
+ pdf_path_final = os.path.join(Dirs.landmarks_armature_overlay_final, ''.join(['landmarks_armature_overlay_final__',str(batch+1), 'of', str(n_batches), '.pdf']))
+
+ ### FINAL
+ # dict_labels = {}
+ fig = plt.figure(figsize=(8.27, 11.69), dpi=dpi) # A4 size, 300 dpi
+ row, col = 0, 0
+ with PdfPages(pdf_path_final) as pdf:
+
+
+
+ for file in os.listdir(dir_components):
+ file_name = str(file.split('.')[0])
+ file_name_parent = file_name.split('__')[0]
+
+ # Project.project_data_list[batch][file_name_parent][component] = []
+
+ if file_name_parent in Project.project_data_list[batch]:
+
+
+
+ if file.endswith(".txt"):
+ with open(os.path.join(dir_components, file), "r") as f:
+ all_points = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ # Project.project_data_list[batch][file_name_parent][component][file_name] = all_points
+
+ height = dimensions_dict[file_name][0]
+ width = dimensions_dict[file_name][1]
+
+ Armature_Skeleton = ArmatureSkeleton(cfg, logger, Dirs, leaf_type, all_points, height, width, dir_temp, file_name)
+ Project = add_armature_skeleton_to_project(cfg, logger, Project, batch, file_name_parent, component, Dirs, leaf_type, all_points, height, width, dir_temp, file_name, Armature_Skeleton)
+ final_add = cv2.cvtColor(Armature_Skeleton.get_final(), cv2.COLOR_BGR2RGB)
+
+ # Add image to the current subplot
+ ax = fig.add_subplot(5, 3, row * 3 + col + 1)
+ ax.imshow(final_add)
+ ax.axis('off')
+
+ col += 1
+ if col == 3:
+ col = 0
+ row += 1
+ if row == 5:
+ row = 0
+ pdf.savefig(fig) # Save the current page
+ fig = plt.figure(figsize=(8.27, 11.69), dpi=300) # Create a new page
+ else:
+ pass
+
+ if row != 0 or col != 0:
+ pdf.savefig(fig) # Save the remaining images on the last page
+
+def add_to_dictionary_from_txt(cfg, logger, Dirs, leaf_type, dir_components, component, Project, dimensions_dict, dir_temp, batch, n_batches):
+ dpi = cfg['leafmachine']['overlay']['overlay_dpi']
+ if leaf_type == 'Landmarks_Whole_Leaves':
+ logger.info(f'Detecting landmarks whole leaves')
+ pdf_path = os.path.join(Dirs.landmarks_whole_leaves_overlay_QC, ''.join(['landmarks_whole_leaves_overlay_QC__',str(batch+1), 'of', str(n_batches), '.pdf']))
+ pdf_path_final = os.path.join(Dirs.landmarks_whole_leaves_overlay_final, ''.join(['landmarks_whole_leaves_overlay_final__',str(batch+1), 'of', str(n_batches), '.pdf']))
+ elif leaf_type == 'Landmarks_Partial_Leaves':
+ logger.info(f'Detecting landmarks partial leaves')
+ pdf_path = os.path.join(Dirs.landmarks_partial_leaves_overlay_QC, ''.join(['landmarks_partial_leaves_overlay_QC__',str(batch+1), 'of', str(n_batches), '.pdf']))
+ pdf_path_final = os.path.join(Dirs.landmarks_partial_leaves_overlay_final, ''.join(['landmarks_partial_leaves_overlay_final__',str(batch+1), 'of', str(n_batches), '.pdf']))
+ elif leaf_type == 'Landmarks_Armature':
+ logger.info(f'Detecting landmarks armature')
+ pdf_path = os.path.join(Dirs.landmarks_armature_overlay_QC, ''.join(['landmarks_armature_overlay_QC__',str(batch+1), 'of', str(n_batches), '.pdf']))
+ pdf_path_final = os.path.join(Dirs.landmarks_armature_overlay_final, ''.join(['landmarks_armature_overlay_final__',str(batch+1), 'of', str(n_batches), '.pdf']))
+
+ ### FINAL
+ # dict_labels = {}
+ fig = plt.figure(figsize=(8.27, 11.69), dpi=dpi) # A4 size, 300 dpi
+ row, col = 0, 0
+ with PdfPages(pdf_path_final) as pdf:
+
+
+
+ for file in os.listdir(dir_components):
+ file_name = str(file.split('.')[0])
+ file_name_parent = file_name.split('__')[0]
+
+ # Project.project_data_list[batch][file_name_parent][component] = []
+
+ if file_name_parent in Project.project_data_list[batch]:
+
+
+
+ if file.endswith(".txt"):
+ with open(os.path.join(dir_components, file), "r") as f:
+ all_points = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ # Project.project_data_list[batch][file_name_parent][component][file_name] = all_points
+
+ height = dimensions_dict[file_name][0]
+ width = dimensions_dict[file_name][1]
+
+ Leaf_Skeleton = LeafSkeleton(cfg, logger, Dirs, leaf_type, all_points, height, width, dir_temp, file_name)
+ Project = add_leaf_skeleton_to_project(cfg, logger, Project, batch, file_name_parent, component, Dirs, leaf_type, all_points, height, width, dir_temp, file_name, Leaf_Skeleton)
+ final_add = cv2.cvtColor(Leaf_Skeleton.get_final(), cv2.COLOR_BGR2RGB)
+
+ # Add image to the current subplot
+ ax = fig.add_subplot(5, 3, row * 3 + col + 1)
+ ax.imshow(final_add)
+ ax.axis('off')
+
+ col += 1
+ if col == 3:
+ col = 0
+ row += 1
+ if row == 5:
+ row = 0
+ pdf.savefig(fig) # Save the current page
+ fig = plt.figure(figsize=(8.27, 11.69), dpi=300) # Create a new page
+ else:
+ pass
+
+ if row != 0 or col != 0:
+ pdf.savefig(fig) # Save the remaining images on the last page
+
+ ### QC
+ '''do_save_QC_pdf = False # TODO refine this
+ if do_save_QC_pdf:
+ # dict_labels = {}
+ fig = plt.figure(figsize=(8.27, 11.69), dpi=dpi) # A4 size, 300 dpi
+ row, col = 0, 0
+ with PdfPages(pdf_path) as pdf:
+
+
+
+ for file in os.listdir(dir_components):
+ file_name = str(file.split('.')[0])
+ file_name_parent = file_name.split('__')[0]
+
+ if file_name_parent in Project.project_data_list[batch]:
+
+ if file.endswith(".txt"):
+ with open(os.path.join(dir_components, file), "r") as f:
+ all_points = [[int(line.split()[0])] + list(map(float, line.split()[1:])) for line in f]
+ Project.project_data_list[batch][file_name_parent][component][file_name] = all_points
+
+ height = dimensions_dict[file_name][0]
+ width = dimensions_dict[file_name][1]
+
+ Leaf_Skeleton = LeafSkeleton(cfg, logger, Dirs, leaf_type, all_points, height, width, dir_temp, file_name)
+ QC_add = cv2.cvtColor(Leaf_Skeleton.get_QC(), cv2.COLOR_BGR2RGB)
+
+ # Add image to the current subplot
+ ax = fig.add_subplot(5, 3, row * 3 + col + 1)
+ ax.imshow(QC_add)
+ ax.axis('off')
+
+ col += 1
+ if col == 3:
+ col = 0
+ row += 1
+ if row == 5:
+ row = 0
+ pdf.savefig(fig) # Save the current page
+ fig = plt.figure(figsize=(8.27, 11.69), dpi=300) # Create a new page
+ else:
+ pass
+
+ if row != 0 or col != 0:
+ pdf.savefig(fig) # Save the remaining images on the last page'''
+
+
+def add_armature_skeleton_to_project(cfg, logger, Project, batch, file_name_parent, component, Dirs, leaf_type, all_points, height, width, dir_temp, file_name, ARM):
+ if ARM.is_complete:
+ try:
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'armature_status': 'complete'}, {'armature': ARM}]})
+ except:
+ Project.project_data_list[batch][file_name_parent][component] = []
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'armature_status': 'complete'}, {'armature': ARM}]})
+
+ else:
+ try:
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'armature_status': 'incomplete'}, {'armature': ARM}]})
+ except:
+ Project.project_data_list[batch][file_name_parent][component] = []
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'armature_status': 'incomplete'}, {'armature': ARM}]})
+
+
+ return Project
+
+
+def add_leaf_skeleton_to_project(cfg, logger, Project, batch, file_name_parent, component, Dirs, leaf_type, all_points, height, width, dir_temp, file_name, LS):
+
+ if LS.is_complete_leaf:
+ try:
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'landmark_status': 'complete_leaf'}, {'landmarks': LS}]})
+ except:
+ Project.project_data_list[batch][file_name_parent][component] = []
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'landmark_status': 'complete_leaf'}, {'landmarks': LS}]})
+ # Project.project_data_list[batch][file_name_parent][component][file_name].update({'landmark_status': 'complete_leaf'})
+ # Project.project_data_list[batch][file_name_parent][component][file_name].update({'landmarks': LS})
+
+ elif LS.is_leaf_no_width:
+ try:
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'landmark_status': 'leaf_no_width'}, {'landmarks': LS}]})
+ except:
+ Project.project_data_list[batch][file_name_parent][component] = []
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'landmark_status': 'leaf_no_width'}, {'landmarks': LS}]})
+ # Project.project_data_list[batch][file_name_parent][component][file_name].update({'landmark_status': 'leaf_no_width'})
+ # Project.project_data_list[batch][file_name_parent][component][file_name].update({'landmarks': LS})
+
+ else:
+ try:
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'landmark_status': 'incomplete'}, {'landmarks': LS}]})
+ except:
+ Project.project_data_list[batch][file_name_parent][component] = []
+ Project.project_data_list[batch][file_name_parent][component].append({file_name: [{'landmark_status': 'incomplete'}, {'landmarks': LS}]})
+
+ # Project.project_data_list[batch][file_name_parent][component][file_name].update({'landmark_status': 'incomplete'})
+ # Project.project_data_list[batch][file_name_parent][component][file_name].update({'landmarks': LS})
+
+ return Project
+
+
+'''
+self.determine_lamina_length('final')
+
+# Lamina tip and base
+if self.has_lamina_tip:
+ cv2.circle(self.image_final, self.lamina_tip, radius=4, color=(0, 255, 0), thickness=2)
+ cv2.circle(self.image_final, self.lamina_tip, radius=2, color=(255, 255, 255), thickness=-1)
+if self.has_lamina_base:
+ cv2.circle(self.image_final, self.lamina_base, radius=4, color=(255, 0, 0), thickness=2)
+ cv2.circle(self.image_final, self.lamina_base, radius=2, color=(255, 255, 255), thickness=-1)
+
+# Apex angle
+# if self.apex_center != []:
+# cv2.circle(self.image_final, self.apex_center, radius=3, color=(0, 255, 0), thickness=-1)
+if self.apex_left != []:
+ cv2.circle(self.image_final, self.apex_left, radius=3, color=(255, 0, 0), thickness=-1)
+if self.apex_right != []:
+ cv2.circle(self.image_final, self.apex_right, radius=3, color=(0, 0, 255), thickness=-1)
+
+# Base angle
+# if self.base_center:
+# cv2.circle(self.image_final, self.base_center, radius=3, color=(0, 255, 0), thickness=-1)
+if self.base_left:
+ cv2.circle(self.image_final, self.base_left, radius=3, color=(255, 0, 0), thickness=-1)
+if self.base_right:
+ cv2.circle(self.image_final, self.base_right, radius=3, color=(0, 0, 255), thickness=-1)
+
+# Draw line of fit
+for point in self.width_infer:
+
+
+'''
+
+
+
+
+
+
+
+
+
+def get_cropped_dimensions(dir_temp):
+ dimensions_dict = {}
+ for file_name in os.listdir(dir_temp):
+ if file_name.endswith(".jpg"):
+ img = cv2.imread(os.path.join(dir_temp, file_name))
+ height, width, channels = img.shape
+ stem = os.path.splitext(file_name)[0]
+ dimensions_dict[stem] = (height, width)
+ return dimensions_dict
+
+def unpack_class_from_components_armature(dict_big, cls, dict_name_yolo, dict_name_location, Project):
+ # Get the dict that contains plant parts, find the whole leaves
+ for filename, value in dict_big.items():
+ if "Detections_Armature_Components" in value:
+ filtered_components = [val for val in value["Detections_Armature_Components"] if val[0] == cls]
+ value[dict_name_yolo] = filtered_components
+
+ for filename, value in dict_big.items():
+ if "Detections_Armature_Components" in value:
+ filtered_components = [val for val in value["Detections_Armature_Components"] if val[0] == cls]
+ height = value['height']
+ width = value['width']
+ converted_list = [[convert_index_to_class_armature(val[0]), int((val[1] * width) - ((val[3] * width) / 2)),
+ int((val[2] * height) - ((val[4] * height) / 2)),
+ int(val[3] * width) + int((val[1] * width) - ((val[3] * width) / 2)),
+ int(val[4] * height) + int((val[2] * height) - ((val[4] * height) / 2))] for val in filtered_components]
+ # Verify that the crops are correct
+ # img = Image.open(os.path.join(Project., '.'.join([filename,'jpg'])))
+ # for d in converted_list:
+ # img_crop = img.crop((d[1], d[2], d[3], d[4]))
+ # img_crop.show()
+ value[dict_name_location] = converted_list
+ # print(dict)
+ return dict_big
+
+def unpack_class_from_components(dict_big, cls, dict_name_yolo, dict_name_location, Project):
+ # Get the dict that contains plant parts, find the whole leaves
+ for filename, value in dict_big.items():
+ if "Detections_Plant_Components" in value:
+ filtered_components = [val for val in value["Detections_Plant_Components"] if val[0] == cls]
+ value[dict_name_yolo] = filtered_components
+
+ for filename, value in dict_big.items():
+ if "Detections_Plant_Components" in value:
+ filtered_components = [val for val in value["Detections_Plant_Components"] if val[0] == cls]
+ height = value['height']
+ width = value['width']
+ converted_list = [[convert_index_to_class(val[0]), int((val[1] * width) - ((val[3] * width) / 2)),
+ int((val[2] * height) - ((val[4] * height) / 2)),
+ int(val[3] * width) + int((val[1] * width) - ((val[3] * width) / 2)),
+ int(val[4] * height) + int((val[2] * height) - ((val[4] * height) / 2))] for val in filtered_components]
+ # Verify that the crops are correct
+ # img = Image.open(os.path.join(Project., '.'.join([filename,'jpg'])))
+ # for d in converted_list:
+ # img_crop = img.crop((d[1], d[2], d[3], d[4]))
+ # img_crop.show()
+ value[dict_name_location] = converted_list
+ # print(dict)
+ return dict_big
+
+
+def crop_images_to_bbox_armature(dict_big, cls, dict_name_cropped, dict_from, Project, Dirs, do_upscale=False, cfg=None):
+ dir_temp = os.path.join(Dirs.landmarks, 'TEMP_landmarks')
+ os.makedirs(dir_temp, exist_ok=True)
+ # For each image, iterate through the whole leaves, segment, report data back to dict_plant_components
+ for filename, value in dict_big.items():
+ value[dict_name_cropped] = []
+ if dict_from in value:
+ bboxes_whole_leaves = [val for val in value[dict_from] if val[0] == convert_index_to_class_armature(cls)]
+ if len(bboxes_whole_leaves) == 0:
+ m = str(''.join(['No objects for class ', convert_index_to_class_armature(0), ' were found']))
+ # Print_Verbose(cfg, 3, m).plain()
+ else:
+ try:
+ img = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename,'jpg'])))
+ # img = cv2.imread(os.path.join(Project, '.'.join([filename,'jpg']))) # Testing
+ except:
+ img = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename,'jpeg'])))
+ # img = cv2.imread(os.path.join(Project, '.'.join([filename,'jpeg']))) # Testing
+
+ for d in bboxes_whole_leaves:
+ # img_crop = img.crop((d[1], d[2], d[3], d[4])) # PIL
+ img_crop = img[d[2]:d[4], d[1]:d[3]]
+ loc = '-'.join([str(d[1]), str(d[2]), str(d[3]), str(d[4])])
+ # value[dict_name_cropped].append({crop_name: img_crop})
+ if do_upscale:
+ upscale_factor = int(cfg['leafmachine']['landmark_detector_armature']['upscale_factor'])
+ if cls == 0:
+ crop_name = '__'.join([filename,f"PRICKLE-{upscale_factor}x",loc])
+ height, width, _ = img_crop.shape
+ img_crop = cv2.resize(img_crop, ((width * upscale_factor), (height * upscale_factor)), interpolation=cv2.INTER_LANCZOS4)
+ else:
+ if cls == 0:
+ crop_name = '__'.join([filename,'PRICKLE',loc])
+
+ cv2.imwrite(os.path.join(dir_temp, '.'.join([crop_name,'jpg'])), img_crop)
+ # cv2.imshow('img_crop', img_crop)
+ # cv2.waitKey(0)
+ # img_crop.show() # PIL
+ return dict_big, dir_temp
+
+
+def crop_images_to_bbox(dict_big, cls, dict_name_cropped, dict_from, Project, Dirs):
+ dir_temp = os.path.join(Dirs.landmarks, 'TEMP_landmarks')
+ os.makedirs(dir_temp, exist_ok=True)
+ # For each image, iterate through the whole leaves, segment, report data back to dict_plant_components
+ for filename, value in dict_big.items():
+ value[dict_name_cropped] = []
+ if dict_from in value:
+ bboxes_whole_leaves = [val for val in value[dict_from] if val[0] == convert_index_to_class(cls)]
+ if len(bboxes_whole_leaves) == 0:
+ m = str(''.join(['No objects for class ', convert_index_to_class(0), ' were found']))
+ # Print_Verbose(cfg, 3, m).plain()
+ else:
+ try:
+ img = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename,'jpg'])))
+ # img = cv2.imread(os.path.join(Project, '.'.join([filename,'jpg']))) # Testing
+ except:
+ img = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename,'jpeg'])))
+ # img = cv2.imread(os.path.join(Project, '.'.join([filename,'jpeg']))) # Testing
+
+ for d in bboxes_whole_leaves:
+ # img_crop = img.crop((d[1], d[2], d[3], d[4])) # PIL
+ img_crop = img[d[2]:d[4], d[1]:d[3]]
+ loc = '-'.join([str(d[1]), str(d[2]), str(d[3]), str(d[4])])
+ if cls == 0:
+ crop_name = '__'.join([filename,'L',loc])
+ elif cls == 1:
+ crop_name = '__'.join([filename,'PL',loc])
+ elif cls == 2:
+ crop_name = '__'.join([filename,'ARM',loc])
+ # value[dict_name_cropped].append({crop_name: img_crop})
+ cv2.imwrite(os.path.join(dir_temp, '.'.join([crop_name,'jpg'])), img_crop)
+ # cv2.imshow('img_crop', img_crop)
+ # cv2.waitKey(0)
+ # img_crop.show() # PIL
+ return dict_big, dir_temp
+
+def convert_index_to_class(ind):
+ mapping = {
+ 0: 'apex_angle',
+ 1: 'base_angle',
+ 2: 'lamina_base',
+ 3: 'lamina_tip',
+ 4: 'lamina_width',
+ 5: 'lobe_tip',
+ 6: 'midvein_trace',
+ 7: 'petiole_tip',
+ 8: 'petiole_trace',
+ }
+ return mapping.get(ind, 'Invalid class').lower()
+
+def convert_index_to_class_armature(ind):
+ mapping = {
+ 0: 'tip',
+ 1: 'middle',
+ 2: 'outer',
+ }
+ return mapping.get(ind, 'Invalid class').lower()
diff --git a/vouchervision/component_detector/data/Argoverse.yaml b/vouchervision/component_detector/data/Argoverse.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..9d114f55dce88a1085a69b5e93619b1bd1e301f5
--- /dev/null
+++ b/vouchervision/component_detector/data/Argoverse.yaml
@@ -0,0 +1,67 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Argoverse-HD dataset (ring-front-center camera) http://www.cs.cmu.edu/~mengtial/proj/streaming/ by Argo AI
+# Example usage: python train.py --data Argoverse.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── Argoverse ← downloads here (31.3 GB)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/Argoverse # dataset root dir
+train: Argoverse-1.1/images/train/ # train images (relative to 'path') 39384 images
+val: Argoverse-1.1/images/val/ # val images (relative to 'path') 15062 images
+test: Argoverse-1.1/images/test/ # test images (optional) https://eval.ai/web/challenges/challenge-page/800/overview
+
+# Classes
+nc: 8 # number of classes
+names: ['person', 'bicycle', 'car', 'motorcycle', 'bus', 'truck', 'traffic_light', 'stop_sign'] # class names
+
+
+# Download script/URL (optional) ---------------------------------------------------------------------------------------
+download: |
+ import json
+
+ from tqdm.auto import tqdm
+ from utils.general import download, Path
+
+
+ def argoverse2yolo(set):
+ labels = {}
+ a = json.load(open(set, "rb"))
+ for annot in tqdm(a['annotations'], desc=f"Converting {set} to YOLOv5 format..."):
+ img_id = annot['image_id']
+ img_name = a['images'][img_id]['name']
+ img_label_name = img_name[:-3] + "txt"
+
+ cls = annot['category_id'] # instance class id
+ x_center, y_center, width, height = annot['bbox']
+ x_center = (x_center + width / 2) / 1920.0 # offset and scale
+ y_center = (y_center + height / 2) / 1200.0 # offset and scale
+ width /= 1920.0 # scale
+ height /= 1200.0 # scale
+
+ img_dir = set.parents[2] / 'Argoverse-1.1' / 'labels' / a['seq_dirs'][a['images'][annot['image_id']]['sid']]
+ if not img_dir.exists():
+ img_dir.mkdir(parents=True, exist_ok=True)
+
+ k = str(img_dir / img_label_name)
+ if k not in labels:
+ labels[k] = []
+ labels[k].append(f"{cls} {x_center} {y_center} {width} {height}\n")
+
+ for k in labels:
+ with open(k, "w") as f:
+ f.writelines(labels[k])
+
+
+ # Download
+ dir = Path('../datasets/Argoverse') # dataset root dir
+ urls = ['https://argoverse-hd.s3.us-east-2.amazonaws.com/Argoverse-HD-Full.zip']
+ download(urls, dir=dir, delete=False)
+
+ # Convert
+ annotations_dir = 'Argoverse-HD/annotations/'
+ (dir / 'Argoverse-1.1' / 'tracking').rename(dir / 'Argoverse-1.1' / 'images') # rename 'tracking' to 'images'
+ for d in "train.json", "val.json":
+ argoverse2yolo(dir / annotations_dir / d) # convert VisDrone annotations to YOLO labels
diff --git a/vouchervision/component_detector/data/GlobalWheat2020.yaml b/vouchervision/component_detector/data/GlobalWheat2020.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..4c43693f1d820bee8e78df630b4300684655cbc8
--- /dev/null
+++ b/vouchervision/component_detector/data/GlobalWheat2020.yaml
@@ -0,0 +1,54 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Global Wheat 2020 dataset http://www.global-wheat.com/ by University of Saskatchewan
+# Example usage: python train.py --data GlobalWheat2020.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── GlobalWheat2020 ← downloads here (7.0 GB)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/GlobalWheat2020 # dataset root dir
+train: # train images (relative to 'path') 3422 images
+ - images/arvalis_1
+ - images/arvalis_2
+ - images/arvalis_3
+ - images/ethz_1
+ - images/rres_1
+ - images/inrae_1
+ - images/usask_1
+val: # val images (relative to 'path') 748 images (WARNING: train set contains ethz_1)
+ - images/ethz_1
+test: # test images (optional) 1276 images
+ - images/utokyo_1
+ - images/utokyo_2
+ - images/nau_1
+ - images/uq_1
+
+# Classes
+nc: 1 # number of classes
+names: ['wheat_head'] # class names
+
+
+# Download script/URL (optional) ---------------------------------------------------------------------------------------
+download: |
+ from utils.general import download, Path
+
+
+ # Download
+ dir = Path(yaml['path']) # dataset root dir
+ urls = ['https://zenodo.org/record/4298502/files/global-wheat-codalab-official.zip',
+ 'https://github.com/ultralytics/yolov5/releases/download/v1.0/GlobalWheat2020_labels.zip']
+ download(urls, dir=dir)
+
+ # Make Directories
+ for p in 'annotations', 'images', 'labels':
+ (dir / p).mkdir(parents=True, exist_ok=True)
+
+ # Move
+ for p in 'arvalis_1', 'arvalis_2', 'arvalis_3', 'ethz_1', 'rres_1', 'inrae_1', 'usask_1', \
+ 'utokyo_1', 'utokyo_2', 'nau_1', 'uq_1':
+ (dir / p).rename(dir / 'images' / p) # move to /images
+ f = (dir / p).with_suffix('.json') # json file
+ if f.exists():
+ f.rename((dir / 'annotations' / p).with_suffix('.json')) # move to /annotations
diff --git a/vouchervision/component_detector/data/Objects365.yaml b/vouchervision/component_detector/data/Objects365.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..334c23c359cfd1e11b01b12cf769facaa2586b79
--- /dev/null
+++ b/vouchervision/component_detector/data/Objects365.yaml
@@ -0,0 +1,114 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Objects365 dataset https://www.objects365.org/ by Megvii
+# Example usage: python train.py --data Objects365.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── Objects365 ← downloads here (712 GB = 367G data + 345G zips)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/Objects365 # dataset root dir
+train: images/train # train images (relative to 'path') 1742289 images
+val: images/val # val images (relative to 'path') 80000 images
+test: # test images (optional)
+
+# Classes
+nc: 365 # number of classes
+names: ['Person', 'Sneakers', 'Chair', 'Other Shoes', 'Hat', 'Car', 'Lamp', 'Glasses', 'Bottle', 'Desk', 'Cup',
+ 'Street Lights', 'Cabinet/shelf', 'Handbag/Satchel', 'Bracelet', 'Plate', 'Picture/Frame', 'Helmet', 'Book',
+ 'Gloves', 'Storage box', 'Boat', 'Leather Shoes', 'Flower', 'Bench', 'Potted Plant', 'Bowl/Basin', 'Flag',
+ 'Pillow', 'Boots', 'Vase', 'Microphone', 'Necklace', 'Ring', 'SUV', 'Wine Glass', 'Belt', 'Monitor/TV',
+ 'Backpack', 'Umbrella', 'Traffic Light', 'Speaker', 'Watch', 'Tie', 'Trash bin Can', 'Slippers', 'Bicycle',
+ 'Stool', 'Barrel/bucket', 'Van', 'Couch', 'Sandals', 'Basket', 'Drum', 'Pen/Pencil', 'Bus', 'Wild Bird',
+ 'High Heels', 'Motorcycle', 'Guitar', 'Carpet', 'Cell Phone', 'Bread', 'Camera', 'Canned', 'Truck',
+ 'Traffic cone', 'Cymbal', 'Lifesaver', 'Towel', 'Stuffed Toy', 'Candle', 'Sailboat', 'Laptop', 'Awning',
+ 'Bed', 'Faucet', 'Tent', 'Horse', 'Mirror', 'Power outlet', 'Sink', 'Apple', 'Air Conditioner', 'Knife',
+ 'Hockey Stick', 'Paddle', 'Pickup Truck', 'Fork', 'Traffic Sign', 'Balloon', 'Tripod', 'Dog', 'Spoon', 'Clock',
+ 'Pot', 'Cow', 'Cake', 'Dinning Table', 'Sheep', 'Hanger', 'Blackboard/Whiteboard', 'Napkin', 'Other Fish',
+ 'Orange/Tangerine', 'Toiletry', 'Keyboard', 'Tomato', 'Lantern', 'Machinery Vehicle', 'Fan',
+ 'Green Vegetables', 'Banana', 'Baseball Glove', 'Airplane', 'Mouse', 'Train', 'Pumpkin', 'Soccer', 'Skiboard',
+ 'Luggage', 'Nightstand', 'Tea pot', 'Telephone', 'Trolley', 'Head Phone', 'Sports Car', 'Stop Sign',
+ 'Dessert', 'Scooter', 'Stroller', 'Crane', 'Remote', 'Refrigerator', 'Oven', 'Lemon', 'Duck', 'Baseball Bat',
+ 'Surveillance Camera', 'Cat', 'Jug', 'Broccoli', 'Piano', 'Pizza', 'Elephant', 'Skateboard', 'Surfboard',
+ 'Gun', 'Skating and Skiing shoes', 'Gas stove', 'Donut', 'Bow Tie', 'Carrot', 'Toilet', 'Kite', 'Strawberry',
+ 'Other Balls', 'Shovel', 'Pepper', 'Computer Box', 'Toilet Paper', 'Cleaning Products', 'Chopsticks',
+ 'Microwave', 'Pigeon', 'Baseball', 'Cutting/chopping Board', 'Coffee Table', 'Side Table', 'Scissors',
+ 'Marker', 'Pie', 'Ladder', 'Snowboard', 'Cookies', 'Radiator', 'Fire Hydrant', 'Basketball', 'Zebra', 'Grape',
+ 'Giraffe', 'Potato', 'Sausage', 'Tricycle', 'Violin', 'Egg', 'Fire Extinguisher', 'Candy', 'Fire Truck',
+ 'Billiards', 'Converter', 'Bathtub', 'Wheelchair', 'Golf Club', 'Briefcase', 'Cucumber', 'Cigar/Cigarette',
+ 'Paint Brush', 'Pear', 'Heavy Truck', 'Hamburger', 'Extractor', 'Extension Cord', 'Tong', 'Tennis Racket',
+ 'Folder', 'American Football', 'earphone', 'Mask', 'Kettle', 'Tennis', 'Ship', 'Swing', 'Coffee Machine',
+ 'Slide', 'Carriage', 'Onion', 'Green beans', 'Projector', 'Frisbee', 'Washing Machine/Drying Machine',
+ 'Chicken', 'Printer', 'Watermelon', 'Saxophone', 'Tissue', 'Toothbrush', 'Ice cream', 'Hot-air balloon',
+ 'Cello', 'French Fries', 'Scale', 'Trophy', 'Cabbage', 'Hot dog', 'Blender', 'Peach', 'Rice', 'Wallet/Purse',
+ 'Volleyball', 'Deer', 'Goose', 'Tape', 'Tablet', 'Cosmetics', 'Trumpet', 'Pineapple', 'Golf Ball',
+ 'Ambulance', 'Parking meter', 'Mango', 'Key', 'Hurdle', 'Fishing Rod', 'Medal', 'Flute', 'Brush', 'Penguin',
+ 'Megaphone', 'Corn', 'Lettuce', 'Garlic', 'Swan', 'Helicopter', 'Green Onion', 'Sandwich', 'Nuts',
+ 'Speed Limit Sign', 'Induction Cooker', 'Broom', 'Trombone', 'Plum', 'Rickshaw', 'Goldfish', 'Kiwi fruit',
+ 'Router/modem', 'Poker Card', 'Toaster', 'Shrimp', 'Sushi', 'Cheese', 'Notepaper', 'Cherry', 'Pliers', 'CD',
+ 'Pasta', 'Hammer', 'Cue', 'Avocado', 'Hamimelon', 'Flask', 'Mushroom', 'Screwdriver', 'Soap', 'Recorder',
+ 'Bear', 'Eggplant', 'Board Eraser', 'Coconut', 'Tape Measure/Ruler', 'Pig', 'Showerhead', 'Globe', 'Chips',
+ 'Steak', 'Crosswalk Sign', 'Stapler', 'Camel', 'Formula 1', 'Pomegranate', 'Dishwasher', 'Crab',
+ 'Hoverboard', 'Meat ball', 'Rice Cooker', 'Tuba', 'Calculator', 'Papaya', 'Antelope', 'Parrot', 'Seal',
+ 'Butterfly', 'Dumbbell', 'Donkey', 'Lion', 'Urinal', 'Dolphin', 'Electric Drill', 'Hair Dryer', 'Egg tart',
+ 'Jellyfish', 'Treadmill', 'Lighter', 'Grapefruit', 'Game board', 'Mop', 'Radish', 'Baozi', 'Target', 'French',
+ 'Spring Rolls', 'Monkey', 'Rabbit', 'Pencil Case', 'Yak', 'Red Cabbage', 'Binoculars', 'Asparagus', 'Barbell',
+ 'Scallop', 'Noddles', 'Comb', 'Dumpling', 'Oyster', 'Table Tennis paddle', 'Cosmetics Brush/Eyeliner Pencil',
+ 'Chainsaw', 'Eraser', 'Lobster', 'Durian', 'Okra', 'Lipstick', 'Cosmetics Mirror', 'Curling', 'Table Tennis']
+
+
+# Download script/URL (optional) ---------------------------------------------------------------------------------------
+download: |
+ from tqdm.auto import tqdm
+
+ from utils.general import Path, check_requirements, download, np, xyxy2xywhn
+
+ check_requirements(('pycocotools>=2.0',))
+ from pycocotools.coco import COCO
+
+ # Make Directories
+ dir = Path(yaml['path']) # dataset root dir
+ for p in 'images', 'labels':
+ (dir / p).mkdir(parents=True, exist_ok=True)
+ for q in 'train', 'val':
+ (dir / p / q).mkdir(parents=True, exist_ok=True)
+
+ # Train, Val Splits
+ for split, patches in [('train', 50 + 1), ('val', 43 + 1)]:
+ print(f"Processing {split} in {patches} patches ...")
+ images, labels = dir / 'images' / split, dir / 'labels' / split
+
+ # Download
+ url = f"https://dorc.ks3-cn-beijing.ksyun.com/data-set/2020Objects365%E6%95%B0%E6%8D%AE%E9%9B%86/{split}/"
+ if split == 'train':
+ download([f'{url}zhiyuan_objv2_{split}.tar.gz'], dir=dir, delete=False) # annotations json
+ download([f'{url}patch{i}.tar.gz' for i in range(patches)], dir=images, curl=True, delete=False, threads=8)
+ elif split == 'val':
+ download([f'{url}zhiyuan_objv2_{split}.json'], dir=dir, delete=False) # annotations json
+ download([f'{url}images/v1/patch{i}.tar.gz' for i in range(15 + 1)], dir=images, curl=True, delete=False, threads=8)
+ download([f'{url}images/v2/patch{i}.tar.gz' for i in range(16, patches)], dir=images, curl=True, delete=False, threads=8)
+
+ # Move
+ for f in tqdm(images.rglob('*.jpg'), desc=f'Moving {split} images'):
+ f.rename(images / f.name) # move to /images/{split}
+
+ # Labels
+ coco = COCO(dir / f'zhiyuan_objv2_{split}.json')
+ names = [x["name"] for x in coco.loadCats(coco.getCatIds())]
+ for cid, cat in enumerate(names):
+ catIds = coco.getCatIds(catNms=[cat])
+ imgIds = coco.getImgIds(catIds=catIds)
+ for im in tqdm(coco.loadImgs(imgIds), desc=f'Class {cid + 1}/{len(names)} {cat}'):
+ width, height = im["width"], im["height"]
+ path = Path(im["file_name"]) # image filename
+ try:
+ with open(labels / path.with_suffix('.txt').name, 'a') as file:
+ annIds = coco.getAnnIds(imgIds=im["id"], catIds=catIds, iscrowd=None)
+ for a in coco.loadAnns(annIds):
+ x, y, w, h = a['bbox'] # bounding box in xywh (xy top-left corner)
+ xyxy = np.array([x, y, x + w, y + h])[None] # pixels(1,4)
+ x, y, w, h = xyxy2xywhn(xyxy, w=width, h=height, clip=True)[0] # normalized and clipped
+ file.write(f"{cid} {x:.5f} {y:.5f} {w:.5f} {h:.5f}\n")
+ except Exception as e:
+ print(e)
diff --git a/vouchervision/component_detector/data/PLANT_Full.yaml b/vouchervision/component_detector/data/PLANT_Full.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..aeeaecf198f6605a280e026468a135ac8637c391
--- /dev/null
+++ b/vouchervision/component_detector/data/PLANT_Full.yaml
@@ -0,0 +1,17 @@
+path: ../datasets/PLANT_Full
+train: images/train
+val: images/val
+test: images/test
+nc: 11
+names:
+- leaf_whole
+- leaf_partial
+- leaflet
+- seed_fruit_one
+- seed_fruit_many
+- flower_one
+- flower_many
+- bud
+- specimen
+- roots
+- wood
\ No newline at end of file
diff --git a/vouchervision/component_detector/data/SKU-110K.yaml b/vouchervision/component_detector/data/SKU-110K.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..2fd689b1bcaca53d69d49725e196b938fcf7d0b2
--- /dev/null
+++ b/vouchervision/component_detector/data/SKU-110K.yaml
@@ -0,0 +1,53 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# SKU-110K retail items dataset https://github.com/eg4000/SKU110K_CVPR19 by Trax Retail
+# Example usage: python train.py --data SKU-110K.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── SKU-110K ← downloads here (13.6 GB)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/SKU-110K # dataset root dir
+train: train.txt # train images (relative to 'path') 8219 images
+val: val.txt # val images (relative to 'path') 588 images
+test: test.txt # test images (optional) 2936 images
+
+# Classes
+nc: 1 # number of classes
+names: ['object'] # class names
+
+
+# Download script/URL (optional) ---------------------------------------------------------------------------------------
+download: |
+ import shutil
+ from tqdm.auto import tqdm
+ from utils.general import np, pd, Path, download, xyxy2xywh
+
+
+ # Download
+ dir = Path(yaml['path']) # dataset root dir
+ parent = Path(dir.parent) # download dir
+ urls = ['http://trax-geometry.s3.amazonaws.com/cvpr_challenge/SKU110K_fixed.tar.gz']
+ download(urls, dir=parent, delete=False)
+
+ # Rename directories
+ if dir.exists():
+ shutil.rmtree(dir)
+ (parent / 'SKU110K_fixed').rename(dir) # rename dir
+ (dir / 'labels').mkdir(parents=True, exist_ok=True) # create labels dir
+
+ # Convert labels
+ names = 'image', 'x1', 'y1', 'x2', 'y2', 'class', 'image_width', 'image_height' # column names
+ for d in 'annotations_train.csv', 'annotations_val.csv', 'annotations_test.csv':
+ x = pd.read_csv(dir / 'annotations' / d, names=names).values # annotations
+ images, unique_images = x[:, 0], np.unique(x[:, 0])
+ with open((dir / d).with_suffix('.txt').__str__().replace('annotations_', ''), 'w') as f:
+ f.writelines(f'./images/{s}\n' for s in unique_images)
+ for im in tqdm(unique_images, desc=f'Converting {dir / d}'):
+ cls = 0 # single-class dataset
+ with open((dir / 'labels' / im).with_suffix('.txt'), 'a') as f:
+ for r in x[images == im]:
+ w, h = r[6], r[7] # image width, height
+ xywh = xyxy2xywh(np.array([[r[1] / w, r[2] / h, r[3] / w, r[4] / h]]))[0] # instance
+ f.write(f"{cls} {xywh[0]:.5f} {xywh[1]:.5f} {xywh[2]:.5f} {xywh[3]:.5f}\n") # write label
diff --git a/vouchervision/component_detector/data/VOC.yaml b/vouchervision/component_detector/data/VOC.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..93a1f181ce8cd2e391949c638f78b3fe9c21ac4d
--- /dev/null
+++ b/vouchervision/component_detector/data/VOC.yaml
@@ -0,0 +1,81 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# PASCAL VOC dataset http://host.robots.ox.ac.uk/pascal/VOC by University of Oxford
+# Example usage: python train.py --data VOC.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── VOC ← downloads here (2.8 GB)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/VOC
+train: # train images (relative to 'path') 16551 images
+ - images/train2012
+ - images/train2007
+ - images/val2012
+ - images/val2007
+val: # val images (relative to 'path') 4952 images
+ - images/test2007
+test: # test images (optional)
+ - images/test2007
+
+# Classes
+nc: 20 # number of classes
+names: ['aeroplane', 'bicycle', 'bird', 'boat', 'bottle', 'bus', 'car', 'cat', 'chair', 'cow', 'diningtable', 'dog',
+ 'horse', 'motorbike', 'person', 'pottedplant', 'sheep', 'sofa', 'train', 'tvmonitor'] # class names
+
+
+# Download script/URL (optional) ---------------------------------------------------------------------------------------
+download: |
+ import xml.etree.ElementTree as ET
+
+ from tqdm.auto import tqdm
+ from utils.general import download, Path
+
+
+ def convert_label(path, lb_path, year, image_id):
+ def convert_box(size, box):
+ dw, dh = 1. / size[0], 1. / size[1]
+ x, y, w, h = (box[0] + box[1]) / 2.0 - 1, (box[2] + box[3]) / 2.0 - 1, box[1] - box[0], box[3] - box[2]
+ return x * dw, y * dh, w * dw, h * dh
+
+ in_file = open(path / f'VOC{year}/Annotations/{image_id}.xml')
+ out_file = open(lb_path, 'w')
+ tree = ET.parse(in_file)
+ root = tree.getroot()
+ size = root.find('size')
+ w = int(size.find('width').text)
+ h = int(size.find('height').text)
+
+ for obj in root.iter('object'):
+ cls = obj.find('name').text
+ if cls in yaml['names'] and not int(obj.find('difficult').text) == 1:
+ xmlbox = obj.find('bndbox')
+ bb = convert_box((w, h), [float(xmlbox.find(x).text) for x in ('xmin', 'xmax', 'ymin', 'ymax')])
+ cls_id = yaml['names'].index(cls) # class id
+ out_file.write(" ".join([str(a) for a in (cls_id, *bb)]) + '\n')
+
+
+ # Download
+ dir = Path(yaml['path']) # dataset root dir
+ url = 'https://github.com/ultralytics/yolov5/releases/download/v1.0/'
+ urls = [url + 'VOCtrainval_06-Nov-2007.zip', # 446MB, 5012 images
+ url + 'VOCtest_06-Nov-2007.zip', # 438MB, 4953 images
+ url + 'VOCtrainval_11-May-2012.zip'] # 1.95GB, 17126 images
+ download(urls, dir=dir / 'images', delete=False, curl=True, threads=3)
+
+ # Convert
+ path = dir / f'images/VOCdevkit'
+ for year, image_set in ('2012', 'train'), ('2012', 'val'), ('2007', 'train'), ('2007', 'val'), ('2007', 'test'):
+ imgs_path = dir / 'images' / f'{image_set}{year}'
+ lbs_path = dir / 'labels' / f'{image_set}{year}'
+ imgs_path.mkdir(exist_ok=True, parents=True)
+ lbs_path.mkdir(exist_ok=True, parents=True)
+
+ with open(path / f'VOC{year}/ImageSets/Main/{image_set}.txt') as f:
+ image_ids = f.read().strip().split()
+ for id in tqdm(image_ids, desc=f'{image_set}{year}'):
+ f = path / f'VOC{year}/JPEGImages/{id}.jpg' # old img path
+ lb_path = (lbs_path / f.name).with_suffix('.txt') # new label path
+ f.rename(imgs_path / f.name) # move image
+ convert_label(path, lb_path, year, id) # convert labels to YOLO format
diff --git a/vouchervision/component_detector/data/VisDrone.yaml b/vouchervision/component_detector/data/VisDrone.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..c38fb2ab769eeaf66506f7d7baacb1e76e360aea
--- /dev/null
+++ b/vouchervision/component_detector/data/VisDrone.yaml
@@ -0,0 +1,61 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# VisDrone2019-DET dataset https://github.com/VisDrone/VisDrone-Dataset by Tianjin University
+# Example usage: python train.py --data VisDrone.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── VisDrone ← downloads here (2.3 GB)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/VisDrone # dataset root dir
+train: VisDrone2019-DET-train/images # train images (relative to 'path') 6471 images
+val: VisDrone2019-DET-val/images # val images (relative to 'path') 548 images
+test: VisDrone2019-DET-test-dev/images # test images (optional) 1610 images
+
+# Classes
+nc: 10 # number of classes
+names: ['pedestrian', 'people', 'bicycle', 'car', 'van', 'truck', 'tricycle', 'awning-tricycle', 'bus', 'motor']
+
+
+# Download script/URL (optional) ---------------------------------------------------------------------------------------
+download: |
+ from utils.general import download, os, Path
+
+ def visdrone2yolo(dir):
+ from PIL import Image
+ from tqdm.auto import tqdm
+
+ def convert_box(size, box):
+ # Convert VisDrone box to YOLO xywh box
+ dw = 1. / size[0]
+ dh = 1. / size[1]
+ return (box[0] + box[2] / 2) * dw, (box[1] + box[3] / 2) * dh, box[2] * dw, box[3] * dh
+
+ (dir / 'labels').mkdir(parents=True, exist_ok=True) # make labels directory
+ pbar = tqdm((dir / 'annotations').glob('*.txt'), desc=f'Converting {dir}')
+ for f in pbar:
+ img_size = Image.open((dir / 'images' / f.name).with_suffix('.jpg')).size
+ lines = []
+ with open(f, 'r') as file: # read annotation.txt
+ for row in [x.split(',') for x in file.read().strip().splitlines()]:
+ if row[4] == '0': # VisDrone 'ignored regions' class 0
+ continue
+ cls = int(row[5]) - 1
+ box = convert_box(img_size, tuple(map(int, row[:4])))
+ lines.append(f"{cls} {' '.join(f'{x:.6f}' for x in box)}\n")
+ with open(str(f).replace(os.sep + 'annotations' + os.sep, os.sep + 'labels' + os.sep), 'w') as fl:
+ fl.writelines(lines) # write label.txt
+
+
+ # Download
+ dir = Path(yaml['path']) # dataset root dir
+ urls = ['https://github.com/ultralytics/yolov5/releases/download/v1.0/VisDrone2019-DET-train.zip',
+ 'https://github.com/ultralytics/yolov5/releases/download/v1.0/VisDrone2019-DET-val.zip',
+ 'https://github.com/ultralytics/yolov5/releases/download/v1.0/VisDrone2019-DET-test-dev.zip',
+ 'https://github.com/ultralytics/yolov5/releases/download/v1.0/VisDrone2019-DET-test-challenge.zip']
+ download(urls, dir=dir, curl=True, threads=4)
+
+ # Convert
+ for d in 'VisDrone2019-DET-train', 'VisDrone2019-DET-val', 'VisDrone2019-DET-test-dev':
+ visdrone2yolo(dir / d) # convert VisDrone annotations to YOLO labels
diff --git a/vouchervision/component_detector/data/coco.yaml b/vouchervision/component_detector/data/coco.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..0c0c4adab05df585c3c38ba07db5e29a03b5a3e0
--- /dev/null
+++ b/vouchervision/component_detector/data/coco.yaml
@@ -0,0 +1,45 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# COCO 2017 dataset http://cocodataset.org by Microsoft
+# Example usage: python train.py --data coco.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── coco ← downloads here (20.1 GB)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/coco # dataset root dir
+train: train2017.txt # train images (relative to 'path') 118287 images
+val: val2017.txt # val images (relative to 'path') 5000 images
+test: test-dev2017.txt # 20288 of 40670 images, submit to https://competitions.codalab.org/competitions/20794
+
+# Classes
+nc: 80 # number of classes
+names: ['person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'traffic light',
+ 'fire hydrant', 'stop sign', 'parking meter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow',
+ 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee',
+ 'skis', 'snowboard', 'sports ball', 'kite', 'baseball bat', 'baseball glove', 'skateboard', 'surfboard',
+ 'tennis racket', 'bottle', 'wine glass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple',
+ 'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza', 'donut', 'cake', 'chair', 'couch',
+ 'potted plant', 'bed', 'dining table', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cell phone',
+ 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddy bear',
+ 'hair drier', 'toothbrush'] # class names
+
+
+# Download script/URL (optional)
+download: |
+ from utils.general import download, Path
+
+
+ # Download labels
+ segments = False # segment or box labels
+ dir = Path(yaml['path']) # dataset root dir
+ url = 'https://github.com/ultralytics/yolov5/releases/download/v1.0/'
+ urls = [url + ('coco2017labels-segments.zip' if segments else 'coco2017labels.zip')] # labels
+ download(urls, dir=dir.parent)
+
+ # Download data
+ urls = ['http://images.cocodataset.org/zips/train2017.zip', # 19G, 118k images
+ 'http://images.cocodataset.org/zips/val2017.zip', # 1G, 5k images
+ 'http://images.cocodataset.org/zips/test2017.zip'] # 7G, 41k images (optional)
+ download(urls, dir=dir / 'images', threads=3)
diff --git a/vouchervision/component_detector/data/coco128.yaml b/vouchervision/component_detector/data/coco128.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..2517d2079257571e90fcdf7e3c8f6e6c035f0b06
--- /dev/null
+++ b/vouchervision/component_detector/data/coco128.yaml
@@ -0,0 +1,30 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# COCO128 dataset https://www.kaggle.com/ultralytics/coco128 (first 128 images from COCO train2017) by Ultralytics
+# Example usage: python train.py --data coco128.yaml
+# parent
+# ├── yolov5
+# └── datasets
+# └── coco128 ← downloads here (7 MB)
+
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/coco128 # dataset root dir
+train: images/train2017 # train images (relative to 'path') 128 images
+val: images/train2017 # val images (relative to 'path') 128 images
+test: # test images (optional)
+
+# Classes
+nc: 80 # number of classes
+names: ['person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'traffic light',
+ 'fire hydrant', 'stop sign', 'parking meter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow',
+ 'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee',
+ 'skis', 'snowboard', 'sports ball', 'kite', 'baseball bat', 'baseball glove', 'skateboard', 'surfboard',
+ 'tennis racket', 'bottle', 'wine glass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple',
+ 'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza', 'donut', 'cake', 'chair', 'couch',
+ 'potted plant', 'bed', 'dining table', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cell phone',
+ 'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddy bear',
+ 'hair drier', 'toothbrush'] # class names
+
+
+# Download script/URL (optional)
+download: https://ultralytics.com/assets/coco128.zip
diff --git a/vouchervision/component_detector/data/hyps/hyp.Objects365.yaml b/vouchervision/component_detector/data/hyps/hyp.Objects365.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..74971740f7c73bf661950f339792b790a26b2b1c
--- /dev/null
+++ b/vouchervision/component_detector/data/hyps/hyp.Objects365.yaml
@@ -0,0 +1,34 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Hyperparameters for Objects365 training
+# python train.py --weights yolov5m.pt --data Objects365.yaml --evolve
+# See Hyperparameter Evolution tutorial for details https://github.com/ultralytics/yolov5#tutorials
+
+lr0: 0.00258
+lrf: 0.17
+momentum: 0.779
+weight_decay: 0.00058
+warmup_epochs: 1.33
+warmup_momentum: 0.86
+warmup_bias_lr: 0.0711
+box: 0.0539
+cls: 0.299
+cls_pw: 0.825
+obj: 0.632
+obj_pw: 1.0
+iou_t: 0.2
+anchor_t: 3.44
+anchors: 3.2
+fl_gamma: 0.0
+hsv_h: 0.0188
+hsv_s: 0.704
+hsv_v: 0.36
+degrees: 0.0
+translate: 0.0902
+scale: 0.491
+shear: 0.0
+perspective: 0.0
+flipud: 0.0
+fliplr: 0.5
+mosaic: 1.0
+mixup: 0.0
+copy_paste: 0.0
diff --git a/vouchervision/component_detector/data/hyps/hyp.VOC.yaml b/vouchervision/component_detector/data/hyps/hyp.VOC.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..0aa4e7d9f8f5162653e3999b04b4636b103c355f
--- /dev/null
+++ b/vouchervision/component_detector/data/hyps/hyp.VOC.yaml
@@ -0,0 +1,40 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Hyperparameters for VOC training
+# python train.py --batch 128 --weights yolov5m6.pt --data VOC.yaml --epochs 50 --img 512 --hyp hyp.scratch-med.yaml --evolve
+# See Hyperparameter Evolution tutorial for details https://github.com/ultralytics/yolov5#tutorials
+
+# YOLOv5 Hyperparameter Evolution Results
+# Best generation: 467
+# Last generation: 996
+# metrics/precision, metrics/recall, metrics/mAP_0.5, metrics/mAP_0.5:0.95, val/box_loss, val/obj_loss, val/cls_loss
+# 0.87729, 0.85125, 0.91286, 0.72664, 0.0076739, 0.0042529, 0.0013865
+
+lr0: 0.00334
+lrf: 0.15135
+momentum: 0.74832
+weight_decay: 0.00025
+warmup_epochs: 3.3835
+warmup_momentum: 0.59462
+warmup_bias_lr: 0.18657
+box: 0.02
+cls: 0.21638
+cls_pw: 0.5
+obj: 0.51728
+obj_pw: 0.67198
+iou_t: 0.2
+anchor_t: 3.3744
+fl_gamma: 0.0
+hsv_h: 0.01041
+hsv_s: 0.54703
+hsv_v: 0.27739
+degrees: 0.0
+translate: 0.04591
+scale: 0.75544
+shear: 0.0
+perspective: 0.0
+flipud: 0.0
+fliplr: 0.5
+mosaic: 0.85834
+mixup: 0.04266
+copy_paste: 0.0
+anchors: 3.412
diff --git a/vouchervision/component_detector/data/hyps/hyp.scratch-high.yaml b/vouchervision/component_detector/data/hyps/hyp.scratch-high.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..123cc8407413e9c130e21a3b5dd8ed33a3632db5
--- /dev/null
+++ b/vouchervision/component_detector/data/hyps/hyp.scratch-high.yaml
@@ -0,0 +1,34 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Hyperparameters for high-augmentation COCO training from scratch
+# python train.py --batch 32 --cfg yolov5m6.yaml --weights '' --data coco.yaml --img 1280 --epochs 300
+# See tutorials for hyperparameter evolution https://github.com/ultralytics/yolov5#tutorials
+
+lr0: 0.01 # initial learning rate (SGD=1E-2, Adam=1E-3)
+lrf: 0.1 # final OneCycleLR learning rate (lr0 * lrf)
+momentum: 0.937 # SGD momentum/Adam beta1
+weight_decay: 0.0005 # optimizer weight decay 5e-4
+warmup_epochs: 3.0 # warmup epochs (fractions ok)
+warmup_momentum: 0.8 # warmup initial momentum
+warmup_bias_lr: 0.1 # warmup initial bias lr
+box: 0.05 # box loss gain
+cls: 0.3 # cls loss gain
+cls_pw: 1.0 # cls BCELoss positive_weight
+obj: 0.7 # obj loss gain (scale with pixels)
+obj_pw: 1.0 # obj BCELoss positive_weight
+iou_t: 0.20 # IoU training threshold
+anchor_t: 4.0 # anchor-multiple threshold
+# anchors: 3 # anchors per output layer (0 to ignore)
+fl_gamma: 0.0 # focal loss gamma (efficientDet default gamma=1.5)
+hsv_h: 0.015 # image HSV-Hue augmentation (fraction)
+hsv_s: 0.7 # image HSV-Saturation augmentation (fraction)
+hsv_v: 0.4 # image HSV-Value augmentation (fraction)
+degrees: 0.0 # image rotation (+/- deg)
+translate: 0.1 # image translation (+/- fraction)
+scale: 0.9 # image scale (+/- gain)
+shear: 0.0 # image shear (+/- deg)
+perspective: 0.0 # image perspective (+/- fraction), range 0-0.001
+flipud: 0.0 # image flip up-down (probability)
+fliplr: 0.5 # image flip left-right (probability)
+mosaic: 1.0 # image mosaic (probability)
+mixup: 0.1 # image mixup (probability)
+copy_paste: 0.1 # segment copy-paste (probability)
diff --git a/vouchervision/component_detector/data/hyps/hyp.scratch-low.yaml b/vouchervision/component_detector/data/hyps/hyp.scratch-low.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..b9ef1d55a3b6ec8873ac87d6f4aa0ca081868bd6
--- /dev/null
+++ b/vouchervision/component_detector/data/hyps/hyp.scratch-low.yaml
@@ -0,0 +1,34 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Hyperparameters for low-augmentation COCO training from scratch
+# python train.py --batch 64 --cfg yolov5n6.yaml --weights '' --data coco.yaml --img 640 --epochs 300 --linear
+# See tutorials for hyperparameter evolution https://github.com/ultralytics/yolov5#tutorials
+
+lr0: 0.01 # initial learning rate (SGD=1E-2, Adam=1E-3)
+lrf: 0.01 # final OneCycleLR learning rate (lr0 * lrf)
+momentum: 0.937 # SGD momentum/Adam beta1
+weight_decay: 0.0005 # optimizer weight decay 5e-4
+warmup_epochs: 3.0 # warmup epochs (fractions ok)
+warmup_momentum: 0.8 # warmup initial momentum
+warmup_bias_lr: 0.1 # warmup initial bias lr
+box: 0.05 # box loss gain
+cls: 0.5 # cls loss gain
+cls_pw: 1.0 # cls BCELoss positive_weight
+obj: 1.0 # obj loss gain (scale with pixels)
+obj_pw: 1.0 # obj BCELoss positive_weight
+iou_t: 0.20 # IoU training threshold
+anchor_t: 4.0 # anchor-multiple threshold
+# anchors: 3 # anchors per output layer (0 to ignore)
+fl_gamma: 0.0 # focal loss gamma (efficientDet default gamma=1.5)
+hsv_h: 0.015 # image HSV-Hue augmentation (fraction)
+hsv_s: 0.7 # image HSV-Saturation augmentation (fraction)
+hsv_v: 0.4 # image HSV-Value augmentation (fraction)
+degrees: 0.0 # image rotation (+/- deg)
+translate: 0.1 # image translation (+/- fraction)
+scale: 0.5 # image scale (+/- gain)
+shear: 0.0 # image shear (+/- deg)
+perspective: 0.0 # image perspective (+/- fraction), range 0-0.001
+flipud: 0.0 # image flip up-down (probability)
+fliplr: 0.5 # image flip left-right (probability)
+mosaic: 1.0 # image mosaic (probability)
+mixup: 0.0 # image mixup (probability)
+copy_paste: 0.0 # segment copy-paste (probability)
diff --git a/vouchervision/component_detector/data/hyps/hyp.scratch-med.yaml b/vouchervision/component_detector/data/hyps/hyp.scratch-med.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d6867d7557bac73db7f8787db60cff4c4c64b440
--- /dev/null
+++ b/vouchervision/component_detector/data/hyps/hyp.scratch-med.yaml
@@ -0,0 +1,34 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Hyperparameters for medium-augmentation COCO training from scratch
+# python train.py --batch 32 --cfg yolov5m6.yaml --weights '' --data coco.yaml --img 1280 --epochs 300
+# See tutorials for hyperparameter evolution https://github.com/ultralytics/yolov5#tutorials
+
+lr0: 0.01 # initial learning rate (SGD=1E-2, Adam=1E-3)
+lrf: 0.1 # final OneCycleLR learning rate (lr0 * lrf)
+momentum: 0.937 # SGD momentum/Adam beta1
+weight_decay: 0.0005 # optimizer weight decay 5e-4
+warmup_epochs: 3.0 # warmup epochs (fractions ok)
+warmup_momentum: 0.8 # warmup initial momentum
+warmup_bias_lr: 0.1 # warmup initial bias lr
+box: 0.05 # box loss gain
+cls: 0.3 # cls loss gain
+cls_pw: 1.0 # cls BCELoss positive_weight
+obj: 0.7 # obj loss gain (scale with pixels)
+obj_pw: 1.0 # obj BCELoss positive_weight
+iou_t: 0.20 # IoU training threshold
+anchor_t: 4.0 # anchor-multiple threshold
+# anchors: 3 # anchors per output layer (0 to ignore)
+fl_gamma: 0.0 # focal loss gamma (efficientDet default gamma=1.5)
+hsv_h: 0.015 # image HSV-Hue augmentation (fraction)
+hsv_s: 0.7 # image HSV-Saturation augmentation (fraction)
+hsv_v: 0.4 # image HSV-Value augmentation (fraction)
+degrees: 0.0 # image rotation (+/- deg)
+translate: 0.1 # image translation (+/- fraction)
+scale: 0.9 # image scale (+/- gain)
+shear: 0.0 # image shear (+/- deg)
+perspective: 0.0 # image perspective (+/- fraction), range 0-0.001
+flipud: 0.0 # image flip up-down (probability)
+fliplr: 0.5 # image flip left-right (probability)
+mosaic: 1.0 # image mosaic (probability)
+mixup: 0.1 # image mixup (probability)
+copy_paste: 0.0 # segment copy-paste (probability)
diff --git a/vouchervision/component_detector/data/images/bus.jpg b/vouchervision/component_detector/data/images/bus.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..2cf0dab1214b3c06668e2c6e3a1666463acfe88c
--- /dev/null
+++ b/vouchervision/component_detector/data/images/bus.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33b198a1d2839bb9ac4c65d61f9e852196793cae9a0781360859425f6022b69c
+size 487438
diff --git a/vouchervision/component_detector/data/images/zidane.jpg b/vouchervision/component_detector/data/images/zidane.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..6d86f9edfce6353b027f16b9df7a973c72e598ba
--- /dev/null
+++ b/vouchervision/component_detector/data/images/zidane.jpg
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:356dad2107bb0254e4e4a81bc1d9c7140043e88569d546e5b404b19bffa77d0a
+size 168949
diff --git a/vouchervision/component_detector/data/scripts/download_weights.sh b/vouchervision/component_detector/data/scripts/download_weights.sh
new file mode 100644
index 0000000000000000000000000000000000000000..e9fa65394178005ba42ad02b91fed2873effb66b
--- /dev/null
+++ b/vouchervision/component_detector/data/scripts/download_weights.sh
@@ -0,0 +1,20 @@
+#!/bin/bash
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Download latest models from https://github.com/ultralytics/yolov5/releases
+# Example usage: bash path/to/download_weights.sh
+# parent
+# └── yolov5
+# ├── yolov5s.pt ← downloads here
+# ├── yolov5m.pt
+# └── ...
+
+python - <= cls >= 0, f'incorrect class index {cls}'
+
+ # Write YOLO label
+ if id not in shapes:
+ shapes[id] = Image.open(file).size
+ box = xyxy2xywhn(box[None].astype(np.float), w=shapes[id][0], h=shapes[id][1], clip=True)
+ with open((labels / id).with_suffix('.txt'), 'a') as f:
+ f.write(f"{cls} {' '.join(f'{x:.6f}' for x in box[0])}\n") # write label.txt
+ except Exception as e:
+ print(f'WARNING: skipping one label for {file}: {e}')
+
+
+ # Download manually from https://challenge.xviewdataset.org
+ dir = Path(yaml['path']) # dataset root dir
+ # urls = ['https://d307kc0mrhucc3.cloudfront.net/train_labels.zip', # train labels
+ # 'https://d307kc0mrhucc3.cloudfront.net/train_images.zip', # 15G, 847 train images
+ # 'https://d307kc0mrhucc3.cloudfront.net/val_images.zip'] # 5G, 282 val images (no labels)
+ # download(urls, dir=dir, delete=False)
+
+ # Convert labels
+ convert_labels(dir / 'xView_train.geojson')
+
+ # Move images
+ images = Path(dir / 'images')
+ images.mkdir(parents=True, exist_ok=True)
+ Path(dir / 'train_images').rename(dir / 'images' / 'train')
+ Path(dir / 'val_images').rename(dir / 'images' / 'val')
+
+ # Split
+ autosplit(dir / 'images' / 'train')
diff --git a/vouchervision/component_detector/detect.py b/vouchervision/component_detector/detect.py
new file mode 100644
index 0000000000000000000000000000000000000000..75c350dbd615a05d62be23f09985860620ebf174
--- /dev/null
+++ b/vouchervision/component_detector/detect.py
@@ -0,0 +1,289 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Run inference on images, videos, directories, streams, etc.
+
+Usage - sources:
+ $ python path/to/detect.py --weights yolov5s.pt --source 0 # webcam
+ img.jpg # image
+ vid.mp4 # video
+ path/ # directory
+ path/*.jpg # glob
+ 'https://youtu.be/Zgi9g1ksQHc' # YouTube
+ 'rtsp://example.com/media.mp4' # RTSP, RTMP, HTTP stream
+
+Usage - formats:
+ $ python path/to/detect.py --weights yolov5s.pt # PyTorch
+ yolov5s.torchscript # TorchScript
+ yolov5s.onnx # ONNX Runtime or OpenCV DNN with --dnn
+ yolov5s.xml # OpenVINO
+ yolov5s.engine # TensorRT
+ yolov5s.mlmodel # CoreML (macOS-only)
+ yolov5s_saved_model # TensorFlow SavedModel
+ yolov5s.pb # TensorFlow GraphDef
+ yolov5s.tflite # TensorFlow Lite
+ yolov5s_edgetpu.tflite # TensorFlow Edge TPU
+"""
+
+import os, argparse, sys, inspect
+from pathlib import Path
+from pandas import read_csv
+import torch
+import torch.backends.cudnn as cudnn
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[0] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+
+from models.common import DetectMultiBackend
+from utils.datasets import IMG_FORMATS, VID_FORMATS, LoadImages, LoadStreams
+from utils.general import (LOGGER, check_file, check_img_size, check_imshow, check_requirements, colorstr, cv2,
+ increment_path, non_max_suppression, print_args, scale_coords, strip_optimizer, xyxy2xywh)
+from utils.plots import Annotator, AnnotatorLandmark, colors, save_one_box
+from utils.torch_utils import select_device, time_sync
+
+@torch.no_grad()
+def run(
+ weights=ROOT / 'yolov5s.pt', # model.pt path(s)
+ source=ROOT / 'data/images', # file/dir/URL/glob, 0 for webcam
+ data=ROOT / 'data/coco128.yaml', # dataset.yaml path
+ anno_type='PREP',
+ ignore_objects_for_overlay = [],
+ mode = None,
+ imgsz=(640, 640), # inference size (height, width)
+ conf_thres=0.25, # confidence threshold
+ iou_thres=0.45, # NMS IOU threshold
+ max_det=3000, # maximum detections per image
+ device='', # cuda device, i.e. 0 or 0,1,2,3 or cpu
+ view_img=False, # show results
+ save_txt=True, # save results to *.txt
+ save_conf=False, # save confidences in --save-txt labels
+ save_crop=False, # save cropped prediction boxes
+ nosave=True, # do not save images/videos
+ classes=None, # filter by class: --class 0, or --class 0 2 3
+ agnostic_nms=False, # class-agnostic NMS
+ augment=False, # augmented inference
+ visualize=False, # visualize features
+ update=False, # update all models
+ project=ROOT / 'runs/detect', # save results to project/name
+ name='exp', # save results to project/name
+ exist_ok=False, # existing project/name ok, do not increment
+ line_thickness=10, # bounding box thickness (pixels)
+ hide_labels=False, # hide labels
+ hide_conf=False, # hide confidences
+ half=False, # use FP16 half-precision inference
+ dnn=False, # use OpenCV DNN for ONNX inference
+ LOGGER=LOGGER,
+):
+ dir_base = os.path.dirname(__file__)
+ if anno_type in ['PREP','Archival_Detector']:
+ COLOR_PROFILE = read_csv(os.path.join(dir_base,'color_profiles','ColorProfile__PREP.csv'))
+ elif anno_type in ['PLANT','Plant_Detector']:
+ COLOR_PROFILE = read_csv(os.path.join(dir_base,'color_profiles','ColorProfile__PLANT.csv'))
+ elif anno_type in ['LANDMARK','Landmark_Detector']:
+ COLOR_PROFILE = read_csv(os.path.join(dir_base,'color_profiles','ColorProfile__LANDMARK.csv'))
+ elif anno_type in ['Arm','ARM','Armature_Detector','Armature']:
+ COLOR_PROFILE = read_csv(os.path.join(dir_base,'color_profiles','ColorProfile__LANDMARK_ARM.csv'))
+ if isinstance(source, list):
+ source_0 = source[0]
+ save_img = not nosave and not source_0.endswith('.txt') # save inference images
+ is_file = Path(source_0).suffix[1:] in (IMG_FORMATS + VID_FORMATS)
+ is_url = source_0.lower().startswith(('rtsp://', 'rtmp://', 'http://', 'https://'))
+ webcam = source_0.isnumeric() or source_0.endswith('.txt') or (is_url and not is_file)
+ # if is_url and is_file:
+ # source = check_file(source) # download
+ else:
+ source = str(source)
+ save_img = not nosave and not source.endswith('.txt') # save inference images
+ is_file = Path(source).suffix[1:] in (IMG_FORMATS + VID_FORMATS)
+ is_url = source.lower().startswith(('rtsp://', 'rtmp://', 'http://', 'https://'))
+ webcam = source.isnumeric() or source.endswith('.txt') or (is_url and not is_file)
+ if is_url and is_file:
+ source = check_file(source) # download
+
+ # Directories
+ if mode is None:
+ save_dir = increment_path(Path(project) / name, exist_ok=exist_ok) # increment run
+ elif mode == 'Landmark':
+ save_dir = Path(project)
+ else:
+ save_dir = Path(project)
+ (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir
+
+ # Load model
+ device = select_device(device)
+ model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half)
+ stride, names, pt = model.stride, model.names, model.pt
+ imgsz = check_img_size(imgsz, s=stride) # check image size
+
+ # Dataloader
+ if isinstance(source, list):
+ LOGGER.info(f'Processing images from {os.path.dirname(source[0])}')
+ else:
+ LOGGER.info(f'Processing images from {source}')
+ if webcam:
+ view_img = check_imshow()
+ cudnn.benchmark = True # set True to speed up constant image size inference
+ dataset = LoadStreams(source, img_size=imgsz, stride=stride, auto=pt)
+ bs = len(dataset) # batch_size
+ else:
+ dataset = LoadImages(source, img_size=imgsz, stride=stride, auto=pt)
+ bs = 1 # batch_size
+ vid_path, vid_writer = [None] * bs, [None] * bs
+
+ # Run inference
+ model.warmup(imgsz=(1 if pt else bs, 3, *imgsz)) # warmup
+ dt, seen = [0.0, 0.0, 0.0], 0
+ for path, im, im0s, vid_cap, s in dataset:
+ t1 = time_sync()
+ im = torch.from_numpy(im).to(device)
+ im = im.half() if model.fp16 else im.float() # uint8 to fp16/32
+ im /= 255 # 0 - 255 to 0.0 - 1.0
+ if len(im.shape) == 3:
+ im = im[None] # expand for batch dim
+ t2 = time_sync()
+ dt[0] += t2 - t1
+
+ # Inference
+ visualize = increment_path(save_dir / Path(path).stem, mkdir=True) if visualize else False
+ pred = model(im, augment=augment, visualize=visualize)
+ t3 = time_sync()
+ dt[1] += t3 - t2
+
+ # NMS
+ pred = non_max_suppression(pred, conf_thres, iou_thres, classes, agnostic_nms, max_det=max_det)
+ dt[2] += time_sync() - t3
+
+ # Second-stage classifier (optional)
+ # pred = utils.general.apply_classifier(pred, classifier_model, im, im0s)
+
+ # Process predictions
+ for i, det in enumerate(pred): # per image
+ seen += 1
+ if webcam: # batch_size >= 1
+ p, im0, frame = path[i], im0s[i].copy(), dataset.count
+ s += f'{i}: '
+ else:
+ p, im0, frame = path, im0s.copy(), getattr(dataset, 'frame', 0)
+
+ p = Path(p) # to Path
+ save_path = str(save_dir / p.name) # im.jpg
+ txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}') # im.txt
+ s += '%gx%g ' % im.shape[2:] # print string
+ gn = torch.tensor(im0.shape)[[1, 0, 1, 0]] # normalization gain whwh
+ imc = im0.copy() if save_crop else im0 # for save_crop
+ if mode == 'Landmark':
+ annotator = AnnotatorLandmark(im0, line_width=line_thickness, example=str(names))
+ else:
+ annotator = Annotator(im0, line_width=line_thickness, example=str(names))
+ if len(det):
+ # Rescale boxes from img_size to im0 size
+ det[:, :4] = scale_coords(im.shape[2:], det[:, :4], im0.shape).round()
+
+ # Print results
+ for c in det[:, -1].unique():
+ n = (det[:, -1] == c).sum() # detections per class
+ s += f"{n} {names[int(c)]}{'s' * (n > 1)}, " # add to string
+
+ # Write results
+ for *xyxy, conf, cls in reversed(det):
+ if save_txt: # Write to file
+ xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
+ line = (cls, *xywh, conf) if save_conf else (cls, *xywh) # label format
+ with open(txt_path + '.txt', 'a') as f:
+ f.write(('%g ' * len(line)).rstrip() % line + '\n')
+
+ if save_img or save_crop or view_img: # Add bbox to image
+ c = int(cls) # integer class
+ label = None if hide_labels else (names[c] if hide_conf else f'{names[c]} {conf:.2f}')
+ # print(label)
+ # print(c)
+ if names[c] not in ignore_objects_for_overlay: #################################################################################### ignore leaf_partial ##################
+ annotator.box_label(xyxy, label, color=colors(c, True))
+ if save_crop:
+ save_one_box(xyxy, imc, file=save_dir / 'crops' / names[c] / f'{p.stem}.jpg', BGR=True)
+
+ # Stream results
+ im0 = annotator.result()
+ if view_img:
+ cv2.imshow(str(p), im0)
+ cv2.waitKey(1) # 1 millisecond
+
+ # Save results (image with detections)
+ if save_img:
+ if dataset.mode == 'image':
+ cv2.imwrite(save_path, im0)
+ else: # 'video' or 'stream'
+ if vid_path[i] != save_path: # new video
+ vid_path[i] = save_path
+ if isinstance(vid_writer[i], cv2.VideoWriter):
+ vid_writer[i].release() # release previous video writer
+ if vid_cap: # video
+ fps = vid_cap.get(cv2.CAP_PROP_FPS)
+ w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
+ h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
+ else: # stream
+ fps, w, h = 30, im0.shape[1], im0.shape[0]
+ save_path = str(Path(save_path).with_suffix('.mp4')) # force *.mp4 suffix on results videos
+ vid_writer[i] = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
+ vid_writer[i].write(im0)
+
+ # Print time (inference-only)
+ LOGGER.info(f'{s}Done. ({t3 - t2:.3f}s)')
+
+ # Print results
+ t = tuple(x / seen * 1E3 for x in dt) # speeds per image
+ LOGGER.info(f'Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {(1, 3, *imgsz)}' % t)
+ if save_txt or save_img:
+ # s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
+ # LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}{s}")
+ s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
+ LOGGER.info(f"Results saved to {save_dir}")
+ LOGGER.info(f"{s}")
+ if update:
+ strip_optimizer(weights) # update model (to fix SourceChangeWarning)
+
+
+def parse_opt():
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--weights', nargs='+', type=str, default=ROOT / 'yolov5s.pt', help='model path(s)')
+ parser.add_argument('--source', type=str, default=ROOT / 'data/images', help='file/dir/URL/glob, 0 for webcam')
+ parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='(optional) dataset.yaml path')
+ parser.add_argument('--imgsz', '--img', '--img-size', nargs='+', type=int, default=[640], help='inference size h,w')
+ parser.add_argument('--conf-thres', type=float, default=0.25, help='confidence threshold')
+ parser.add_argument('--iou-thres', type=float, default=0.45, help='NMS IoU threshold')
+ parser.add_argument('--max-det', type=int, default=1000, help='maximum detections per image')
+ parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--view-img', action='store_true', help='show results')
+ parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
+ parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
+ parser.add_argument('--save-crop', action='store_true', help='save cropped prediction boxes')
+ parser.add_argument('--nosave', action='store_true', help='do not save images/videos')
+ parser.add_argument('--classes', nargs='+', type=int, help='filter by class: --classes 0, or --classes 0 2 3')
+ parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
+ parser.add_argument('--augment', action='store_true', help='augmented inference')
+ parser.add_argument('--visualize', action='store_true', help='visualize features')
+ parser.add_argument('--update', action='store_true', help='update all models')
+ parser.add_argument('--project', default=ROOT / 'runs/detect', help='save results to project/name')
+ parser.add_argument('--name', default='exp', help='save results to project/name')
+ parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+ parser.add_argument('--line-thickness', default=10, type=int, help='bounding box thickness (pixels)')
+ parser.add_argument('--hide-labels', default=False, action='store_true', help='hide labels')
+ parser.add_argument('--hide-conf', default=False, action='store_true', help='hide confidences')
+ parser.add_argument('--half', action='store_true', help='use FP16 half-precision inference')
+ parser.add_argument('--dnn', action='store_true', help='use OpenCV DNN for ONNX inference')
+ opt = parser.parse_args()
+ opt.imgsz *= 2 if len(opt.imgsz) == 1 else 1 # expand
+ print_args(vars(opt))
+ return opt
+
+
+def main(opt):
+ check_requirements(exclude=('tensorboard', 'thop'))
+ run(**vars(opt))
+
+
+if __name__ == "__main__":
+ opt = parse_opt()
+ main(opt)
diff --git a/vouchervision/component_detector/export.py b/vouchervision/component_detector/export.py
new file mode 100644
index 0000000000000000000000000000000000000000..a84a9bfee5e74df100d662fdea9681834e7c15cd
--- /dev/null
+++ b/vouchervision/component_detector/export.py
@@ -0,0 +1,612 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Export a YOLOv5 PyTorch model to other formats. TensorFlow exports authored by https://github.com/zldrobit
+
+Format | `export.py --include` | Model
+--- | --- | ---
+PyTorch | - | yolov5s.pt
+TorchScript | `torchscript` | yolov5s.torchscript
+ONNX | `onnx` | yolov5s.onnx
+OpenVINO | `openvino` | yolov5s_openvino_model/
+TensorRT | `engine` | yolov5s.engine
+CoreML | `coreml` | yolov5s.mlmodel
+TensorFlow SavedModel | `saved_model` | yolov5s_saved_model/
+TensorFlow GraphDef | `pb` | yolov5s.pb
+TensorFlow Lite | `tflite` | yolov5s.tflite
+TensorFlow Edge TPU | `edgetpu` | yolov5s_edgetpu.tflite
+TensorFlow.js | `tfjs` | yolov5s_web_model/
+
+Requirements:
+ $ pip install -r requirements.txt coremltools onnx onnx-simplifier onnxruntime openvino-dev tensorflow-cpu # CPU
+ $ pip install -r requirements.txt coremltools onnx onnx-simplifier onnxruntime-gpu openvino-dev tensorflow # GPU
+
+Usage:
+ $ python path/to/export.py --weights yolov5s.pt --include torchscript onnx openvino engine coreml tflite ...
+
+Inference:
+ $ python path/to/detect.py --weights yolov5s.pt # PyTorch
+ yolov5s.torchscript # TorchScript
+ yolov5s.onnx # ONNX Runtime or OpenCV DNN with --dnn
+ yolov5s.xml # OpenVINO
+ yolov5s.engine # TensorRT
+ yolov5s.mlmodel # CoreML (macOS-only)
+ yolov5s_saved_model # TensorFlow SavedModel
+ yolov5s.pb # TensorFlow GraphDef
+ yolov5s.tflite # TensorFlow Lite
+ yolov5s_edgetpu.tflite # TensorFlow Edge TPU
+
+TensorFlow.js:
+ $ cd .. && git clone https://github.com/zldrobit/tfjs-yolov5-example.git && cd tfjs-yolov5-example
+ $ npm install
+ $ ln -s ../../yolov5/yolov5s_web_model public/yolov5s_web_model
+ $ npm start
+"""
+
+import argparse
+import json
+import os
+import platform
+import subprocess
+import sys
+import time
+import warnings
+from pathlib import Path
+
+import pandas as pd
+import torch
+from torch.utils.mobile_optimizer import optimize_for_mobile
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[0] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+if platform.system() != 'Windows':
+ ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+
+from models.experimental import attempt_load
+from models.yolo import Detect
+from utils.datasets import LoadImages
+from utils.general import (LOGGER, check_dataset, check_img_size, check_requirements, check_version, colorstr,
+ file_size, print_args, url2file)
+from utils.torch_utils import select_device
+
+
+def export_formats():
+ # YOLOv5 export formats
+ x = [
+ ['PyTorch', '-', '.pt', True],
+ ['TorchScript', 'torchscript', '.torchscript', True],
+ ['ONNX', 'onnx', '.onnx', True],
+ ['OpenVINO', 'openvino', '_openvino_model', False],
+ ['TensorRT', 'engine', '.engine', True],
+ ['CoreML', 'coreml', '.mlmodel', False],
+ ['TensorFlow SavedModel', 'saved_model', '_saved_model', True],
+ ['TensorFlow GraphDef', 'pb', '.pb', True],
+ ['TensorFlow Lite', 'tflite', '.tflite', False],
+ ['TensorFlow Edge TPU', 'edgetpu', '_edgetpu.tflite', False],
+ ['TensorFlow.js', 'tfjs', '_web_model', False],]
+ return pd.DataFrame(x, columns=['Format', 'Argument', 'Suffix', 'GPU'])
+
+
+def export_torchscript(model, im, file, optimize, prefix=colorstr('TorchScript:')):
+ # YOLOv5 TorchScript model export
+ try:
+ LOGGER.info(f'\n{prefix} starting export with torch {torch.__version__}...')
+ f = file.with_suffix('.torchscript')
+
+ ts = torch.jit.trace(model, im, strict=False)
+ d = {"shape": im.shape, "stride": int(max(model.stride)), "names": model.names}
+ extra_files = {'config.txt': json.dumps(d)} # torch._C.ExtraFilesMap()
+ if optimize: # https://pytorch.org/tutorials/recipes/mobile_interpreter.html
+ optimize_for_mobile(ts)._save_for_lite_interpreter(str(f), _extra_files=extra_files)
+ else:
+ ts.save(str(f), _extra_files=extra_files)
+
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'{prefix} export failure: {e}')
+ #https://github.com/ultralytics/yolov5/issues/10490
+ # try:
+ # LOGGER.info(f'\n{prefix} starting export with torch {torch.version}...')
+ # fl = file.with_suffix('.torchscript.ptl')
+ # ts = torch.jit.trace(model, im, strict=False)
+ # d = {"shape": im.shape, "stride": int(max(model.stride)), "names": model.names}
+ # extra_files = {'config.txt': json.dumps(d)} # torch._C.ExtraFilesMap()
+ # if optimize:
+ # optimize_for_mobile(ts)._save_for_lite_interpreter(str(fl), _extra_files=extra_files)
+ # else:
+ # ts.save(str(fl), _extra_files=extra_files)
+ # return fl, None
+ # except Exception as e:
+ # LOGGER.info(f'{prefix} export failure: {e}')
+ # return None, e
+
+
+def export_onnx(model, im, file, opset, train, dynamic, simplify, prefix=colorstr('ONNX:')):
+ # YOLOv5 ONNX export
+ try:
+ check_requirements(('onnx',))
+ import onnx
+
+ LOGGER.info(f'\n{prefix} starting export with onnx {onnx.__version__}...')
+ f = file.with_suffix('.onnx')
+
+ torch.onnx.export(
+ model,
+ im,
+ f,
+ verbose=False,
+ opset_version=opset,
+ training=torch.onnx.TrainingMode.TRAINING if train else torch.onnx.TrainingMode.EVAL,
+ do_constant_folding=not train,
+ input_names=['images'],
+ output_names=['output'],
+ dynamic_axes={
+ 'images': {
+ 0: 'batch',
+ 2: 'height',
+ 3: 'width'}, # shape(1,3,640,640)
+ 'output': {
+ 0: 'batch',
+ 1: 'anchors'} # shape(1,25200,85)
+ } if dynamic else None)
+
+ # Checks
+ model_onnx = onnx.load(f) # load onnx model
+ onnx.checker.check_model(model_onnx) # check onnx model
+
+ # Metadata
+ d = {'stride': int(max(model.stride)), 'names': model.names}
+ for k, v in d.items():
+ meta = model_onnx.metadata_props.add()
+ meta.key, meta.value = k, str(v)
+ onnx.save(model_onnx, f)
+
+ # Simplify
+ if simplify:
+ try:
+ check_requirements(('onnx-simplifier',))
+ import onnxsim
+
+ LOGGER.info(f'{prefix} simplifying with onnx-simplifier {onnxsim.__version__}...')
+ model_onnx, check = onnxsim.simplify(model_onnx,
+ dynamic_input_shape=dynamic,
+ input_shapes={'images': list(im.shape)} if dynamic else None)
+ assert check, 'assert check failed'
+ onnx.save(model_onnx, f)
+ except Exception as e:
+ LOGGER.info(f'{prefix} simplifier failure: {e}')
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'{prefix} export failure: {e}')
+
+
+def export_openvino(model, im, file, prefix=colorstr('OpenVINO:')):
+ # YOLOv5 OpenVINO export
+ try:
+ check_requirements(('openvino-dev',)) # requires openvino-dev: https://pypi.org/project/openvino-dev/
+ import openvino.inference_engine as ie
+
+ LOGGER.info(f'\n{prefix} starting export with openvino {ie.__version__}...')
+ f = str(file).replace('.pt', '_openvino_model' + os.sep)
+
+ cmd = f"mo --input_model {file.with_suffix('.onnx')} --output_dir {f}"
+ subprocess.check_output(cmd, shell=True)
+
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+
+
+def export_coreml(model, im, file, int8, half, prefix=colorstr('CoreML:')):
+ # YOLOv5 CoreML export
+ try:
+ check_requirements(('coremltools',))
+ import coremltools as ct
+
+ LOGGER.info(f'\n{prefix} starting export with coremltools {ct.__version__}...')
+ f = file.with_suffix('.mlmodel')
+
+ ts = torch.jit.trace(model, im, strict=False) # TorchScript model
+ ct_model = ct.convert(ts, inputs=[ct.ImageType('image', shape=im.shape, scale=1 / 255, bias=[0, 0, 0])])
+ bits, mode = (8, 'kmeans_lut') if int8 else (16, 'linear') if half else (32, None)
+ if bits < 32:
+ if platform.system() == 'Darwin': # quantization only supported on macOS
+ with warnings.catch_warnings():
+ warnings.filterwarnings("ignore", category=DeprecationWarning) # suppress numpy==1.20 float warning
+ ct_model = ct.models.neural_network.quantization_utils.quantize_weights(ct_model, bits, mode)
+ else:
+ print(f'{prefix} quantization only supported on macOS, skipping...')
+ ct_model.save(f)
+
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return ct_model, f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+ return None, None
+
+
+def export_engine(model, im, file, train, half, simplify, workspace=4, verbose=False, prefix=colorstr('TensorRT:')):
+ # YOLOv5 TensorRT export https://developer.nvidia.com/tensorrt
+ try:
+ assert im.device.type != 'cpu', 'export running on CPU but must be on GPU, i.e. `python export.py --device 0`'
+ if platform.system() == 'Linux':
+ check_requirements(('nvidia-tensorrt',), cmds=('-U --index-url https://pypi.ngc.nvidia.com',))
+ import tensorrt as trt
+
+ if trt.__version__[0] == '7': # TensorRT 7 handling https://github.com/ultralytics/yolov5/issues/6012
+ grid = model.model[-1].anchor_grid
+ model.model[-1].anchor_grid = [a[..., :1, :1, :] for a in grid]
+ export_onnx(model, im, file, 12, train, False, simplify) # opset 12
+ model.model[-1].anchor_grid = grid
+ else: # TensorRT >= 8
+ check_version(trt.__version__, '8.0.0', hard=True) # require tensorrt>=8.0.0
+ export_onnx(model, im, file, 13, train, False, simplify) # opset 13
+ onnx = file.with_suffix('.onnx')
+
+ LOGGER.info(f'\n{prefix} starting export with TensorRT {trt.__version__}...')
+ assert onnx.exists(), f'failed to export ONNX file: {onnx}'
+ f = file.with_suffix('.engine') # TensorRT engine file
+ logger = trt.Logger(trt.Logger.INFO)
+ if verbose:
+ logger.min_severity = trt.Logger.Severity.VERBOSE
+
+ builder = trt.Builder(logger)
+ config = builder.create_builder_config()
+ config.max_workspace_size = workspace * 1 << 30
+ # config.set_memory_pool_limit(trt.MemoryPoolType.WORKSPACE, workspace << 30) # fix TRT 8.4 deprecation notice
+
+ flag = (1 << int(trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH))
+ network = builder.create_network(flag)
+ parser = trt.OnnxParser(network, logger)
+ if not parser.parse_from_file(str(onnx)):
+ raise RuntimeError(f'failed to load ONNX file: {onnx}')
+
+ inputs = [network.get_input(i) for i in range(network.num_inputs)]
+ outputs = [network.get_output(i) for i in range(network.num_outputs)]
+ LOGGER.info(f'{prefix} Network Description:')
+ for inp in inputs:
+ LOGGER.info(f'{prefix}\tinput "{inp.name}" with shape {inp.shape} and dtype {inp.dtype}')
+ for out in outputs:
+ LOGGER.info(f'{prefix}\toutput "{out.name}" with shape {out.shape} and dtype {out.dtype}')
+
+ LOGGER.info(f'{prefix} building FP{16 if builder.platform_has_fast_fp16 else 32} engine in {f}')
+ if builder.platform_has_fast_fp16:
+ config.set_flag(trt.BuilderFlag.FP16)
+ with builder.build_engine(network, config) as engine, open(f, 'wb') as t:
+ t.write(engine.serialize())
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+
+
+def export_saved_model(model,
+ im,
+ file,
+ dynamic,
+ tf_nms=False,
+ agnostic_nms=False,
+ topk_per_class=100,
+ topk_all=100,
+ iou_thres=0.45,
+ conf_thres=0.25,
+ keras=False,
+ prefix=colorstr('TensorFlow SavedModel:')):
+ # YOLOv5 TensorFlow SavedModel export
+ try:
+ import tensorflow as tf
+ from tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2
+
+ from models.tf import TFDetect, TFModel
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflow {tf.__version__}...')
+ f = str(file).replace('.pt', '_saved_model')
+ batch_size, ch, *imgsz = list(im.shape) # BCHW
+
+ tf_model = TFModel(cfg=model.yaml, model=model, nc=model.nc, imgsz=imgsz)
+ im = tf.zeros((batch_size, *imgsz, ch)) # BHWC order for TensorFlow
+ _ = tf_model.predict(im, tf_nms, agnostic_nms, topk_per_class, topk_all, iou_thres, conf_thres)
+ inputs = tf.keras.Input(shape=(*imgsz, ch), batch_size=None if dynamic else batch_size)
+ outputs = tf_model.predict(inputs, tf_nms, agnostic_nms, topk_per_class, topk_all, iou_thres, conf_thres)
+ keras_model = tf.keras.Model(inputs=inputs, outputs=outputs)
+ keras_model.trainable = False
+ keras_model.summary()
+ if keras:
+ keras_model.save(f, save_format='tf')
+ else:
+ spec = tf.TensorSpec(keras_model.inputs[0].shape, keras_model.inputs[0].dtype)
+ m = tf.function(lambda x: keras_model(x)) # full model
+ m = m.get_concrete_function(spec)
+ frozen_func = convert_variables_to_constants_v2(m)
+ tfm = tf.Module()
+ tfm.__call__ = tf.function(lambda x: frozen_func(x)[:4] if tf_nms else frozen_func(x)[0], [spec])
+ tfm.__call__(im)
+ tf.saved_model.save(tfm,
+ f,
+ options=tf.saved_model.SaveOptions(experimental_custom_gradients=False)
+ if check_version(tf.__version__, '2.6') else tf.saved_model.SaveOptions())
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return keras_model, f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+ return None, None
+
+
+def export_pb(keras_model, im, file, prefix=colorstr('TensorFlow GraphDef:')):
+ # YOLOv5 TensorFlow GraphDef *.pb export https://github.com/leimao/Frozen_Graph_TensorFlow
+ try:
+ import tensorflow as tf
+ from tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflow {tf.__version__}...')
+ f = file.with_suffix('.pb')
+
+ m = tf.function(lambda x: keras_model(x)) # full model
+ m = m.get_concrete_function(tf.TensorSpec(keras_model.inputs[0].shape, keras_model.inputs[0].dtype))
+ frozen_func = convert_variables_to_constants_v2(m)
+ frozen_func.graph.as_graph_def()
+ tf.io.write_graph(graph_or_graph_def=frozen_func.graph, logdir=str(f.parent), name=f.name, as_text=False)
+
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+
+
+def export_tflite(keras_model, im, file, int8, data, nms, agnostic_nms, prefix=colorstr('TensorFlow Lite:')):
+ # YOLOv5 TensorFlow Lite export
+ try:
+ import tensorflow as tf
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflow {tf.__version__}...')
+ batch_size, ch, *imgsz = list(im.shape) # BCHW
+ f = str(file).replace('.pt', '-fp16.tflite')
+
+ converter = tf.lite.TFLiteConverter.from_keras_model(keras_model)
+ converter.target_spec.supported_ops = [tf.lite.OpsSet.TFLITE_BUILTINS]
+ converter.target_spec.supported_types = [tf.float16]
+ converter.optimizations = [tf.lite.Optimize.DEFAULT]
+ if int8:
+ from models.tf import representative_dataset_gen
+ dataset = LoadImages(check_dataset(data)['train'], img_size=imgsz, auto=False) # representative data
+ converter.representative_dataset = lambda: representative_dataset_gen(dataset, ncalib=100)
+ converter.target_spec.supported_ops = [tf.lite.OpsSet.TFLITE_BUILTINS_INT8]
+ converter.target_spec.supported_types = []
+ converter.inference_input_type = tf.uint8 # or tf.int8
+ converter.inference_output_type = tf.uint8 # or tf.int8
+ converter.experimental_new_quantizer = True
+ f = str(file).replace('.pt', '-int8.tflite')
+ if nms or agnostic_nms:
+ converter.target_spec.supported_ops.append(tf.lite.OpsSet.SELECT_TF_OPS)
+
+ tflite_model = converter.convert()
+ open(f, "wb").write(tflite_model)
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+
+
+def export_edgetpu(keras_model, im, file, prefix=colorstr('Edge TPU:')):
+ # YOLOv5 Edge TPU export https://coral.ai/docs/edgetpu/models-intro/
+ try:
+ cmd = 'edgetpu_compiler --version'
+ help_url = 'https://coral.ai/docs/edgetpu/compiler/'
+ assert platform.system() == 'Linux', f'export only supported on Linux. See {help_url}'
+ if subprocess.run(cmd + ' >/dev/null', shell=True).returncode != 0:
+ LOGGER.info(f'\n{prefix} export requires Edge TPU compiler. Attempting install from {help_url}')
+ sudo = subprocess.run('sudo --version >/dev/null', shell=True).returncode == 0 # sudo installed on system
+ for c in (
+ 'curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add -',
+ 'echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list',
+ 'sudo apt-get update', 'sudo apt-get install edgetpu-compiler'):
+ subprocess.run(c if sudo else c.replace('sudo ', ''), shell=True, check=True)
+ ver = subprocess.run(cmd, shell=True, capture_output=True, check=True).stdout.decode().split()[-1]
+
+ LOGGER.info(f'\n{prefix} starting export with Edge TPU compiler {ver}...')
+ f = str(file).replace('.pt', '-int8_edgetpu.tflite') # Edge TPU model
+ f_tfl = str(file).replace('.pt', '-int8.tflite') # TFLite model
+
+ cmd = f"edgetpu_compiler -s -o {file.parent} {f_tfl}"
+ subprocess.run(cmd, shell=True, check=True)
+
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+
+
+def export_tfjs(keras_model, im, file, prefix=colorstr('TensorFlow.js:')):
+ # YOLOv5 TensorFlow.js export
+ try:
+ check_requirements(('tensorflowjs',))
+ import re
+
+ import tensorflowjs as tfjs
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflowjs {tfjs.__version__}...')
+ f = str(file).replace('.pt', '_web_model') # js dir
+ f_pb = file.with_suffix('.pb') # *.pb path
+ f_json = f + '/model.json' # *.json path
+
+ cmd = f'tensorflowjs_converter --input_format=tf_frozen_model ' \
+ f'--output_node_names="Identity,Identity_1,Identity_2,Identity_3" {f_pb} {f}'
+ subprocess.run(cmd, shell=True)
+
+ with open(f_json) as j:
+ json = j.read()
+ with open(f_json, 'w') as j: # sort JSON Identity_* in ascending order
+ subst = re.sub(
+ r'{"outputs": {"Identity.?.?": {"name": "Identity.?.?"}, '
+ r'"Identity.?.?": {"name": "Identity.?.?"}, '
+ r'"Identity.?.?": {"name": "Identity.?.?"}, '
+ r'"Identity.?.?": {"name": "Identity.?.?"}}}', r'{"outputs": {"Identity": {"name": "Identity"}, '
+ r'"Identity_1": {"name": "Identity_1"}, '
+ r'"Identity_2": {"name": "Identity_2"}, '
+ r'"Identity_3": {"name": "Identity_3"}}}', json)
+ j.write(subst)
+
+ LOGGER.info(f'{prefix} export success, saved as {f} ({file_size(f):.1f} MB)')
+ return f
+ except Exception as e:
+ LOGGER.info(f'\n{prefix} export failure: {e}')
+
+
+@torch.no_grad()
+def run(
+ data=ROOT / 'data/coco128.yaml', # 'dataset.yaml path'
+ weights=ROOT / 'yolov5s.pt', # weights path
+ imgsz=(640, 640), # image (height, width)
+ batch_size=1, # batch size
+ device='cpu', # cuda device, i.e. 0 or 0,1,2,3 or cpu
+ include=('torchscript', 'onnx'), # include formats
+ half=False, # FP16 half-precision export
+ inplace=False, # set YOLOv5 Detect() inplace=True
+ train=False, # model.train() mode
+ optimize=False, # TorchScript: optimize for mobile
+ int8=False, # CoreML/TF INT8 quantization
+ dynamic=False, # ONNX/TF: dynamic axes
+ simplify=False, # ONNX: simplify model
+ opset=12, # ONNX: opset version
+ verbose=False, # TensorRT: verbose log
+ workspace=4, # TensorRT: workspace size (GB)
+ nms=False, # TF: add NMS to model
+ agnostic_nms=False, # TF: add agnostic NMS to model
+ topk_per_class=100, # TF.js NMS: topk per class to keep
+ topk_all=100, # TF.js NMS: topk for all classes to keep
+ iou_thres=0.45, # TF.js NMS: IoU threshold
+ conf_thres=0.25, # TF.js NMS: confidence threshold
+):
+ t = time.time()
+ include = [x.lower() for x in include] # to lowercase
+ formats = tuple(export_formats()['Argument'][1:]) # --include arguments
+ flags = [x in include for x in formats]
+ assert sum(flags) == len(include), f'ERROR: Invalid --include {include}, valid --include arguments are {formats}'
+ jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs = flags # export booleans
+ file = Path(url2file(weights) if str(weights).startswith(('http:/', 'https:/')) else weights) # PyTorch weights
+
+ # Load PyTorch model
+ device = select_device(device)
+ if half:
+ assert device.type != 'cpu' or coreml, '--half only compatible with GPU export, i.e. use --device 0'
+ model = attempt_load(weights, map_location=device, inplace=True, fuse=True) # load FP32 model
+ nc, names = model.nc, model.names # number of classes, class names
+
+ # Checks
+ imgsz *= 2 if len(imgsz) == 1 else 1 # expand
+ assert nc == len(names), f'Model class count {nc} != len(names) {len(names)}'
+
+ # Input
+ gs = int(max(model.stride)) # grid size (max stride)
+ imgsz = [check_img_size(x, gs) for x in imgsz] # verify img_size are gs-multiples
+ im = torch.zeros(batch_size, 3, *imgsz).to(device) # image size(1,3,320,192) BCHW iDetection
+
+ # Update model
+ if half and not coreml:
+ im, model = im.half(), model.half() # to FP16
+ model.train() if train else model.eval() # training mode = no Detect() layer grid construction
+ for k, m in model.named_modules():
+ if isinstance(m, Detect):
+ m.inplace = inplace
+ m.onnx_dynamic = dynamic
+ m.export = True
+
+ for _ in range(2):
+ y = model(im) # dry runs
+ shape = tuple(y[0].shape) # model output shape
+ LOGGER.info(f"\n{colorstr('PyTorch:')} starting from {file} with output shape {shape} ({file_size(file):.1f} MB)")
+
+ # Exports
+ f = [''] * 10 # exported filenames
+ warnings.filterwarnings(action='ignore', category=torch.jit.TracerWarning) # suppress TracerWarning
+ if jit:
+ f[0] = export_torchscript(model, im, file, optimize)
+ if engine: # TensorRT required before ONNX
+ f[1] = export_engine(model, im, file, train, half, simplify, workspace, verbose)
+ if onnx or xml: # OpenVINO requires ONNX
+ f[2] = export_onnx(model, im, file, opset, train, dynamic, simplify)
+ if xml: # OpenVINO
+ f[3] = export_openvino(model, im, file)
+ if coreml:
+ _, f[4] = export_coreml(model, im, file, int8, half)
+
+ # TensorFlow Exports
+ if any((saved_model, pb, tflite, edgetpu, tfjs)):
+ if int8 or edgetpu: # TFLite --int8 bug https://github.com/ultralytics/yolov5/issues/5707
+ check_requirements(('flatbuffers==1.12',)) # required before `import tensorflow`
+ assert not (tflite and tfjs), 'TFLite and TF.js models must be exported separately, please pass only one type.'
+ model, f[5] = export_saved_model(model.cpu(),
+ im,
+ file,
+ dynamic,
+ tf_nms=nms or agnostic_nms or tfjs,
+ agnostic_nms=agnostic_nms or tfjs,
+ topk_per_class=topk_per_class,
+ topk_all=topk_all,
+ conf_thres=conf_thres,
+ iou_thres=iou_thres) # keras model
+ if pb or tfjs: # pb prerequisite to tfjs
+ f[6] = export_pb(model, im, file)
+ if tflite or edgetpu:
+ f[7] = export_tflite(model, im, file, int8=int8 or edgetpu, data=data, nms=nms, agnostic_nms=agnostic_nms)
+ if edgetpu:
+ f[8] = export_edgetpu(model, im, file)
+ if tfjs:
+ f[9] = export_tfjs(model, im, file)
+
+ # Finish
+ f = [str(x) for x in f if x] # filter out '' and None
+ if any(f):
+ LOGGER.info(f'\nExport complete ({time.time() - t:.2f}s)'
+ f"\nResults saved to {colorstr('bold', file.parent.resolve())}"
+ f"\nDetect: python detect.py --weights {f[-1]}"
+ f"\nPyTorch Hub: model = torch.hub.load('ultralytics/yolov5', 'custom', '{f[-1]}')"
+ f"\nValidate: python val.py --weights {f[-1]}"
+ f"\nVisualize: https://netron.app")
+ return f # return list of exported files/dirs
+
+
+def parse_opt():
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='dataset.yaml path')
+ parser.add_argument('--weights', nargs='+', type=str, default=ROOT / 'yolov5s.pt', help='model.pt path(s)')
+ parser.add_argument('--imgsz', '--img', '--img-size', nargs='+', type=int, default=[640, 640], help='image (h, w)')
+ parser.add_argument('--batch-size', type=int, default=1, help='batch size')
+ parser.add_argument('--device', default='cpu', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--half', action='store_true', help='FP16 half-precision export')
+ parser.add_argument('--inplace', action='store_true', help='set YOLOv5 Detect() inplace=True')
+ parser.add_argument('--train', action='store_true', help='model.train() mode')
+ parser.add_argument('--optimize', action='store_true', help='TorchScript: optimize for mobile')
+ parser.add_argument('--int8', action='store_true', help='CoreML/TF INT8 quantization')
+ parser.add_argument('--dynamic', action='store_true', help='ONNX/TF: dynamic axes')
+ parser.add_argument('--simplify', action='store_true', help='ONNX: simplify model')
+ parser.add_argument('--opset', type=int, default=12, help='ONNX: opset version')
+ parser.add_argument('--verbose', action='store_true', help='TensorRT: verbose log')
+ parser.add_argument('--workspace', type=int, default=4, help='TensorRT: workspace size (GB)')
+ parser.add_argument('--nms', action='store_true', help='TF: add NMS to model')
+ parser.add_argument('--agnostic-nms', action='store_true', help='TF: add agnostic NMS to model')
+ parser.add_argument('--topk-per-class', type=int, default=100, help='TF.js NMS: topk per class to keep')
+ parser.add_argument('--topk-all', type=int, default=100, help='TF.js NMS: topk for all classes to keep')
+ parser.add_argument('--iou-thres', type=float, default=0.45, help='TF.js NMS: IoU threshold')
+ parser.add_argument('--conf-thres', type=float, default=0.25, help='TF.js NMS: confidence threshold')
+ parser.add_argument('--include',
+ nargs='+',
+ default=['torchscript', 'onnx'],
+ help='torchscript, onnx, openvino, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs')
+ opt = parser.parse_args()
+ print_args(vars(opt))
+ return opt
+
+
+def main(opt):
+ for opt.weights in (opt.weights if isinstance(opt.weights, list) else [opt.weights]):
+ run(**vars(opt))
+
+
+if __name__ == "__main__":
+ opt = parse_opt()
+ main(opt)
diff --git a/vouchervision/component_detector/export_torchscript.py b/vouchervision/component_detector/export_torchscript.py
new file mode 100644
index 0000000000000000000000000000000000000000..40733df90d47d82f8ad8cff926795ad9c028b164
--- /dev/null
+++ b/vouchervision/component_detector/export_torchscript.py
@@ -0,0 +1,864 @@
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+Export a YOLOv5 PyTorch model to other formats. TensorFlow exports authored by https://github.com/zldrobit
+
+Format | `export.py --include` | Model
+--- | --- | ---
+PyTorch | - | yolov5s.pt
+TorchScript | `torchscript` | yolov5s.torchscript
+ONNX | `onnx` | yolov5s.onnx
+OpenVINO | `openvino` | yolov5s_openvino_model/
+TensorRT | `engine` | yolov5s.engine
+CoreML | `coreml` | yolov5s.mlmodel
+TensorFlow SavedModel | `saved_model` | yolov5s_saved_model/
+TensorFlow GraphDef | `pb` | yolov5s.pb
+TensorFlow Lite | `tflite` | yolov5s.tflite
+TensorFlow Edge TPU | `edgetpu` | yolov5s_edgetpu.tflite
+TensorFlow.js | `tfjs` | yolov5s_web_model/
+PaddlePaddle | `paddle` | yolov5s_paddle_model/
+
+Requirements:
+ $ pip install -r requirements.txt coremltools onnx onnx-simplifier onnxruntime openvino-dev tensorflow-cpu # CPU
+ $ pip install -r requirements.txt coremltools onnx onnx-simplifier onnxruntime-gpu openvino-dev tensorflow # GPU
+
+Usage:
+ $ python export.py --weights yolov5s.pt --include torchscript onnx openvino engine coreml tflite ...
+
+Inference:
+ $ python detect.py --weights yolov5s.pt # PyTorch
+ yolov5s.torchscript # TorchScript
+ yolov5s.onnx # ONNX Runtime or OpenCV DNN with --dnn
+ yolov5s_openvino_model # OpenVINO
+ yolov5s.engine # TensorRT
+ yolov5s.mlmodel # CoreML (macOS-only)
+ yolov5s_saved_model # TensorFlow SavedModel
+ yolov5s.pb # TensorFlow GraphDef
+ yolov5s.tflite # TensorFlow Lite
+ yolov5s_edgetpu.tflite # TensorFlow Edge TPU
+ yolov5s_paddle_model # PaddlePaddle
+
+TensorFlow.js:
+ $ cd .. && git clone https://github.com/zldrobit/tfjs-yolov5-example.git && cd tfjs-yolov5-example
+ $ npm install
+ $ ln -s ../../yolov5/yolov5s_web_model public/yolov5s_web_model
+ $ npm start
+"""
+
+import argparse
+import contextlib
+import json
+import os
+import platform
+import re
+import subprocess
+import sys
+import time
+import warnings
+from pathlib import Path
+
+import pandas as pd
+import torch
+from torch.utils.mobile_optimizer import optimize_for_mobile
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[0] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+if platform.system() != 'Windows':
+ ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+
+from models.experimental import attempt_load
+from models.yolo_torchscript import ClassificationModel, Detect, DetectionModel, SegmentationModel
+from utils.dataloaders import LoadImages
+from utils.general_torchscript import (LOGGER, Profile, check_dataset, check_img_size, check_requirements, check_version,
+ check_yaml, colorstr, file_size, get_default_args, print_args, url2file, yaml_save)
+from utils.torch_utils_torchscript import select_device, smart_inference_mode
+
+MACOS = platform.system() == 'Darwin' # macOS environment
+
+
+class iOSModel(torch.nn.Module):
+
+ def __init__(self, model, im):
+ super().__init__()
+ b, c, h, w = im.shape # batch, channel, height, width
+ self.model = model
+ self.nc = model.nc # number of classes
+ if w == h:
+ self.normalize = 1. / w
+ else:
+ self.normalize = torch.tensor([1. / w, 1. / h, 1. / w, 1. / h]) # broadcast (slower, smaller)
+ # np = model(im)[0].shape[1] # number of points
+ # self.normalize = torch.tensor([1. / w, 1. / h, 1. / w, 1. / h]).expand(np, 4) # explicit (faster, larger)
+
+ def forward(self, x):
+ xywh, conf, cls = self.model(x)[0].squeeze().split((4, 1, self.nc), 1)
+ return cls * conf, xywh * self.normalize # confidence (3780, 80), coordinates (3780, 4)
+
+
+def export_formats():
+ # YOLOv5 export formats
+ x = [
+ ['PyTorch', '-', '.pt', True, True],
+ ['TorchScript', 'torchscript', '.torchscript', True, True],
+ ['ONNX', 'onnx', '.onnx', True, True],
+ ['OpenVINO', 'openvino', '_openvino_model', True, False],
+ ['TensorRT', 'engine', '.engine', False, True],
+ ['CoreML', 'coreml', '.mlmodel', True, False],
+ ['TensorFlow SavedModel', 'saved_model', '_saved_model', True, True],
+ ['TensorFlow GraphDef', 'pb', '.pb', True, True],
+ ['TensorFlow Lite', 'tflite', '.tflite', True, False],
+ ['TensorFlow Edge TPU', 'edgetpu', '_edgetpu.tflite', False, False],
+ ['TensorFlow.js', 'tfjs', '_web_model', False, False],
+ ['PaddlePaddle', 'paddle', '_paddle_model', True, True], ]
+ return pd.DataFrame(x, columns=['Format', 'Argument', 'Suffix', 'CPU', 'GPU'])
+
+
+def try_export(inner_func):
+ # YOLOv5 export decorator, i..e @try_export
+ inner_args = get_default_args(inner_func)
+
+ def outer_func(*args, **kwargs):
+ prefix = inner_args['prefix']
+ try:
+ with Profile() as dt:
+ f, model = inner_func(*args, **kwargs)
+ LOGGER.info(f'{prefix} export success ✅ {dt.t:.1f}s, saved as {f} ({file_size(f):.1f} MB)')
+ return f, model
+ except Exception as e:
+ LOGGER.info(f'{prefix} export failure ❌ {dt.t:.1f}s: {e}')
+ return None, None
+
+ return outer_func
+
+
+@try_export
+def export_torchscript(model, im, file, optimize, prefix=colorstr('TorchScript:')):
+ # YOLOv5 TorchScript model export
+ LOGGER.info(f'\n{prefix} starting export with torch {torch.__version__}...')
+ f = file.with_suffix('.torchscript')
+
+ ts = torch.jit.trace(model, im, strict=False)
+ d = {'shape': im.shape, 'stride': int(max(model.stride)), 'names': model.names}
+ extra_files = {'config.txt': json.dumps(d)} # torch._C.ExtraFilesMap()
+ if optimize: # https://pytorch.org/tutorials/recipes/mobile_interpreter.html
+ optimize_for_mobile(ts)._save_for_lite_interpreter(str(f), _extra_files=extra_files)
+ else:
+ # ts.save(str(f), _extra_files=extra_files)
+ optimize_for_mobile(ts)._save_for_lite_interpreter(str(f), _extra_files=extra_files)
+ return f, None
+
+
+@try_export
+def export_onnx(model, im, file, opset, dynamic, simplify, prefix=colorstr('ONNX:')):
+ # YOLOv5 ONNX export
+ check_requirements('onnx>=1.12.0')
+ import onnx
+
+ LOGGER.info(f'\n{prefix} starting export with onnx {onnx.__version__}...')
+ f = file.with_suffix('.onnx')
+
+ output_names = ['output0', 'output1'] if isinstance(model, SegmentationModel) else ['output0']
+ if dynamic:
+ dynamic = {'images': {0: 'batch', 2: 'height', 3: 'width'}} # shape(1,3,640,640)
+ if isinstance(model, SegmentationModel):
+ dynamic['output0'] = {0: 'batch', 1: 'anchors'} # shape(1,25200,85)
+ dynamic['output1'] = {0: 'batch', 2: 'mask_height', 3: 'mask_width'} # shape(1,32,160,160)
+ elif isinstance(model, DetectionModel):
+ dynamic['output0'] = {0: 'batch', 1: 'anchors'} # shape(1,25200,85)
+
+ torch.onnx.export(
+ model.cpu() if dynamic else model, # --dynamic only compatible with cpu
+ im.cpu() if dynamic else im,
+ f,
+ verbose=False,
+ opset_version=opset,
+ do_constant_folding=True, # WARNING: DNN inference with torch>=1.12 may require do_constant_folding=False
+ input_names=['images'],
+ output_names=output_names,
+ dynamic_axes=dynamic or None)
+
+ # Checks
+ model_onnx = onnx.load(f) # load onnx model
+ onnx.checker.check_model(model_onnx) # check onnx model
+
+ # Metadata
+ d = {'stride': int(max(model.stride)), 'names': model.names}
+ for k, v in d.items():
+ meta = model_onnx.metadata_props.add()
+ meta.key, meta.value = k, str(v)
+ onnx.save(model_onnx, f)
+
+ # Simplify
+ if simplify:
+ try:
+ cuda = torch.cuda.is_available()
+ check_requirements(('onnxruntime-gpu' if cuda else 'onnxruntime', 'onnx-simplifier>=0.4.1'))
+ import onnxsim
+
+ LOGGER.info(f'{prefix} simplifying with onnx-simplifier {onnxsim.__version__}...')
+ model_onnx, check = onnxsim.simplify(model_onnx)
+ assert check, 'assert check failed'
+ onnx.save(model_onnx, f)
+ except Exception as e:
+ LOGGER.info(f'{prefix} simplifier failure: {e}')
+ return f, model_onnx
+
+
+@try_export
+def export_openvino(file, metadata, half, int8, data, prefix=colorstr('OpenVINO:')):
+ # YOLOv5 OpenVINO export
+ check_requirements('openvino-dev>=2023.0') # requires openvino-dev: https://pypi.org/project/openvino-dev/
+ import openvino.runtime as ov # noqa
+ from openvino.tools import mo # noqa
+
+ LOGGER.info(f'\n{prefix} starting export with openvino {ov.__version__}...')
+ f = str(file).replace(file.suffix, f'_openvino_model{os.sep}')
+ f_onnx = file.with_suffix('.onnx')
+ f_ov = str(Path(f) / file.with_suffix('.xml').name)
+ if int8:
+ check_requirements('nncf>=2.4.0') # requires at least version 2.4.0 to use the post-training quantization
+ import nncf
+ import numpy as np
+ from openvino.runtime import Core
+
+ from utils.dataloaders import create_dataloader
+ core = Core()
+ onnx_model = core.read_model(f_onnx) # export
+
+ def prepare_input_tensor(image: np.ndarray):
+ input_tensor = image.astype(np.float32) # uint8 to fp16/32
+ input_tensor /= 255.0 # 0 - 255 to 0.0 - 1.0
+
+ if input_tensor.ndim == 3:
+ input_tensor = np.expand_dims(input_tensor, 0)
+ return input_tensor
+
+ def gen_dataloader(yaml_path, task='train', imgsz=640, workers=4):
+ data_yaml = check_yaml(yaml_path)
+ data = check_dataset(data_yaml)
+ dataloader = create_dataloader(data[task],
+ imgsz=imgsz,
+ batch_size=1,
+ stride=32,
+ pad=0.5,
+ single_cls=False,
+ rect=False,
+ workers=workers)[0]
+ return dataloader
+
+ # noqa: F811
+
+ def transform_fn(data_item):
+ """
+ Quantization transform function. Extracts and preprocess input data from dataloader item for quantization.
+ Parameters:
+ data_item: Tuple with data item produced by DataLoader during iteration
+ Returns:
+ input_tensor: Input data for quantization
+ """
+ img = data_item[0].numpy()
+ input_tensor = prepare_input_tensor(img)
+ return input_tensor
+
+ ds = gen_dataloader(data)
+ quantization_dataset = nncf.Dataset(ds, transform_fn)
+ ov_model = nncf.quantize(onnx_model, quantization_dataset, preset=nncf.QuantizationPreset.MIXED)
+ else:
+ ov_model = mo.convert_model(f_onnx, model_name=file.stem, framework='onnx', compress_to_fp16=half) # export
+
+ ov.serialize(ov_model, f_ov) # save
+ yaml_save(Path(f) / file.with_suffix('.yaml').name, metadata) # add metadata.yaml
+ return f, None
+
+
+@try_export
+def export_paddle(model, im, file, metadata, prefix=colorstr('PaddlePaddle:')):
+ # YOLOv5 Paddle export
+ check_requirements(('paddlepaddle', 'x2paddle'))
+ import x2paddle
+ from x2paddle.convert import pytorch2paddle
+
+ LOGGER.info(f'\n{prefix} starting export with X2Paddle {x2paddle.__version__}...')
+ f = str(file).replace('.pt', f'_paddle_model{os.sep}')
+
+ pytorch2paddle(module=model, save_dir=f, jit_type='trace', input_examples=[im]) # export
+ yaml_save(Path(f) / file.with_suffix('.yaml').name, metadata) # add metadata.yaml
+ return f, None
+
+
+@try_export
+def export_coreml(model, im, file, int8, half, nms, prefix=colorstr('CoreML:')):
+ # YOLOv5 CoreML export
+ check_requirements('coremltools')
+ import coremltools as ct
+
+ LOGGER.info(f'\n{prefix} starting export with coremltools {ct.__version__}...')
+ f = file.with_suffix('.mlmodel')
+
+ if nms:
+ model = iOSModel(model, im)
+ ts = torch.jit.trace(model, im, strict=False) # TorchScript model
+ ct_model = ct.convert(ts, inputs=[ct.ImageType('image', shape=im.shape, scale=1 / 255, bias=[0, 0, 0])])
+ bits, mode = (8, 'kmeans_lut') if int8 else (16, 'linear') if half else (32, None)
+ if bits < 32:
+ if MACOS: # quantization only supported on macOS
+ with warnings.catch_warnings():
+ warnings.filterwarnings('ignore', category=DeprecationWarning) # suppress numpy==1.20 float warning
+ ct_model = ct.models.neural_network.quantization_utils.quantize_weights(ct_model, bits, mode)
+ else:
+ print(f'{prefix} quantization only supported on macOS, skipping...')
+ ct_model.save(f)
+ return f, ct_model
+
+
+@try_export
+def export_engine(model, im, file, half, dynamic, simplify, workspace=4, verbose=False, prefix=colorstr('TensorRT:')):
+ # YOLOv5 TensorRT export https://developer.nvidia.com/tensorrt
+ assert im.device.type != 'cpu', 'export running on CPU but must be on GPU, i.e. `python export.py --device 0`'
+ try:
+ import tensorrt as trt
+ except Exception:
+ if platform.system() == 'Linux':
+ check_requirements('nvidia-tensorrt', cmds='-U --index-url https://pypi.ngc.nvidia.com')
+ import tensorrt as trt
+
+ if trt.__version__[0] == '7': # TensorRT 7 handling https://github.com/ultralytics/yolov5/issues/6012
+ grid = model.model[-1].anchor_grid
+ model.model[-1].anchor_grid = [a[..., :1, :1, :] for a in grid]
+ export_onnx(model, im, file, 12, dynamic, simplify) # opset 12
+ model.model[-1].anchor_grid = grid
+ else: # TensorRT >= 8
+ check_version(trt.__version__, '8.0.0', hard=True) # require tensorrt>=8.0.0
+ export_onnx(model, im, file, 12, dynamic, simplify) # opset 12
+ onnx = file.with_suffix('.onnx')
+
+ LOGGER.info(f'\n{prefix} starting export with TensorRT {trt.__version__}...')
+ assert onnx.exists(), f'failed to export ONNX file: {onnx}'
+ f = file.with_suffix('.engine') # TensorRT engine file
+ logger = trt.Logger(trt.Logger.INFO)
+ if verbose:
+ logger.min_severity = trt.Logger.Severity.VERBOSE
+
+ builder = trt.Builder(logger)
+ config = builder.create_builder_config()
+ config.max_workspace_size = workspace * 1 << 30
+ # config.set_memory_pool_limit(trt.MemoryPoolType.WORKSPACE, workspace << 30) # fix TRT 8.4 deprecation notice
+
+ flag = (1 << int(trt.NetworkDefinitionCreationFlag.EXPLICIT_BATCH))
+ network = builder.create_network(flag)
+ parser = trt.OnnxParser(network, logger)
+ if not parser.parse_from_file(str(onnx)):
+ raise RuntimeError(f'failed to load ONNX file: {onnx}')
+
+ inputs = [network.get_input(i) for i in range(network.num_inputs)]
+ outputs = [network.get_output(i) for i in range(network.num_outputs)]
+ for inp in inputs:
+ LOGGER.info(f'{prefix} input "{inp.name}" with shape{inp.shape} {inp.dtype}')
+ for out in outputs:
+ LOGGER.info(f'{prefix} output "{out.name}" with shape{out.shape} {out.dtype}')
+
+ if dynamic:
+ if im.shape[0] <= 1:
+ LOGGER.warning(f'{prefix} WARNING ⚠️ --dynamic model requires maximum --batch-size argument')
+ profile = builder.create_optimization_profile()
+ for inp in inputs:
+ profile.set_shape(inp.name, (1, *im.shape[1:]), (max(1, im.shape[0] // 2), *im.shape[1:]), im.shape)
+ config.add_optimization_profile(profile)
+
+ LOGGER.info(f'{prefix} building FP{16 if builder.platform_has_fast_fp16 and half else 32} engine as {f}')
+ if builder.platform_has_fast_fp16 and half:
+ config.set_flag(trt.BuilderFlag.FP16)
+ with builder.build_engine(network, config) as engine, open(f, 'wb') as t:
+ t.write(engine.serialize())
+ return f, None
+
+
+@try_export
+def export_saved_model(model,
+ im,
+ file,
+ dynamic,
+ tf_nms=False,
+ agnostic_nms=False,
+ topk_per_class=100,
+ topk_all=100,
+ iou_thres=0.45,
+ conf_thres=0.25,
+ keras=False,
+ prefix=colorstr('TensorFlow SavedModel:')):
+ # YOLOv5 TensorFlow SavedModel export
+ try:
+ import tensorflow as tf
+ except Exception:
+ check_requirements(f"tensorflow{'' if torch.cuda.is_available() else '-macos' if MACOS else '-cpu'}")
+ import tensorflow as tf
+ from tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2
+
+ from models.tf import TFModel
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflow {tf.__version__}...')
+ f = str(file).replace('.pt', '_saved_model')
+ batch_size, ch, *imgsz = list(im.shape) # BCHW
+
+ tf_model = TFModel(cfg=model.yaml, model=model, nc=model.nc, imgsz=imgsz)
+ im = tf.zeros((batch_size, *imgsz, ch)) # BHWC order for TensorFlow
+ _ = tf_model.predict(im, tf_nms, agnostic_nms, topk_per_class, topk_all, iou_thres, conf_thres)
+ inputs = tf.keras.Input(shape=(*imgsz, ch), batch_size=None if dynamic else batch_size)
+ outputs = tf_model.predict(inputs, tf_nms, agnostic_nms, topk_per_class, topk_all, iou_thres, conf_thres)
+ keras_model = tf.keras.Model(inputs=inputs, outputs=outputs)
+ keras_model.trainable = False
+ keras_model.summary()
+ if keras:
+ keras_model.save(f, save_format='tf')
+ else:
+ spec = tf.TensorSpec(keras_model.inputs[0].shape, keras_model.inputs[0].dtype)
+ m = tf.function(lambda x: keras_model(x)) # full model
+ m = m.get_concrete_function(spec)
+ frozen_func = convert_variables_to_constants_v2(m)
+ tfm = tf.Module()
+ tfm.__call__ = tf.function(lambda x: frozen_func(x)[:4] if tf_nms else frozen_func(x), [spec])
+ tfm.__call__(im)
+ tf.saved_model.save(tfm,
+ f,
+ options=tf.saved_model.SaveOptions(experimental_custom_gradients=False) if check_version(
+ tf.__version__, '2.6') else tf.saved_model.SaveOptions())
+ return f, keras_model
+
+
+@try_export
+def export_pb(keras_model, file, prefix=colorstr('TensorFlow GraphDef:')):
+ # YOLOv5 TensorFlow GraphDef *.pb export https://github.com/leimao/Frozen_Graph_TensorFlow
+ import tensorflow as tf
+ from tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflow {tf.__version__}...')
+ f = file.with_suffix('.pb')
+
+ m = tf.function(lambda x: keras_model(x)) # full model
+ m = m.get_concrete_function(tf.TensorSpec(keras_model.inputs[0].shape, keras_model.inputs[0].dtype))
+ frozen_func = convert_variables_to_constants_v2(m)
+ frozen_func.graph.as_graph_def()
+ tf.io.write_graph(graph_or_graph_def=frozen_func.graph, logdir=str(f.parent), name=f.name, as_text=False)
+ return f, None
+
+
+@try_export
+def export_tflite(keras_model, im, file, int8, data, nms, agnostic_nms, prefix=colorstr('TensorFlow Lite:')):
+ # YOLOv5 TensorFlow Lite export
+ import tensorflow as tf
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflow {tf.__version__}...')
+ batch_size, ch, *imgsz = list(im.shape) # BCHW
+ f = str(file).replace('.pt', '-fp16.tflite')
+
+ converter = tf.lite.TFLiteConverter.from_keras_model(keras_model)
+ converter.target_spec.supported_ops = [tf.lite.OpsSet.TFLITE_BUILTINS]
+ converter.target_spec.supported_types = [tf.float16]
+ converter.optimizations = [tf.lite.Optimize.DEFAULT]
+ if int8:
+ from models.tf import representative_dataset_gen
+ dataset = LoadImages(check_dataset(check_yaml(data))['train'], img_size=imgsz, auto=False)
+ converter.representative_dataset = lambda: representative_dataset_gen(dataset, ncalib=100)
+ converter.target_spec.supported_ops = [tf.lite.OpsSet.TFLITE_BUILTINS_INT8]
+ converter.target_spec.supported_types = []
+ converter.inference_input_type = tf.uint8 # or tf.int8
+ converter.inference_output_type = tf.uint8 # or tf.int8
+ converter.experimental_new_quantizer = True
+ f = str(file).replace('.pt', '-int8.tflite')
+ if nms or agnostic_nms:
+ converter.target_spec.supported_ops.append(tf.lite.OpsSet.SELECT_TF_OPS)
+
+ tflite_model = converter.convert()
+ open(f, 'wb').write(tflite_model)
+ return f, None
+
+
+@try_export
+def export_edgetpu(file, prefix=colorstr('Edge TPU:')):
+ # YOLOv5 Edge TPU export https://coral.ai/docs/edgetpu/models-intro/
+ cmd = 'edgetpu_compiler --version'
+ help_url = 'https://coral.ai/docs/edgetpu/compiler/'
+ assert platform.system() == 'Linux', f'export only supported on Linux. See {help_url}'
+ if subprocess.run(f'{cmd} > /dev/null 2>&1', shell=True).returncode != 0:
+ LOGGER.info(f'\n{prefix} export requires Edge TPU compiler. Attempting install from {help_url}')
+ sudo = subprocess.run('sudo --version >/dev/null', shell=True).returncode == 0 # sudo installed on system
+ for c in (
+ 'curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add -',
+ 'echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list',
+ 'sudo apt-get update', 'sudo apt-get install edgetpu-compiler'):
+ subprocess.run(c if sudo else c.replace('sudo ', ''), shell=True, check=True)
+ ver = subprocess.run(cmd, shell=True, capture_output=True, check=True).stdout.decode().split()[-1]
+
+ LOGGER.info(f'\n{prefix} starting export with Edge TPU compiler {ver}...')
+ f = str(file).replace('.pt', '-int8_edgetpu.tflite') # Edge TPU model
+ f_tfl = str(file).replace('.pt', '-int8.tflite') # TFLite model
+
+ subprocess.run([
+ 'edgetpu_compiler',
+ '-s',
+ '-d',
+ '-k',
+ '10',
+ '--out_dir',
+ str(file.parent),
+ f_tfl, ], check=True)
+ return f, None
+
+
+@try_export
+def export_tfjs(file, int8, prefix=colorstr('TensorFlow.js:')):
+ # YOLOv5 TensorFlow.js export
+ check_requirements('tensorflowjs')
+ import tensorflowjs as tfjs
+
+ LOGGER.info(f'\n{prefix} starting export with tensorflowjs {tfjs.__version__}...')
+ f = str(file).replace('.pt', '_web_model') # js dir
+ f_pb = file.with_suffix('.pb') # *.pb path
+ f_json = f'{f}/model.json' # *.json path
+
+ args = [
+ 'tensorflowjs_converter',
+ '--input_format=tf_frozen_model',
+ '--quantize_uint8' if int8 else '',
+ '--output_node_names=Identity,Identity_1,Identity_2,Identity_3',
+ str(f_pb),
+ str(f), ]
+ subprocess.run([arg for arg in args if arg], check=True)
+
+ json = Path(f_json).read_text()
+ with open(f_json, 'w') as j: # sort JSON Identity_* in ascending order
+ subst = re.sub(
+ r'{"outputs": {"Identity.?.?": {"name": "Identity.?.?"}, '
+ r'"Identity.?.?": {"name": "Identity.?.?"}, '
+ r'"Identity.?.?": {"name": "Identity.?.?"}, '
+ r'"Identity.?.?": {"name": "Identity.?.?"}}}', r'{"outputs": {"Identity": {"name": "Identity"}, '
+ r'"Identity_1": {"name": "Identity_1"}, '
+ r'"Identity_2": {"name": "Identity_2"}, '
+ r'"Identity_3": {"name": "Identity_3"}}}', json)
+ j.write(subst)
+ return f, None
+
+
+def add_tflite_metadata(file, metadata, num_outputs):
+ # Add metadata to *.tflite models per https://www.tensorflow.org/lite/models/convert/metadata
+ with contextlib.suppress(ImportError):
+ # check_requirements('tflite_support')
+ from tflite_support import flatbuffers
+ from tflite_support import metadata as _metadata
+ from tflite_support import metadata_schema_py_generated as _metadata_fb
+
+ tmp_file = Path('/tmp/meta.txt')
+ with open(tmp_file, 'w') as meta_f:
+ meta_f.write(str(metadata))
+
+ model_meta = _metadata_fb.ModelMetadataT()
+ label_file = _metadata_fb.AssociatedFileT()
+ label_file.name = tmp_file.name
+ model_meta.associatedFiles = [label_file]
+
+ subgraph = _metadata_fb.SubGraphMetadataT()
+ subgraph.inputTensorMetadata = [_metadata_fb.TensorMetadataT()]
+ subgraph.outputTensorMetadata = [_metadata_fb.TensorMetadataT()] * num_outputs
+ model_meta.subgraphMetadata = [subgraph]
+
+ b = flatbuffers.Builder(0)
+ b.Finish(model_meta.Pack(b), _metadata.MetadataPopulator.METADATA_FILE_IDENTIFIER)
+ metadata_buf = b.Output()
+
+ populator = _metadata.MetadataPopulator.with_model_file(file)
+ populator.load_metadata_buffer(metadata_buf)
+ populator.load_associated_files([str(tmp_file)])
+ populator.populate()
+ tmp_file.unlink()
+
+
+def pipeline_coreml(model, im, file, names, y, prefix=colorstr('CoreML Pipeline:')):
+ # YOLOv5 CoreML pipeline
+ import coremltools as ct
+ from PIL import Image
+
+ print(f'{prefix} starting pipeline with coremltools {ct.__version__}...')
+ batch_size, ch, h, w = list(im.shape) # BCHW
+ t = time.time()
+
+ # YOLOv5 Output shapes
+ spec = model.get_spec()
+ out0, out1 = iter(spec.description.output)
+ if platform.system() == 'Darwin':
+ img = Image.new('RGB', (w, h)) # img(192 width, 320 height)
+ # img = torch.zeros((*opt.img_size, 3)).numpy() # img size(320,192,3) iDetection
+ out = model.predict({'image': img})
+ out0_shape, out1_shape = out[out0.name].shape, out[out1.name].shape
+ else: # linux and windows can not run model.predict(), get sizes from pytorch output y
+ s = tuple(y[0].shape)
+ out0_shape, out1_shape = (s[1], s[2] - 5), (s[1], 4) # (3780, 80), (3780, 4)
+
+ # Checks
+ nx, ny = spec.description.input[0].type.imageType.width, spec.description.input[0].type.imageType.height
+ na, nc = out0_shape
+ # na, nc = out0.type.multiArrayType.shape # number anchors, classes
+ assert len(names) == nc, f'{len(names)} names found for nc={nc}' # check
+
+ # Define output shapes (missing)
+ out0.type.multiArrayType.shape[:] = out0_shape # (3780, 80)
+ out1.type.multiArrayType.shape[:] = out1_shape # (3780, 4)
+ # spec.neuralNetwork.preprocessing[0].featureName = '0'
+
+ # Flexible input shapes
+ # from coremltools.models.neural_network import flexible_shape_utils
+ # s = [] # shapes
+ # s.append(flexible_shape_utils.NeuralNetworkImageSize(320, 192))
+ # s.append(flexible_shape_utils.NeuralNetworkImageSize(640, 384)) # (height, width)
+ # flexible_shape_utils.add_enumerated_image_sizes(spec, feature_name='image', sizes=s)
+ # r = flexible_shape_utils.NeuralNetworkImageSizeRange() # shape ranges
+ # r.add_height_range((192, 640))
+ # r.add_width_range((192, 640))
+ # flexible_shape_utils.update_image_size_range(spec, feature_name='image', size_range=r)
+
+ # Print
+ print(spec.description)
+
+ # Model from spec
+ model = ct.models.MLModel(spec)
+
+ # 3. Create NMS protobuf
+ nms_spec = ct.proto.Model_pb2.Model()
+ nms_spec.specificationVersion = 5
+ for i in range(2):
+ decoder_output = model._spec.description.output[i].SerializeToString()
+ nms_spec.description.input.add()
+ nms_spec.description.input[i].ParseFromString(decoder_output)
+ nms_spec.description.output.add()
+ nms_spec.description.output[i].ParseFromString(decoder_output)
+
+ nms_spec.description.output[0].name = 'confidence'
+ nms_spec.description.output[1].name = 'coordinates'
+
+ output_sizes = [nc, 4]
+ for i in range(2):
+ ma_type = nms_spec.description.output[i].type.multiArrayType
+ ma_type.shapeRange.sizeRanges.add()
+ ma_type.shapeRange.sizeRanges[0].lowerBound = 0
+ ma_type.shapeRange.sizeRanges[0].upperBound = -1
+ ma_type.shapeRange.sizeRanges.add()
+ ma_type.shapeRange.sizeRanges[1].lowerBound = output_sizes[i]
+ ma_type.shapeRange.sizeRanges[1].upperBound = output_sizes[i]
+ del ma_type.shape[:]
+
+ nms = nms_spec.nonMaximumSuppression
+ nms.confidenceInputFeatureName = out0.name # 1x507x80
+ nms.coordinatesInputFeatureName = out1.name # 1x507x4
+ nms.confidenceOutputFeatureName = 'confidence'
+ nms.coordinatesOutputFeatureName = 'coordinates'
+ nms.iouThresholdInputFeatureName = 'iouThreshold'
+ nms.confidenceThresholdInputFeatureName = 'confidenceThreshold'
+ nms.iouThreshold = 0.45
+ nms.confidenceThreshold = 0.25
+ nms.pickTop.perClass = True
+ nms.stringClassLabels.vector.extend(names.values())
+ nms_model = ct.models.MLModel(nms_spec)
+
+ # 4. Pipeline models together
+ pipeline = ct.models.pipeline.Pipeline(input_features=[('image', ct.models.datatypes.Array(3, ny, nx)),
+ ('iouThreshold', ct.models.datatypes.Double()),
+ ('confidenceThreshold', ct.models.datatypes.Double())],
+ output_features=['confidence', 'coordinates'])
+ pipeline.add_model(model)
+ pipeline.add_model(nms_model)
+
+ # Correct datatypes
+ pipeline.spec.description.input[0].ParseFromString(model._spec.description.input[0].SerializeToString())
+ pipeline.spec.description.output[0].ParseFromString(nms_model._spec.description.output[0].SerializeToString())
+ pipeline.spec.description.output[1].ParseFromString(nms_model._spec.description.output[1].SerializeToString())
+
+ # Update metadata
+ pipeline.spec.specificationVersion = 5
+ pipeline.spec.description.metadata.versionString = 'https://github.com/ultralytics/yolov5'
+ pipeline.spec.description.metadata.shortDescription = 'https://github.com/ultralytics/yolov5'
+ pipeline.spec.description.metadata.author = 'glenn.jocher@ultralytics.com'
+ pipeline.spec.description.metadata.license = 'https://github.com/ultralytics/yolov5/blob/master/LICENSE'
+ pipeline.spec.description.metadata.userDefined.update({
+ 'classes': ','.join(names.values()),
+ 'iou_threshold': str(nms.iouThreshold),
+ 'confidence_threshold': str(nms.confidenceThreshold)})
+
+ # Save the model
+ f = file.with_suffix('.mlmodel') # filename
+ model = ct.models.MLModel(pipeline.spec)
+ model.input_description['image'] = 'Input image'
+ model.input_description['iouThreshold'] = f'(optional) IOU Threshold override (default: {nms.iouThreshold})'
+ model.input_description['confidenceThreshold'] = \
+ f'(optional) Confidence Threshold override (default: {nms.confidenceThreshold})'
+ model.output_description['confidence'] = 'Boxes × Class confidence (see user-defined metadata "classes")'
+ model.output_description['coordinates'] = 'Boxes × [x, y, width, height] (relative to image size)'
+ model.save(f) # pipelined
+ print(f'{prefix} pipeline success ({time.time() - t:.2f}s), saved as {f} ({file_size(f):.1f} MB)')
+
+
+@smart_inference_mode()
+def run(
+ data=ROOT / 'data/coco128.yaml', # 'dataset.yaml path'
+ weights=ROOT / 'yolov5s.pt', # weights path
+ imgsz=(640, 640), # image (height, width)
+ batch_size=1, # batch size
+ device='cpu', # cuda device, i.e. 0 or 0,1,2,3 or cpu
+ include=('torchscript', 'onnx'), # include formats
+ half=False, # FP16 half-precision export
+ inplace=False, # set YOLOv5 Detect() inplace=True
+ keras=False, # use Keras
+ optimize=False, # TorchScript: optimize for mobile
+ int8=False, # CoreML/TF INT8 quantization
+ dynamic=False, # ONNX/TF/TensorRT: dynamic axes
+ simplify=False, # ONNX: simplify model
+ opset=12, # ONNX: opset version
+ verbose=False, # TensorRT: verbose log
+ workspace=4, # TensorRT: workspace size (GB)
+ nms=False, # TF: add NMS to model
+ agnostic_nms=False, # TF: add agnostic NMS to model
+ topk_per_class=100, # TF.js NMS: topk per class to keep
+ topk_all=100, # TF.js NMS: topk for all classes to keep
+ iou_thres=0.45, # TF.js NMS: IoU threshold
+ conf_thres=0.25, # TF.js NMS: confidence threshold
+):
+ t = time.time()
+ include = [x.lower() for x in include] # to lowercase
+ fmts = tuple(export_formats()['Argument'][1:]) # --include arguments
+ flags = [x in include for x in fmts]
+ assert sum(flags) == len(include), f'ERROR: Invalid --include {include}, valid --include arguments are {fmts}'
+ jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs, paddle = flags # export booleans
+ file = Path(url2file(weights) if str(weights).startswith(('http:/', 'https:/')) else weights) # PyTorch weights
+
+ # Load PyTorch model
+ device = select_device(device)
+ if half:
+ assert device.type != 'cpu' or coreml, '--half only compatible with GPU export, i.e. use --device 0'
+ assert not dynamic, '--half not compatible with --dynamic, i.e. use either --half or --dynamic but not both'
+ model = attempt_load(weights, map_location=device, inplace=True, fuse=True) # load FP32 model
+
+ # Checks
+ imgsz *= 2 if len(imgsz) == 1 else 1 # expand
+ if optimize:
+ assert device.type == 'cpu', '--optimize not compatible with cuda devices, i.e. use --device cpu'
+
+ # Input
+ gs = int(max(model.stride)) # grid size (max stride)
+ imgsz = [check_img_size(x, gs) for x in imgsz] # verify img_size are gs-multiples
+ im = torch.zeros(batch_size, 3, *imgsz).to(device) # image size(1,3,320,192) BCHW iDetection
+
+ # Update model
+ model.eval()
+ for k, m in model.named_modules():
+ if isinstance(m, Detect):
+ m.inplace = inplace
+ m.dynamic = dynamic
+ m.export = True
+
+ for _ in range(2):
+ y = model(im) # dry runs
+ if half and not coreml:
+ im, model = im.half(), model.half() # to FP16
+ shape = tuple((y[0] if isinstance(y, tuple) else y).shape) # model output shape
+ metadata = {'stride': int(max(model.stride)), 'names': model.names} # model metadata
+ LOGGER.info(f"\n{colorstr('PyTorch:')} starting from {file} with output shape {shape} ({file_size(file):.1f} MB)")
+
+ # Exports
+ f = [''] * len(fmts) # exported filenames
+ warnings.filterwarnings(action='ignore', category=torch.jit.TracerWarning) # suppress TracerWarning
+ if jit: # TorchScript
+ f[0], _ = export_torchscript(model, im, file, optimize)
+ if engine: # TensorRT required before ONNX
+ f[1], _ = export_engine(model, im, file, half, dynamic, simplify, workspace, verbose)
+ if onnx or xml: # OpenVINO requires ONNX
+ f[2], _ = export_onnx(model, im, file, opset, dynamic, simplify)
+ if xml: # OpenVINO
+ f[3], _ = export_openvino(file, metadata, half, int8, data)
+ if coreml: # CoreML
+ f[4], ct_model = export_coreml(model, im, file, int8, half, nms)
+ if nms:
+ pipeline_coreml(ct_model, im, file, model.names, y)
+ if any((saved_model, pb, tflite, edgetpu, tfjs)): # TensorFlow formats
+ assert not tflite or not tfjs, 'TFLite and TF.js models must be exported separately, please pass only one type.'
+ assert not isinstance(model, ClassificationModel), 'ClassificationModel export to TF formats not yet supported.'
+ f[5], s_model = export_saved_model(model.cpu(),
+ im,
+ file,
+ dynamic,
+ tf_nms=nms or agnostic_nms or tfjs,
+ agnostic_nms=agnostic_nms or tfjs,
+ topk_per_class=topk_per_class,
+ topk_all=topk_all,
+ iou_thres=iou_thres,
+ conf_thres=conf_thres,
+ keras=keras)
+ if pb or tfjs: # pb prerequisite to tfjs
+ f[6], _ = export_pb(s_model, file)
+ if tflite or edgetpu:
+ f[7], _ = export_tflite(s_model, im, file, int8 or edgetpu, data=data, nms=nms, agnostic_nms=agnostic_nms)
+ if edgetpu:
+ f[8], _ = export_edgetpu(file)
+ add_tflite_metadata(f[8] or f[7], metadata, num_outputs=len(s_model.outputs))
+ if tfjs:
+ f[9], _ = export_tfjs(file, int8)
+ if paddle: # PaddlePaddle
+ f[10], _ = export_paddle(model, im, file, metadata)
+
+ # Finish
+ f = [str(x) for x in f if x] # filter out '' and None
+ if any(f):
+ cls, det, seg = (isinstance(model, x) for x in (ClassificationModel, DetectionModel, SegmentationModel)) # type
+ det &= not seg # segmentation models inherit from SegmentationModel(DetectionModel)
+ dir = Path('segment' if seg else 'classify' if cls else '')
+ h = '--half' if half else '' # --half FP16 inference arg
+ s = '# WARNING ⚠️ ClassificationModel not yet supported for PyTorch Hub AutoShape inference' if cls else \
+ '# WARNING ⚠️ SegmentationModel not yet supported for PyTorch Hub AutoShape inference' if seg else ''
+ LOGGER.info(f'\nExport complete ({time.time() - t:.1f}s)'
+ f"\nResults saved to {colorstr('bold', file.parent.resolve())}"
+ f"\nDetect: python {dir / ('detect.py' if det else 'predict.py')} --weights {f[-1]} {h}"
+ f"\nValidate: python {dir / 'val.py'} --weights {f[-1]} {h}"
+ f"\nPyTorch Hub: model = torch.hub.load('ultralytics/yolov5', 'custom', '{f[-1]}') {s}"
+ f'\nVisualize: https://netron.app')
+ return f # return list of exported files/dirs
+
+
+def parse_opt(known=False):
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='dataset.yaml path')
+ parser.add_argument('--weights', nargs='+', type=str, default=ROOT / 'yolov5s.pt', help='model.pt path(s)')
+ parser.add_argument('--imgsz', '--img', '--img-size', nargs='+', type=int, default=[640, 640], help='image (h, w)')
+ parser.add_argument('--batch-size', type=int, default=1, help='batch size')
+ parser.add_argument('--device', default='cpu', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--half', action='store_true', help='FP16 half-precision export')
+ parser.add_argument('--inplace', action='store_true', help='set YOLOv5 Detect() inplace=True')
+ parser.add_argument('--keras', action='store_true', help='TF: use Keras')
+ parser.add_argument('--optimize', action='store_true', help='TorchScript: optimize for mobile')
+ parser.add_argument('--int8', action='store_true', help='CoreML/TF/OpenVINO INT8 quantization')
+ parser.add_argument('--dynamic', action='store_true', help='ONNX/TF/TensorRT: dynamic axes')
+ parser.add_argument('--simplify', action='store_true', help='ONNX: simplify model')
+ parser.add_argument('--opset', type=int, default=17, help='ONNX: opset version')
+ parser.add_argument('--verbose', action='store_true', help='TensorRT: verbose log')
+ parser.add_argument('--workspace', type=int, default=4, help='TensorRT: workspace size (GB)')
+ parser.add_argument('--nms', action='store_true', help='TF: add NMS to model')
+ parser.add_argument('--agnostic-nms', action='store_true', help='TF: add agnostic NMS to model')
+ parser.add_argument('--topk-per-class', type=int, default=100, help='TF.js NMS: topk per class to keep')
+ parser.add_argument('--topk-all', type=int, default=100, help='TF.js NMS: topk for all classes to keep')
+ parser.add_argument('--iou-thres', type=float, default=0.45, help='TF.js NMS: IoU threshold')
+ parser.add_argument('--conf-thres', type=float, default=0.25, help='TF.js NMS: confidence threshold')
+ parser.add_argument(
+ '--include',
+ nargs='+',
+ default=['torchscript'],
+ help='torchscript, onnx, openvino, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs, paddle')
+ opt = parser.parse_known_args()[0] if known else parser.parse_args()
+ print_args(vars(opt))
+ return opt
+
+
+def main(opt):
+ for opt.weights in (opt.weights if isinstance(opt.weights, list) else [opt.weights]):
+ run(**vars(opt))
+
+
+if __name__ == '__main__':
+ opt = parse_opt()
+ main(opt)
\ No newline at end of file
diff --git a/vouchervision/component_detector/hubconf.py b/vouchervision/component_detector/hubconf.py
new file mode 100644
index 0000000000000000000000000000000000000000..4e05149026b33192fbd745e1406226827be8c38a
--- /dev/null
+++ b/vouchervision/component_detector/hubconf.py
@@ -0,0 +1,146 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+PyTorch Hub models https://pytorch.org/hub/ultralytics_yolov5/
+
+Usage:
+ import torch
+ model = torch.hub.load('ultralytics/yolov5', 'yolov5s')
+ model = torch.hub.load('ultralytics/yolov5:master', 'custom', 'path/to/yolov5s.onnx') # file from branch
+"""
+
+import torch
+
+
+def _create(name, pretrained=True, channels=3, classes=80, autoshape=True, verbose=True, device=None):
+ """Creates or loads a YOLOv5 model
+
+ Arguments:
+ name (str): model name 'yolov5s' or path 'path/to/best.pt'
+ pretrained (bool): load pretrained weights into the model
+ channels (int): number of input channels
+ classes (int): number of model classes
+ autoshape (bool): apply YOLOv5 .autoshape() wrapper to model
+ verbose (bool): print all information to screen
+ device (str, torch.device, None): device to use for model parameters
+
+ Returns:
+ YOLOv5 model
+ """
+ from pathlib import Path
+
+ from models.common import AutoShape, DetectMultiBackend
+ from models.yolo import Model
+ from utils.downloads import attempt_download
+ from utils.general import LOGGER, check_requirements, intersect_dicts, logging
+ from utils.torch_utils import select_device
+
+ if not verbose:
+ LOGGER.setLevel(logging.WARNING)
+
+ check_requirements(exclude=('tensorboard', 'thop', 'opencv-python'))
+ name = Path(name)
+ path = name.with_suffix('.pt') if name.suffix == '' else name # checkpoint path
+ try:
+ device = select_device(('0' if torch.cuda.is_available() else 'cpu') if device is None else device)
+
+ if pretrained and channels == 3 and classes == 80:
+ model = DetectMultiBackend(path, device=device) # download/load FP32 model
+ # model = models.experimental.attempt_load(path, map_location=device) # download/load FP32 model
+ else:
+ cfg = list((Path(__file__).parent / 'models').rglob(f'{path.stem}.yaml'))[0] # model.yaml path
+ model = Model(cfg, channels, classes) # create model
+ if pretrained:
+ ckpt = torch.load(attempt_download(path), map_location=device) # load
+ csd = ckpt['model'].float().state_dict() # checkpoint state_dict as FP32
+ csd = intersect_dicts(csd, model.state_dict(), exclude=['anchors']) # intersect
+ model.load_state_dict(csd, strict=False) # load
+ if len(ckpt['model'].names) == classes:
+ model.names = ckpt['model'].names # set class names attribute
+ if autoshape:
+ model = AutoShape(model) # for file/URI/PIL/cv2/np inputs and NMS
+ return model.to(device)
+
+ except Exception as e:
+ help_url = 'https://github.com/ultralytics/yolov5/issues/36'
+ s = f'{e}. Cache may be out of date, try `force_reload=True` or see {help_url} for help.'
+ raise Exception(s) from e
+
+
+def custom(path='path/to/model.pt', autoshape=True, _verbose=True, device=None):
+ # YOLOv5 custom or local model
+ return _create(path, autoshape=autoshape, verbose=_verbose, device=device)
+
+
+def yolov5n(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-nano model https://github.com/ultralytics/yolov5
+ return _create('yolov5n', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5s(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-small model https://github.com/ultralytics/yolov5
+ return _create('yolov5s', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5m(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-medium model https://github.com/ultralytics/yolov5
+ return _create('yolov5m', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5l(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-large model https://github.com/ultralytics/yolov5
+ return _create('yolov5l', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5x(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-xlarge model https://github.com/ultralytics/yolov5
+ return _create('yolov5x', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5n6(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-nano-P6 model https://github.com/ultralytics/yolov5
+ return _create('yolov5n6', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5s6(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-small-P6 model https://github.com/ultralytics/yolov5
+ return _create('yolov5s6', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5m6(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-medium-P6 model https://github.com/ultralytics/yolov5
+ return _create('yolov5m6', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5l6(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-large-P6 model https://github.com/ultralytics/yolov5
+ return _create('yolov5l6', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+def yolov5x6(pretrained=True, channels=3, classes=80, autoshape=True, _verbose=True, device=None):
+ # YOLOv5-xlarge-P6 model https://github.com/ultralytics/yolov5
+ return _create('yolov5x6', pretrained, channels, classes, autoshape, _verbose, device)
+
+
+if __name__ == '__main__':
+ model = _create(name='yolov5s', pretrained=True, channels=3, classes=80, autoshape=True, verbose=True)
+ # model = custom(path='path/to/model.pt') # custom
+
+ # Verify inference
+ from pathlib import Path
+
+ import numpy as np
+ from PIL import Image
+
+ from utils.general import cv2
+
+ imgs = [
+ 'data/images/zidane.jpg', # filename
+ Path('data/images/zidane.jpg'), # Path
+ 'https://ultralytics.com/images/zidane.jpg', # URI
+ cv2.imread('data/images/bus.jpg')[:, :, ::-1], # OpenCV
+ Image.open('data/images/bus.jpg'), # PIL
+ np.zeros((320, 640, 3))] # numpy
+
+ results = model(imgs, size=320) # batched inference
+ results.print()
+ results.save()
diff --git a/vouchervision/component_detector/landmark_processing.py b/vouchervision/component_detector/landmark_processing.py
new file mode 100644
index 0000000000000000000000000000000000000000..6338d2bc4aeae2a8c6c77601b5756640ebf6350a
--- /dev/null
+++ b/vouchervision/component_detector/landmark_processing.py
@@ -0,0 +1,1956 @@
+import os, math, cv2, random
+import numpy as np
+from itertools import combinations
+from PIL import Image
+from dataclasses import dataclass, field
+from typing import List, Dict
+
+@dataclass()
+class LeafSkeleton:
+ cfg: str
+ Dirs: str
+ leaf_type: str
+ all_points: list
+ dir_temp: str
+ file_name: str
+ width: int
+ height: int
+ logger: object
+
+ do_show_QC_images: bool = False
+ do_save_QC_images: bool = False
+
+ classes: float = None
+ points_list: float = None
+
+ image: float = None
+
+ ordered_midvein: float = None
+ midvein_fit: float = None
+ midvein_fit_points: float = None
+ ordered_midvein_length: float = None
+ has_midvein = False
+
+ is_split = False
+
+ ordered_petiole: float = None
+ ordered_petiole_length: float = None
+ has_ordered_petiole = False
+
+ has_apex: bool = False
+ apex_left: float = None
+ apex_right: float = None
+ apex_center: float = None
+ apex_angle_type: str = 'NA'
+ apex_angle_degrees: float = None
+
+ has_base: bool = False
+ base_left: float = None
+ base_right: float = None
+ base_center: float = None
+ base_angle_type: str = 'NA'
+ base_angle_degrees: float = None
+
+ has_lamina_tip: bool = False
+ lamina_tip: float = None
+
+ has_lamina_base: bool = False
+ lamina_base: float = None
+
+ has_lamina_length: bool = False
+ lamina_fit: float = None
+ lamina_length: float = None
+
+ has_width: bool = False
+ lamina_width: float = None
+ width_left: float = None
+ width_right: float = None
+
+ has_lobes: bool = False
+ lobe_count: float = None
+ lobes: float = None
+
+ def __init__(self, cfg, logger, Dirs, leaf_type, all_points, height, width, dir_temp, file_name) -> None:
+ # Store the necessary arguments as instance attributes
+ self.cfg = cfg
+ self.Dirs = Dirs
+ self.leaf_type = leaf_type
+ self.all_points = all_points
+ self.height = height
+ self.width = width
+ self.dir_temp = dir_temp
+ self.file_name = file_name
+
+ logger.name = f'[{leaf_type} - {file_name}]'
+ self.logger = logger
+
+ self.init_lists_dicts()
+
+ # Setup
+ self.set_cfg_values()
+ self.define_landmark_classes()
+
+ self.setup_QC_image()
+ self.setup_final_image()
+
+ self.parse_all_points()
+ self.convert_YOLO_bbox_to_point()
+
+ # Start with ordering the midvein and petiole
+ self.order_midvein()
+ self.order_petiole()
+ # print(self.ordered_midvein)
+
+ # Split the image using the midvein IF has_midvein == True
+ self.split_image_by_midvein()
+
+ # Process angles IF is_split == True. Need orientation to pick the appropriate pts for angle calcs
+ self.determine_apex()
+ self.determine_base()
+
+ self.determine_lamina_tip()
+ self.determine_lamina_base()
+ self.determine_lamina_length('QC')
+
+ self.determine_width()
+
+ self.determine_lobes()
+ self.determine_petiole() # straight length of petiole vs. ordered_petiole length which is tracing the petiole
+
+ self.restrictions()
+
+ # creates self.is_complete_leaf = False and self.is_leaf_no_width = False
+ # can add less restrictive options later, but for now only very complete leaves will pass
+ self.redo_measurements()
+
+ self.create_final_image()
+ self.translate_measurements_to_full_image()
+
+ self.show_QC_image()
+ self.show_final_image()
+ # self.save_QC_image()
+ # print('hi')
+
+ def get(self, attribute, default=None):
+ return getattr(self, attribute, default)
+
+ def split_image_by_midvein(self):
+
+ if self.has_midvein:
+ n_fit = 1
+
+ # Convert the points to a numpy array
+ points_arr = np.array(self.ordered_midvein)
+
+ # Fit a line to the points
+ self.midvein_fit = np.polyfit(points_arr[:, 0], points_arr[:, 1], n_fit)
+
+ if len(self.midvein_fit) < 1:
+ self.midvein_fit = None
+ else:
+ # Plot a sample of points from along the line
+ max_dim = max(self.height, self.width)
+ if max_dim < 400:
+ num_points = 40
+ elif max_dim < 1000:
+ num_points = 80
+ else:
+ num_points = 120
+
+ # Get the endpoints of the line segment that lies within the bounds of the image
+ x1 = 0
+ y1 = int(self.midvein_fit[0] * x1 + self.midvein_fit[1])
+ x2 = self.width - 1
+ y2 = int(self.midvein_fit[0] * x2 + self.midvein_fit[1])
+ denom = self.midvein_fit[0]
+ if denom == 0:
+ denom = 0.0000000001
+ if y1 < 0:
+ y1 = 0
+ x1 = int((y1 - self.midvein_fit[1]) / denom)
+ if y2 >= self.height:
+ y2 = self.height - 1
+ x2 = int((y2 - self.midvein_fit[1]) / denom)
+
+ # Sample num_points points along the line segment within the bounds of the image
+ x_vals = np.linspace(x1, x2, num_points)
+ y_vals = self.midvein_fit[0] * x_vals + self.midvein_fit[1]
+
+ # Remove any points that are outside the bounds of the image
+ indices = np.where((y_vals >= 0) & (y_vals < self.height))[0]
+ x_vals = x_vals[indices]
+ y_vals = y_vals[indices]
+
+ # Recompute y-values using the line equation and updated x-values
+ y_vals = self.midvein_fit[0] * x_vals + self.midvein_fit[1]
+
+ self.midvein_fit_points = np.column_stack((x_vals, y_vals))
+ self.is_split = True
+
+ # Draw line of fit
+ for point in self.midvein_fit_points:
+ cv2.circle(self.image, tuple(point.astype(int)), radius=1, color=(255, 255, 255), thickness=-1)
+
+
+ '''def split_image_by_midvein(self): # cubic
+ if self.file_name == 'B_774373024_Ebenaceae_Diospyros_glutinifera__L__469-164-888-632':
+ print('hi')
+ if self.has_midvein:
+ n_fit = 3
+
+ # Convert the points to a numpy array
+ points_arr = np.array(self.ordered_midvein)
+
+ # Fit a curve to the points
+ self.midvein_fit = np.polyfit(points_arr[:, 0], points_arr[:, 1], n_fit)
+
+ # Plot a sample of points from along the curve
+ max_dim = max(self.height, self.width)
+ if max_dim < 400:
+ num_points = 40
+ elif max_dim < 1000:
+ num_points = 80
+ else:
+ num_points = 120
+
+ # Get the endpoints of the curve segment that lies within the bounds of the image
+ x1 = 0
+ y1 = int(self.midvein_fit[0] * x1**3 + self.midvein_fit[1] * x1**2 + self.midvein_fit[2] * x1 + self.midvein_fit[3])
+ x2 = self.width - 1
+ y2 = int(self.midvein_fit[0] * x2**3 + self.midvein_fit[1] * x2**2 + self.midvein_fit[2] * x2 + self.midvein_fit[3])
+
+ # Sample num_points y-values that are evenly spaced within the bounds of the image
+ y_vals = np.linspace(0, self.height - 1, num_points)
+
+ # Compute the corresponding x-values using the polynomial
+ p = np.poly1d(self.midvein_fit)
+ x_vals = np.zeros(num_points)
+ for i, y in enumerate(y_vals):
+ roots = p - y
+ real_roots = roots.r[np.isreal(roots.r)].real
+ x_val = real_roots[(real_roots >= 0) & (real_roots < self.width)]
+ if len(x_val) > 0:
+ x_vals[i] = x_val[0]
+
+ # Remove any points that are outside the bounds of the image
+ indices = np.where((y_vals > 0) & (y_vals < self.height-1))[0]
+ x_vals = x_vals[indices]
+ y_vals = y_vals[indices]
+
+ # Recompute y-values using the polynomial and updated x-values
+ y_vals = self.midvein_fit[0] * x_vals**3 + self.midvein_fit[1] * x_vals**2 + self.midvein_fit[2] * x_vals + self.midvein_fit[3]
+
+ self.midvein_fit_points = np.column_stack((x_vals, y_vals))
+ self.is_split = True'''
+
+
+
+
+
+
+
+
+ def determine_apex(self):
+ if self.is_split:
+ can_get_angle = False
+ if 'apex_angle' in self.points_list:
+ if 'lamina_tip' in self.points_list:
+ self.apex_center, self.points_list['apex_angle'] = self.get_closest_point_to_sampled_points(self.points_list['apex_angle'], self.points_list['lamina_tip'])
+ can_get_angle = True
+ elif self.midvein_fit_points.shape[0] > 0:
+ self.apex_center, self.points_list['apex_angle'] = self.get_closest_point_to_sampled_points(self.points_list['apex_angle'], self.midvein_fit_points)
+ can_get_angle = True
+
+ if can_get_angle:
+ left = []
+ right = []
+ for point in self.points_list['apex_angle']:
+ loc = self.point_position_relative_to_line(point, self.midvein_fit)
+ if loc == 'right':
+ right.append(point)
+ elif loc == 'left':
+ left.append(point)
+
+ if (left == []) or (right == []):
+ self.has_apex = False
+ if (left == []) and (right != []):
+ self.apex_right, right = self.get_far_point(right, self.apex_center)
+ self.apex_left = None
+ elif (right == []) and (left != []):
+ self.apex_left, left = self.get_far_point(left, self.apex_center)
+ self.apex_right = None
+ else:
+ self.apex_left = None
+ self.apex_right = None
+ else:
+ self.has_apex = True
+ self.apex_left, left = self.get_far_point(left, self.apex_center)
+ self.apex_right, right = self.get_far_point(right, self.apex_center)
+
+ # print(self.points_list['apex_angle'])
+ # print(f'apex_center: {self.apex_center} apex_left: {self.apex_left} apex_right: {self.apex_right}')
+ self.logger.debug(f"[apex_angle_list] {self.points_list['apex_angle']}")
+ self.logger.debug(f"[apex_center] {self.apex_center} [apex_left] {self.apex_left} [apex_right] {self.apex_right}")
+
+ if self.has_apex:
+ self.apex_angle_type, self.apex_angle_degrees = self.determine_reflex(self.apex_left, self.apex_right, self.apex_center)
+ # print(f'angle_type {self.apex_angle_type} angle {self.apex_angle_degrees}')
+ self.logger.debug(f"[angle_type] {self.apex_angle_type} [angle] {self.apex_angle_degrees}")
+ else:
+ self.apex_angle_type = 'NA'
+ self.apex_angle_degrees = None
+ self.logger.debug(f"[angle_type] {self.apex_angle_type} [angle] {self.apex_angle_degrees}")
+
+
+ if self.has_apex:
+ if self.apex_center is not None:
+ cv2.circle(self.image, self.apex_center, radius=3, color=(0, 255, 0), thickness=-1)
+ if self.apex_left is not None:
+ cv2.circle(self.image, self.apex_left, radius=3, color=(255, 0, 0), thickness=-1)
+ if self.apex_right is not None:
+ cv2.circle(self.image, self.apex_right, radius=3, color=(0, 0, 255), thickness=-1)
+
+ def determine_apex_redo(self):
+ self.logger.debug(f"[apex_angle_list REDO] ")
+ self.logger.debug(f"[apex_center REDO] {self.apex_center} [apex_left] {self.apex_left} [apex_right] {self.apex_right}")
+
+ if self.has_apex:
+ self.apex_angle_type, self.apex_angle_degrees = self.determine_reflex(self.apex_left, self.apex_right, self.apex_center)
+ self.logger.debug(f"[angle_type REDO] {self.apex_angle_type} [angle] {self.apex_angle_degrees}")
+ else:
+ self.apex_angle_type = 'NA'
+ self.apex_angle_degrees = None
+ self.logger.debug(f"[angle_type REDO] {self.apex_angle_type} [angle] {self.apex_angle_degrees}")
+
+
+ if self.has_apex:
+ if self.apex_center is not None:
+ cv2.circle(self.image, self.apex_center, radius=11, color=(0, 255, 0), thickness=2)
+ if self.apex_left is not None:
+ cv2.circle(self.image, self.apex_left, radius=3, color=(255, 0, 0), thickness=-1)
+ if self.apex_right is not None:
+ cv2.circle(self.image, self.apex_right, radius=3, color=(0, 0, 255), thickness=-1)
+
+ def determine_base_redo(self):
+ self.logger.debug(f"[base_angle_list REDO] ")
+ self.logger.debug(f"[base_center REDO] {self.base_center} [base_left] {self.base_left} [base_right] {self.base_right}")
+
+ if self.has_base:
+ self.base_angle_type, self.base_angle_degrees = self.determine_reflex(self.base_left, self.base_right, self.base_center)
+ self.logger.debug(f"[angle_type REDO] {self.base_angle_type} [angle] {self.base_angle_degrees}")
+ else:
+ self.base_angle_type = 'NA'
+ self.base_angle_degrees = None
+ self.logger.debug(f"[angle_type REDO] {self.base_angle_type} [angle] {self.base_angle_degrees}")
+
+
+ if self.has_base:
+ if self.base_center is not None:
+ cv2.circle(self.image, self.base_center, radius=11, color=(0, 255, 0), thickness=2)
+ if self.base_left is not None:
+ cv2.circle(self.image, self.base_left, radius=3, color=(255, 0, 0), thickness=-1)
+ if self.base_right is not None:
+ cv2.circle(self.image, self.base_right, radius=3, color=(0, 0, 255), thickness=-1)
+
+ def determine_base(self):
+ if self.is_split:
+ can_get_angle = False
+ if 'base_angle' in self.points_list:
+ if 'lamina_base' in self.points_list:
+ self.base_center, self.points_list['base_angle'] = self.get_closest_point_to_sampled_points(self.points_list['base_angle'], self.points_list['lamina_base'])
+ can_get_angle = True
+ elif self.midvein_fit_points.shape[0] > 0:
+ self.base_center, self.points_list['base_angle'] = self.get_closest_point_to_sampled_points(self.points_list['base_angle'], self.midvein_fit_points)
+ can_get_angle = True
+
+ if can_get_angle:
+ left = []
+ right = []
+ for point in self.points_list['base_angle']:
+ loc = self.point_position_relative_to_line(point, self.midvein_fit)
+ if loc == 'right':
+ right.append(point)
+ elif loc == 'left':
+ left.append(point)
+
+ if (left == []) or (right == []):
+ self.has_base = False
+ if (left == []) and (right != []):
+ self.base_right, right = self.get_far_point(right, self.base_center)
+ self.base_left = None
+ elif (right == []) and (left != []):
+ self.base_left, left = self.get_far_point(left, self.base_center)
+ self.base_right = None
+ else:
+ self.base_left = None
+ self.base_right = None
+ else:
+ self.has_base = True
+ self.base_left, left = self.get_far_point(left, self.base_center)
+ self.base_right, right = self.get_far_point(right, self.base_center)
+
+ # print(self.points_list['base_angle'])
+ # print(f'base_center: {self.base_center} base_left: {self.base_left} base_right: {self.base_right}')
+ self.logger.debug(f"[base_angle_list] {self.points_list['base_angle']}")
+ self.logger.debug(f"[base_center] {self.base_center} [base_left] {self.base_left} [base_right] {self.base_right}")
+
+
+ if self.has_base:
+ self.base_angle_type, self.base_angle_degrees = self.determine_reflex(self.base_left, self.base_right, self.base_center)
+ # print(f'angle_type {self.base_angle_type} angle {self.base_angle_degrees}')
+ self.logger.debug(f"[angle_type] {self.base_angle_type} [angle] {self.base_angle_degrees}")
+ else:
+ self.base_angle_type = 'NA'
+ self.base_angle_degrees = None
+ self.logger.debug(f"[angle_type] {self.base_angle_type} [angle] {self.base_angle_degrees}")
+
+ if self.has_base:
+ if self.base_center:
+ cv2.circle(self.image, self.base_center, radius=3, color=(0, 255, 0), thickness=-1)
+ if self.base_left:
+ cv2.circle(self.image, self.base_left, radius=3, color=(255, 0, 0), thickness=-1)
+ if self.base_right:
+ cv2.circle(self.image, self.base_right, radius=3, color=(0, 0, 255), thickness=-1)
+
+ def determine_lamina_tip(self):
+ if 'lamina_tip' in self.points_list:
+ self.has_lamina_tip = True
+ if self.apex_center:
+ self.lamina_tip, self.lamina_tip_alternate = self.get_closest_point_to_sampled_points(self.points_list['lamina_tip'], self.apex_center)
+ elif len(self.midvein_fit_points) > 0:
+ self.lamina_tip, self.lamina_tip_alternate = self.get_closest_point_to_sampled_points(self.points_list['lamina_tip'], self.midvein_fit_points)
+ else:
+ if len(self.points_list['lamina_tip']) == 1:
+ self.lamina_tip = self.points_list['lamina_tip'][0]
+ self.lamina_tip_alternate = None
+ else: # blindly choose the most "central points"
+ centroid = tuple(np.mean(self.points_list['lamina_tip'], axis=0))
+ self.lamina_tip = min(self.points_list['lamina_tip'], key=lambda p: np.linalg.norm(np.array(p) - np.array(centroid)))
+ self.lamina_tip_alternate = None # TODO finish this
+
+ # if lamina_tip is closer to midvein_fit_points, then apex_center = lamina_tip
+ if self.apex_center and (len(self.midvein_fit_points) > 0):
+ d_apex = self.calc_min_distance(self.apex_center, self.midvein_fit_points)
+ d_lamina = self.calc_min_distance(self.lamina_tip, self.midvein_fit_points)
+ if d_lamina < d_apex:
+ cv2.circle(self.image, self.apex_center, radius=5, color=(255, 255, 255), thickness=3) # white hollow, indicates switch
+ cv2.circle(self.image, self.lamina_tip, radius=3, color=(0, 255, 0), thickness=-1) # repaint the point, indicates switch
+ self.apex_center = self.lamina_tip
+ if self.has_apex:
+ self.apex_angle_type, self.apex_angle_degrees = self.determine_reflex(self.apex_left, self.apex_right, self.apex_center)
+ else:
+ if self.apex_center:
+ self.has_lamina_tip = True
+ self.lamina_tip = self.apex_center
+ self.lamina_tip_alternate = None
+
+ if self.lamina_tip:
+ cv2.circle(self.image, self.lamina_tip, radius=5, color=(255, 0, 230), thickness=2) # pink solid
+ if self.lamina_tip_alternate:
+ for pt in self.lamina_tip_alternate:
+ cv2.circle(self.image, pt, radius=3, color=(255, 0, 230), thickness=-1) # pink hollow
+
+ def determine_lamina_base(self):
+ if 'lamina_base' in self.points_list:
+ self.has_lamina_base = True
+ if self.base_center:
+ self.lamina_base, self.lamina_base_alternate = self.get_closest_point_to_sampled_points(self.points_list['lamina_base'], self.base_center)
+ elif len(self.midvein_fit_points) > 0:
+ self.lamina_base, self.lamina_base_alternate = self.get_closest_point_to_sampled_points(self.points_list['lamina_base'], self.midvein_fit_points)
+ else:
+ if len(self.points_list['lamina_base']) == 1:
+ self.lamina_base = self.points_list['lamina_base'][0]
+ self.lamina_base_alternate = None
+ else: # blindly choose the most "central points"
+ centroid = tuple(np.mean(self.points_list['lamina_base'], axis=0))
+ self.lamina_base = min(self.points_list['lamina_base'], key=lambda p: np.linalg.norm(np.array(p) - np.array(centroid)))
+ self.lamina_base_alternate = None
+
+ # if has_lamina_tip is closer to midvein_fit_points, then base_center = has_lamina_tip
+ if self.base_center and (len(self.midvein_fit_points) > 0):
+ d_base = self.calc_min_distance(self.base_center, self.midvein_fit_points)
+ d_lamina = self.calc_min_distance(self.lamina_base, self.midvein_fit_points)
+ if d_lamina < d_base:
+ cv2.circle(self.image, self.base_center, radius=5, color=(255, 255, 255), thickness=3) # white hollow, indicates switch
+ cv2.circle(self.image, self.lamina_base, radius=3, color=(0, 255, 0), thickness=-1) # repaint the point, indicates switch
+ self.base_center = self.lamina_base
+ if self.has_base:
+ self.base_angle_type, self.base_angle_degrees = self.determine_reflex(self.base_left, self.base_right, self.base_center)
+ else:
+ if self.base_center:
+ self.has_lamina_base = True
+ self.lamina_base = self.base_center
+ self.lamina_base_alternate = None
+
+ if self.lamina_base:
+ cv2.circle(self.image, self.lamina_base, radius=5, color=(0, 100, 255), thickness=2) # orange
+ if self.lamina_base_alternate:
+ for pt in self.lamina_base_alternate:
+ cv2.circle(self.image, pt, radius=3, color=(0, 100, 255), thickness=-1) # orange hollow
+
+ def determine_lamina_length(self, QC_or_final):
+ if self.has_lamina_base and self.has_lamina_tip:
+ self.lamina_length = self.distance(self.lamina_base, self.lamina_tip)
+ ends = np.array([self.lamina_base, self.lamina_tip])
+ self.lamina_fit = np.polyfit(ends[:, 0], ends[:, 1], 1)
+ self.has_lamina_length = True
+ # r_base = 0
+ r_base = 16
+ # col = (0, 100, 0)
+ col = (0, 0, 0)
+ if QC_or_final == 'QC':
+ cv2.line(self.image, self.lamina_base, self.lamina_tip, col, 2 + r_base)
+ else:
+ cv2.line(self.image_final, self.lamina_base, self.lamina_tip, col, 2 + r_base)
+ else:
+ col = (0, 0, 0)
+ r_base = 16
+ if self.has_lamina_base and (not self.has_lamina_tip) and self.has_apex: # lamina base and apex center
+ self.lamina_length = self.distance(self.lamina_base, self.apex_center)
+ ends = np.array([self.lamina_base, self.apex_center])
+ self.lamina_fit = np.polyfit(ends[:, 0], ends[:, 1], 1)
+ self.has_lamina_length = True
+ if QC_or_final == 'QC':
+ cv2.line(self.image, self.lamina_base, self.apex_center, col, 2 + r_base)
+ else:
+ cv2.line(self.image, self.lamina_base, self.apex_center, col, 2 + r_base)
+ elif self.has_lamina_tip and (not self.has_lamina_base) and self.has_base: # lamina tip and base center
+ self.lamina_length = self.distance(self.lamina_tip, self.base_center)
+ ends = np.array([self.lamina_tip, self.apex_center])
+ self.lamina_fit = np.polyfit(ends[:, 0], ends[:, 1], 1)
+ self.has_lamina_length = True
+ if QC_or_final == 'QC':
+ cv2.line(self.image, self.lamina_tip, self.apex_center, col, 2 + r_base)
+ else:
+ cv2.line(self.image, self.lamina_tip, self.apex_center, col, 2 + r_base)
+ elif (not self.has_lamina_tip) and (not self.has_lamina_base) and self.has_apex and self.has_base: # apex center and base center
+ self.lamina_length = self.distance(self.apex_center, self.base_center)
+ ends = np.array([self.base_center, self.apex_center])
+ self.lamina_fit = np.polyfit(ends[:, 0], ends[:, 1], 1)
+ self.has_lamina_length = True
+ if QC_or_final == 'QC':
+ cv2.line(self.image, self.base_center, self.apex_center, col, 2 + r_base)
+ else:
+ cv2.line(self.image, self.base_center, self.apex_center, col, 2 + r_base) # 0, 175, 200
+ else:
+ self.lamina_length = None
+ self.lamina_fit = None
+ self.has_lamina_length = False
+
+ def determine_width(self):
+ if (('lamina_width' in self.points_list) and ((self.midvein_fit is not None and len(self.midvein_fit) > 0) or (self.lamina_fit is not None))):
+ left = []
+ right = []
+ if len(self.midvein_fit) > 0: # try using the midvein as a reference first
+ for point in self.points_list['lamina_width']:
+ loc = self.point_position_relative_to_line(point, self.midvein_fit)
+
+ if loc == 'right':
+ right.append(point)
+ elif loc == 'left':
+ left.append(point)
+ elif len(self.lamina_fit) > 0: # then try just the lamina tip/base
+ for point in self.points_list['lamina_width']:
+ loc = self.point_position_relative_to_line(point, self.lamina_fit)
+
+ if loc == 'right':
+ right.append(point)
+ elif loc == 'left':
+ left.append(point)
+ else:
+ self.has_width = False
+ self.width_left = None
+ self.width_right = None
+ self.lamina_width = None
+
+ if (left == []) or (right == []) or not self.has_width:
+ self.has_width = False
+ self.width_left = None
+ self.width_right = None
+ self.lamina_width = None
+ else:
+ self.has_width = True
+ if len(self.midvein_fit) > 0:
+ self.width_left, self.width_right = self.find_most_orthogonal_vectors(left, right, self.midvein_fit)
+ self.lamina_width = self.distance(self.width_left, self.width_right)
+ self.order_points_plot([self.width_left, self.width_right], 'lamina_width', 'QC')
+ else: # get shortest width if the nidvein is absent for comparison
+ self.width_left, self.width_right = self.find_min_width(left, right)
+ self.lamina_width = self.distance(self.width_left, self.width_right)
+ self.order_points_plot([self.width_left, self.width_right], 'lamina_width_alt', 'QC')
+ else:
+ self.has_width = False
+ self.width_left = None
+ self.width_right = None
+ self.lamina_width = None
+
+ def determine_lobes(self):
+ if 'lobe_tip' in self.points_list:
+ self.has_lobes = True
+ self.lobe_count = len(self.points_list['lobe_tip'])
+ self.lobes = self.points_list['lobe_tip']
+ for lobe in self.lobes:
+ cv2.circle(self.image, tuple(lobe), radius=6, color=(0, 255, 255), thickness=3)
+
+ def determine_petiole(self):
+ if 'petiole_tip' in self.points_list:
+ self.has_petiole_tip = True
+
+ if len(self.points_list['petiole_tip']) == 1:
+ self.petiole_tip = self.points_list['petiole_tip'][0]
+ self.petiole_tip_alternate = None
+ else: # blindly choose the most "central points"
+ centroid = tuple(np.mean(self.points_list['petiole_tip'], axis=0))
+ self.petiole_tip = min(self.points_list['petiole_tip'], key=lambda p: np.linalg.norm(np.array(p) - np.array(centroid)))
+ self.petiole_tip_alternate = None
+
+ # Straight length of petiole points
+ if self.has_ordered_petiole:
+ self.petiole_tip_opposite, self.petiole_tip_alternate = self.get_far_point(self.ordered_petiole, self.petiole_tip)
+ self.petiole_length = self.distance(self.petiole_tip_opposite, self.petiole_tip)
+ self.order_points_plot([self.petiole_tip_opposite, self.petiole_tip], 'petiole_tip', 'QC')
+ else:
+ self.petiole_tip_opposite = None
+ self.petiole_length = None
+
+ # Straight length of petiole tip to lamina base
+ if self.lamina_base is not None:
+ self.petiole_length_to_lamina_base = self.distance(self.lamina_base, self.petiole_tip)
+ self.petiole_tip_opposite_alternate = self.lamina_base
+ self.order_points_plot([self.petiole_tip_opposite_alternate, self.petiole_tip], 'petiole_tip_alt', 'QC')
+ elif self.base_center:
+ self.petiole_length_to_lamina_base = self.distance(self.base_center, self.petiole_tip)
+ self.petiole_tip_opposite_alternate = self.base_center
+ self.order_points_plot([self.petiole_tip_opposite_alternate, self.petiole_tip], 'petiole_tip_alt', 'QC')
+ else:
+ self.petiole_length_to_lamina_base = None
+ self.petiole_tip_opposite_alternate = None
+
+ def redo_measurements(self):
+ if self.has_width:
+ self.lamina_width = self.distance(self.width_left, self.width_right)
+
+ if self.has_ordered_petiole:
+ self.ordered_petiole_length, self.ordered_petiole = self.get_length_of_ordered_points(self.ordered_petiole, 'petiole_trace')
+
+ if self.has_midvein:
+ self.ordered_midvein_length, self.ordered_midvein = self.get_length_of_ordered_points(self.ordered_midvein, 'midvein_trace')
+
+ if self.has_apex:
+ self.apex_angle_type, self.apex_angle_degrees = self.determine_reflex(self.apex_left, self.apex_right, self.apex_center)
+
+ if self.has_base:
+ self.base_angle_type, self.base_angle_degrees = self.determine_reflex(self.base_left, self.base_right, self.base_center)
+
+ self.determine_lamina_length('final') # Calling just in case, should already be updated
+
+ def translate_measurements_to_full_image(self):
+ loc = self.file_name.split('__')[-1]
+ self.add_x = int(loc.split('-')[0])
+ self.add_y = int(loc.split('-')[1])
+
+ if self.has_base:
+ self.t_base_center = [self.base_center[0] + self.add_x, self.base_center[1] + self.add_y]
+ self.t_base_left = [self.base_left[0] + self.add_x, self.base_left[1] + self.add_y]
+ self.t_base_right = [self.base_right[0] + self.add_x, self.base_right[1] + self.add_y]
+
+ if self.has_apex:
+ self.t_apex_center = [self.apex_center[0] + self.add_x, self.apex_center[1] + self.add_y]
+ self.t_apex_left = [self.apex_left[0] + self.add_x, self.apex_left[1] + self.add_y]
+ self.t_apex_right = [self.apex_right[0] + self.add_x, self.apex_right[1] + self.add_y]
+
+ if self.has_lamina_base:
+ self.t_lamina_base = [self.lamina_base[0] + self.add_x, self.lamina_base[1] + self.add_y]
+ if self.has_lamina_tip:
+ self.t_lamina_tip = [self.lamina_tip[0] + self.add_x, self.lamina_tip[1] + self.add_y]
+
+ if self.has_lobes:
+ self.t_lobes = []
+ for point in self.lobes:
+ new_x = int(point[0]) + self.add_x
+ new_y = int(point[1]) + self.add_y
+ new_point = [new_x, new_y]
+ self.t_lobes.append(new_point)
+
+ if self.has_midvein:
+ self.t_midvein_fit_points = []
+ for point in self.midvein_fit_points:
+ new_x = int(point[0]) + self.add_x
+ new_y = int(point[1]) + self.add_y
+ new_point = [new_x, new_y]
+ self.t_midvein_fit_points.append(new_point)
+
+ self.t_midvein = []
+ for point in self.ordered_midvein:
+ new_x = int(point[0]) + self.add_x
+ new_y = int(point[1]) + self.add_y
+ new_point = [new_x, new_y]
+ self.t_midvein.append(new_point)
+
+ if self.has_ordered_petiole:
+ self.t_petiole = []
+ for point in self.ordered_petiole:
+ new_x = int(point[0]) + self.add_x
+ new_y = int(point[1]) + self.add_y
+ new_point = [new_x, new_y]
+ self.t_petiole.append(new_point)
+
+ if self.has_width:
+ self.t_width_left = [self.width_left[0] + self.add_x, self.width_left[1] + self.add_y]
+ self.t_width_right = [self.width_right[0] + self.add_x, self.width_right[1] + self.add_y]
+
+ if self.width_infer is not None:
+ self.t_width_infer = []
+ for point in self.width_infer:
+ new_x = int(point[0]) + self.add_x
+ new_y = int(point[1]) + self.add_y
+ new_point = [new_x, new_y]
+ self.t_width_infer.append(new_point)
+
+ def create_final_image(self):
+ self.is_complete_leaf = False ###########################################################################################################################################################
+ self.is_leaf_no_width = False
+ # r_base = 0
+ r_base = 16
+ if (self.has_apex and self.has_base and self.has_ordered_petiole and self.has_midvein and self.has_width):
+ self.is_complete_leaf = True
+
+ self.order_points_plot([self.width_left, self.width_right], 'lamina_width', 'final')
+ self.order_points_plot(self.ordered_midvein, 'midvein_trace', 'final')
+ self.order_points_plot(self.ordered_petiole, 'petiole_trace', 'final')
+ self.order_points_plot([self.apex_left, self.apex_center, self.apex_right], self.apex_angle_type, 'final')
+ self.order_points_plot([self.base_left, self.base_center, self.base_right], self.base_angle_type, 'final')
+
+ self.determine_lamina_length('final') # try
+
+
+ # Lamina tip and base
+ if self.has_lamina_tip:
+ cv2.circle(self.image_final, self.lamina_tip, radius=4 + r_base, color=(0, 255, 0), thickness=2)
+ cv2.circle(self.image_final, self.lamina_tip, radius=2 + r_base, color=(255, 255, 255), thickness=-1)
+ if self.has_lamina_base:
+ cv2.circle(self.image_final, self.lamina_base, radius=4 + r_base, color=(255, 0, 0), thickness=2)
+ cv2.circle(self.image_final, self.lamina_base, radius=2 + r_base, color=(255, 255, 255), thickness=-1)
+
+ # Apex angle
+ # if self.apex_center != []:
+ # cv2.circle(self.image_final, self.apex_center, radius=3, color=(0, 255, 0), thickness=-1)
+ if self.apex_left is not None:
+ cv2.circle(self.image_final, self.apex_left, radius=3 + r_base, color=(255, 0, 0), thickness=-1)
+ if self.apex_right is not None:
+ cv2.circle(self.image_final, self.apex_right, radius=3 + r_base, color=(0, 0, 255), thickness=-1)
+
+ # Base angle
+ # if self.base_center:
+ # cv2.circle(self.image_final, self.base_center, radius=3, color=(0, 255, 0), thickness=-1)
+ if self.base_left:
+ cv2.circle(self.image_final, self.base_left, radius=3 + r_base, color=(255, 0, 0), thickness=-1)
+ if self.base_right:
+ cv2.circle(self.image_final, self.base_right, radius=3 + r_base, color=(0, 0, 255), thickness=-1)
+
+ # Lobes
+ if self.has_lobes:
+ for lobe in self.lobes:
+ cv2.circle(self.image, tuple(lobe), radius=6 + r_base, color=(0, 255, 255), thickness=3)
+
+ elif self.has_apex and self.has_base and self.has_ordered_petiole and self.has_midvein and (not self.has_width):
+ self.is_leaf_no_width = True
+
+ self.order_points_plot(self.ordered_midvein, 'midvein_trace', 'final')
+ self.order_points_plot(self.ordered_petiole, 'petiole_trace', 'final')
+ self.order_points_plot([self.apex_left, self.apex_center, self.apex_right], self.apex_angle_type, 'final')
+ self.order_points_plot([self.base_left, self.base_center, self.base_right], self.base_angle_type, 'final')
+
+ self.determine_lamina_length('final')
+
+ # Lamina tip and base
+ if self.has_lamina_tip:
+ cv2.circle(self.image_final, self.lamina_tip, radius=4 + r_base, color=(0, 255, 0), thickness=2)
+ cv2.circle(self.image_final, self.lamina_tip, radius=2 + r_base, color=(255, 255, 255), thickness=-1)
+ if self.has_lamina_base:
+ cv2.circle(self.image_final, self.lamina_base, radius=4 + r_base, color=(255, 0, 0), thickness=2)
+ cv2.circle(self.image_final, self.lamina_base, radius=2 + r_base, color=(255, 255, 255), thickness=-1)
+
+ # Apex angle
+ # if self.apex_center != []:
+ # cv2.circle(self.image_final, self.apex_center, radius=3, color=(0, 255, 0), thickness=-1)
+ if self.apex_left is not None:
+ cv2.circle(self.image_final, self.apex_left, radius=3 + r_base, color=(255, 0, 0), thickness=-1)
+ if self.apex_right is not None:
+ cv2.circle(self.image_final, self.apex_right, radius=3 + r_base, color=(0, 0, 255), thickness=-1)
+
+ # Base angle
+ # if self.base_center:
+ # cv2.circle(self.image_final, self.base_center, radius=3, color=(0, 255, 0), thickness=-1)
+ if self.base_left:
+ cv2.circle(self.image_final, self.base_left, radius=3 + r_base, color=(255, 0, 0), thickness=-1)
+ if self.base_right:
+ cv2.circle(self.image_final, self.base_right, radius=3 + r_base, color=(0, 0, 255), thickness=-1)
+
+ # Draw line of fit
+ for point in self.width_infer:
+ point[0] = np.clip(point[0], 0, self.width - 1)
+ point[1] = np.clip(point[1], 0, self.height - 1)
+ cv2.circle(self.image_final, tuple(point.astype(int)), radius=4 + r_base, color=(0, 0, 255), thickness=-1)
+
+ # Lobes
+ if self.has_lobes:
+ for lobe in self.lobes:
+ cv2.circle(self.image, tuple(lobe), radius=6 + r_base, color=(0, 255, 255), thickness=3)
+
+
+
+
+
+ def restrictions(self):
+ # self.check_tips()
+ self.connect_midvein_to_tips()
+ self.connect_petiole_to_midvein()
+ self.check_crossing_width()
+
+ def check_tips(self): # TODO need to check the sides to prevent base from ending up on the tip side. just need to check which side of the oredered list to pull from
+ if max([self.height, self.width]) < 200:
+ scale_factor = 0.25
+ elif max([self.height, self.width]) < 500:
+ scale_factor = 0.5
+ else:
+ scale_factor = 1
+
+ if self.has_lamina_base:
+ second_last_dir = np.array(self.ordered_midvein[-1]) - np.array(self.lamina_base)
+
+ end_vector_mag = np.linalg.norm(second_last_dir)
+ avg_dist = np.mean([np.linalg.norm(np.array(self.ordered_midvein[i])-np.array(self.ordered_midvein[i-1])) for i in range(1, len(self.ordered_midvein))])
+
+ if (end_vector_mag > (scale_factor * 0.01 * avg_dist * len(self.ordered_midvein))):
+ self.lamina_base = self.ordered_midvein[-1]
+ cv2.circle(self.image, self.lamina_base, radius=4, color=(0, 0, 0), thickness=-1)
+ cv2.circle(self.image, self.lamina_base, radius=8, color=(0, 0, 255), thickness=2)
+ self.logger.debug(f'Check Tips - lamina base - made lamina base the last midvein point')
+ else:
+ self.logger.debug(f'Check Tips - lamina base - kept lamina base')
+
+
+ if self.has_lamina_tip:
+ second_last_dir = np.array(self.ordered_midvein[0]) - np.array(self.lamina_tip)
+
+ end_vector_mag = np.linalg.norm(second_last_dir)
+ avg_dist = np.mean([np.linalg.norm(np.array(self.ordered_midvein[i])-np.array(self.ordered_midvein[i-1])) for i in range(1, len(self.ordered_midvein))])
+
+ if (end_vector_mag > (scale_factor * 0.01 * avg_dist * len(self.ordered_midvein))):
+ self.lamina_tip = self.ordered_midvein[-1]
+ cv2.circle(self.image, self.lamina_tip, radius=4, color=(0, 0, 0), thickness=-1)
+ cv2.circle(self.image, self.lamina_tip, radius=8, color=(0, 0, 255), thickness=2)
+ self.logger.debug(f'Check Tips - lamina tip - made lamina tip the first midvein point')
+ else:
+ self.logger.debug(f'Check Tips - lamina tip - kept lamina tip')
+
+ def connect_midvein_to_tips(self):
+ self.logger.debug(f'Restrictions [Midvein Connect] - connect_midvein_to_tips()')
+ if self.has_midvein:
+ if self.has_lamina_tip:
+ original_lamina_tip = self.lamina_tip
+
+ start_or_end = self.add_tip(self.lamina_tip)
+ self.logger.debug(f'Restrictions [Midvein Connect] - Lamina tip [{self.lamina_tip}]')
+
+
+ self.ordered_midvein, move_midvein = self.check_momentum_complex(self.ordered_midvein, True, start_or_end)
+ if move_midvein: # the tip changed the momentum too much
+ self.logger.debug(f'Restrictions [Midvein Connect] - REDO APEX ANGLE - SWAP LAMINA TIP FOR FIRST MIDVEIN POINT')
+ # get midvein point cloases to tip
+ # new_endpoint_side, _ = self.get_closest_point_to_sampled_points(self.ordered_midvein, original_lamina_tip)
+ # new_endpoint, _ = self.get_closest_point_to_sampled_points([self.ordered_midvein[0], self.ordered_midvein[-1]], new_endpoint_side)
+
+ # change the apex to new endpoint
+ self.lamina_tip = self.ordered_midvein[0]
+ self.apex_center = self.ordered_midvein[0]
+ self.determine_lamina_length('QC')
+
+ self.determine_apex_redo()
+
+ # cv2.imshow('img', self.image)
+ # cv2.waitKey(0)
+ # self.order_points_plot(self.ordered_midvein, 'midvein_trace')
+ self.logger.debug(f'Restrictions [Midvein Connect] - connected lamina tip to midvein')
+ else:
+ self.logger.debug(f'Restrictions [Midvein Connect] - lacks lamina tip')
+
+
+
+ if self.has_lamina_base:
+ original_lamina_base = self.lamina_base
+
+ start_or_end = self.add_tip(self.lamina_base)
+ self.logger.debug(f'Restrictions [Midvein Connect] - Lamina base [{self.lamina_base}]')
+
+ self.ordered_midvein, move_midvein = self.check_momentum_complex(self.ordered_midvein, True, start_or_end)
+ if move_midvein: # the tip changed the momentum too much
+ self.logger.debug(f'Restrictions [Midvein Connect] - REDO BASE ANGLE - SWAP LAMINA BASE FOR LAST MIDVEIN POINT')
+
+ # get midvein point cloases to tip
+ # new_endpoint_side, _ = self.get_closest_point_to_sampled_points(self.ordered_midvein, original_lamina_base)
+ # new_endpoint, _ = self.get_closest_point_to_sampled_points([self.ordered_midvein[0], self.ordered_midvein[-1]], new_endpoint_side)
+
+ # change the apex to new endpoint
+ self.lamina_base = self.ordered_midvein[-1]
+ self.base_center = self.ordered_midvein[-1]
+ self.determine_lamina_length('QC')
+
+ self.determine_base_redo()
+
+ # self.order_points_plot(self.ordered_midvein, 'midvein_trace')
+ self.logger.debug(f'Restrictions [Midvein Connect] - connected lamina base to midvein')
+ else:
+ self.logger.debug(f'Restrictions [Midvein Connect] - lacks lamina base')
+
+
+ def connect_petiole_to_midvein(self):
+ if self.has_ordered_petiole and self.has_midvein:
+ if len(self.ordered_petiole) > 0 and len(self.ordered_midvein) > 0:
+ # Find the closest pair of points between ordered_petiole and ordered_midvein
+ min_dist = np.inf
+ closest_petiole_idx = None
+ closest_midvein_idx = None
+
+ for i, petiole_point in enumerate(self.ordered_petiole):
+ for j, midvein_point in enumerate(self.ordered_midvein):
+ # Convert petiole_point and midvein_point to NumPy arrays
+ petiole_point = np.array(petiole_point)
+ midvein_point = np.array(midvein_point)
+
+ # Calculate the distance between the two points
+ dist = np.linalg.norm(petiole_point - midvein_point)
+ if dist < min_dist:
+ min_dist = dist
+ closest_petiole_idx = i
+ closest_midvein_idx = j
+
+ # Calculate the midpoint between the closest points
+ petiole_point = self.ordered_petiole[closest_petiole_idx]
+ midvein_point = self.ordered_midvein[closest_midvein_idx]
+ midpoint = (int((petiole_point[0] + midvein_point[0]) / 2), int((petiole_point[1] + midvein_point[1]) / 2))
+
+ # Determine whether the midpoint should be added to the beginning or end of each list
+ petiole_dist_to_end = np.linalg.norm(np.array(self.ordered_petiole[closest_petiole_idx]) - np.array(self.ordered_petiole[-1]))
+ midvein_dist_to_end = np.linalg.norm(np.array(self.ordered_midvein[closest_midvein_idx]) - np.array(self.ordered_midvein[-1]))
+
+ if (petiole_dist_to_end < midvein_dist_to_end):
+ # Add the midpoint to the end of the petiole list and the beginning of the midvein list
+ self.ordered_midvein.insert(0, midpoint)
+ self.ordered_petiole.append(midpoint)
+ self.lamina_base = midpoint
+ cv2.circle(self.image, self.lamina_base, radius=4, color=(0, 255, 0), thickness=-1)
+ cv2.circle(self.image, self.lamina_base, radius=6, color=(0, 0, 0), thickness=2)
+ else:
+ # Add the midpoint to the end of the midvein list and the beginning of the petiole list
+ self.ordered_petiole.insert(0, midpoint)
+ self.ordered_midvein.append(midpoint)
+ self.lamina_base = midpoint
+ cv2.circle(self.image, self.lamina_base, radius=4, color=(0, 255, 0), thickness=-1)
+ cv2.circle(self.image, self.lamina_base, radius=6, color=(0, 0, 0), thickness=2)
+ # If the momentum changed, then move the apex/base centers to the begninning/end of the new midvein.
+ # self.ordered_midvein, move_midvein = self.check_momentum(self.ordered_midvein, True)
+ # self.ordered_petiole, move_petiole = self.check_momentum(self.ordered_petiole, True)
+
+ # if move_midvein or move_petiole:
+ # self.logger.debug(f'')
+
+ self.order_points_plot(self.ordered_midvein, 'midvein_trace', 'QC')
+ self.order_points_plot(self.ordered_petiole, 'petiole_trace', 'QC')
+
+
+
+ def check_crossing_width(self):
+ self.logger.debug(f'Restrictions [Crossing Width Line] - check_crossing_width()')
+ self.width_infer = None
+
+ if self.has_width:
+ self.logger.debug(f'Restrictions [Crossing Width Line] - has width')
+ # Given two points
+ x1, y1 = self.width_left
+ x2, y2 = self.width_right
+
+ # Calculate the slope and y-intercept
+ denom = (x2 - x1)
+ if denom == 0:
+ denom = 0.00000000001
+ m = (y2 - y1) / denom
+ b = y1 - m * x1
+ line_params = [m, b]
+
+ self.restrict_by_width_relation(line_params)
+
+ elif not self.has_width:
+ # generate approximate width line
+ self.logger.debug(f'Restrictions [Crossing Width Line] - infer width')
+ if self.has_apex and self.has_base:
+ line_params = self.infer_width_relation()
+ self.restrict_by_width_relation(line_params)
+
+ else:
+ self.has_ordered_petiole = False
+ self.has_apex = False
+ self.has_base = False
+ self.has_valid_apex_loc = False
+ self.has_valid_base_loc = False
+ self.logger.debug(f'Restrictions [Crossing Width Line] - CANNOT VALIDATE APEX, BASE, PETIOLE LOCATIONS')
+
+ else:
+ self.logger.debug(f'Restrictions [Crossing Width Line] - width fail *** ERROR ***')
+
+ def infer_width_relation(self):
+ top = [np.array((self.apex_center[0], self.apex_center[1])), np.array((self.apex_left[0], self.apex_left[1])), np.array((self.apex_right[0], self.apex_right[1]))]
+ bottom = [np.array((self.base_center[0], self.base_center[1])), np.array((self.base_left[0], self.base_left[1])), np.array((self.base_right[0], self.base_right[1]))]
+ if self.has_ordered_petiole:
+ bottom = bottom + [np.array(pt) for pt in self.ordered_petiole]
+
+ if self.has_midvein:
+ midvein = np.array(self.ordered_midvein)
+ self.logger.debug(f'Restrictions [Crossing Width Line] - infer width - using midvein points')
+ else:
+ self.logger.debug(f'Restrictions [Crossing Width Line] - infer width - estimating midvein points')
+ x_increment = (centroid2[0] - centroid1[0]) / 11
+ y_increment = (centroid2[1] - centroid1[1]) / 11
+ midvein = []
+ for i in range(1, 11):
+ x = centroid1[0] + i * x_increment
+ y = centroid1[1] + i * y_increment
+ midvein.append([x, y])
+
+ # find the centroids of each group of points
+ centroid1 = np.mean(top, axis=0)
+ centroid2 = np.mean(bottom, axis=0)
+
+ # calculate the midpoint between the centroids
+ midpoint = (centroid1 + centroid2) / 2
+
+ # calculate the vector between the centroids
+ centroid_vector = centroid2 - centroid1
+
+ # calculate the vector perpendicular to the centroid vector
+ perp_vector = np.array([-centroid_vector[1], centroid_vector[0]])
+
+ # normalize the perpendicular vector
+ perp_unit_vector = perp_vector / np.linalg.norm(perp_vector)
+
+ # define the length of the line segment
+ # line_segment_length = np.linalg.norm(centroid_vector) / 2
+
+ # calculate the maximum length of the line segment that can be drawn inside the image
+ max_line_segment_length = min(midpoint[0], midpoint[1], self.width - midpoint[0], self.height - midpoint[1])
+
+ # calculate the step size
+ step_size = max_line_segment_length / 5
+
+ # generate 10 points along the line that is perpendicular to the centroid vector and goes through the midpoint
+ points = []
+ for i in range(-5, 6):
+ point = midpoint + i * step_size * perp_unit_vector
+ points.append(point)
+
+ # find the equation of the line passing through the midpoint and with the perpendicular unit vector as the slope
+ b = midpoint[1] - perp_unit_vector[1] * midpoint[0]
+ if perp_unit_vector[0] == 0:
+ denom = 0.0000000001
+ else:
+ denom = perp_unit_vector[0]
+ m = perp_unit_vector[1] / denom
+
+ self.width_infer = points
+ # Draw line of fit
+ for point in points:
+ point[0] = np.clip(point[0], 0, self.width - 1)
+ point[1] = np.clip(point[1], 0, self.height - 1)
+ cv2.circle(self.image, tuple(point.astype(int)), radius=2, color=(0, 0, 255), thickness=-1)
+
+ return [m, b]
+
+
+ def restrict_by_width_relation(self, line_params):
+ '''
+ Are the tips on the same side
+ '''
+ if self.has_lamina_base and self.has_lamina_tip:
+ loc_tip = self.point_position_relative_to_line(self.lamina_tip, line_params)
+ loc_base = self.point_position_relative_to_line(self.lamina_base, line_params)
+
+ if loc_tip == loc_base:
+ self.has_lamina_base = False
+ self.has_lamina_tip = False
+
+ cv2.circle(self.image, self.lamina_tip, radius=5, color=(0, 0, 0), thickness=2) # pink solid
+ cv2.circle(self.image, self.lamina_base, radius=5, color=(0, 0, 0), thickness=2) # purple
+
+ self.logger.debug(f'Restrictions [Lamina Tip/Base] - fail - Lamina tip and base are on same side')
+ else:
+ self.logger.debug(f'Restrictions [Lamina Tip/Base] - pass - Lamina tip and base are on opposite side')
+
+ '''
+ are all apex and base values on their respecitive sides?
+ '''
+ self.has_valid_apex_loc = False
+ self.has_valid_base_loc = False
+ apex_side = 'NA'
+ base_side = 'NA'
+ if self.has_apex:
+ loc_left = self.point_position_relative_to_line(self.apex_left, line_params)
+ loc_right = self.point_position_relative_to_line(self.apex_right, line_params)
+ loc_center = self.point_position_relative_to_line(self.apex_center, line_params)
+ if loc_left == loc_right == loc_center: # all the same
+ apex_side = loc_center
+ self.has_valid_apex_loc = True
+ else:
+ self.has_valid_apex_loc = False
+ self.logger.debug(f'Restrictions [Angles] - has_valid_apex_loc = False, apex loc crosses width')
+ else:
+ self.logger.debug(f'Restrictions [Angles] - has_valid_apex_loc = False, no apex')
+
+ if self.has_base:
+ loc_left_b = self.point_position_relative_to_line(self.base_left, line_params)
+ loc_right_b = self.point_position_relative_to_line(self.base_right, line_params)
+ loc_center_b = self.point_position_relative_to_line(self.base_center, line_params)
+ if loc_left_b == loc_right_b == loc_center_b: # all the same
+ base_side = loc_center_b
+ self.has_valid_base_loc = True
+ else:
+ self.logger.debug(f'Restrictions [Angles] - has_valid_base_loc = False, base loc crosses width')
+ else:
+ self.logger.debug(f'Restrictions [Angles] - has_valid_base_loc = False')
+
+ if self.has_valid_apex_loc and self.has_valid_base_loc and (base_side != apex_side):
+ self.logger.debug(f'Restrictions [Angles] - pass - apex and base')
+ elif (base_side == apex_side) and (self.has_apex) and (self.has_base):
+ self.has_valid_apex_loc = False
+ self.has_valid_base_loc = False
+ ### This is most restrictive
+ self.has_apex = False
+ self.has_base = False
+
+ self.order_points_plot([self.apex_left, self.apex_center, self.apex_right], 'failed_angle', 'QC')
+ self.order_points_plot([self.base_left, self.base_center, self.base_right], 'failed_angle', 'QC')
+
+ self.logger.debug(f'Restrictions [Angles] - fail - apex and base')
+ elif (not self.has_valid_apex_loc) and (self.has_apex):
+ self.has_apex = False
+ self.order_points_plot([self.apex_left, self.apex_center, self.apex_right], 'failed_angle', 'QC')
+ self.logger.debug(f'Restrictions [Angles] - fail - apex')
+
+ elif (not self.has_valid_base_loc) and (self.has_base):
+ self.has_base = False
+ self.order_points_plot([self.base_left, self.base_center, self.base_right], 'failed_angle', 'QC')
+ self.logger.debug(f'Restrictions [Angles] - fail - base')
+ else:
+ self.logger.debug(f'Restrictions [Angles] - no change')
+
+
+ '''
+ does the petiole cross the width loc?
+ '''
+ if self.has_ordered_petiole:
+ petiole_check = []
+ for point in self.ordered_petiole:
+ check_val = self.point_position_relative_to_line(point, line_params)
+ petiole_check.append(check_val)
+ petiole_check = list(set(petiole_check))
+ self.logger.debug(f'Restrictions [Petiole] - petiole set = {petiole_check}')
+
+ if len(petiole_check) == 1:
+ self.has_ordered_petiole = True # Keep the petiole
+ petiole_check = petiole_check[0]
+ self.logger.debug(f'Restrictions [Petiole] - petiole does not cross width - pass')
+ else:
+ self.has_ordered_petiole = False # Reject the petiole, it crossed the center
+ self.logger.debug(f'Restrictions [Petiole] - petiole does cross width - fail')
+ else:
+ self.logger.debug(f'Restrictions [Petiole] - has_ordered_petiole = False')
+
+ '''
+ Is the lamina base on the same side as the petiole?
+ happens after the other checks...
+ '''
+ if self.has_lamina_base and self.has_lamina_tip and self.has_ordered_petiole:
+ # base is not on the same side as petiole, swap IF base and tip are already opposite
+ if loc_base != petiole_check:
+ if loc_base != loc_tip: # make sure that the tips are on opposite sides, if yes, swap the base and tip
+ hold_data = self.lamina_tip
+ self.lamina_tip = self.lamina_base
+ self.lamina_base = hold_data
+
+ cv2.circle(self.image, self.lamina_tip, radius=9, color=(255, 0, 230), thickness=2) # pink solid
+ cv2.circle(self.image, self.lamina_base, radius=9, color=(0, 100, 255), thickness=2) # purple
+
+ self.logger.debug(f'Restrictions [Petiole/Lamina Tip Same Side] - pass - swapped lamina tip and lamina base')
+ else:
+ self.has_lamina_base = False
+ self.has_lamina_tip = False
+ self.logger.debug(f'Restrictions [Petiole/Lamina Tip Same Side] - fail - lamina base not on same side as petiole, base and tip are on same side')
+ else: # base is on correct side
+ if loc_base == loc_tip: # base and tip are on the same side. error
+ self.has_lamina_base = False
+ self.has_lamina_tip = False
+ self.logger.debug(f'Restrictions [Petiole/Lamina Tip Same Side] - fail - base and tip are on the same side, but base and petiole are ok')
+ else:
+ self.logger.debug(f'Restrictions [Petiole/Lamina Tip Same Side] - pass - no swap')
+
+
+ def add_tip(self, tip):
+ # Calculate the distances between the first and last points in midvein and the new point
+ dist_start = math.dist(self.ordered_midvein[0], tip)
+ dist_end = math.dist(self.ordered_midvein[-1], tip)
+
+ # Append tip to the beginning of the list if it's closer to the first point, otherwise append it to the end of the list
+ if dist_start < dist_end:
+ self.ordered_midvein.insert(0, tip)
+ start_or_end = 'start'
+ self.logger.debug(f'Restrictions [Midvein Connect] - tip added to beginning of ordered_midvein')
+ else:
+ self.ordered_midvein.append(tip)
+ start_or_end = 'end'
+ self.logger.debug(f'Restrictions [Midvein Connect] - tip added to end of ordered_midvein')
+ return start_or_end
+
+ def find_min_width(self, left, right):
+ left_vectors = np.array(left)[:, np.newaxis, :]
+ right_vectors = np.array(right)[np.newaxis, :, :]
+ distances = np.linalg.norm(left_vectors - right_vectors, axis=2)
+ indices = np.unravel_index(np.argmin(distances), distances.shape)
+ return left[indices[0]], right[indices[1]]
+
+ def find_most_orthogonal_vectors(self, left, right, midvein_fit):
+ left_vectors = np.array(left)[:, np.newaxis, :] - np.array(right)[np.newaxis, :, :]
+ right_vectors = -left_vectors
+ midvein_vector = np.array(midvein_fit[-1]) - np.array(midvein_fit[0])
+ midvein_vector /= np.linalg.norm(midvein_vector)
+
+ dot_products = np.abs(np.sum(left_vectors * midvein_vector, axis=2)) + np.abs(np.sum(right_vectors * midvein_vector, axis=2))
+ indices = np.unravel_index(np.argmax(dot_products), dot_products.shape)
+ return left[indices[0]], right[indices[1]]
+
+ def determine_reflex(self, apex_left, apex_right, apex_center):
+ vector_left_to_center = np.array([apex_center[0] - apex_left[0], apex_center[1] - apex_left[1]])
+ vector_right_to_center = np.array([apex_center[0] - apex_right[0], apex_center[1] - apex_right[1]])
+
+ # Calculate the vector pointing to the average midvein trace value
+ midvein_trace_arr = np.array([(x, y) for x, y in self.ordered_midvein])
+ midvein_trace_avg = midvein_trace_arr.mean(axis=0)
+ vector_to_midvein_trace = midvein_trace_avg - np.array(apex_center)
+
+ # Determine whether the angle is reflex or not
+ if np.dot(vector_left_to_center, vector_to_midvein_trace) > 0 and np.dot(vector_right_to_center, vector_to_midvein_trace) > 0:
+ angle_type = 'reflex'
+ else:
+ angle_type = 'not_reflex'
+
+ angle = self.calculate_angle(apex_left, apex_center, apex_right)
+ if angle_type == 'reflex':
+ angle = 360 - angle
+
+ self.order_points_plot([apex_left, apex_center, apex_right], angle_type, 'QC')
+
+ return angle_type, angle
+
+ def calculate_angle(self, p1, p2, p3):
+ # Calculate the vectors between the points
+ v1 = (p1[0] - p2[0], p1[1] - p2[1])
+ v2 = (p3[0] - p2[0], p3[1] - p2[1])
+
+ # Calculate the dot product and magnitudes of the vectors
+ dot_product = v1[0]*v2[0] + v1[1]*v2[1]
+ mag_v1 = math.sqrt(v1[0]**2 + v1[1]**2)
+ mag_v2 = math.sqrt(v2[0]**2 + v2[1]**2)
+
+ if mag_v1 == 0:
+ mag_v1 = 0.000000001
+ if mag_v2 == 0:
+ mag_v2 = 0.000000001
+ # Calculate the cosine of the angle
+ denom = (mag_v1 * mag_v2)
+ if denom == 0:
+ denom = 0.000000001
+ cos_angle = dot_product / denom
+
+ # Calculate the angle in radians and degrees
+ angle_rad = math.acos(min(max(cos_angle, -1), 1))
+
+ angle_deg = math.degrees(angle_rad)
+
+ return angle_deg
+
+ def calc_min_distance(self, point, reference_points):
+ # Convert the points and reference points to numpy arrays
+ points_arr = np.atleast_2d(point)
+ reference_arr = np.array(reference_points)
+
+ # Calculate the distances between each point in "points" and each point in "reference_points"
+ dists = np.linalg.norm(points_arr[:, np.newaxis, :] - reference_arr, axis=2)
+ distance = np.min(dists, axis=1)
+ return distance
+
+
+ def get_closest_point_to_sampled_points(self, points, reference_points):
+ # Convert the points and reference points to numpy arrays
+ points_arr = np.array(points)
+ reference_arr = np.array(reference_points)
+
+ # Calculate the distances between each point in "points" and each point in "reference_points"
+ dists = np.linalg.norm(points_arr[:, np.newaxis, :] - reference_arr, axis=2)
+ distances = np.min(dists, axis=1)
+
+ # Get the index of the closest point
+ closest_idx = np.argmin(distances)
+
+ # Remove the closest point from the list of points
+ return points.pop(closest_idx), points
+
+ def get_far_point(self, points, reference_point):
+ # Calculate the distances between each point and the reference points
+ distances = [math.dist(point, reference_point) for point in points]
+
+ # Get the index of the closest point
+ closest_idx = distances.index(max(distances))
+ far_point = points.pop(closest_idx)
+
+ # Remove the closest point from the list of points
+ return far_point, points
+
+ '''def point_position_relative_to_line(self, point, line_params):
+ # Extract the cubic coefficients from the line parameters
+ a, b, c, d = line_params
+
+ # Determine the x-coordinate of the point where it intersects with the line
+ # We solve the cubic equation ax^3 + bx^2 + cx + d = y for x, given y = point[1]
+ f = lambda x: a*x**3 + b*x**2 + c*x + d - point[1]
+ roots = np.roots([a, b, c, d-point[1]])
+ real_roots = roots[np.isreal(roots)].real
+ if len(real_roots) == 0:
+ return "left" # point is below the curve
+ x_intersection = real_roots[0]
+
+ # Determine the midpoint of the line
+ mid_x = self.width / 2
+ mid_y = self.height / 2
+
+ # Determine if the point is to the left or right of the line
+ if self.height > self.width:
+ if point[0] < x_intersection:
+ return "left"
+ else:
+ return "right"
+ else:
+ if point[1] < a*mid_x**3 + b*mid_x**2 + c*mid_x + d:
+ return "left"
+ else:
+ return "right"'''
+
+ def point_position_relative_to_line(self, point, line_params):
+ # Extract the slope and y-intercept from the line parameters
+ slope, y_intercept = line_params
+
+ if (slope == 0.0) or (slope == 0):
+ slope = 0.00000000000001
+
+ # Determine the x-coordinate of the point where it intersects with the line
+ x_intersection = (point[1] - y_intercept) / slope
+
+ # Determine the midpoint of the line
+ mid_x = self.width / 2
+ mid_y = self.height / 2
+
+ # Determine if the point is to the left or right of the line
+ if self.height > self.width:
+ if point[0] < x_intersection:
+ return "left"
+ else:
+ return "right"
+ else:
+ if point[1] < slope * (point[0] - mid_x) + mid_y:
+ return "left" #below
+ else:
+ return "right" #above
+
+ def rotate_points(self, points, angle_cw):
+ # Calculate the center of the image
+ center_x = self.width / 2
+ center_y = self.height / 2
+
+ # Translate the points to the center
+ translated_points = [(point[0]-center_x, point[1]-center_y) for point in points]
+
+ # Convert the angle to radians
+ angle_cw = math.radians(angle_cw)
+
+ # Rotate the points
+ rotated_points = [(round(point[0]*math.cos(angle_cw)-point[1]*math.sin(angle_cw)), round(point[0]*math.sin(angle_cw)+point[1]*math.cos(angle_cw))) for point in translated_points]
+
+ # Translate the points back to the original origin
+ return [(point[0]+center_x, point[1]+center_y) for point in rotated_points]
+
+
+ def order_petiole(self):
+ if 'petiole_trace' in self.points_list:
+ if len(self.points_list['petiole_trace']) >= 5:
+ self.logger.debug(f"Ordered Petiole - Raw list contains {len(self.points_list['petiole_trace'])} points - using momentum")
+ self.ordered_petiole = self.order_points(self.points_list['petiole_trace'])
+ self.ordered_petiole = self.remove_duplicate_points(self.ordered_petiole)
+
+ self.ordered_petiole = self.check_momentum(self.ordered_petiole, False)
+
+ self.order_points_plot(self.ordered_petiole, 'petiole_trace', 'QC')
+ self.ordered_petiole_length, self.ordered_petiole = self.get_length_of_ordered_points(self.ordered_petiole, 'petiole_trace')
+ self.has_ordered_petiole = True
+ elif len(self.points_list['petiole_trace']) >= 2:
+ self.logger.debug(f"Ordered Petiole - Raw list contains {len(self.points_list['petiole_trace'])} points - SKIPPING momentum")
+ self.ordered_petiole = self.order_points(self.points_list['petiole_trace'])
+ self.ordered_petiole = self.remove_duplicate_points(self.ordered_petiole)
+
+ self.order_points_plot(self.ordered_petiole, 'petiole_trace', 'QC')
+ self.ordered_petiole_length, self.ordered_petiole = self.get_length_of_ordered_points(self.ordered_petiole, 'petiole_trace')
+ self.has_ordered_petiole = True
+ else:
+ self.logger.debug(f"Ordered Petiole - Raw list contains {len(self.points_list['petiole_trace'])} points - SKIPPING PETIOLE")
+
+ def order_midvein(self):
+ if 'midvein_trace' in self.points_list:
+ if len(self.points_list['midvein_trace']) >= 5:
+ self.logger.debug(f"Ordered Midvein - Raw list contains {len(self.points_list['midvein_trace'])} points - using momentum")
+ self.ordered_midvein = self.order_points(self.points_list['midvein_trace'])
+ self.ordered_midvein = self.remove_duplicate_points(self.ordered_midvein)
+
+ self.ordered_midvein = self.check_momentum(self.ordered_midvein, False)
+
+ self.order_points_plot(self.ordered_midvein, 'midvein_trace', 'QC')
+ self.ordered_midvein_length, self.ordered_midvein = self.get_length_of_ordered_points(self.ordered_midvein, 'midvein_trace')
+ self.has_midvein = True
+ else:
+ self.logger.debug(f"Ordered Midvein - Raw list contains {len(self.points_list['midvein_trace'])} points - SKIPPING MIDVEIN")
+
+
+ def check_momentum(self, coords, info):
+ original_coords = coords
+ # find middle index of coordinates
+ mid_idx = len(coords) // 2
+
+ # set up variables for running average
+ running_avg = np.array(coords[mid_idx-1])
+ avg_count = 1
+
+ # iterate over coordinates to check momentum change
+ prev_vec = np.array(coords[mid_idx-1]) - np.array(coords[mid_idx-2])
+ cur_idx = mid_idx - 1
+ while cur_idx >= 0:
+ cur_vec = np.array(coords[cur_idx]) - np.array(coords[cur_idx-1])
+
+ # add current point to running average
+ running_avg = (running_avg * avg_count + np.array(coords[cur_idx])) / (avg_count + 1)
+ avg_count += 1
+
+ # check for momentum change
+ if self.check_momentum_change(prev_vec, cur_vec):
+ break
+
+ prev_vec = cur_vec
+ cur_idx -= 1
+
+ # use running average to check for momentum change
+ cur_vec = np.array(coords[cur_idx]) - running_avg
+ if self.check_momentum_change(prev_vec, cur_vec):
+ cur_idx += 1
+
+ prev_vec = np.array(coords[mid_idx+1]) - np.array(coords[mid_idx])
+ cur_idx2 = mid_idx + 1
+ while cur_idx2 < len(coords):
+
+ # check if current index is out of range
+ if cur_idx2 >= len(coords):
+ break
+
+ cur_vec = np.array(coords[cur_idx2]) - np.array(coords[cur_idx2-1])
+
+ # add current point to running average
+ running_avg = (running_avg * avg_count + np.array(coords[cur_idx2])) / (avg_count + 1)
+ avg_count += 1
+
+ # check for momentum change
+ if self.check_momentum_change(prev_vec, cur_vec):
+ break
+
+ prev_vec = cur_vec
+ cur_idx2 += 1
+
+ # use running average to check for momentum change
+ if cur_idx2 < len(coords):
+ cur_vec = np.array(coords[cur_idx2]) - running_avg
+ if self.check_momentum_change(prev_vec, cur_vec):
+ cur_idx2 -= 1
+
+ # remove problematic points and subsequent points from list of coordinates
+ new_coords = coords[:cur_idx2] + coords[mid_idx:cur_idx2:-1]
+ if info:
+ return new_coords, len(original_coords) != len(new_coords)
+ else:
+ return new_coords
+
+ # define function to check for momentum change
+ def check_momentum_change(self, prev_vec, cur_vec):
+ dot_product = np.dot(prev_vec, cur_vec)
+ prev_norm = np.linalg.norm(prev_vec)
+ cur_norm = np.linalg.norm(cur_vec)
+ denom = (prev_norm * cur_norm)
+ if denom == 0:
+ denom = 0.0000000001
+ cos_theta = dot_product / denom
+ theta = np.arccos(cos_theta)
+ return abs(theta) > np.pi / 2
+
+ '''def check_momentum_complex(self, coords, info, start_or_end):
+ original_coords = coords
+ # find middle index of coordinates
+ mid_idx = len(coords) // 2
+
+ # get directional vectors for first-middle, middle-last, and second-first and second-last pairs of points
+ first_middle_dir = np.array(coords[1]) - np.array(coords[0])
+ middle_last_dir = np.array(coords[-1]) - np.array(coords[-2])
+ second_first_dir = np.array(coords[1]) - np.array(coords[2])
+ second_last_dir = np.array(coords[-1]) - np.array(coords[-3])
+
+ if start_or_end == 'end':
+ # check directional change for first-middle vector
+ cur_idx = 2
+ while cur_idx < len(coords):
+ cur_vec = np.array(coords[cur_idx]) - np.array(coords[cur_idx-1])
+ if self.check_momentum_change_complex(first_middle_dir, cur_vec):
+ break
+ cur_idx += 1
+
+ cur_idx2 = len(coords) - 2
+
+ elif start_or_end == 'start':
+ # check directional change for last-middle vector
+ cur_idx2 = len(coords)-3
+ while cur_idx2 >= 0:
+ cur_vec = np.array(coords[cur_idx2]) - np.array(coords[cur_idx2+1])
+ if self.check_momentum_change_complex(middle_last_dir, cur_vec):
+ break
+ cur_idx2 -= 1
+
+ cur_idx = 1
+
+ # check directional change for second-first and second-last vectors
+ second_first_change = self.check_momentum_change_complex(second_first_dir, first_middle_dir)
+ second_last_change = self.check_momentum_change_complex(second_last_dir, middle_last_dir)
+
+ # remove problematic points and subsequent points from list of coordinates
+ if cur_idx <= cur_idx2:
+ new_coords = coords[:cur_idx+1] + coords[cur_idx2:mid_idx:-1] + coords[cur_idx+1:cur_idx2+1]
+ else:
+ new_coords = coords[:mid_idx+1] + coords[cur_idx2:cur_idx:-1] + coords[mid_idx+1:cur_idx2+1]
+
+ self.logger.debug(f'Original midvein points - {self.ordered_midvein}')
+ self.logger.debug(f'Momentum midvein points - {new_coords}')
+ if info:
+ return new_coords, len(original_coords) != len(new_coords) or second_first_change or second_last_change
+ else:
+ return new_coords'''
+
+ def check_momentum_complex(self, coords, info, start_or_end): # Works, but removes ALL points after momentum change
+ original_coords = coords
+
+ if max([self.height, self.width]) < 200:
+ scale_factor = 0.25
+ elif max([self.height, self.width]) < 500:
+ scale_factor = 0.5
+ else:
+ scale_factor = 1
+ self.logger.debug(f'Scale factor - [{scale_factor}]')
+
+ # find middle index of coordinates
+ mid_idx = len(coords) // 2
+
+ # get directional vectors for first-middle, middle-last, and second-first and second-last pairs of points
+ first_middle_dir = np.array(coords[1]) - np.array(coords[mid_idx])
+ middle_last_dir = np.array(coords[mid_idx]) - np.array(coords[-2])
+ second_first_dir = np.array(coords[1]) - np.array(coords[0])
+ second_last_dir = np.array(coords[-1]) - np.array(coords[-2])
+
+ if start_or_end == 'end':
+ # check directional change for first-middle vector
+ cur_idx_list = []
+ cur_idx = 2
+ while cur_idx < len(coords):
+ cur_vec = np.array(coords[cur_idx]) - np.array(coords[cur_idx-1])
+ if self.check_momentum_change_complex(first_middle_dir, cur_vec):
+ # break
+ cur_idx_list.append(cur_idx)
+ cur_idx += 1
+ if len(cur_idx_list) > 0:
+ cur_idx = max(cur_idx_list)
+ else:
+ cur_idx = len(coords)
+ # remove problematic points and subsequent points from list of coordinates
+ end_vector_mag = np.linalg.norm(second_last_dir)
+ avg_dist = np.mean([np.linalg.norm(np.array(coords[i])-np.array(coords[i-1])) for i in range(1, len(coords))])
+ new_coords = coords
+
+ if (end_vector_mag > (scale_factor * 0.01 * avg_dist * len(new_coords))) and (len(cur_idx_list) > 0):
+ # new_coords = coords[:cur_idx+1] + coords[-2:cur_idx:-1][::-1] #coords[-2:cur_idx:-1]
+ new_coords = coords[:len(new_coords)-1]# + coords[-2:cur_idx:-1][::-1] #coords[-2:cur_idx:-1]
+ self.logger.debug(f'Momentum - removing last point')
+ else:
+ self.logger.debug(f'Momentum - change not detected, no change')
+
+
+
+ elif start_or_end == 'start':
+ # check directional change for last-middle vector
+ cur_idx2_list = []
+ cur_idx2 = len(coords)-3
+ while cur_idx2 >= 0:
+ cur_vec = np.array(coords[cur_idx2]) - np.array(coords[cur_idx2+1])
+ if self.check_momentum_change_complex(middle_last_dir, cur_vec):
+ # break
+ cur_idx2_list.append(cur_idx2)
+ cur_idx2 -= 1
+ if len(cur_idx2_list) > 0:
+ cur_idx2 = min(cur_idx2_list)
+ else:
+ cur_idx2 = 0
+ # remove problematic points and subsequent points from list of coordinates
+ new_coords = coords
+ start_vector_mag = np.linalg.norm(second_first_dir)
+ avg_dist = np.mean([np.linalg.norm(np.array(coords[i])-np.array(coords[i-1])) for i in range(1, len(coords))])
+
+ if (start_vector_mag > (scale_factor * 0.01 * avg_dist * len(new_coords))) and (len(cur_idx2_list) > 0):
+ # new_coords = coords[:mid_idx+1] + coords[cur_idx2:mid_idx:-1][::-1] # #coords[cur_idx2:mid_idx:-1]
+ new_coords = coords[1:]#ur_idx2-1] + coords[mid_idx+1:]
+ # new_coords = coords[cur_idx2:mid_idx+1][::-1] + coords[mid_idx+1:]
+ self.logger.debug(f'Momentum - removing first point')
+ else:
+ self.logger.debug(f'Momentum - change not detected, no change')
+ else:
+ print('hi')
+
+ # check directional change for second-first and second-last vectors
+ # second_first_change = self.check_momentum_change_complex(second_first_dir, first_middle_dir)
+ # second_last_change = self.check_momentum_change_complex(second_last_dir, middle_last_dir)
+
+ self.logger.debug(f'Original midvein points complex - {start_or_end} - {self.ordered_midvein}')
+ self.logger.debug(f'Momentum midvein points complex - {start_or_end} - {new_coords}')
+ if info:
+ return new_coords, len(original_coords) != len(new_coords) #or second_first_change or second_last_change
+ else:
+ return new_coords
+
+ '''def check_momentum_complex(self, coords, info, start_or_end): # does not seem to work
+ original_coords = coords
+
+ # get directional vectors for first-middle, middle-last, and second-first and second-last pairs of points
+ first_middle_dir = np.array(coords[1]) - np.array(coords[0])
+ middle_last_dir = np.array(coords[-1]) - np.array(coords[-2])
+ second_first_dir = np.array(coords[1]) - np.array(coords[2])
+ second_last_dir = np.array(coords[-1]) - np.array(coords[-3])
+
+ # calculate running average momentum and check if endpoints are different
+ if start_or_end == 'end':
+ end_vector_mag = np.linalg.norm(first_middle_dir)
+ avg_dist = np.mean([np.linalg.norm(np.array(coords[i])-np.array(coords[i-1])) for i in range(1, len(coords))])
+ running_avg_momentum = np.mean([np.linalg.norm(np.array(coords[i])-np.array(coords[i-1])) for i in range(len(coords)-10, len(coords))])
+ endpoint_diff = np.linalg.norm(np.array(coords[-1])-self.ordered_midvein[-1]) > 0.1*running_avg_momentum
+ elif start_or_end == 'start':
+ start_vector_mag = np.linalg.norm(middle_last_dir)
+ avg_dist = np.mean([np.linalg.norm(np.array(coords[i])-np.array(coords[i-1])) for i in range(1, len(coords))])
+ running_avg_momentum = np.mean([np.linalg.norm(np.array(coords[i])-np.array(coords[i-1])) for i in range(10)])
+ endpoint_diff = np.linalg.norm(np.array(coords[0])-self.ordered_midvein[0]) > 0.1*running_avg_momentum
+
+ # remove problematic points and subsequent points from list of coordinates
+ if start_or_end == 'end' and endpoint_diff:
+ cur_idx = 2
+ while cur_idx < len(coords):
+ cur_vec = np.array(coords[cur_idx]) - np.array(coords[cur_idx-1])
+ if self.check_momentum_change_complex(first_middle_dir, cur_vec):
+ break
+ cur_idx += 1
+ new_coords = coords[:cur_idx+1] + coords[-2:cur_idx:-1][::-1]
+ elif start_or_end == 'start' and endpoint_diff:
+ cur_idx2 = len(coords)-3
+ while cur_idx2 >= 0:
+ cur_vec = np.array(coords[cur_idx2]) - np.array(coords[cur_idx2+1])
+ if self.check_momentum_change_complex(middle_last_dir, cur_vec):
+ break
+ cur_idx2 -= 1
+ new_coords = coords[:1] + coords[cur_idx2:0:-1][::-1]
+ else:
+ new_coords = coords
+
+ # check directional change for second-first and second-last vectors
+ second_first_change = self.check_momentum_change_complex(second_first_dir, first_middle_dir)
+ second_last_change = self.check_momentum_change_complex(second_last_dir, middle_last_dir)
+
+ self.logger.debug(f'Original midvein points - {self.ordered_midvein}')
+ self.logger.debug(f'Momentum midvein points - {new_coords}')
+ if info:
+ return new_coords, len(original_coords) != len(new_coords) #or second_first_change or second_last_change or endpoint_diff
+ else:
+ return new_coords'''
+
+
+
+
+
+
+
+
+ # define function to check for momentum change
+ def check_momentum_change_complex(self, prev_vec, cur_vec):
+ dot_product = np.dot(prev_vec, cur_vec)
+ prev_norm = np.linalg.norm(prev_vec)
+ cur_norm = np.linalg.norm(cur_vec)
+ denom = (prev_norm * cur_norm)
+ if denom == 0:
+ denom = 0.0000000001
+ cos_theta = dot_product / denom
+ theta = np.arccos(cos_theta)
+ return abs(theta) > np.pi / 2
+
+
+
+ def remove_duplicate_points(self, points):
+ unique_set = set()
+ new_list = []
+
+ for item in points:
+ if item not in unique_set:
+ unique_set.add(item)
+ new_list.append(item)
+ return new_list
+
+ def order_points_plot(self, points, version, QC_or_final):
+ # thk_base = 0
+ thk_base = 16
+
+ if version == 'midvein_trace':
+ # color = (0, 255, 0)
+ color = (0, 255, 255) # yellow
+ thick = 2 + thk_base
+ elif version == 'petiole_trace':
+ color = (255, 255, 0)
+ thick = 2 + thk_base
+ elif version == 'lamina_width':
+ color = (0, 0, 255)
+ thick = 2 + thk_base
+ elif version == 'lamina_width_alt':
+ color = (100, 100, 255)
+ thick = 2 + thk_base
+ elif version == 'not_reflex':
+ color = (200, 0, 123)
+ thick = 3 + thk_base
+ elif version == 'reflex':
+ color = (0, 120, 200)
+ thick = 3 + thk_base
+ elif version == 'petiole_tip_alt':
+ color = (255, 55, 100)
+ thick = 1 + thk_base
+ elif version == 'petiole_tip':
+ color = (100, 255, 55)
+ thick = 1 + thk_base
+ elif version == 'failed_angle':
+ color = (0, 0, 0)
+ thick = 3 + thk_base
+ # Convert the points to a numpy array and round to integer values
+ points_arr = np.round(np.array(points)).astype(int)
+
+ # Draw a green line connecting all of the points
+ if QC_or_final == 'QC':
+ for i in range(len(points_arr) - 1):
+ cv2.line(self.image, tuple(points_arr[i]), tuple(points_arr[i+1]), color, thick)
+ else:
+ for i in range(len(points_arr) - 1):
+ cv2.line(self.image_final, tuple(points_arr[i]), tuple(points_arr[i+1]), color, thick)
+
+
+
+
+ def get_length_of_ordered_points(self, points, name):
+ # if self.file_name == 'B_774373631_Ebenaceae_Diospyros_buxifolia__L__438-687-578-774':
+ # print('hi')
+ total_length = 0
+ total_length_first_pass = 0
+ for i in range(len(points) - 1):
+ x1, y1 = points[i]
+ x2, y2 = points[i+1]
+ segment_length = math.sqrt((x2-x1)**2 + (y2-y1)**2)
+ total_length_first_pass += segment_length
+ cutoff = total_length_first_pass / 2
+ # print(f'Total length of {name}: {total_length_first_pass}')
+ # print(f'points length {len(points)}')
+ self.logger.debug(f"Total length of {name}: {total_length_first_pass}")
+ self.logger.debug(f"Points length {len(points)}")
+
+
+ # If there are more than 2 points, this will exclude extreme outliers, or
+ # misordered points that don't belong
+ if len(points) > 2:
+ pop_ind = []
+ for i in range(len(points) - 1):
+ x1, y1 = points[i]
+ x2, y2 = points[i+1]
+ segment_length = math.sqrt((x2-x1)**2 + (y2-y1)**2)
+ if segment_length < cutoff:
+ total_length += segment_length
+ else:
+ pop_ind.append(i)
+
+ for exclude in pop_ind:
+ points.pop(exclude)
+ # print(f'Total length of {name}: {total_length}')
+ # print(f'Excluded {len(pop_ind)} points')
+ # print(f'points length {len(points)}')
+ self.logger.debug(f"Total length of {name}: {total_length}")
+ self.logger.debug(f"Excluded {len(pop_ind)} points")
+ self.logger.debug(f"Points length {len(points)}")
+
+ else:
+ total_length = total_length_first_pass
+
+ return total_length, points
+
+ def convert_YOLO_bbox_to_point(self):
+ for point_type, bbox in self.points_list.items():
+ xy_points = []
+ for point in bbox:
+ x = point[0]
+ y = point[1]
+ w = point[2]
+ h = point[3]
+ x1 = int((x - w/2) * self.width)
+ y1 = int((y - h/2) * self.height)
+ x2 = int((x + w/2) * self.width)
+ y2 = int((y + h/2) * self.height)
+ xy_points.append((int((x1+x2)/2), int((y1+y2)/2)))
+ self.points_list[point_type] = xy_points
+
+ def parse_all_points(self):
+ points_list = {}
+
+ for sublist in self.all_points:
+ key = sublist[0]
+ value = sublist[1:]
+
+ key = self.swap_number_for_string(key)
+
+ if key not in points_list:
+ points_list[key] = []
+ points_list[key].append(value)
+
+ # print(points_list)
+ self.points_list = points_list
+
+ def swap_number_for_string(self, key):
+ for k, v in self.classes.items():
+ if v == key:
+ return k
+ return key
+
+ def distance(self, point1, point2):
+ x1, y1 = point1
+ x2, y2 = point2
+ return math.sqrt((x2 - x1) ** 2 + (y2 - y1) ** 2)
+
+ def order_points(self, points):
+ # height = max(points, key=lambda point: point[1])[1] - min(points, key=lambda point: point[1])[1]
+ # width = max(points, key=lambda point: point[0])[0] - min(points, key=lambda point: point[0])[0]
+
+ if self.height > self.width:
+ start_point = min(points, key=lambda point: point[1])
+ end_point = max(filter(lambda point: point[0] == max(points, key=lambda point: point[0])[0], points), key=lambda point: point[1])
+ else:
+ start_point = min(points, key=lambda point: point[0])
+ end_point = max(filter(lambda point: point[1] == max(points, key=lambda point: point[1])[1], points), key=lambda point: point[0])
+
+ tour = [start_point]
+ unvisited = set(points) - {start_point}
+
+ while unvisited:
+ nearest = min(unvisited, key=lambda point: self.distance(tour[-1], point))
+ tour.append(nearest)
+ unvisited.remove(nearest)
+
+ tour.append(end_point)
+ return tour
+
+ def define_landmark_classes(self):
+ self.classes = {
+ 'apex_angle': 0,
+ 'base_angle': 1,
+ 'lamina_base': 2,
+ 'lamina_tip': 3,
+ 'lamina_width': 4,
+ 'lobe_tip': 5,
+ 'midvein_trace': 6,
+ 'petiole_tip': 7,
+ 'petiole_trace': 8
+ }
+
+ def set_cfg_values(self):
+ self.do_show_QC_images = self.cfg['leafmachine']['landmark_detector']['do_show_QC_images']
+ self.do_save_QC_images = self.cfg['leafmachine']['landmark_detector']['do_save_QC_images']
+ self.do_show_final_images = self.cfg['leafmachine']['landmark_detector']['do_show_final_images']
+ self.do_save_final_images = self.cfg['leafmachine']['landmark_detector']['do_save_final_images']
+
+ def setup_QC_image(self):
+ self.image = cv2.imread(os.path.join(self.dir_temp, '.'.join([self.file_name, 'jpg'])))
+
+ if self.leaf_type == 'Landmarks_Whole_Leaves':
+ self.path_QC_image = os.path.join(self.Dirs.landmarks_whole_leaves_overlay_QC, '.'.join([self.file_name, 'jpg']))
+ elif self.leaf_type == 'Landmarks_Partial_Leaves':
+ self.path_QC_image = os.path.join(self.Dirs.landmarks_partial_leaves_overlay_QC, '.'.join([self.file_name, 'jpg']))
+
+ def setup_final_image(self):
+ self.image_final = cv2.imread(os.path.join(self.dir_temp, '.'.join([self.file_name, 'jpg'])))
+
+ if self.leaf_type == 'Landmarks_Whole_Leaves':
+ self.path_image_final = os.path.join(self.Dirs.landmarks_whole_leaves_overlay_final, '.'.join([self.file_name, 'jpg']))
+ elif self.leaf_type == 'Landmarks_Partial_Leaves':
+ self.path_image_final = os.path.join(self.Dirs.landmarks_partial_leaves_overlay_final, '.'.join([self.file_name, 'jpg']))
+
+ def show_QC_image(self):
+ if self.do_show_QC_images:
+ cv2.imshow('QC image', self.image)
+ cv2.waitKey(0)
+
+ def show_final_image(self):
+ if self.do_show_final_images:
+ cv2.imshow('Final image', self.image_final)
+ cv2.waitKey(0)
+
+ def save_QC_image(self):
+ if self.do_save_QC_images:
+ cv2.imwrite(self.path_QC_image, self.image)
+
+ def get_QC(self):
+ return self.image
+
+ def get_final(self):
+ return self.image_final
+
+ def init_lists_dicts(self):
+ # Initialize all lists and dictionaries
+ self.classes = {}
+ self.points_list = []
+ self.image = []
+ self.ordered_midvein = []
+ self.midvein_fit = []
+ self.midvein_fit_points = []
+ self.ordered_petiole = []
+ self.apex_left = self.apex_left or None
+ self.apex_right = self.apex_right or None
+ self.apex_center = self.apex_center or None
+ self.base_left = self.base_left or None
+ self.base_right = self.base_right or None
+ self.base_center = self.base_center or None
+ self.lamina_tip = self.lamina_tip or None
+ self.lamina_base = self.lamina_base or None
+ self.width_left = self.width_left or None
+ self.width_right = self.width_right or None
+
+
diff --git a/vouchervision/component_detector/models/__init__.py b/vouchervision/component_detector/models/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/vouchervision/component_detector/models/common.py b/vouchervision/component_detector/models/common.py
new file mode 100644
index 0000000000000000000000000000000000000000..5119881e683f507967dcf4ff318706419b660518
--- /dev/null
+++ b/vouchervision/component_detector/models/common.py
@@ -0,0 +1,703 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Common modules
+"""
+
+import json
+import math
+import platform
+import warnings
+from collections import OrderedDict, namedtuple
+from copy import copy
+from pathlib import Path
+
+import cv2
+import numpy as np
+import pandas as pd
+import requests
+import torch
+import torch.nn as nn
+import yaml
+from PIL import Image
+from torch.cuda import amp
+
+from utils.datasets import exif_transpose, letterbox
+from utils.general import (LOGGER, check_requirements, check_suffix, check_version, colorstr, increment_path,
+ make_divisible, non_max_suppression, scale_coords, xywh2xyxy, xyxy2xywh)
+from utils.plots import Annotator, colors, save_one_box
+from utils.torch_utils import copy_attr, time_sync
+
+
+def autopad(k, p=None): # kernel, padding
+ # Pad to 'same'
+ if p is None:
+ p = k // 2 if isinstance(k, int) else (x // 2 for x in k) # auto-pad
+ return p
+
+
+class Conv(nn.Module):
+ # Standard convolution
+ def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups
+ super().__init__()
+ self.conv = nn.Conv2d(c1, c2, k, s, autopad(k, p), groups=g, bias=False)
+ self.bn = nn.BatchNorm2d(c2)
+ self.act = nn.SiLU() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
+
+ def forward(self, x):
+ return self.act(self.bn(self.conv(x)))
+
+ def forward_fuse(self, x):
+ return self.act(self.conv(x))
+
+
+class DWConv(Conv):
+ # Depth-wise convolution class
+ def __init__(self, c1, c2, k=1, s=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups
+ super().__init__(c1, c2, k, s, g=math.gcd(c1, c2), act=act)
+
+
+class TransformerLayer(nn.Module):
+ # Transformer layer https://arxiv.org/abs/2010.11929 (LayerNorm layers removed for better performance)
+ def __init__(self, c, num_heads):
+ super().__init__()
+ self.q = nn.Linear(c, c, bias=False)
+ self.k = nn.Linear(c, c, bias=False)
+ self.v = nn.Linear(c, c, bias=False)
+ self.ma = nn.MultiheadAttention(embed_dim=c, num_heads=num_heads)
+ self.fc1 = nn.Linear(c, c, bias=False)
+ self.fc2 = nn.Linear(c, c, bias=False)
+
+ def forward(self, x):
+ x = self.ma(self.q(x), self.k(x), self.v(x))[0] + x
+ x = self.fc2(self.fc1(x)) + x
+ return x
+
+
+class TransformerBlock(nn.Module):
+ # Vision Transformer https://arxiv.org/abs/2010.11929
+ def __init__(self, c1, c2, num_heads, num_layers):
+ super().__init__()
+ self.conv = None
+ if c1 != c2:
+ self.conv = Conv(c1, c2)
+ self.linear = nn.Linear(c2, c2) # learnable position embedding
+ self.tr = nn.Sequential(*(TransformerLayer(c2, num_heads) for _ in range(num_layers)))
+ self.c2 = c2
+
+ def forward(self, x):
+ if self.conv is not None:
+ x = self.conv(x)
+ b, _, w, h = x.shape
+ p = x.flatten(2).permute(2, 0, 1)
+ return self.tr(p + self.linear(p)).permute(1, 2, 0).reshape(b, self.c2, w, h)
+
+
+class Bottleneck(nn.Module):
+ # Standard bottleneck
+ def __init__(self, c1, c2, shortcut=True, g=1, e=0.5): # ch_in, ch_out, shortcut, groups, expansion
+ super().__init__()
+ c_ = int(c2 * e) # hidden channels
+ self.cv1 = Conv(c1, c_, 1, 1)
+ self.cv2 = Conv(c_, c2, 3, 1, g=g)
+ self.add = shortcut and c1 == c2
+
+ def forward(self, x):
+ return x + self.cv2(self.cv1(x)) if self.add else self.cv2(self.cv1(x))
+
+
+class BottleneckCSP(nn.Module):
+ # CSP Bottleneck https://github.com/WongKinYiu/CrossStagePartialNetworks
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5): # ch_in, ch_out, number, shortcut, groups, expansion
+ super().__init__()
+ c_ = int(c2 * e) # hidden channels
+ self.cv1 = Conv(c1, c_, 1, 1)
+ self.cv2 = nn.Conv2d(c1, c_, 1, 1, bias=False)
+ self.cv3 = nn.Conv2d(c_, c_, 1, 1, bias=False)
+ self.cv4 = Conv(2 * c_, c2, 1, 1)
+ self.bn = nn.BatchNorm2d(2 * c_) # applied to cat(cv2, cv3)
+ self.act = nn.SiLU()
+ self.m = nn.Sequential(*(Bottleneck(c_, c_, shortcut, g, e=1.0) for _ in range(n)))
+
+ def forward(self, x):
+ y1 = self.cv3(self.m(self.cv1(x)))
+ y2 = self.cv2(x)
+ return self.cv4(self.act(self.bn(torch.cat((y1, y2), 1))))
+
+
+class C3(nn.Module):
+ # CSP Bottleneck with 3 convolutions
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5): # ch_in, ch_out, number, shortcut, groups, expansion
+ super().__init__()
+ c_ = int(c2 * e) # hidden channels
+ self.cv1 = Conv(c1, c_, 1, 1)
+ self.cv2 = Conv(c1, c_, 1, 1)
+ self.cv3 = Conv(2 * c_, c2, 1) # optional act=FReLU(c2)
+ self.m = nn.Sequential(*(Bottleneck(c_, c_, shortcut, g, e=1.0) for _ in range(n)))
+ # self.m = nn.Sequential(*(CrossConv(c_, c_, 3, 1, g, 1.0, shortcut) for _ in range(n)))
+
+ def forward(self, x):
+ return self.cv3(torch.cat((self.m(self.cv1(x)), self.cv2(x)), 1))
+
+
+class C3TR(C3):
+ # C3 module with TransformerBlock()
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5):
+ super().__init__(c1, c2, n, shortcut, g, e)
+ c_ = int(c2 * e)
+ self.m = TransformerBlock(c_, c_, 4, n)
+
+
+class C3SPP(C3):
+ # C3 module with SPP()
+ def __init__(self, c1, c2, k=(5, 9, 13), n=1, shortcut=True, g=1, e=0.5):
+ super().__init__(c1, c2, n, shortcut, g, e)
+ c_ = int(c2 * e)
+ self.m = SPP(c_, c_, k)
+
+
+class C3Ghost(C3):
+ # C3 module with GhostBottleneck()
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5):
+ super().__init__(c1, c2, n, shortcut, g, e)
+ c_ = int(c2 * e) # hidden channels
+ self.m = nn.Sequential(*(GhostBottleneck(c_, c_) for _ in range(n)))
+
+
+class SPP(nn.Module):
+ # Spatial Pyramid Pooling (SPP) layer https://arxiv.org/abs/1406.4729
+ def __init__(self, c1, c2, k=(5, 9, 13)):
+ super().__init__()
+ c_ = c1 // 2 # hidden channels
+ self.cv1 = Conv(c1, c_, 1, 1)
+ self.cv2 = Conv(c_ * (len(k) + 1), c2, 1, 1)
+ self.m = nn.ModuleList([nn.MaxPool2d(kernel_size=x, stride=1, padding=x // 2) for x in k])
+
+ def forward(self, x):
+ x = self.cv1(x)
+ with warnings.catch_warnings():
+ warnings.simplefilter('ignore') # suppress torch 1.9.0 max_pool2d() warning
+ return self.cv2(torch.cat([x] + [m(x) for m in self.m], 1))
+
+
+class SPPF(nn.Module):
+ # Spatial Pyramid Pooling - Fast (SPPF) layer for YOLOv5 by Glenn Jocher
+ def __init__(self, c1, c2, k=5): # equivalent to SPP(k=(5, 9, 13))
+ super().__init__()
+ c_ = c1 // 2 # hidden channels
+ self.cv1 = Conv(c1, c_, 1, 1)
+ self.cv2 = Conv(c_ * 4, c2, 1, 1)
+ self.m = nn.MaxPool2d(kernel_size=k, stride=1, padding=k // 2)
+
+ def forward(self, x):
+ x = self.cv1(x)
+ with warnings.catch_warnings():
+ warnings.simplefilter('ignore') # suppress torch 1.9.0 max_pool2d() warning
+ y1 = self.m(x)
+ y2 = self.m(y1)
+ return self.cv2(torch.cat((x, y1, y2, self.m(y2)), 1))
+
+
+class Focus(nn.Module):
+ # Focus wh information into c-space
+ def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True): # ch_in, ch_out, kernel, stride, padding, groups
+ super().__init__()
+ self.conv = Conv(c1 * 4, c2, k, s, p, g, act)
+ # self.contract = Contract(gain=2)
+
+ def forward(self, x): # x(b,c,w,h) -> y(b,4c,w/2,h/2)
+ return self.conv(torch.cat((x[..., ::2, ::2], x[..., 1::2, ::2], x[..., ::2, 1::2], x[..., 1::2, 1::2]), 1))
+ # return self.conv(self.contract(x))
+
+
+class GhostConv(nn.Module):
+ # Ghost Convolution https://github.com/huawei-noah/ghostnet
+ def __init__(self, c1, c2, k=1, s=1, g=1, act=True): # ch_in, ch_out, kernel, stride, groups
+ super().__init__()
+ c_ = c2 // 2 # hidden channels
+ self.cv1 = Conv(c1, c_, k, s, None, g, act)
+ self.cv2 = Conv(c_, c_, 5, 1, None, c_, act)
+
+ def forward(self, x):
+ y = self.cv1(x)
+ return torch.cat((y, self.cv2(y)), 1)
+
+
+class GhostBottleneck(nn.Module):
+ # Ghost Bottleneck https://github.com/huawei-noah/ghostnet
+ def __init__(self, c1, c2, k=3, s=1): # ch_in, ch_out, kernel, stride
+ super().__init__()
+ c_ = c2 // 2
+ self.conv = nn.Sequential(
+ GhostConv(c1, c_, 1, 1), # pw
+ DWConv(c_, c_, k, s, act=False) if s == 2 else nn.Identity(), # dw
+ GhostConv(c_, c2, 1, 1, act=False)) # pw-linear
+ self.shortcut = nn.Sequential(DWConv(c1, c1, k, s, act=False), Conv(c1, c2, 1, 1,
+ act=False)) if s == 2 else nn.Identity()
+
+ def forward(self, x):
+ return self.conv(x) + self.shortcut(x)
+
+
+class Contract(nn.Module):
+ # Contract width-height into channels, i.e. x(1,64,80,80) to x(1,256,40,40)
+ def __init__(self, gain=2):
+ super().__init__()
+ self.gain = gain
+
+ def forward(self, x):
+ b, c, h, w = x.size() # assert (h / s == 0) and (W / s == 0), 'Indivisible gain'
+ s = self.gain
+ x = x.view(b, c, h // s, s, w // s, s) # x(1,64,40,2,40,2)
+ x = x.permute(0, 3, 5, 1, 2, 4).contiguous() # x(1,2,2,64,40,40)
+ return x.view(b, c * s * s, h // s, w // s) # x(1,256,40,40)
+
+
+class Expand(nn.Module):
+ # Expand channels into width-height, i.e. x(1,64,80,80) to x(1,16,160,160)
+ def __init__(self, gain=2):
+ super().__init__()
+ self.gain = gain
+
+ def forward(self, x):
+ b, c, h, w = x.size() # assert C / s ** 2 == 0, 'Indivisible gain'
+ s = self.gain
+ x = x.view(b, s, s, c // s ** 2, h, w) # x(1,2,2,16,80,80)
+ x = x.permute(0, 3, 4, 1, 5, 2).contiguous() # x(1,16,80,2,80,2)
+ return x.view(b, c // s ** 2, h * s, w * s) # x(1,16,160,160)
+
+
+class Concat(nn.Module):
+ # Concatenate a list of tensors along dimension
+ def __init__(self, dimension=1):
+ super().__init__()
+ self.d = dimension
+
+ def forward(self, x):
+ return torch.cat(x, self.d)
+
+
+class DetectMultiBackend(nn.Module):
+ # YOLOv5 MultiBackend class for python inference on various backends
+ def __init__(self, weights='yolov5s.pt', device=torch.device('cpu'), dnn=False, data=None, fp16=False):
+ # Usage:
+ # PyTorch: weights = *.pt
+ # TorchScript: *.torchscript
+ # ONNX Runtime: *.onnx
+ # ONNX OpenCV DNN: *.onnx with --dnn
+ # OpenVINO: *.xml
+ # CoreML: *.mlmodel
+ # TensorRT: *.engine
+ # TensorFlow SavedModel: *_saved_model
+ # TensorFlow GraphDef: *.pb
+ # TensorFlow Lite: *.tflite
+ # TensorFlow Edge TPU: *_edgetpu.tflite
+ from models.experimental import attempt_download, attempt_load # scoped to avoid circular import
+
+ super().__init__()
+ w = str(weights[0] if isinstance(weights, list) else weights)
+ pt, jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs = self.model_type(w) # get backend
+ stride, names = 32, [f'class{i}' for i in range(1000)] # assign defaults
+ w = attempt_download(w) # download if not local
+ fp16 &= (pt or jit or onnx or engine) and device.type != 'cpu' # FP16
+ if data: # data.yaml path (optional)
+ with open(data, errors='ignore') as f:
+ names = yaml.safe_load(f)['names'] # class names
+
+ if pt: # PyTorch
+ model = attempt_load(weights if isinstance(weights, list) else w, map_location=device)
+ stride = max(int(model.stride.max()), 32) # model stride
+ names = model.module.names if hasattr(model, 'module') else model.names # get class names
+ model.half() if fp16 else model.float()
+ self.model = model # explicitly assign for to(), cpu(), cuda(), half()
+ elif jit: # TorchScript
+ LOGGER.info(f'Loading {w} for TorchScript inference...')
+ extra_files = {'config.txt': ''} # model metadata
+ model = torch.jit.load(w, _extra_files=extra_files)
+ model.half() if fp16 else model.float()
+ if extra_files['config.txt']:
+ d = json.loads(extra_files['config.txt']) # extra_files dict
+ stride, names = int(d['stride']), d['names']
+ elif dnn: # ONNX OpenCV DNN
+ LOGGER.info(f'Loading {w} for ONNX OpenCV DNN inference...')
+ check_requirements(('opencv-python>=4.5.4',))
+ net = cv2.dnn.readNetFromONNX(w)
+ elif onnx: # ONNX Runtime
+ LOGGER.info(f'Loading {w} for ONNX Runtime inference...')
+ cuda = torch.cuda.is_available()
+ check_requirements(('onnx', 'onnxruntime-gpu' if cuda else 'onnxruntime'))
+ import onnxruntime
+ providers = ['CUDAExecutionProvider', 'CPUExecutionProvider'] if cuda else ['CPUExecutionProvider']
+ session = onnxruntime.InferenceSession(w, providers=providers)
+ meta = session.get_modelmeta().custom_metadata_map # metadata
+ if 'stride' in meta:
+ stride, names = int(meta['stride']), eval(meta['names'])
+ elif xml: # OpenVINO
+ LOGGER.info(f'Loading {w} for OpenVINO inference...')
+ check_requirements(('openvino-dev',)) # requires openvino-dev: https://pypi.org/project/openvino-dev/
+ import openvino.inference_engine as ie
+ core = ie.IECore()
+ if not Path(w).is_file(): # if not *.xml
+ w = next(Path(w).glob('*.xml')) # get *.xml file from *_openvino_model dir
+ network = core.read_network(model=w, weights=Path(w).with_suffix('.bin')) # *.xml, *.bin paths
+ executable_network = core.load_network(network, device_name='CPU', num_requests=1)
+ elif engine: # TensorRT
+ LOGGER.info(f'Loading {w} for TensorRT inference...')
+ import tensorrt as trt # https://developer.nvidia.com/nvidia-tensorrt-download
+ check_version(trt.__version__, '7.0.0', hard=True) # require tensorrt>=7.0.0
+ Binding = namedtuple('Binding', ('name', 'dtype', 'shape', 'data', 'ptr'))
+ logger = trt.Logger(trt.Logger.INFO)
+ with open(w, 'rb') as f, trt.Runtime(logger) as runtime:
+ model = runtime.deserialize_cuda_engine(f.read())
+ bindings = OrderedDict()
+ fp16 = False # default updated below
+ for index in range(model.num_bindings):
+ name = model.get_binding_name(index)
+ dtype = trt.nptype(model.get_binding_dtype(index))
+ shape = tuple(model.get_binding_shape(index))
+ data = torch.from_numpy(np.empty(shape, dtype=np.dtype(dtype))).to(device)
+ bindings[name] = Binding(name, dtype, shape, data, int(data.data_ptr()))
+ if model.binding_is_input(index) and dtype == np.float16:
+ fp16 = True
+ binding_addrs = OrderedDict((n, d.ptr) for n, d in bindings.items())
+ context = model.create_execution_context()
+ batch_size = bindings['images'].shape[0]
+ elif coreml: # CoreML
+ LOGGER.info(f'Loading {w} for CoreML inference...')
+ import coremltools as ct
+ model = ct.models.MLModel(w)
+ else: # TensorFlow (SavedModel, GraphDef, Lite, Edge TPU)
+ if saved_model: # SavedModel
+ LOGGER.info(f'Loading {w} for TensorFlow SavedModel inference...')
+ import tensorflow as tf
+ keras = False # assume TF1 saved_model
+ model = tf.keras.models.load_model(w) if keras else tf.saved_model.load(w)
+ elif pb: # GraphDef https://www.tensorflow.org/guide/migrate#a_graphpb_or_graphpbtxt
+ LOGGER.info(f'Loading {w} for TensorFlow GraphDef inference...')
+ import tensorflow as tf
+
+ def wrap_frozen_graph(gd, inputs, outputs):
+ x = tf.compat.v1.wrap_function(lambda: tf.compat.v1.import_graph_def(gd, name=""), []) # wrapped
+ ge = x.graph.as_graph_element
+ return x.prune(tf.nest.map_structure(ge, inputs), tf.nest.map_structure(ge, outputs))
+
+ gd = tf.Graph().as_graph_def() # graph_def
+ with open(w, 'rb') as f:
+ gd.ParseFromString(f.read())
+ frozen_func = wrap_frozen_graph(gd, inputs="x:0", outputs="Identity:0")
+ elif tflite or edgetpu: # https://www.tensorflow.org/lite/guide/python#install_tensorflow_lite_for_python
+ try: # https://coral.ai/docs/edgetpu/tflite-python/#update-existing-tf-lite-code-for-the-edge-tpu
+ from tflite_runtime.interpreter import Interpreter, load_delegate
+ except ImportError:
+ import tensorflow as tf
+ Interpreter, load_delegate = tf.lite.Interpreter, tf.lite.experimental.load_delegate,
+ if edgetpu: # Edge TPU https://coral.ai/software/#edgetpu-runtime
+ LOGGER.info(f'Loading {w} for TensorFlow Lite Edge TPU inference...')
+ delegate = {
+ 'Linux': 'libedgetpu.so.1',
+ 'Darwin': 'libedgetpu.1.dylib',
+ 'Windows': 'edgetpu.dll'}[platform.system()]
+ interpreter = Interpreter(model_path=w, experimental_delegates=[load_delegate(delegate)])
+ else: # Lite
+ LOGGER.info(f'Loading {w} for TensorFlow Lite inference...')
+ interpreter = Interpreter(model_path=w) # load TFLite model
+ interpreter.allocate_tensors() # allocate
+ input_details = interpreter.get_input_details() # inputs
+ output_details = interpreter.get_output_details() # outputs
+ elif tfjs:
+ raise Exception('ERROR: YOLOv5 TF.js inference is not supported')
+ self.__dict__.update(locals()) # assign all variables to self
+
+ def forward(self, im, augment=False, visualize=False, val=False):
+ # YOLOv5 MultiBackend inference
+ b, ch, h, w = im.shape # batch, channel, height, width
+ if self.pt: # PyTorch
+ y = self.model(im, augment=augment, visualize=visualize)[0]
+ elif self.jit: # TorchScript
+ y = self.model(im)[0]
+ elif self.dnn: # ONNX OpenCV DNN
+ im = im.cpu().numpy() # torch to numpy
+ self.net.setInput(im)
+ y = self.net.forward()
+ elif self.onnx: # ONNX Runtime
+ im = im.cpu().numpy() # torch to numpy
+ y = self.session.run([self.session.get_outputs()[0].name], {self.session.get_inputs()[0].name: im})[0]
+ elif self.xml: # OpenVINO
+ im = im.cpu().numpy() # FP32
+ desc = self.ie.TensorDesc(precision='FP32', dims=im.shape, layout='NCHW') # Tensor Description
+ request = self.executable_network.requests[0] # inference request
+ request.set_blob(blob_name='images', blob=self.ie.Blob(desc, im)) # name=next(iter(request.input_blobs))
+ request.infer()
+ y = request.output_blobs['output'].buffer # name=next(iter(request.output_blobs))
+ elif self.engine: # TensorRT
+ assert im.shape == self.bindings['images'].shape, (im.shape, self.bindings['images'].shape)
+ self.binding_addrs['images'] = int(im.data_ptr())
+ self.context.execute_v2(list(self.binding_addrs.values()))
+ y = self.bindings['output'].data
+ elif self.coreml: # CoreML
+ im = im.permute(0, 2, 3, 1).cpu().numpy() # torch BCHW to numpy BHWC shape(1,320,192,3)
+ im = Image.fromarray((im[0] * 255).astype('uint8'))
+ # im = im.resize((192, 320), Image.ANTIALIAS)
+ y = self.model.predict({'image': im}) # coordinates are xywh normalized
+ if 'confidence' in y:
+ box = xywh2xyxy(y['coordinates'] * [[w, h, w, h]]) # xyxy pixels
+ conf, cls = y['confidence'].max(1), y['confidence'].argmax(1).astype(np.float)
+ y = np.concatenate((box, conf.reshape(-1, 1), cls.reshape(-1, 1)), 1)
+ else:
+ k = 'var_' + str(sorted(int(k.replace('var_', '')) for k in y)[-1]) # output key
+ y = y[k] # output
+ else: # TensorFlow (SavedModel, GraphDef, Lite, Edge TPU)
+ im = im.permute(0, 2, 3, 1).cpu().numpy() # torch BCHW to numpy BHWC shape(1,320,192,3)
+ if self.saved_model: # SavedModel
+ y = (self.model(im, training=False) if self.keras else self.model(im)).numpy()
+ elif self.pb: # GraphDef
+ y = self.frozen_func(x=self.tf.constant(im)).numpy()
+ else: # Lite or Edge TPU
+ input, output = self.input_details[0], self.output_details[0]
+ int8 = input['dtype'] == np.uint8 # is TFLite quantized uint8 model
+ if int8:
+ scale, zero_point = input['quantization']
+ im = (im / scale + zero_point).astype(np.uint8) # de-scale
+ self.interpreter.set_tensor(input['index'], im)
+ self.interpreter.invoke()
+ y = self.interpreter.get_tensor(output['index'])
+ if int8:
+ scale, zero_point = output['quantization']
+ y = (y.astype(np.float32) - zero_point) * scale # re-scale
+ y[..., :4] *= [w, h, w, h] # xywh normalized to pixels
+
+ if isinstance(y, np.ndarray):
+ y = torch.tensor(y, device=self.device)
+ return (y, []) if val else y
+
+ def warmup(self, imgsz=(1, 3, 640, 640)):
+ # Warmup model by running inference once
+ if any((self.pt, self.jit, self.onnx, self.engine, self.saved_model, self.pb)): # warmup types
+ if self.device.type != 'cpu': # only warmup GPU models
+ im = torch.zeros(*imgsz, dtype=torch.half if self.fp16 else torch.float, device=self.device) # input
+ for _ in range(2 if self.jit else 1): #
+ self.forward(im) # warmup
+
+ @staticmethod
+ def model_type(p='path/to/model.pt'):
+ # Return model type from model path, i.e. path='path/to/model.onnx' -> type=onnx
+ from export import export_formats
+ suffixes = list(export_formats().Suffix) + ['.xml'] # export suffixes
+ check_suffix(p, suffixes) # checks
+ p = Path(p).name # eliminate trailing separators
+ pt, jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs, xml2 = (s in p for s in suffixes)
+ xml |= xml2 # *_openvino_model or *.xml
+ tflite &= not edgetpu # *.tflite
+ return pt, jit, onnx, xml, engine, coreml, saved_model, pb, tflite, edgetpu, tfjs
+
+
+class AutoShape(nn.Module):
+ # YOLOv5 input-robust model wrapper for passing cv2/np/PIL/torch inputs. Includes preprocessing, inference and NMS
+ conf = 0.25 # NMS confidence threshold
+ iou = 0.45 # NMS IoU threshold
+ agnostic = False # NMS class-agnostic
+ multi_label = False # NMS multiple labels per box
+ classes = None # (optional list) filter by class, i.e. = [0, 15, 16] for COCO persons, cats and dogs
+ max_det = 1000 # maximum number of detections per image
+ amp = False # Automatic Mixed Precision (AMP) inference
+
+ def __init__(self, model):
+ super().__init__()
+ LOGGER.info('Adding AutoShape... ')
+ copy_attr(self, model, include=('yaml', 'nc', 'hyp', 'names', 'stride', 'abc'), exclude=()) # copy attributes
+ self.dmb = isinstance(model, DetectMultiBackend) # DetectMultiBackend() instance
+ self.pt = not self.dmb or model.pt # PyTorch model
+ self.model = model.eval()
+
+ def _apply(self, fn):
+ # Apply to(), cpu(), cuda(), half() to model tensors that are not parameters or registered buffers
+ self = super()._apply(fn)
+ if self.pt:
+ m = self.model.model.model[-1] if self.dmb else self.model.model[-1] # Detect()
+ m.stride = fn(m.stride)
+ m.grid = list(map(fn, m.grid))
+ if isinstance(m.anchor_grid, list):
+ m.anchor_grid = list(map(fn, m.anchor_grid))
+ return self
+
+ @torch.no_grad()
+ def forward(self, imgs, size=640, augment=False, profile=False):
+ # Inference from various sources. For height=640, width=1280, RGB images example inputs are:
+ # file: imgs = 'data/images/zidane.jpg' # str or PosixPath
+ # URI: = 'https://ultralytics.com/images/zidane.jpg'
+ # OpenCV: = cv2.imread('image.jpg')[:,:,::-1] # HWC BGR to RGB x(640,1280,3)
+ # PIL: = Image.open('image.jpg') or ImageGrab.grab() # HWC x(640,1280,3)
+ # numpy: = np.zeros((640,1280,3)) # HWC
+ # torch: = torch.zeros(16,3,320,640) # BCHW (scaled to size=640, 0-1 values)
+ # multiple: = [Image.open('image1.jpg'), Image.open('image2.jpg'), ...] # list of images
+
+ t = [time_sync()]
+ p = next(self.model.parameters()) if self.pt else torch.zeros(1, device=self.model.device) # for device, type
+ autocast = self.amp and (p.device.type != 'cpu') # Automatic Mixed Precision (AMP) inference
+ if isinstance(imgs, torch.Tensor): # torch
+ with amp.autocast(autocast):
+ return self.model(imgs.to(p.device).type_as(p), augment, profile) # inference
+
+ # Pre-process
+ n, imgs = (len(imgs), list(imgs)) if isinstance(imgs, (list, tuple)) else (1, [imgs]) # number, list of images
+ shape0, shape1, files = [], [], [] # image and inference shapes, filenames
+ for i, im in enumerate(imgs):
+ f = f'image{i}' # filename
+ if isinstance(im, (str, Path)): # filename or uri
+ im, f = Image.open(requests.get(im, stream=True).raw if str(im).startswith('http') else im), im
+ im = np.asarray(exif_transpose(im))
+ elif isinstance(im, Image.Image): # PIL Image
+ im, f = np.asarray(exif_transpose(im)), getattr(im, 'filename', f) or f
+ files.append(Path(f).with_suffix('.jpg').name)
+ if im.shape[0] < 5: # image in CHW
+ im = im.transpose((1, 2, 0)) # reverse dataloader .transpose(2, 0, 1)
+ im = im[..., :3] if im.ndim == 3 else np.tile(im[..., None], 3) # enforce 3ch input
+ s = im.shape[:2] # HWC
+ shape0.append(s) # image shape
+ g = (size / max(s)) # gain
+ shape1.append([y * g for y in s])
+ imgs[i] = im if im.data.contiguous else np.ascontiguousarray(im) # update
+ shape1 = [make_divisible(x, self.stride) if self.pt else size for x in np.array(shape1).max(0)] # inf shape
+ x = [letterbox(im, shape1, auto=False)[0] for im in imgs] # pad
+ x = np.ascontiguousarray(np.array(x).transpose((0, 3, 1, 2))) # stack and BHWC to BCHW
+ x = torch.from_numpy(x).to(p.device).type_as(p) / 255 # uint8 to fp16/32
+ t.append(time_sync())
+
+ with amp.autocast(autocast):
+ # Inference
+ y = self.model(x, augment, profile) # forward
+ t.append(time_sync())
+
+ # Post-process
+ y = non_max_suppression(y if self.dmb else y[0],
+ self.conf,
+ self.iou,
+ self.classes,
+ self.agnostic,
+ self.multi_label,
+ max_det=self.max_det) # NMS
+ for i in range(n):
+ scale_coords(shape1, y[i][:, :4], shape0[i])
+
+ t.append(time_sync())
+ return Detections(imgs, y, files, t, self.names, x.shape)
+
+
+class Detections:
+ # YOLOv5 detections class for inference results
+ def __init__(self, imgs, pred, files, times=(0, 0, 0, 0), names=None, shape=None):
+ super().__init__()
+ d = pred[0].device # device
+ gn = [torch.tensor([*(im.shape[i] for i in [1, 0, 1, 0]), 1, 1], device=d) for im in imgs] # normalizations
+ self.imgs = imgs # list of images as numpy arrays
+ self.pred = pred # list of tensors pred[0] = (xyxy, conf, cls)
+ self.names = names # class names
+ self.files = files # image filenames
+ self.times = times # profiling times
+ self.xyxy = pred # xyxy pixels
+ self.xywh = [xyxy2xywh(x) for x in pred] # xywh pixels
+ self.xyxyn = [x / g for x, g in zip(self.xyxy, gn)] # xyxy normalized
+ self.xywhn = [x / g for x, g in zip(self.xywh, gn)] # xywh normalized
+ self.n = len(self.pred) # number of images (batch size)
+ self.t = tuple((times[i + 1] - times[i]) * 1000 / self.n for i in range(3)) # timestamps (ms)
+ self.s = shape # inference BCHW shape
+
+ def display(self, pprint=False, show=False, save=False, crop=False, render=False, labels=True, save_dir=Path('')):
+ crops = []
+ for i, (im, pred) in enumerate(zip(self.imgs, self.pred)):
+ s = f'image {i + 1}/{len(self.pred)}: {im.shape[0]}x{im.shape[1]} ' # string
+ if pred.shape[0]:
+ for c in pred[:, -1].unique():
+ n = (pred[:, -1] == c).sum() # detections per class
+ s += f"{n} {self.names[int(c)]}{'s' * (n > 1)}, " # add to string
+ if show or save or render or crop:
+ annotator = Annotator(im, example=str(self.names))
+ for *box, conf, cls in reversed(pred): # xyxy, confidence, class
+ label = f'{self.names[int(cls)]} {conf:.2f}'
+ if crop:
+ file = save_dir / 'crops' / self.names[int(cls)] / self.files[i] if save else None
+ crops.append({
+ 'box': box,
+ 'conf': conf,
+ 'cls': cls,
+ 'label': label,
+ 'im': save_one_box(box, im, file=file, save=save)})
+ else: # all others
+ annotator.box_label(box, label if labels else '', color=colors(cls))
+ im = annotator.im
+ else:
+ s += '(no detections)'
+
+ im = Image.fromarray(im.astype(np.uint8)) if isinstance(im, np.ndarray) else im # from np
+ if pprint:
+ print(s.rstrip(', '))
+ if show:
+ im.show(self.files[i]) # show
+ if save:
+ f = self.files[i]
+ im.save(save_dir / f) # save
+ if i == self.n - 1:
+ LOGGER.info(f"Saved {self.n} image{'s' * (self.n > 1)} to {colorstr('bold', save_dir)}")
+ if render:
+ self.imgs[i] = np.asarray(im)
+ if crop:
+ if save:
+ LOGGER.info(f'Saved results to {save_dir}\n')
+ return crops
+
+ def print(self):
+ self.display(pprint=True) # print results
+ print(f'Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {tuple(self.s)}' % self.t)
+
+ def show(self, labels=True):
+ self.display(show=True, labels=labels) # show results
+
+ def save(self, labels=True, save_dir='runs/detect/exp'):
+ save_dir = increment_path(save_dir, exist_ok=save_dir != 'runs/detect/exp', mkdir=True) # increment save_dir
+ self.display(save=True, labels=labels, save_dir=save_dir) # save results
+
+ def crop(self, save=True, save_dir='runs/detect/exp'):
+ save_dir = increment_path(save_dir, exist_ok=save_dir != 'runs/detect/exp', mkdir=True) if save else None
+ return self.display(crop=True, save=save, save_dir=save_dir) # crop results
+
+ def render(self, labels=True):
+ self.display(render=True, labels=labels) # render results
+ return self.imgs
+
+ def pandas(self):
+ # return detections as pandas DataFrames, i.e. print(results.pandas().xyxy[0])
+ new = copy(self) # return copy
+ ca = 'xmin', 'ymin', 'xmax', 'ymax', 'confidence', 'class', 'name' # xyxy columns
+ cb = 'xcenter', 'ycenter', 'width', 'height', 'confidence', 'class', 'name' # xywh columns
+ for k, c in zip(['xyxy', 'xyxyn', 'xywh', 'xywhn'], [ca, ca, cb, cb]):
+ a = [[x[:5] + [int(x[5]), self.names[int(x[5])]] for x in x.tolist()] for x in getattr(self, k)] # update
+ setattr(new, k, [pd.DataFrame(x, columns=c) for x in a])
+ return new
+
+ def tolist(self):
+ # return a list of Detections objects, i.e. 'for result in results.tolist():'
+ r = range(self.n) # iterable
+ x = [Detections([self.imgs[i]], [self.pred[i]], [self.files[i]], self.times, self.names, self.s) for i in r]
+ # for d in x:
+ # for k in ['imgs', 'pred', 'xyxy', 'xyxyn', 'xywh', 'xywhn']:
+ # setattr(d, k, getattr(d, k)[0]) # pop out of list
+ return x
+
+ def __len__(self):
+ return self.n # override len(results)
+
+ def __str__(self):
+ self.print() # override print(results)
+ return ''
+
+
+class Classify(nn.Module):
+ # Classification head, i.e. x(b,c1,20,20) to x(b,c2)
+ def __init__(self, c1, c2, k=1, s=1, p=None, g=1): # ch_in, ch_out, kernel, stride, padding, groups
+ super().__init__()
+ self.aap = nn.AdaptiveAvgPool2d(1) # to x(b,c1,1,1)
+ self.conv = nn.Conv2d(c1, c2, k, s, autopad(k, p), groups=g) # to x(b,c2,1,1)
+ self.flat = nn.Flatten()
+
+ def forward(self, x):
+ z = torch.cat([self.aap(y) for y in (x if isinstance(x, list) else [x])], 1) # cat if list
+ return self.flat(self.conv(z)) # flatten to x(b,c2)
diff --git a/vouchervision/component_detector/models/experimental.py b/vouchervision/component_detector/models/experimental.py
new file mode 100644
index 0000000000000000000000000000000000000000..b8d4d70d26e806ae41c7accc4fdbb7edb99e9de9
--- /dev/null
+++ b/vouchervision/component_detector/models/experimental.py
@@ -0,0 +1,122 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Experimental modules
+"""
+import math
+
+import numpy as np
+import torch
+import torch.nn as nn
+
+from models.common import Conv
+from utils.downloads import attempt_download
+
+
+class CrossConv(nn.Module):
+ # Cross Convolution Downsample
+ def __init__(self, c1, c2, k=3, s=1, g=1, e=1.0, shortcut=False):
+ # ch_in, ch_out, kernel, stride, groups, expansion, shortcut
+ super().__init__()
+ c_ = int(c2 * e) # hidden channels
+ self.cv1 = Conv(c1, c_, (1, k), (1, s))
+ self.cv2 = Conv(c_, c2, (k, 1), (s, 1), g=g)
+ self.add = shortcut and c1 == c2
+
+ def forward(self, x):
+ return x + self.cv2(self.cv1(x)) if self.add else self.cv2(self.cv1(x))
+
+
+class Sum(nn.Module):
+ # Weighted sum of 2 or more layers https://arxiv.org/abs/1911.09070
+ def __init__(self, n, weight=False): # n: number of inputs
+ super().__init__()
+ self.weight = weight # apply weights boolean
+ self.iter = range(n - 1) # iter object
+ if weight:
+ self.w = nn.Parameter(-torch.arange(1.0, n) / 2, requires_grad=True) # layer weights
+
+ def forward(self, x):
+ y = x[0] # no weight
+ if self.weight:
+ w = torch.sigmoid(self.w) * 2
+ for i in self.iter:
+ y = y + x[i + 1] * w[i]
+ else:
+ for i in self.iter:
+ y = y + x[i + 1]
+ return y
+
+
+class MixConv2d(nn.Module):
+ # Mixed Depth-wise Conv https://arxiv.org/abs/1907.09595
+ def __init__(self, c1, c2, k=(1, 3), s=1, equal_ch=True): # ch_in, ch_out, kernel, stride, ch_strategy
+ super().__init__()
+ n = len(k) # number of convolutions
+ if equal_ch: # equal c_ per group
+ i = torch.linspace(0, n - 1E-6, c2).floor() # c2 indices
+ c_ = [(i == g).sum() for g in range(n)] # intermediate channels
+ else: # equal weight.numel() per group
+ b = [c2] + [0] * n
+ a = np.eye(n + 1, n, k=-1)
+ a -= np.roll(a, 1, axis=1)
+ a *= np.array(k) ** 2
+ a[0] = 1
+ c_ = np.linalg.lstsq(a, b, rcond=None)[0].round() # solve for equal weight indices, ax = b
+
+ self.m = nn.ModuleList([
+ nn.Conv2d(c1, int(c_), k, s, k // 2, groups=math.gcd(c1, int(c_)), bias=False) for k, c_ in zip(k, c_)])
+ self.bn = nn.BatchNorm2d(c2)
+ self.act = nn.SiLU()
+
+ def forward(self, x):
+ return self.act(self.bn(torch.cat([m(x) for m in self.m], 1)))
+
+
+class Ensemble(nn.ModuleList):
+ # Ensemble of models
+ def __init__(self):
+ super().__init__()
+
+ def forward(self, x, augment=False, profile=False, visualize=False):
+ y = []
+ for module in self:
+ y.append(module(x, augment, profile, visualize)[0])
+ # y = torch.stack(y).max(0)[0] # max ensemble
+ # y = torch.stack(y).mean(0) # mean ensemble
+ y = torch.cat(y, 1) # nms ensemble
+ return y, None # inference, train output
+
+
+def attempt_load(weights, map_location=None, inplace=True, fuse=True):
+ from models.yolo import Detect, Model
+
+ # Loads an ensemble of models weights=[a,b,c] or a single model weights=[a] or weights=a
+ model = Ensemble()
+ for w in weights if isinstance(weights, list) else [weights]:
+ ckpt = torch.load(attempt_download(w), map_location=map_location) # load
+ ckpt = (ckpt.get('ema') or ckpt['model']).float() # FP32 model
+ model.append(ckpt.fuse().eval() if fuse else ckpt.eval()) # fused or un-fused model in eval mode
+
+ # Compatibility updates
+ for m in model.modules():
+ t = type(m)
+ if t in (nn.Hardswish, nn.LeakyReLU, nn.ReLU, nn.ReLU6, nn.SiLU, Detect, Model):
+ m.inplace = inplace # torch 1.7.0 compatibility
+ if t is Detect:
+ if not isinstance(m.anchor_grid, list): # new Detect Layer compatibility
+ delattr(m, 'anchor_grid')
+ setattr(m, 'anchor_grid', [torch.zeros(1)] * m.nl)
+ elif t is Conv:
+ m._non_persistent_buffers_set = set() # torch 1.6.0 compatibility
+ elif t is nn.Upsample and not hasattr(m, 'recompute_scale_factor'):
+ m.recompute_scale_factor = None # torch 1.11.0 compatibility
+
+ if len(model) == 1:
+ return model[-1] # return model
+ else:
+ print(f'Ensemble created with {weights}\n')
+ for k in 'names', 'nc', 'yaml':
+ setattr(model, k, getattr(model[0], k))
+ model.stride = model[torch.argmax(torch.tensor([m.stride.max() for m in model])).int()].stride # max stride
+ assert all(model[0].nc == m.nc for m in model), f'Models have different class counts: {[m.nc for m in model]}'
+ return model # return ensemble
diff --git a/vouchervision/component_detector/models/hub/anchors.yaml b/vouchervision/component_detector/models/hub/anchors.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..e4d7beb06e07f295eaf58b1ebb2430a67997d2d4
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/anchors.yaml
@@ -0,0 +1,59 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+# Default anchors for COCO data
+
+
+# P5 -------------------------------------------------------------------------------------------------------------------
+# P5-640:
+anchors_p5_640:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+
+# P6 -------------------------------------------------------------------------------------------------------------------
+# P6-640: thr=0.25: 0.9964 BPR, 5.54 anchors past thr, n=12, img_size=640, metric_all=0.281/0.716-mean/best, past_thr=0.469-mean: 9,11, 21,19, 17,41, 43,32, 39,70, 86,64, 65,131, 134,130, 120,265, 282,180, 247,354, 512,387
+anchors_p6_640:
+ - [9,11, 21,19, 17,41] # P3/8
+ - [43,32, 39,70, 86,64] # P4/16
+ - [65,131, 134,130, 120,265] # P5/32
+ - [282,180, 247,354, 512,387] # P6/64
+
+# P6-1280: thr=0.25: 0.9950 BPR, 5.55 anchors past thr, n=12, img_size=1280, metric_all=0.281/0.714-mean/best, past_thr=0.468-mean: 19,27, 44,40, 38,94, 96,68, 86,152, 180,137, 140,301, 303,264, 238,542, 436,615, 739,380, 925,792
+anchors_p6_1280:
+ - [19,27, 44,40, 38,94] # P3/8
+ - [96,68, 86,152, 180,137] # P4/16
+ - [140,301, 303,264, 238,542] # P5/32
+ - [436,615, 739,380, 925,792] # P6/64
+
+# P6-1920: thr=0.25: 0.9950 BPR, 5.55 anchors past thr, n=12, img_size=1920, metric_all=0.281/0.714-mean/best, past_thr=0.468-mean: 28,41, 67,59, 57,141, 144,103, 129,227, 270,205, 209,452, 455,396, 358,812, 653,922, 1109,570, 1387,1187
+anchors_p6_1920:
+ - [28,41, 67,59, 57,141] # P3/8
+ - [144,103, 129,227, 270,205] # P4/16
+ - [209,452, 455,396, 358,812] # P5/32
+ - [653,922, 1109,570, 1387,1187] # P6/64
+
+
+# P7 -------------------------------------------------------------------------------------------------------------------
+# P7-640: thr=0.25: 0.9962 BPR, 6.76 anchors past thr, n=15, img_size=640, metric_all=0.275/0.733-mean/best, past_thr=0.466-mean: 11,11, 13,30, 29,20, 30,46, 61,38, 39,92, 78,80, 146,66, 79,163, 149,150, 321,143, 157,303, 257,402, 359,290, 524,372
+anchors_p7_640:
+ - [11,11, 13,30, 29,20] # P3/8
+ - [30,46, 61,38, 39,92] # P4/16
+ - [78,80, 146,66, 79,163] # P5/32
+ - [149,150, 321,143, 157,303] # P6/64
+ - [257,402, 359,290, 524,372] # P7/128
+
+# P7-1280: thr=0.25: 0.9968 BPR, 6.71 anchors past thr, n=15, img_size=1280, metric_all=0.273/0.732-mean/best, past_thr=0.463-mean: 19,22, 54,36, 32,77, 70,83, 138,71, 75,173, 165,159, 148,334, 375,151, 334,317, 251,626, 499,474, 750,326, 534,814, 1079,818
+anchors_p7_1280:
+ - [19,22, 54,36, 32,77] # P3/8
+ - [70,83, 138,71, 75,173] # P4/16
+ - [165,159, 148,334, 375,151] # P5/32
+ - [334,317, 251,626, 499,474] # P6/64
+ - [750,326, 534,814, 1079,818] # P7/128
+
+# P7-1920: thr=0.25: 0.9968 BPR, 6.71 anchors past thr, n=15, img_size=1920, metric_all=0.273/0.732-mean/best, past_thr=0.463-mean: 29,34, 81,55, 47,115, 105,124, 207,107, 113,259, 247,238, 222,500, 563,227, 501,476, 376,939, 749,711, 1126,489, 801,1222, 1618,1227
+anchors_p7_1920:
+ - [29,34, 81,55, 47,115] # P3/8
+ - [105,124, 207,107, 113,259] # P4/16
+ - [247,238, 222,500, 563,227] # P5/32
+ - [501,476, 376,939, 749,711] # P6/64
+ - [1126,489, 801,1222, 1618,1227] # P7/128
diff --git a/vouchervision/component_detector/models/hub/yolov3-spp.yaml b/vouchervision/component_detector/models/hub/yolov3-spp.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..c66982158ce82d4e4ed7241c469b6f0166f0db49
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov3-spp.yaml
@@ -0,0 +1,51 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# darknet53 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [32, 3, 1]], # 0
+ [-1, 1, Conv, [64, 3, 2]], # 1-P1/2
+ [-1, 1, Bottleneck, [64]],
+ [-1, 1, Conv, [128, 3, 2]], # 3-P2/4
+ [-1, 2, Bottleneck, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 5-P3/8
+ [-1, 8, Bottleneck, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 7-P4/16
+ [-1, 8, Bottleneck, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P5/32
+ [-1, 4, Bottleneck, [1024]], # 10
+ ]
+
+# YOLOv3-SPP head
+head:
+ [[-1, 1, Bottleneck, [1024, False]],
+ [-1, 1, SPP, [512, [5, 9, 13]]],
+ [-1, 1, Conv, [1024, 3, 1]],
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, Conv, [1024, 3, 1]], # 15 (P5/32-large)
+
+ [-2, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P4
+ [-1, 1, Bottleneck, [512, False]],
+ [-1, 1, Bottleneck, [512, False]],
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, Conv, [512, 3, 1]], # 22 (P4/16-medium)
+
+ [-2, 1, Conv, [128, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P3
+ [-1, 1, Bottleneck, [256, False]],
+ [-1, 2, Bottleneck, [256, False]], # 27 (P3/8-small)
+
+ [[27, 22, 15], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov3-tiny.yaml b/vouchervision/component_detector/models/hub/yolov3-tiny.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..b28b443152485e39dcf690d18c403780c898bfab
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov3-tiny.yaml
@@ -0,0 +1,41 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [10,14, 23,27, 37,58] # P4/16
+ - [81,82, 135,169, 344,319] # P5/32
+
+# YOLOv3-tiny backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [16, 3, 1]], # 0
+ [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 1-P1/2
+ [-1, 1, Conv, [32, 3, 1]],
+ [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 3-P2/4
+ [-1, 1, Conv, [64, 3, 1]],
+ [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 5-P3/8
+ [-1, 1, Conv, [128, 3, 1]],
+ [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 7-P4/16
+ [-1, 1, Conv, [256, 3, 1]],
+ [-1, 1, nn.MaxPool2d, [2, 2, 0]], # 9-P5/32
+ [-1, 1, Conv, [512, 3, 1]],
+ [-1, 1, nn.ZeroPad2d, [[0, 1, 0, 1]]], # 11
+ [-1, 1, nn.MaxPool2d, [2, 1, 0]], # 12
+ ]
+
+# YOLOv3-tiny head
+head:
+ [[-1, 1, Conv, [1024, 3, 1]],
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, Conv, [512, 3, 1]], # 15 (P5/32-large)
+
+ [-2, 1, Conv, [128, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P4
+ [-1, 1, Conv, [256, 3, 1]], # 19 (P4/16-medium)
+
+ [[19, 15], 1, Detect, [nc, anchors]], # Detect(P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov3.yaml b/vouchervision/component_detector/models/hub/yolov3.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d1ef91290a8d261ccaf3a9663802e78b6b4e7542
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov3.yaml
@@ -0,0 +1,51 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# darknet53 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [32, 3, 1]], # 0
+ [-1, 1, Conv, [64, 3, 2]], # 1-P1/2
+ [-1, 1, Bottleneck, [64]],
+ [-1, 1, Conv, [128, 3, 2]], # 3-P2/4
+ [-1, 2, Bottleneck, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 5-P3/8
+ [-1, 8, Bottleneck, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 7-P4/16
+ [-1, 8, Bottleneck, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P5/32
+ [-1, 4, Bottleneck, [1024]], # 10
+ ]
+
+# YOLOv3 head
+head:
+ [[-1, 1, Bottleneck, [1024, False]],
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, Conv, [1024, 3, 1]],
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, Conv, [1024, 3, 1]], # 15 (P5/32-large)
+
+ [-2, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P4
+ [-1, 1, Bottleneck, [512, False]],
+ [-1, 1, Bottleneck, [512, False]],
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, Conv, [512, 3, 1]], # 22 (P4/16-medium)
+
+ [-2, 1, Conv, [128, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P3
+ [-1, 1, Bottleneck, [256, False]],
+ [-1, 2, Bottleneck, [256, False]], # 27 (P3/8-small)
+
+ [[27, 22, 15], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5-bifpn.yaml b/vouchervision/component_detector/models/hub/yolov5-bifpn.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..504815f5cfa03329618c4a1801f16ce68ec666e0
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5-bifpn.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 BiFPN head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14, 6], 1, Concat, [1]], # cat P4 <--- BiFPN change
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5-fpn.yaml b/vouchervision/component_detector/models/hub/yolov5-fpn.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..a23e9c6fbf9f7f00c9e7f2a24bc8513a9d5717ea
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5-fpn.yaml
@@ -0,0 +1,42 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 FPN head
+head:
+ [[-1, 3, C3, [1024, False]], # 10 (P5/32-large)
+
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 3, C3, [512, False]], # 14 (P4/16-medium)
+
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 3, C3, [256, False]], # 18 (P3/8-small)
+
+ [[18, 14, 10], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5-p2.yaml b/vouchervision/component_detector/models/hub/yolov5-p2.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..554117dda59aca4a016b2ff42851d39cdc34f714
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5-p2.yaml
@@ -0,0 +1,54 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors: 3 # AutoAnchor evolves 3 anchors per P output layer
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head with (P2, P3, P4, P5) outputs
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [128, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 2], 1, Concat, [1]], # cat backbone P2
+ [-1, 1, C3, [128, False]], # 21 (P2/4-xsmall)
+
+ [-1, 1, Conv, [128, 3, 2]],
+ [[-1, 18], 1, Concat, [1]], # cat head P3
+ [-1, 3, C3, [256, False]], # 24 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 27 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 30 (P5/32-large)
+
+ [[21, 24, 27, 30], 1, Detect, [nc, anchors]], # Detect(P2, P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5-p34.yaml b/vouchervision/component_detector/models/hub/yolov5-p34.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..dbf0f850083ebf546ae7fc367be029297c174da1
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5-p34.yaml
@@ -0,0 +1,41 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.33 # model depth multiple
+width_multiple: 0.50 # layer channel multiple
+anchors: 3 # AutoAnchor evolves 3 anchors per P output layer
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [ [ -1, 1, Conv, [ 64, 6, 2, 2 ] ], # 0-P1/2
+ [ -1, 1, Conv, [ 128, 3, 2 ] ], # 1-P2/4
+ [ -1, 3, C3, [ 128 ] ],
+ [ -1, 1, Conv, [ 256, 3, 2 ] ], # 3-P3/8
+ [ -1, 6, C3, [ 256 ] ],
+ [ -1, 1, Conv, [ 512, 3, 2 ] ], # 5-P4/16
+ [ -1, 9, C3, [ 512 ] ],
+ [ -1, 1, Conv, [ 1024, 3, 2 ] ], # 7-P5/32
+ [ -1, 3, C3, [ 1024 ] ],
+ [ -1, 1, SPPF, [ 1024, 5 ] ], # 9
+ ]
+
+# YOLOv5 v6.0 head with (P3, P4) outputs
+head:
+ [ [ -1, 1, Conv, [ 512, 1, 1 ] ],
+ [ -1, 1, nn.Upsample, [ None, 2, 'nearest' ] ],
+ [ [ -1, 6 ], 1, Concat, [ 1 ] ], # cat backbone P4
+ [ -1, 3, C3, [ 512, False ] ], # 13
+
+ [ -1, 1, Conv, [ 256, 1, 1 ] ],
+ [ -1, 1, nn.Upsample, [ None, 2, 'nearest' ] ],
+ [ [ -1, 4 ], 1, Concat, [ 1 ] ], # cat backbone P3
+ [ -1, 3, C3, [ 256, False ] ], # 17 (P3/8-small)
+
+ [ -1, 1, Conv, [ 256, 3, 2 ] ],
+ [ [ -1, 14 ], 1, Concat, [ 1 ] ], # cat head P4
+ [ -1, 3, C3, [ 512, False ] ], # 20 (P4/16-medium)
+
+ [ [ 17, 20 ], 1, Detect, [ nc, anchors ] ], # Detect(P3, P4)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5-p6.yaml b/vouchervision/component_detector/models/hub/yolov5-p6.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..a17202f22044c0546bd9373ea58bd21c06b1d334
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5-p6.yaml
@@ -0,0 +1,56 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors: 3 # AutoAnchor evolves 3 anchors per P output layer
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [768, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [768]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P6/64
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 11
+ ]
+
+# YOLOv5 v6.0 head with (P3, P4, P5, P6) outputs
+head:
+ [[-1, 1, Conv, [768, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P5
+ [-1, 3, C3, [768, False]], # 15
+
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 19
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 23 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 20], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 26 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 16], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [768, False]], # 29 (P5/32-large)
+
+ [-1, 1, Conv, [768, 3, 2]],
+ [[-1, 12], 1, Concat, [1]], # cat head P6
+ [-1, 3, C3, [1024, False]], # 32 (P6/64-xlarge)
+
+ [[23, 26, 29, 32], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5, P6)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5-p7.yaml b/vouchervision/component_detector/models/hub/yolov5-p7.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..edd7d13a34a6c40e94d900ecce8ca64ae11bf5a1
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5-p7.yaml
@@ -0,0 +1,67 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors: 3 # AutoAnchor evolves 3 anchors per P output layer
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [768, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [768]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P6/64
+ [-1, 3, C3, [1024]],
+ [-1, 1, Conv, [1280, 3, 2]], # 11-P7/128
+ [-1, 3, C3, [1280]],
+ [-1, 1, SPPF, [1280, 5]], # 13
+ ]
+
+# YOLOv5 v6.0 head with (P3, P4, P5, P6, P7) outputs
+head:
+ [[-1, 1, Conv, [1024, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 10], 1, Concat, [1]], # cat backbone P6
+ [-1, 3, C3, [1024, False]], # 17
+
+ [-1, 1, Conv, [768, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P5
+ [-1, 3, C3, [768, False]], # 21
+
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 25
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 29 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 26], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 32 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 22], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [768, False]], # 35 (P5/32-large)
+
+ [-1, 1, Conv, [768, 3, 2]],
+ [[-1, 18], 1, Concat, [1]], # cat head P6
+ [-1, 3, C3, [1024, False]], # 38 (P6/64-xlarge)
+
+ [-1, 1, Conv, [1024, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P7
+ [-1, 3, C3, [1280, False]], # 41 (P7/128-xxlarge)
+
+ [[29, 32, 35, 38, 41], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5, P6, P7)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5-panet.yaml b/vouchervision/component_detector/models/hub/yolov5-panet.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..ccfbf900691c5738b4705d2ce7944171b6152c98
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5-panet.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 PANet head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5l6.yaml b/vouchervision/component_detector/models/hub/yolov5l6.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..632c2cb699e3cf261da462ec7dd20c0ffb7aaad3
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5l6.yaml
@@ -0,0 +1,60 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [19,27, 44,40, 38,94] # P3/8
+ - [96,68, 86,152, 180,137] # P4/16
+ - [140,301, 303,264, 238,542] # P5/32
+ - [436,615, 739,380, 925,792] # P6/64
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [768, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [768]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P6/64
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 11
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [768, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P5
+ [-1, 3, C3, [768, False]], # 15
+
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 19
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 23 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 20], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 26 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 16], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [768, False]], # 29 (P5/32-large)
+
+ [-1, 1, Conv, [768, 3, 2]],
+ [[-1, 12], 1, Concat, [1]], # cat head P6
+ [-1, 3, C3, [1024, False]], # 32 (P6/64-xlarge)
+
+ [[23, 26, 29, 32], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5, P6)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5m6.yaml b/vouchervision/component_detector/models/hub/yolov5m6.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..ecc53fd68ba6421b4fe63d6693b6563ecaa0e981
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5m6.yaml
@@ -0,0 +1,60 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.67 # model depth multiple
+width_multiple: 0.75 # layer channel multiple
+anchors:
+ - [19,27, 44,40, 38,94] # P3/8
+ - [96,68, 86,152, 180,137] # P4/16
+ - [140,301, 303,264, 238,542] # P5/32
+ - [436,615, 739,380, 925,792] # P6/64
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [768, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [768]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P6/64
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 11
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [768, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P5
+ [-1, 3, C3, [768, False]], # 15
+
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 19
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 23 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 20], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 26 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 16], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [768, False]], # 29 (P5/32-large)
+
+ [-1, 1, Conv, [768, 3, 2]],
+ [[-1, 12], 1, Concat, [1]], # cat head P6
+ [-1, 3, C3, [1024, False]], # 32 (P6/64-xlarge)
+
+ [[23, 26, 29, 32], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5, P6)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5n6.yaml b/vouchervision/component_detector/models/hub/yolov5n6.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..0c0c71d32551789d57e5f44fd936636ecb4e3414
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5n6.yaml
@@ -0,0 +1,60 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.33 # model depth multiple
+width_multiple: 0.25 # layer channel multiple
+anchors:
+ - [19,27, 44,40, 38,94] # P3/8
+ - [96,68, 86,152, 180,137] # P4/16
+ - [140,301, 303,264, 238,542] # P5/32
+ - [436,615, 739,380, 925,792] # P6/64
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [768, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [768]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P6/64
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 11
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [768, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P5
+ [-1, 3, C3, [768, False]], # 15
+
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 19
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 23 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 20], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 26 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 16], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [768, False]], # 29 (P5/32-large)
+
+ [-1, 1, Conv, [768, 3, 2]],
+ [[-1, 12], 1, Concat, [1]], # cat head P6
+ [-1, 3, C3, [1024, False]], # 32 (P6/64-xlarge)
+
+ [[23, 26, 29, 32], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5, P6)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5s-ghost.yaml b/vouchervision/component_detector/models/hub/yolov5s-ghost.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..ff9519c3f1aa354f512ddab8b23e861d0f3de6c6
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5s-ghost.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.33 # model depth multiple
+width_multiple: 0.50 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, GhostConv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3Ghost, [128]],
+ [-1, 1, GhostConv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3Ghost, [256]],
+ [-1, 1, GhostConv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3Ghost, [512]],
+ [-1, 1, GhostConv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3Ghost, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, GhostConv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3Ghost, [512, False]], # 13
+
+ [-1, 1, GhostConv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3Ghost, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, GhostConv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3Ghost, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, GhostConv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3Ghost, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5s-transformer.yaml b/vouchervision/component_detector/models/hub/yolov5s-transformer.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..100d7c447527f1116e0edb3e1c096904fe3302f1
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5s-transformer.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.33 # model depth multiple
+width_multiple: 0.50 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3TR, [1024]], # 9 <--- C3TR() Transformer module
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5s6.yaml b/vouchervision/component_detector/models/hub/yolov5s6.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..a28fb559482b25a41531517a68f08253f08edb0f
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5s6.yaml
@@ -0,0 +1,60 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.33 # model depth multiple
+width_multiple: 0.50 # layer channel multiple
+anchors:
+ - [19,27, 44,40, 38,94] # P3/8
+ - [96,68, 86,152, 180,137] # P4/16
+ - [140,301, 303,264, 238,542] # P5/32
+ - [436,615, 739,380, 925,792] # P6/64
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [768, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [768]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P6/64
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 11
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [768, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P5
+ [-1, 3, C3, [768, False]], # 15
+
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 19
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 23 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 20], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 26 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 16], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [768, False]], # 29 (P5/32-large)
+
+ [-1, 1, Conv, [768, 3, 2]],
+ [[-1, 12], 1, Concat, [1]], # cat head P6
+ [-1, 3, C3, [1024, False]], # 32 (P6/64-xlarge)
+
+ [[23, 26, 29, 32], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5, P6)
+ ]
diff --git a/vouchervision/component_detector/models/hub/yolov5x6.yaml b/vouchervision/component_detector/models/hub/yolov5x6.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..ba795c4aad319b94db0fb4fd6961e9ef0cac207a
--- /dev/null
+++ b/vouchervision/component_detector/models/hub/yolov5x6.yaml
@@ -0,0 +1,60 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.33 # model depth multiple
+width_multiple: 1.25 # layer channel multiple
+anchors:
+ - [19,27, 44,40, 38,94] # P3/8
+ - [96,68, 86,152, 180,137] # P4/16
+ - [140,301, 303,264, 238,542] # P5/32
+ - [436,615, 739,380, 925,792] # P6/64
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [768, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [768]],
+ [-1, 1, Conv, [1024, 3, 2]], # 9-P6/64
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 11
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [768, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 8], 1, Concat, [1]], # cat backbone P5
+ [-1, 3, C3, [768, False]], # 15
+
+ [-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 19
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 23 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 20], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 26 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 16], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [768, False]], # 29 (P5/32-large)
+
+ [-1, 1, Conv, [768, 3, 2]],
+ [[-1, 12], 1, Concat, [1]], # cat head P6
+ [-1, 3, C3, [1024, False]], # 32 (P6/64-xlarge)
+
+ [[23, 26, 29, 32], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5, P6)
+ ]
diff --git a/vouchervision/component_detector/models/tf.py b/vouchervision/component_detector/models/tf.py
new file mode 100644
index 0000000000000000000000000000000000000000..04b1cd378f18e7c7dedc3abad512a18098a13025
--- /dev/null
+++ b/vouchervision/component_detector/models/tf.py
@@ -0,0 +1,498 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+TensorFlow, Keras and TFLite versions of YOLOv5
+Authored by https://github.com/zldrobit in PR https://github.com/ultralytics/yolov5/pull/1127
+
+Usage:
+ $ python models/tf.py --weights yolov5s.pt
+
+Export:
+ $ python path/to/export.py --weights yolov5s.pt --include saved_model pb tflite tfjs
+"""
+
+import argparse
+import sys
+from copy import deepcopy
+from pathlib import Path
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+# ROOT = ROOT.relative_to(Path.cwd()) # relative
+
+import numpy as np
+import tensorflow as tf
+import torch
+import torch.nn as nn
+from tensorflow import keras
+
+from models.common import C3, SPP, SPPF, Bottleneck, BottleneckCSP, Concat, Conv, DWConv, Focus, autopad
+from models.experimental import CrossConv, MixConv2d, attempt_load
+from models.yolo import Detect
+from utils.activations import SiLU
+from utils.general import LOGGER, make_divisible, print_args
+
+
+class TFBN(keras.layers.Layer):
+ # TensorFlow BatchNormalization wrapper
+ def __init__(self, w=None):
+ super().__init__()
+ self.bn = keras.layers.BatchNormalization(
+ beta_initializer=keras.initializers.Constant(w.bias.numpy()),
+ gamma_initializer=keras.initializers.Constant(w.weight.numpy()),
+ moving_mean_initializer=keras.initializers.Constant(w.running_mean.numpy()),
+ moving_variance_initializer=keras.initializers.Constant(w.running_var.numpy()),
+ epsilon=w.eps)
+
+ def call(self, inputs):
+ return self.bn(inputs)
+
+
+class TFPad(keras.layers.Layer):
+
+ def __init__(self, pad):
+ super().__init__()
+ self.pad = tf.constant([[0, 0], [pad, pad], [pad, pad], [0, 0]])
+
+ def call(self, inputs):
+ return tf.pad(inputs, self.pad, mode='constant', constant_values=0)
+
+
+class TFConv(keras.layers.Layer):
+ # Standard convolution
+ def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None):
+ # ch_in, ch_out, weights, kernel, stride, padding, groups
+ super().__init__()
+ assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument"
+ assert isinstance(k, int), "Convolution with multiple kernels are not allowed."
+ # TensorFlow convolution padding is inconsistent with PyTorch (e.g. k=3 s=2 'SAME' padding)
+ # see https://stackoverflow.com/questions/52975843/comparing-conv2d-with-padding-between-tensorflow-and-pytorch
+
+ conv = keras.layers.Conv2D(
+ c2,
+ k,
+ s,
+ 'SAME' if s == 1 else 'VALID',
+ use_bias=False if hasattr(w, 'bn') else True,
+ kernel_initializer=keras.initializers.Constant(w.conv.weight.permute(2, 3, 1, 0).numpy()),
+ bias_initializer='zeros' if hasattr(w, 'bn') else keras.initializers.Constant(w.conv.bias.numpy()))
+ self.conv = conv if s == 1 else keras.Sequential([TFPad(autopad(k, p)), conv])
+ self.bn = TFBN(w.bn) if hasattr(w, 'bn') else tf.identity
+
+ # YOLOv5 activations
+ if isinstance(w.act, nn.LeakyReLU):
+ self.act = (lambda x: keras.activations.relu(x, alpha=0.1)) if act else tf.identity
+ elif isinstance(w.act, nn.Hardswish):
+ self.act = (lambda x: x * tf.nn.relu6(x + 3) * 0.166666667) if act else tf.identity
+ elif isinstance(w.act, (nn.SiLU, SiLU)):
+ self.act = (lambda x: keras.activations.swish(x)) if act else tf.identity
+ else:
+ raise Exception(f'no matching TensorFlow activation found for {w.act}')
+
+ def call(self, inputs):
+ return self.act(self.bn(self.conv(inputs)))
+
+
+class TFFocus(keras.layers.Layer):
+ # Focus wh information into c-space
+ def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None):
+ # ch_in, ch_out, kernel, stride, padding, groups
+ super().__init__()
+ self.conv = TFConv(c1 * 4, c2, k, s, p, g, act, w.conv)
+
+ def call(self, inputs): # x(b,w,h,c) -> y(b,w/2,h/2,4c)
+ # inputs = inputs / 255 # normalize 0-255 to 0-1
+ return self.conv(
+ tf.concat(
+ [inputs[:, ::2, ::2, :], inputs[:, 1::2, ::2, :], inputs[:, ::2, 1::2, :], inputs[:, 1::2, 1::2, :]],
+ 3))
+
+
+class TFBottleneck(keras.layers.Layer):
+ # Standard bottleneck
+ def __init__(self, c1, c2, shortcut=True, g=1, e=0.5, w=None): # ch_in, ch_out, shortcut, groups, expansion
+ super().__init__()
+ c_ = int(c2 * e) # hidden channels
+ self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+ self.cv2 = TFConv(c_, c2, 3, 1, g=g, w=w.cv2)
+ self.add = shortcut and c1 == c2
+
+ def call(self, inputs):
+ return inputs + self.cv2(self.cv1(inputs)) if self.add else self.cv2(self.cv1(inputs))
+
+
+class TFConv2d(keras.layers.Layer):
+ # Substitution for PyTorch nn.Conv2D
+ def __init__(self, c1, c2, k, s=1, g=1, bias=True, w=None):
+ super().__init__()
+ assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument"
+ self.conv = keras.layers.Conv2D(
+ c2,
+ k,
+ s,
+ 'VALID',
+ use_bias=bias,
+ kernel_initializer=keras.initializers.Constant(w.weight.permute(2, 3, 1, 0).numpy()),
+ bias_initializer=keras.initializers.Constant(w.bias.numpy()) if bias else None,
+ )
+
+ def call(self, inputs):
+ return self.conv(inputs)
+
+
+class TFBottleneckCSP(keras.layers.Layer):
+ # CSP Bottleneck https://github.com/WongKinYiu/CrossStagePartialNetworks
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None):
+ # ch_in, ch_out, number, shortcut, groups, expansion
+ super().__init__()
+ c_ = int(c2 * e) # hidden channels
+ self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+ self.cv2 = TFConv2d(c1, c_, 1, 1, bias=False, w=w.cv2)
+ self.cv3 = TFConv2d(c_, c_, 1, 1, bias=False, w=w.cv3)
+ self.cv4 = TFConv(2 * c_, c2, 1, 1, w=w.cv4)
+ self.bn = TFBN(w.bn)
+ self.act = lambda x: keras.activations.swish(x)
+ self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)])
+
+ def call(self, inputs):
+ y1 = self.cv3(self.m(self.cv1(inputs)))
+ y2 = self.cv2(inputs)
+ return self.cv4(self.act(self.bn(tf.concat((y1, y2), axis=3))))
+
+
+class TFC3(keras.layers.Layer):
+ # CSP Bottleneck with 3 convolutions
+ def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None):
+ # ch_in, ch_out, number, shortcut, groups, expansion
+ super().__init__()
+ c_ = int(c2 * e) # hidden channels
+ self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+ self.cv2 = TFConv(c1, c_, 1, 1, w=w.cv2)
+ self.cv3 = TFConv(2 * c_, c2, 1, 1, w=w.cv3)
+ self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)])
+
+ def call(self, inputs):
+ return self.cv3(tf.concat((self.m(self.cv1(inputs)), self.cv2(inputs)), axis=3))
+
+
+class TFSPP(keras.layers.Layer):
+ # Spatial pyramid pooling layer used in YOLOv3-SPP
+ def __init__(self, c1, c2, k=(5, 9, 13), w=None):
+ super().__init__()
+ c_ = c1 // 2 # hidden channels
+ self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+ self.cv2 = TFConv(c_ * (len(k) + 1), c2, 1, 1, w=w.cv2)
+ self.m = [keras.layers.MaxPool2D(pool_size=x, strides=1, padding='SAME') for x in k]
+
+ def call(self, inputs):
+ x = self.cv1(inputs)
+ return self.cv2(tf.concat([x] + [m(x) for m in self.m], 3))
+
+
+class TFSPPF(keras.layers.Layer):
+ # Spatial pyramid pooling-Fast layer
+ def __init__(self, c1, c2, k=5, w=None):
+ super().__init__()
+ c_ = c1 // 2 # hidden channels
+ self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+ self.cv2 = TFConv(c_ * 4, c2, 1, 1, w=w.cv2)
+ self.m = keras.layers.MaxPool2D(pool_size=k, strides=1, padding='SAME')
+
+ def call(self, inputs):
+ x = self.cv1(inputs)
+ y1 = self.m(x)
+ y2 = self.m(y1)
+ return self.cv2(tf.concat([x, y1, y2, self.m(y2)], 3))
+
+
+class TFDetect(keras.layers.Layer):
+ # TF YOLOv5 Detect layer
+ def __init__(self, nc=80, anchors=(), ch=(), imgsz=(640, 640), w=None): # detection layer
+ super().__init__()
+ self.stride = tf.convert_to_tensor(w.stride.numpy(), dtype=tf.float32)
+ self.nc = nc # number of classes
+ self.no = nc + 5 # number of outputs per anchor
+ self.nl = len(anchors) # number of detection layers
+ self.na = len(anchors[0]) // 2 # number of anchors
+ self.grid = [tf.zeros(1)] * self.nl # init grid
+ self.anchors = tf.convert_to_tensor(w.anchors.numpy(), dtype=tf.float32)
+ self.anchor_grid = tf.reshape(self.anchors * tf.reshape(self.stride, [self.nl, 1, 1]), [self.nl, 1, -1, 1, 2])
+ self.m = [TFConv2d(x, self.no * self.na, 1, w=w.m[i]) for i, x in enumerate(ch)]
+ self.training = False # set to False after building model
+ self.imgsz = imgsz
+ for i in range(self.nl):
+ ny, nx = self.imgsz[0] // self.stride[i], self.imgsz[1] // self.stride[i]
+ self.grid[i] = self._make_grid(nx, ny)
+
+ def call(self, inputs):
+ z = [] # inference output
+ x = []
+ for i in range(self.nl):
+ x.append(self.m[i](inputs[i]))
+ # x(bs,20,20,255) to x(bs,3,20,20,85)
+ ny, nx = self.imgsz[0] // self.stride[i], self.imgsz[1] // self.stride[i]
+ x[i] = tf.reshape(x[i], [-1, ny * nx, self.na, self.no])
+
+ if not self.training: # inference
+ y = tf.sigmoid(x[i])
+ grid = tf.transpose(self.grid[i], [0, 2, 1, 3]) - 0.5
+ anchor_grid = tf.transpose(self.anchor_grid[i], [0, 2, 1, 3]) * 4
+ xy = (y[..., 0:2] * 2 + grid) * self.stride[i] # xy
+ wh = y[..., 2:4] ** 2 * anchor_grid
+ # Normalize xywh to 0-1 to reduce calibration error
+ xy /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32)
+ wh /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32)
+ y = tf.concat([xy, wh, y[..., 4:]], -1)
+ z.append(tf.reshape(y, [-1, self.na * ny * nx, self.no]))
+
+ return tf.transpose(x, [0, 2, 1, 3]) if self.training else (tf.concat(z, 1), x)
+
+ @staticmethod
+ def _make_grid(nx=20, ny=20):
+ # yv, xv = torch.meshgrid([torch.arange(ny), torch.arange(nx)])
+ # return torch.stack((xv, yv), 2).view((1, 1, ny, nx, 2)).float()
+ xv, yv = tf.meshgrid(tf.range(nx), tf.range(ny))
+ return tf.cast(tf.reshape(tf.stack([xv, yv], 2), [1, 1, ny * nx, 2]), dtype=tf.float32)
+
+
+class TFUpsample(keras.layers.Layer):
+ # TF version of torch.nn.Upsample()
+ def __init__(self, size, scale_factor, mode, w=None): # warning: all arguments needed including 'w'
+ super().__init__()
+ assert scale_factor == 2, "scale_factor must be 2"
+ self.upsample = lambda x: tf.image.resize(x, (x.shape[1] * 2, x.shape[2] * 2), method=mode)
+ # self.upsample = keras.layers.UpSampling2D(size=scale_factor, interpolation=mode)
+ # with default arguments: align_corners=False, half_pixel_centers=False
+ # self.upsample = lambda x: tf.raw_ops.ResizeNearestNeighbor(images=x,
+ # size=(x.shape[1] * 2, x.shape[2] * 2))
+
+ def call(self, inputs):
+ return self.upsample(inputs)
+
+
+class TFConcat(keras.layers.Layer):
+ # TF version of torch.concat()
+ def __init__(self, dimension=1, w=None):
+ super().__init__()
+ assert dimension == 1, "convert only NCHW to NHWC concat"
+ self.d = 3
+
+ def call(self, inputs):
+ return tf.concat(inputs, self.d)
+
+
+def parse_model(d, ch, model, imgsz): # model_dict, input_channels(3)
+ LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}")
+ anchors, nc, gd, gw = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple']
+ na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors # number of anchors
+ no = na * (nc + 5) # number of outputs = anchors * (classes + 5)
+
+ layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out
+ for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']): # from, number, module, args
+ m_str = m
+ m = eval(m) if isinstance(m, str) else m # eval strings
+ for j, a in enumerate(args):
+ try:
+ args[j] = eval(a) if isinstance(a, str) else a # eval strings
+ except NameError:
+ pass
+
+ n = max(round(n * gd), 1) if n > 1 else n # depth gain
+ if m in [nn.Conv2d, Conv, Bottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv, BottleneckCSP, C3]:
+ c1, c2 = ch[f], args[0]
+ c2 = make_divisible(c2 * gw, 8) if c2 != no else c2
+
+ args = [c1, c2, *args[1:]]
+ if m in [BottleneckCSP, C3]:
+ args.insert(2, n)
+ n = 1
+ elif m is nn.BatchNorm2d:
+ args = [ch[f]]
+ elif m is Concat:
+ c2 = sum(ch[-1 if x == -1 else x + 1] for x in f)
+ elif m is Detect:
+ args.append([ch[x + 1] for x in f])
+ if isinstance(args[1], int): # number of anchors
+ args[1] = [list(range(args[1] * 2))] * len(f)
+ args.append(imgsz)
+ else:
+ c2 = ch[f]
+
+ tf_m = eval('TF' + m_str.replace('nn.', ''))
+ m_ = keras.Sequential([tf_m(*args, w=model.model[i][j]) for j in range(n)]) if n > 1 \
+ else tf_m(*args, w=model.model[i]) # module
+
+ torch_m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args) # module
+ t = str(m)[8:-2].replace('__main__.', '') # module type
+ np = sum(x.numel() for x in torch_m_.parameters()) # number params
+ m_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number params
+ LOGGER.info(f'{i:>3}{str(f):>18}{str(n):>3}{np:>10} {t:<40}{str(args):<30}') # print
+ save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist
+ layers.append(m_)
+ ch.append(c2)
+ return keras.Sequential(layers), sorted(save)
+
+
+class TFModel:
+ # TF YOLOv5 model
+ def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, model=None, imgsz=(640, 640)): # model, channels, classes
+ super().__init__()
+ if isinstance(cfg, dict):
+ self.yaml = cfg # model dict
+ else: # is *.yaml
+ import yaml # for torch hub
+ self.yaml_file = Path(cfg).name
+ with open(cfg) as f:
+ self.yaml = yaml.load(f, Loader=yaml.FullLoader) # model dict
+
+ # Define model
+ if nc and nc != self.yaml['nc']:
+ LOGGER.info(f"Overriding {cfg} nc={self.yaml['nc']} with nc={nc}")
+ self.yaml['nc'] = nc # override yaml value
+ self.model, self.savelist = parse_model(deepcopy(self.yaml), ch=[ch], model=model, imgsz=imgsz)
+
+ def predict(self,
+ inputs,
+ tf_nms=False,
+ agnostic_nms=False,
+ topk_per_class=100,
+ topk_all=100,
+ iou_thres=0.45,
+ conf_thres=0.25):
+ y = [] # outputs
+ x = inputs
+ for i, m in enumerate(self.model.layers):
+ if m.f != -1: # if not from previous layer
+ x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers
+
+ x = m(x) # run
+ y.append(x if m.i in self.savelist else None) # save output
+
+ # Add TensorFlow NMS
+ if tf_nms:
+ boxes = self._xywh2xyxy(x[0][..., :4])
+ probs = x[0][:, :, 4:5]
+ classes = x[0][:, :, 5:]
+ scores = probs * classes
+ if agnostic_nms:
+ nms = AgnosticNMS()((boxes, classes, scores), topk_all, iou_thres, conf_thres)
+ return nms, x[1]
+ else:
+ boxes = tf.expand_dims(boxes, 2)
+ nms = tf.image.combined_non_max_suppression(boxes,
+ scores,
+ topk_per_class,
+ topk_all,
+ iou_thres,
+ conf_thres,
+ clip_boxes=False)
+ return nms, x[1]
+
+ return x[0] # output only first tensor [1,6300,85] = [xywh, conf, class0, class1, ...]
+ # x = x[0][0] # [x(1,6300,85), ...] to x(6300,85)
+ # xywh = x[..., :4] # x(6300,4) boxes
+ # conf = x[..., 4:5] # x(6300,1) confidences
+ # cls = tf.reshape(tf.cast(tf.argmax(x[..., 5:], axis=1), tf.float32), (-1, 1)) # x(6300,1) classes
+ # return tf.concat([conf, cls, xywh], 1)
+
+ @staticmethod
+ def _xywh2xyxy(xywh):
+ # Convert nx4 boxes from [x, y, w, h] to [x1, y1, x2, y2] where xy1=top-left, xy2=bottom-right
+ x, y, w, h = tf.split(xywh, num_or_size_splits=4, axis=-1)
+ return tf.concat([x - w / 2, y - h / 2, x + w / 2, y + h / 2], axis=-1)
+
+
+class AgnosticNMS(keras.layers.Layer):
+ # TF Agnostic NMS
+ def call(self, input, topk_all, iou_thres, conf_thres):
+ # wrap map_fn to avoid TypeSpec related error https://stackoverflow.com/a/65809989/3036450
+ return tf.map_fn(lambda x: self._nms(x, topk_all, iou_thres, conf_thres),
+ input,
+ fn_output_signature=(tf.float32, tf.float32, tf.float32, tf.int32),
+ name='agnostic_nms')
+
+ @staticmethod
+ def _nms(x, topk_all=100, iou_thres=0.45, conf_thres=0.25): # agnostic NMS
+ boxes, classes, scores = x
+ class_inds = tf.cast(tf.argmax(classes, axis=-1), tf.float32)
+ scores_inp = tf.reduce_max(scores, -1)
+ selected_inds = tf.image.non_max_suppression(boxes,
+ scores_inp,
+ max_output_size=topk_all,
+ iou_threshold=iou_thres,
+ score_threshold=conf_thres)
+ selected_boxes = tf.gather(boxes, selected_inds)
+ padded_boxes = tf.pad(selected_boxes,
+ paddings=[[0, topk_all - tf.shape(selected_boxes)[0]], [0, 0]],
+ mode="CONSTANT",
+ constant_values=0.0)
+ selected_scores = tf.gather(scores_inp, selected_inds)
+ padded_scores = tf.pad(selected_scores,
+ paddings=[[0, topk_all - tf.shape(selected_boxes)[0]]],
+ mode="CONSTANT",
+ constant_values=-1.0)
+ selected_classes = tf.gather(class_inds, selected_inds)
+ padded_classes = tf.pad(selected_classes,
+ paddings=[[0, topk_all - tf.shape(selected_boxes)[0]]],
+ mode="CONSTANT",
+ constant_values=-1.0)
+ valid_detections = tf.shape(selected_inds)[0]
+ return padded_boxes, padded_scores, padded_classes, valid_detections
+
+
+def representative_dataset_gen(dataset, ncalib=100):
+ # Representative dataset generator for use with converter.representative_dataset, returns a generator of np arrays
+ for n, (path, img, im0s, vid_cap, string) in enumerate(dataset):
+ input = np.transpose(img, [1, 2, 0])
+ input = np.expand_dims(input, axis=0).astype(np.float32)
+ input /= 255
+ yield [input]
+ if n >= ncalib:
+ break
+
+
+def run(
+ weights=ROOT / 'yolov5s.pt', # weights path
+ imgsz=(640, 640), # inference size h,w
+ batch_size=1, # batch size
+ dynamic=False, # dynamic batch size
+):
+ # PyTorch model
+ im = torch.zeros((batch_size, 3, *imgsz)) # BCHW image
+ model = attempt_load(weights, map_location=torch.device('cpu'), inplace=True, fuse=False)
+ _ = model(im) # inference
+ model.info()
+
+ # TensorFlow model
+ im = tf.zeros((batch_size, *imgsz, 3)) # BHWC image
+ tf_model = TFModel(cfg=model.yaml, model=model, nc=model.nc, imgsz=imgsz)
+ _ = tf_model.predict(im) # inference
+
+ # Keras model
+ im = keras.Input(shape=(*imgsz, 3), batch_size=None if dynamic else batch_size)
+ keras_model = keras.Model(inputs=im, outputs=tf_model.predict(im))
+ keras_model.summary()
+
+ LOGGER.info('PyTorch, TensorFlow and Keras models successfully verified.\nUse export.py for TF model export.')
+
+
+def parse_opt():
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--weights', type=str, default=ROOT / 'yolov5s.pt', help='weights path')
+ parser.add_argument('--imgsz', '--img', '--img-size', nargs='+', type=int, default=[640], help='inference size h,w')
+ parser.add_argument('--batch-size', type=int, default=1, help='batch size')
+ parser.add_argument('--dynamic', action='store_true', help='dynamic batch size')
+ opt = parser.parse_args()
+ opt.imgsz *= 2 if len(opt.imgsz) == 1 else 1 # expand
+ print_args(vars(opt))
+ return opt
+
+
+def main(opt):
+ run(**vars(opt))
+
+
+if __name__ == "__main__":
+ opt = parse_opt()
+ main(opt)
diff --git a/vouchervision/component_detector/models/yolo.py b/vouchervision/component_detector/models/yolo.py
new file mode 100644
index 0000000000000000000000000000000000000000..f072aeeb8eacbe69730ca28db7e738c86345e0ed
--- /dev/null
+++ b/vouchervision/component_detector/models/yolo.py
@@ -0,0 +1,335 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+YOLO-specific modules
+
+Usage:
+ $ python path/to/models/yolo.py --cfg yolov5s.yaml
+"""
+
+import argparse
+import os
+import platform
+import sys
+from copy import deepcopy
+from pathlib import Path
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+if platform.system() != 'Windows':
+ ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+
+from models.common import *
+from models.experimental import *
+from utils.autoanchor import check_anchor_order
+from utils.general import LOGGER, check_version, check_yaml, make_divisible, print_args
+from utils.plots import feature_visualization
+from utils.torch_utils import (fuse_conv_and_bn, initialize_weights, model_info, profile, scale_img, select_device,
+ time_sync)
+
+try:
+ import thop # for FLOPs computation
+except ImportError:
+ thop = None
+
+
+class Detect(nn.Module):
+ stride = None # strides computed during build
+ onnx_dynamic = False # ONNX export parameter
+ export = False # export mode
+
+ def __init__(self, nc=80, anchors=(), ch=(), inplace=True): # detection layer
+ super().__init__()
+ self.nc = nc # number of classes
+ self.no = nc + 5 # number of outputs per anchor
+ self.nl = len(anchors) # number of detection layers
+ self.na = len(anchors[0]) // 2 # number of anchors
+ self.grid = [torch.zeros(1)] * self.nl # init grid
+ self.anchor_grid = [torch.zeros(1)] * self.nl # init anchor grid
+ self.register_buffer('anchors', torch.tensor(anchors).float().view(self.nl, -1, 2)) # shape(nl,na,2)
+ self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch) # output conv
+ self.inplace = inplace # use in-place ops (e.g. slice assignment)
+
+ def forward(self, x):
+ z = [] # inference output
+ for i in range(self.nl):
+ x[i] = self.m[i](x[i]) # conv
+ bs, _, ny, nx = x[i].shape # x(bs,255,20,20) to x(bs,3,20,20,85)
+ x[i] = x[i].view(bs, self.na, self.no, ny, nx).permute(0, 1, 3, 4, 2).contiguous()
+
+ if not self.training: # inference
+ if self.onnx_dynamic or self.grid[i].shape[2:4] != x[i].shape[2:4]:
+ self.grid[i], self.anchor_grid[i] = self._make_grid(nx, ny, i)
+
+ y = x[i].sigmoid()
+ if self.inplace:
+ y[..., 0:2] = (y[..., 0:2] * 2 + self.grid[i]) * self.stride[i] # xy
+ y[..., 2:4] = (y[..., 2:4] * 2) ** 2 * self.anchor_grid[i] # wh
+ else: # for YOLOv5 on AWS Inferentia https://github.com/ultralytics/yolov5/pull/2953
+ xy, wh, conf = y.split((2, 2, self.nc + 1), 4) # y.tensor_split((2, 4, 5), 4) # torch 1.8.0
+ xy = (xy * 2 + self.grid[i]) * self.stride[i] # xy
+ wh = (wh * 2) ** 2 * self.anchor_grid[i] # wh
+ y = torch.cat((xy, wh, conf), 4)
+ z.append(y.view(bs, -1, self.no))
+
+ return x if self.training else (torch.cat(z, 1),) if self.export else (torch.cat(z, 1), x)
+
+ def _make_grid(self, nx=20, ny=20, i=0):
+ d = self.anchors[i].device
+ t = self.anchors[i].dtype
+ shape = 1, self.na, ny, nx, 2 # grid shape
+ y, x = torch.arange(ny, device=d, dtype=t), torch.arange(nx, device=d, dtype=t)
+ if check_version(torch.__version__, '1.10.0'): # torch>=1.10.0 meshgrid workaround for torch>=0.7 compatibility
+ yv, xv = torch.meshgrid(y, x, indexing='ij')
+ else:
+ yv, xv = torch.meshgrid(y, x)
+ grid = torch.stack((xv, yv), 2).expand(shape) - 0.5 # add grid offset, i.e. y = 2.0 * x - 0.5
+ anchor_grid = (self.anchors[i] * self.stride[i]).view((1, self.na, 1, 1, 2)).expand(shape)
+ return grid, anchor_grid
+
+
+class Model(nn.Module):
+ # YOLOv5 model
+ def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, anchors=None): # model, input channels, number of classes
+ super().__init__()
+ if isinstance(cfg, dict):
+ self.yaml = cfg # model dict
+ else: # is *.yaml
+ import yaml # for torch hub
+ self.yaml_file = Path(cfg).name
+ with open(cfg, encoding='ascii', errors='ignore') as f:
+ self.yaml = yaml.safe_load(f) # model dict
+
+ # Define model
+ ch = self.yaml['ch'] = self.yaml.get('ch', ch) # input channels
+ if nc and nc != self.yaml['nc']:
+ LOGGER.info(f"Overriding model.yaml nc={self.yaml['nc']} with nc={nc}")
+ self.yaml['nc'] = nc # override yaml value
+ if anchors:
+ LOGGER.info(f'Overriding model.yaml anchors with anchors={anchors}')
+ self.yaml['anchors'] = round(anchors) # override yaml value
+ self.model, self.save = parse_model(deepcopy(self.yaml), ch=[ch]) # model, savelist
+ self.names = [str(i) for i in range(self.yaml['nc'])] # default names
+ self.inplace = self.yaml.get('inplace', True)
+
+ # Build strides, anchors
+ m = self.model[-1] # Detect()
+ if isinstance(m, Detect):
+ s = 256 # 2x min stride
+ m.inplace = self.inplace
+ m.stride = torch.tensor([s / x.shape[-2] for x in self.forward(torch.zeros(1, ch, s, s))]) # forward
+ check_anchor_order(m) # must be in pixel-space (not grid-space)
+ m.anchors /= m.stride.view(-1, 1, 1)
+ self.stride = m.stride
+ self._initialize_biases() # only run once
+
+ # Init weights, biases
+ initialize_weights(self)
+ self.info()
+ LOGGER.info('')
+
+ def forward(self, x, augment=False, profile=False, visualize=False):
+ if augment:
+ return self._forward_augment(x) # augmented inference, None
+ return self._forward_once(x, profile, visualize) # single-scale inference, train
+
+ def _forward_augment(self, x):
+ img_size = x.shape[-2:] # height, width
+ s = [1, 0.83, 0.67] # scales
+ f = [None, 3, None] # flips (2-ud, 3-lr)
+ y = [] # outputs
+ for si, fi in zip(s, f):
+ xi = scale_img(x.flip(fi) if fi else x, si, gs=int(self.stride.max()))
+ yi = self._forward_once(xi)[0] # forward
+ # cv2.imwrite(f'img_{si}.jpg', 255 * xi[0].cpu().numpy().transpose((1, 2, 0))[:, :, ::-1]) # save
+ yi = self._descale_pred(yi, fi, si, img_size)
+ y.append(yi)
+ y = self._clip_augmented(y) # clip augmented tails
+ return torch.cat(y, 1), None # augmented inference, train
+
+ def _forward_once(self, x, profile=False, visualize=False):
+ y, dt = [], [] # outputs
+ for m in self.model:
+ if m.f != -1: # if not from previous layer
+ x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers
+ if profile:
+ self._profile_one_layer(m, x, dt)
+ x = m(x) # run
+ y.append(x if m.i in self.save else None) # save output
+ if visualize:
+ feature_visualization(x, m.type, m.i, save_dir=visualize)
+ return x
+
+ def _descale_pred(self, p, flips, scale, img_size):
+ # de-scale predictions following augmented inference (inverse operation)
+ if self.inplace:
+ p[..., :4] /= scale # de-scale
+ if flips == 2:
+ p[..., 1] = img_size[0] - p[..., 1] # de-flip ud
+ elif flips == 3:
+ p[..., 0] = img_size[1] - p[..., 0] # de-flip lr
+ else:
+ x, y, wh = p[..., 0:1] / scale, p[..., 1:2] / scale, p[..., 2:4] / scale # de-scale
+ if flips == 2:
+ y = img_size[0] - y # de-flip ud
+ elif flips == 3:
+ x = img_size[1] - x # de-flip lr
+ p = torch.cat((x, y, wh, p[..., 4:]), -1)
+ return p
+
+ def _clip_augmented(self, y):
+ # Clip YOLOv5 augmented inference tails
+ nl = self.model[-1].nl # number of detection layers (P3-P5)
+ g = sum(4 ** x for x in range(nl)) # grid points
+ e = 1 # exclude layer count
+ i = (y[0].shape[1] // g) * sum(4 ** x for x in range(e)) # indices
+ y[0] = y[0][:, :-i] # large
+ i = (y[-1].shape[1] // g) * sum(4 ** (nl - 1 - x) for x in range(e)) # indices
+ y[-1] = y[-1][:, i:] # small
+ return y
+
+ def _profile_one_layer(self, m, x, dt):
+ c = isinstance(m, Detect) # is final layer, copy input as inplace fix
+ o = thop.profile(m, inputs=(x.copy() if c else x,), verbose=False)[0] / 1E9 * 2 if thop else 0 # FLOPs
+ t = time_sync()
+ for _ in range(10):
+ m(x.copy() if c else x)
+ dt.append((time_sync() - t) * 100)
+ if m == self.model[0]:
+ LOGGER.info(f"{'time (ms)':>10s} {'GFLOPs':>10s} {'params':>10s} {'module'}")
+ LOGGER.info(f'{dt[-1]:10.2f} {o:10.2f} {m.np:10.0f} {m.type}')
+ if c:
+ LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s} Total")
+
+ def _initialize_biases(self, cf=None): # initialize biases into Detect(), cf is class frequency
+ # https://arxiv.org/abs/1708.02002 section 3.3
+ # cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1.
+ m = self.model[-1] # Detect() module
+ for mi, s in zip(m.m, m.stride): # from
+ b = mi.bias.view(m.na, -1) # conv.bias(255) to (3,85)
+ b.data[:, 4] += math.log(8 / (640 / s) ** 2) # obj (8 objects per 640 image)
+ b.data[:, 5:] += math.log(0.6 / (m.nc - 0.999999)) if cf is None else torch.log(cf / cf.sum()) # cls
+ mi.bias = torch.nn.Parameter(b.view(-1), requires_grad=True)
+
+ def _print_biases(self):
+ m = self.model[-1] # Detect() module
+ for mi in m.m: # from
+ b = mi.bias.detach().view(m.na, -1).T # conv.bias(255) to (3,85)
+ LOGGER.info(
+ ('%6g Conv2d.bias:' + '%10.3g' * 6) % (mi.weight.shape[1], *b[:5].mean(1).tolist(), b[5:].mean()))
+
+ # def _print_weights(self):
+ # for m in self.model.modules():
+ # if type(m) is Bottleneck:
+ # LOGGER.info('%10.3g' % (m.w.detach().sigmoid() * 2)) # shortcut weights
+
+ def fuse(self): # fuse model Conv2d() + BatchNorm2d() layers
+ LOGGER.info('Fusing layers... ')
+ for m in self.model.modules():
+ if isinstance(m, (Conv, DWConv)) and hasattr(m, 'bn'):
+ m.conv = fuse_conv_and_bn(m.conv, m.bn) # update conv
+ delattr(m, 'bn') # remove batchnorm
+ m.forward = m.forward_fuse # update forward
+ self.info()
+ return self
+
+ def info(self, verbose=False, img_size=640): # print model information
+ model_info(self, verbose, img_size)
+
+ def _apply(self, fn):
+ # Apply to(), cpu(), cuda(), half() to model tensors that are not parameters or registered buffers
+ self = super()._apply(fn)
+ m = self.model[-1] # Detect()
+ if isinstance(m, Detect):
+ m.stride = fn(m.stride)
+ m.grid = list(map(fn, m.grid))
+ if isinstance(m.anchor_grid, list):
+ m.anchor_grid = list(map(fn, m.anchor_grid))
+ return self
+
+
+def parse_model(d, ch): # model_dict, input_channels(3)
+ LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}")
+ anchors, nc, gd, gw = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple']
+ na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors # number of anchors
+ no = na * (nc + 5) # number of outputs = anchors * (classes + 5)
+
+ layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out
+ for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']): # from, number, module, args
+ m = eval(m) if isinstance(m, str) else m # eval strings
+ for j, a in enumerate(args):
+ try:
+ args[j] = eval(a) if isinstance(a, str) else a # eval strings
+ except NameError:
+ pass
+
+ n = n_ = max(round(n * gd), 1) if n > 1 else n # depth gain
+ if m in (Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv,
+ BottleneckCSP, C3, C3TR, C3SPP, C3Ghost):
+ c1, c2 = ch[f], args[0]
+ if c2 != no: # if not output
+ c2 = make_divisible(c2 * gw, 8)
+
+ args = [c1, c2, *args[1:]]
+ if m in [BottleneckCSP, C3, C3TR, C3Ghost]:
+ args.insert(2, n) # number of repeats
+ n = 1
+ elif m is nn.BatchNorm2d:
+ args = [ch[f]]
+ elif m is Concat:
+ c2 = sum(ch[x] for x in f)
+ elif m is Detect:
+ args.append([ch[x] for x in f])
+ if isinstance(args[1], int): # number of anchors
+ args[1] = [list(range(args[1] * 2))] * len(f)
+ elif m is Contract:
+ c2 = ch[f] * args[0] ** 2
+ elif m is Expand:
+ c2 = ch[f] // args[0] ** 2
+ else:
+ c2 = ch[f]
+
+ m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args) # module
+ t = str(m)[8:-2].replace('__main__.', '') # module type
+ np = sum(x.numel() for x in m_.parameters()) # number params
+ m_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number params
+ LOGGER.info(f'{i:>3}{str(f):>18}{n_:>3}{np:10.0f} {t:<40}{str(args):<30}') # print
+ save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist
+ layers.append(m_)
+ if i == 0:
+ ch = []
+ ch.append(c2)
+ return nn.Sequential(*layers), sorted(save)
+
+
+if __name__ == '__main__':
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--cfg', type=str, default='yolov5s.yaml', help='model.yaml')
+ parser.add_argument('--batch-size', type=int, default=1, help='total batch size for all GPUs')
+ parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--profile', action='store_true', help='profile model speed')
+ parser.add_argument('--line-profile', action='store_true', help='profile model speed layer by layer')
+ parser.add_argument('--test', action='store_true', help='test all yolo*.yaml')
+ opt = parser.parse_args()
+ opt.cfg = check_yaml(opt.cfg) # check YAML
+ print_args(vars(opt))
+ device = select_device(opt.device)
+
+ # Create model
+ im = torch.rand(opt.batch_size, 3, 640, 640).to(device)
+ model = Model(opt.cfg).to(device)
+
+ # Options
+ if opt.line_profile: # profile layer by layer
+ _ = model(im, profile=True)
+
+ elif opt.profile: # profile forward-backward
+ results = profile(input=im, ops=[model], n=3)
+
+ elif opt.test: # test all models
+ for cfg in Path(ROOT / 'models').rglob('yolo*.yaml'):
+ try:
+ _ = Model(cfg)
+ except Exception as e:
+ print(f'Error in {cfg}: {e}')
diff --git a/vouchervision/component_detector/models/yolo_torchscript.py b/vouchervision/component_detector/models/yolo_torchscript.py
new file mode 100644
index 0000000000000000000000000000000000000000..ef7351969a3f1914a54b16d5a5f2748d70baf1ae
--- /dev/null
+++ b/vouchervision/component_detector/models/yolo_torchscript.py
@@ -0,0 +1,391 @@
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+YOLO-specific modules
+
+Usage:
+ $ python models/yolo.py --cfg yolov5s.yaml
+"""
+
+import argparse
+import contextlib
+import os
+import platform
+import sys
+from copy import deepcopy
+from pathlib import Path
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+if platform.system() != 'Windows':
+ ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+
+from models.common import * # noqa
+from models.experimental import * # noqa
+from utils.autoanchor import check_anchor_order
+from utils.general import LOGGER, check_version, check_yaml, make_divisible, print_args
+from utils.plots import feature_visualization
+from utils.torch_utils import (fuse_conv_and_bn, initialize_weights, model_info, profile, scale_img, select_device,
+ time_sync)
+
+try:
+ import thop # for FLOPs computation
+except ImportError:
+ thop = None
+
+
+class Detect(nn.Module):
+ # YOLOv5 Detect head for detection models
+ stride = None # strides computed during build
+ dynamic = False # force grid reconstruction
+ export = False # export mode
+
+ def __init__(self, nc=80, anchors=(), ch=(), inplace=True): # detection layer
+ super().__init__()
+ self.nc = nc # number of classes
+ self.no = nc + 5 # number of outputs per anchor
+ self.nl = len(anchors) # number of detection layers
+ self.na = len(anchors[0]) // 2 # number of anchors
+ self.grid = [torch.empty(0) for _ in range(self.nl)] # init grid
+ self.anchor_grid = [torch.empty(0) for _ in range(self.nl)] # init anchor grid
+ self.register_buffer('anchors', torch.tensor(anchors).float().view(self.nl, -1, 2)) # shape(nl,na,2)
+ self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch) # output conv
+ self.inplace = inplace # use inplace ops (e.g. slice assignment)
+
+ def forward(self, x):
+ z = [] # inference output
+ for i in range(self.nl):
+ x[i] = self.m[i](x[i]) # conv
+ bs, _, ny, nx = x[i].shape # x(bs,255,20,20) to x(bs,3,20,20,85)
+ x[i] = x[i].view(bs, self.na, self.no, ny, nx).permute(0, 1, 3, 4, 2).contiguous()
+
+ if not self.training: # inference
+ if self.dynamic or self.grid[i].shape[2:4] != x[i].shape[2:4]:
+ self.grid[i], self.anchor_grid[i] = self._make_grid(nx, ny, i)
+
+ if isinstance(self, Segment): # (boxes + masks)
+ xy, wh, conf, mask = x[i].split((2, 2, self.nc + 1, self.no - self.nc - 5), 4)
+ xy = (xy.sigmoid() * 2 + self.grid[i]) * self.stride[i] # xy
+ wh = (wh.sigmoid() * 2) ** 2 * self.anchor_grid[i] # wh
+ y = torch.cat((xy, wh, conf.sigmoid(), mask), 4)
+ else: # Detect (boxes only)
+ xy, wh, conf = x[i].sigmoid().split((2, 2, self.nc + 1), 4)
+ xy = (xy * 2 + self.grid[i]) * self.stride[i] # xy
+ wh = (wh * 2) ** 2 * self.anchor_grid[i] # wh
+ y = torch.cat((xy, wh, conf), 4)
+ z.append(y.view(bs, self.na * nx * ny, self.no))
+
+ return x if self.training else (torch.cat(z, 1), ) if self.export else (torch.cat(z, 1), x)
+
+ def _make_grid(self, nx=20, ny=20, i=0, torch_1_10=check_version(torch.__version__, '1.10.0')):
+ d = self.anchors[i].device
+ t = self.anchors[i].dtype
+ shape = 1, self.na, ny, nx, 2 # grid shape
+ y, x = torch.arange(ny, device=d, dtype=t), torch.arange(nx, device=d, dtype=t)
+ yv, xv = torch.meshgrid(y, x, indexing='ij') if torch_1_10 else torch.meshgrid(y, x) # torch>=0.7 compatibility
+ grid = torch.stack((xv, yv), 2).expand(shape) - 0.5 # add grid offset, i.e. y = 2.0 * x - 0.5
+ anchor_grid = (self.anchors[i] * self.stride[i]).view((1, self.na, 1, 1, 2)).expand(shape)
+ return grid, anchor_grid
+
+
+class Segment(Detect):
+ # YOLOv5 Segment head for segmentation models
+ def __init__(self, nc=80, anchors=(), nm=32, npr=256, ch=(), inplace=True):
+ super().__init__(nc, anchors, ch, inplace)
+ self.nm = nm # number of masks
+ self.npr = npr # number of protos
+ self.no = 5 + nc + self.nm # number of outputs per anchor
+ self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch) # output conv
+ self.proto = Proto(ch[0], self.npr, self.nm) # protos
+ self.detect = Detect.forward
+
+ def forward(self, x):
+ p = self.proto(x[0])
+ x = self.detect(self, x)
+ return (x, p) if self.training else (x[0], p) if self.export else (x[0], p, x[1])
+
+
+class BaseModel(nn.Module):
+ # YOLOv5 base model
+ def forward(self, x, profile=False, visualize=False):
+ return self._forward_once(x, profile, visualize) # single-scale inference, train
+
+ def _forward_once(self, x, profile=False, visualize=False):
+ y, dt = [], [] # outputs
+ for m in self.model:
+ if m.f != -1: # if not from previous layer
+ x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers
+ if profile:
+ self._profile_one_layer(m, x, dt)
+ x = m(x) # run
+ y.append(x if m.i in self.save else None) # save output
+ if visualize:
+ feature_visualization(x, m.type, m.i, save_dir=visualize)
+ return x
+
+ def _profile_one_layer(self, m, x, dt):
+ c = m == self.model[-1] # is final layer, copy input as inplace fix
+ o = thop.profile(m, inputs=(x.copy() if c else x, ), verbose=False)[0] / 1E9 * 2 if thop else 0 # FLOPs
+ t = time_sync()
+ for _ in range(10):
+ m(x.copy() if c else x)
+ dt.append((time_sync() - t) * 100)
+ if m == self.model[0]:
+ LOGGER.info(f"{'time (ms)':>10s} {'GFLOPs':>10s} {'params':>10s} module")
+ LOGGER.info(f'{dt[-1]:10.2f} {o:10.2f} {m.np:10.0f} {m.type}')
+ if c:
+ LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s} Total")
+
+ def fuse(self): # fuse model Conv2d() + BatchNorm2d() layers
+ LOGGER.info('Fusing layers... ')
+ for m in self.model.modules():
+ if isinstance(m, (Conv, DWConv)) and hasattr(m, 'bn'):
+ m.conv = fuse_conv_and_bn(m.conv, m.bn) # update conv
+ delattr(m, 'bn') # remove batchnorm
+ m.forward = m.forward_fuse # update forward
+ self.info()
+ return self
+
+ def info(self, verbose=False, img_size=640): # print model information
+ model_info(self, verbose, img_size)
+
+ def _apply(self, fn):
+ # Apply to(), cpu(), cuda(), half() to model tensors that are not parameters or registered buffers
+ self = super()._apply(fn)
+ m = self.model[-1] # Detect()
+ if isinstance(m, (Detect, Segment)):
+ m.stride = fn(m.stride)
+ m.grid = list(map(fn, m.grid))
+ if isinstance(m.anchor_grid, list):
+ m.anchor_grid = list(map(fn, m.anchor_grid))
+ return self
+
+
+class DetectionModel(BaseModel):
+ # YOLOv5 detection model
+ def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, anchors=None): # model, input channels, number of classes
+ super().__init__()
+ if isinstance(cfg, dict):
+ self.yaml = cfg # model dict
+ else: # is *.yaml
+ import yaml # for torch hub
+ self.yaml_file = Path(cfg).name
+ with open(cfg, encoding='ascii', errors='ignore') as f:
+ self.yaml = yaml.safe_load(f) # model dict
+
+ # Define model
+ ch = self.yaml['ch'] = self.yaml.get('ch', ch) # input channels
+ if nc and nc != self.yaml['nc']:
+ LOGGER.info(f"Overriding model.yaml nc={self.yaml['nc']} with nc={nc}")
+ self.yaml['nc'] = nc # override yaml value
+ if anchors:
+ LOGGER.info(f'Overriding model.yaml anchors with anchors={anchors}')
+ self.yaml['anchors'] = round(anchors) # override yaml value
+ self.model, self.save = parse_model(deepcopy(self.yaml), ch=[ch]) # model, savelist
+ self.names = [str(i) for i in range(self.yaml['nc'])] # default names
+ self.inplace = self.yaml.get('inplace', True)
+
+ # Build strides, anchors
+ m = self.model[-1] # Detect()
+ if isinstance(m, (Detect, Segment)):
+ s = 256 # 2x min stride
+ m.inplace = self.inplace
+ forward = lambda x: self.forward(x)[0] if isinstance(m, Segment) else self.forward(x)
+ m.stride = torch.tensor([s / x.shape[-2] for x in forward(torch.zeros(1, ch, s, s))]) # forward
+ check_anchor_order(m)
+ m.anchors /= m.stride.view(-1, 1, 1)
+ self.stride = m.stride
+ self._initialize_biases() # only run once
+
+ # Init weights, biases
+ initialize_weights(self)
+ self.info()
+ LOGGER.info('')
+
+ def forward(self, x, augment=False, profile=False, visualize=False):
+ if augment:
+ return self._forward_augment(x) # augmented inference, None
+ return self._forward_once(x, profile, visualize) # single-scale inference, train
+
+ def _forward_augment(self, x):
+ img_size = x.shape[-2:] # height, width
+ s = [1, 0.83, 0.67] # scales
+ f = [None, 3, None] # flips (2-ud, 3-lr)
+ y = [] # outputs
+ for si, fi in zip(s, f):
+ xi = scale_img(x.flip(fi) if fi else x, si, gs=int(self.stride.max()))
+ yi = self._forward_once(xi)[0] # forward
+ # cv2.imwrite(f'img_{si}.jpg', 255 * xi[0].cpu().numpy().transpose((1, 2, 0))[:, :, ::-1]) # save
+ yi = self._descale_pred(yi, fi, si, img_size)
+ y.append(yi)
+ y = self._clip_augmented(y) # clip augmented tails
+ return torch.cat(y, 1), None # augmented inference, train
+
+ def _descale_pred(self, p, flips, scale, img_size):
+ # de-scale predictions following augmented inference (inverse operation)
+ if self.inplace:
+ p[..., :4] /= scale # de-scale
+ if flips == 2:
+ p[..., 1] = img_size[0] - p[..., 1] # de-flip ud
+ elif flips == 3:
+ p[..., 0] = img_size[1] - p[..., 0] # de-flip lr
+ else:
+ x, y, wh = p[..., 0:1] / scale, p[..., 1:2] / scale, p[..., 2:4] / scale # de-scale
+ if flips == 2:
+ y = img_size[0] - y # de-flip ud
+ elif flips == 3:
+ x = img_size[1] - x # de-flip lr
+ p = torch.cat((x, y, wh, p[..., 4:]), -1)
+ return p
+
+ def _clip_augmented(self, y):
+ # Clip YOLOv5 augmented inference tails
+ nl = self.model[-1].nl # number of detection layers (P3-P5)
+ g = sum(4 ** x for x in range(nl)) # grid points
+ e = 1 # exclude layer count
+ i = (y[0].shape[1] // g) * sum(4 ** x for x in range(e)) # indices
+ y[0] = y[0][:, :-i] # large
+ i = (y[-1].shape[1] // g) * sum(4 ** (nl - 1 - x) for x in range(e)) # indices
+ y[-1] = y[-1][:, i:] # small
+ return y
+
+ def _initialize_biases(self, cf=None): # initialize biases into Detect(), cf is class frequency
+ # https://arxiv.org/abs/1708.02002 section 3.3
+ # cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1.
+ m = self.model[-1] # Detect() module
+ for mi, s in zip(m.m, m.stride): # from
+ b = mi.bias.view(m.na, -1) # conv.bias(255) to (3,85)
+ b.data[:, 4] += math.log(8 / (640 / s) ** 2) # obj (8 objects per 640 image)
+ b.data[:, 5:5 + m.nc] += math.log(0.6 / (m.nc - 0.99999)) if cf is None else torch.log(cf / cf.sum()) # cls
+ mi.bias = torch.nn.Parameter(b.view(-1), requires_grad=True)
+
+
+Model = DetectionModel # retain YOLOv5 'Model' class for backwards compatibility
+
+
+class SegmentationModel(DetectionModel):
+ # YOLOv5 segmentation model
+ def __init__(self, cfg='yolov5s-seg.yaml', ch=3, nc=None, anchors=None):
+ super().__init__(cfg, ch, nc, anchors)
+
+
+class ClassificationModel(BaseModel):
+ # YOLOv5 classification model
+ def __init__(self, cfg=None, model=None, nc=1000, cutoff=10): # yaml, model, number of classes, cutoff index
+ super().__init__()
+ self._from_detection_model(model, nc, cutoff) if model is not None else self._from_yaml(cfg)
+
+ def _from_detection_model(self, model, nc=1000, cutoff=10):
+ # Create a YOLOv5 classification model from a YOLOv5 detection model
+ if isinstance(model, DetectMultiBackend):
+ model = model.model # unwrap DetectMultiBackend
+ model.model = model.model[:cutoff] # backbone
+ m = model.model[-1] # last layer
+ ch = m.conv.in_channels if hasattr(m, 'conv') else m.cv1.conv.in_channels # ch into module
+ c = Classify(ch, nc) # Classify()
+ c.i, c.f, c.type = m.i, m.f, 'models.common.Classify' # index, from, type
+ model.model[-1] = c # replace
+ self.model = model.model
+ self.stride = model.stride
+ self.save = []
+ self.nc = nc
+
+ def _from_yaml(self, cfg):
+ # Create a YOLOv5 classification model from a *.yaml file
+ self.model = None
+
+
+def parse_model(d, ch): # model_dict, input_channels(3)
+ # Parse a YOLOv5 model.yaml dictionary
+ LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}")
+ anchors, nc, gd, gw, act = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple'], d.get('activation')
+ if act:
+ Conv.default_act = eval(act) # redefine default activation, i.e. Conv.default_act = nn.SiLU()
+ LOGGER.info(f"{colorstr('activation:')} {act}") # print
+ na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors # number of anchors
+ no = na * (nc + 5) # number of outputs = anchors * (classes + 5)
+
+ layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out
+ for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']): # from, number, module, args
+ m = eval(m) if isinstance(m, str) else m # eval strings
+ for j, a in enumerate(args):
+ with contextlib.suppress(NameError):
+ args[j] = eval(a) if isinstance(a, str) else a # eval strings
+
+ n = n_ = max(round(n * gd), 1) if n > 1 else n # depth gain
+ if m in {
+ Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv,
+ BottleneckCSP, C3, C3TR, C3SPP, C3Ghost, nn.ConvTranspose2d, DWConvTranspose2d, C3x}:
+ c1, c2 = ch[f], args[0]
+ if c2 != no: # if not output
+ c2 = make_divisible(c2 * gw, 8)
+
+ args = [c1, c2, *args[1:]]
+ if m in {BottleneckCSP, C3, C3TR, C3Ghost, C3x}:
+ args.insert(2, n) # number of repeats
+ n = 1
+ elif m is nn.BatchNorm2d:
+ args = [ch[f]]
+ elif m is Concat:
+ c2 = sum(ch[x] for x in f)
+ # TODO: channel, gw, gd
+ elif m in {Detect, Segment}:
+ args.append([ch[x] for x in f])
+ if isinstance(args[1], int): # number of anchors
+ args[1] = [list(range(args[1] * 2))] * len(f)
+ if m is Segment:
+ args[3] = make_divisible(args[3] * gw, 8)
+ elif m is Contract:
+ c2 = ch[f] * args[0] ** 2
+ elif m is Expand:
+ c2 = ch[f] // args[0] ** 2
+ else:
+ c2 = ch[f]
+
+ m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args) # module
+ t = str(m)[8:-2].replace('__main__.', '') # module type
+ np = sum(x.numel() for x in m_.parameters()) # number params
+ m_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number params
+ LOGGER.info(f'{i:>3}{str(f):>18}{n_:>3}{np:10.0f} {t:<40}{str(args):<30}') # print
+ save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist
+ layers.append(m_)
+ if i == 0:
+ ch = []
+ ch.append(c2)
+ return nn.Sequential(*layers), sorted(save)
+
+
+if __name__ == '__main__':
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--cfg', type=str, default='yolov5s.yaml', help='model.yaml')
+ parser.add_argument('--batch-size', type=int, default=1, help='total batch size for all GPUs')
+ parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--profile', action='store_true', help='profile model speed')
+ parser.add_argument('--line-profile', action='store_true', help='profile model speed layer by layer')
+ parser.add_argument('--test', action='store_true', help='test all yolo*.yaml')
+ opt = parser.parse_args()
+ opt.cfg = check_yaml(opt.cfg) # check YAML
+ print_args(vars(opt))
+ device = select_device(opt.device)
+
+ # Create model
+ im = torch.rand(opt.batch_size, 3, 640, 640).to(device)
+ model = Model(opt.cfg).to(device)
+
+ # Options
+ if opt.line_profile: # profile layer by layer
+ model(im, profile=True)
+
+ elif opt.profile: # profile forward-backward
+ results = profile(input=im, ops=[model], n=3)
+
+ elif opt.test: # test all models
+ for cfg in Path(ROOT / 'models').rglob('yolo*.yaml'):
+ try:
+ _ = Model(cfg)
+ except Exception as e:
+ print(f'Error in {cfg}: {e}')
+
+ else: # report fused model summary
+ model.fuse()
\ No newline at end of file
diff --git a/vouchervision/component_detector/models/yolov5l.yaml b/vouchervision/component_detector/models/yolov5l.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..ce8a5de46a2785f5537c09fe27f3077c057bb4f3
--- /dev/null
+++ b/vouchervision/component_detector/models/yolov5l.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.0 # model depth multiple
+width_multiple: 1.0 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/yolov5m.yaml b/vouchervision/component_detector/models/yolov5m.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..ad13ab370ff6532931284a0193959afba214f6f4
--- /dev/null
+++ b/vouchervision/component_detector/models/yolov5m.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.67 # model depth multiple
+width_multiple: 0.75 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/yolov5n.yaml b/vouchervision/component_detector/models/yolov5n.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..8a28a40d6e20383727da1a9eed180c9e13ee89fd
--- /dev/null
+++ b/vouchervision/component_detector/models/yolov5n.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.33 # model depth multiple
+width_multiple: 0.25 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/yolov5s.yaml b/vouchervision/component_detector/models/yolov5s.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..f35beabb1e1c76f9ec2cad0cb7adbce76f6b7c4c
--- /dev/null
+++ b/vouchervision/component_detector/models/yolov5s.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 0.33 # model depth multiple
+width_multiple: 0.50 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/models/yolov5x.yaml b/vouchervision/component_detector/models/yolov5x.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..f617a027d8a20a2b7c2a4b415da0941c02aeb3a3
--- /dev/null
+++ b/vouchervision/component_detector/models/yolov5x.yaml
@@ -0,0 +1,48 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80 # number of classes
+depth_multiple: 1.33 # model depth multiple
+width_multiple: 1.25 # layer channel multiple
+anchors:
+ - [10,13, 16,30, 33,23] # P3/8
+ - [30,61, 62,45, 59,119] # P4/16
+ - [116,90, 156,198, 373,326] # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+ # [from, number, module, args]
+ [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2
+ [-1, 1, Conv, [128, 3, 2]], # 1-P2/4
+ [-1, 3, C3, [128]],
+ [-1, 1, Conv, [256, 3, 2]], # 3-P3/8
+ [-1, 6, C3, [256]],
+ [-1, 1, Conv, [512, 3, 2]], # 5-P4/16
+ [-1, 9, C3, [512]],
+ [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32
+ [-1, 3, C3, [1024]],
+ [-1, 1, SPPF, [1024, 5]], # 9
+ ]
+
+# YOLOv5 v6.0 head
+head:
+ [[-1, 1, Conv, [512, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 6], 1, Concat, [1]], # cat backbone P4
+ [-1, 3, C3, [512, False]], # 13
+
+ [-1, 1, Conv, [256, 1, 1]],
+ [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+ [[-1, 4], 1, Concat, [1]], # cat backbone P3
+ [-1, 3, C3, [256, False]], # 17 (P3/8-small)
+
+ [-1, 1, Conv, [256, 3, 2]],
+ [[-1, 14], 1, Concat, [1]], # cat head P4
+ [-1, 3, C3, [512, False]], # 20 (P4/16-medium)
+
+ [-1, 1, Conv, [512, 3, 2]],
+ [[-1, 10], 1, Concat, [1]], # cat head P5
+ [-1, 3, C3, [1024, False]], # 23 (P5/32-large)
+
+ [[17, 20, 23], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5)
+ ]
diff --git a/vouchervision/component_detector/runYOLOv5forDir.py b/vouchervision/component_detector/runYOLOv5forDir.py
new file mode 100644
index 0000000000000000000000000000000000000000..b225f483bc5ebc95fdb20f3ea0dada9e94fda971
--- /dev/null
+++ b/vouchervision/component_detector/runYOLOv5forDir.py
@@ -0,0 +1,212 @@
+# Run yolov5 on dir
+import os
+from os import walk
+import pandas as pd
+import shutil
+import subprocess
+
+from detect import *
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+from machine.general_utils import make_file_names_valid
+
+# pip install cython matplotlib tqdm scipy ipython ninja yacs opencv-python ffmpeg opencv-contrib-python Pillow scikit-image scikit-learn lmfit imutils pyyaml jupyterlab==3
+# pip install torch==1.7.1+cu101 torchvision==0.8.2+cu101 torchaudio==0.7.2 -f https://download.pytorch.org/whl/torch_stable.html
+
+class bcolors:
+ HEADER = '\033[95m'
+ OKBLUE = '\033[94m'
+ OKCYAN = '\033[96m'
+ OKGREEN = '\033[92m'
+ WARNING = '\033[93m'
+ FAIL = '\033[91m'
+ ENDC = '\033[0m'
+ BOLD = '\033[1m'
+ UNDERLINE = '\033[4m'
+
+
+def validateDir(dir):
+ if not os.path.exists(dir):
+ os.makedirs(dir)
+
+def readDetailedSampleJPGs(dirDetailed):
+ f = []
+ nameList = []
+ for (dirpath, dirnames, filenames) in walk(dirDetailed):
+ f.extend(filenames)
+ break
+ # print(f)
+ type(f)
+ fileList = pd.DataFrame(f)
+ fileList.columns = ['fname']
+ df = fileList['fname'].str.split('_', expand=True)
+ family = df[2]
+ genus = df[3]
+ species = df[4]
+ species = species.str.split('.', expand=True)[0]
+ nameList = family + "_" + genus + '_' + species
+ nameList = pd.DataFrame(nameList)
+ nameList = pd.DataFrame(nameList[0].unique())
+ nameList = nameList.dropna()
+ nameList.columns = ['fullname']
+ return fileList, nameList
+
+def saveDetailedBySpecies(fileList, nameList):
+
+ for species in nameList['fullname']:
+ dirSpecies = os.path.join(dirDetailedBySpecies,species)
+ validateDir(dirSpecies)
+
+ ind = fileList['fname'].str.contains(species, regex=False)
+ speciesFiles = fileList['fname'][ind]
+ if len(speciesFiles) == 50:
+ print(f"{bcolors.BOLD}Species: {species} >>> Number of Images: {len(speciesFiles)}{bcolors.ENDC}")
+ else:
+ print(f"{bcolors.WARNING}Species: {species} >>> Number of Images: {len(speciesFiles)}{bcolors.ENDC}")
+
+ for img in speciesFiles:
+ print(f"{bcolors.BOLD} Copied: {img}{bcolors.ENDC}")
+ shutil.copy(os.path.join(dirDetailed,img), os.path.join(dirSpecies,img))
+
+def runYOLOforDirOfFolders(dirDetailedBySpecies,dirOutBase,nosave,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS,ANNO_TYPE):
+ if INCLUDE_SUBDIRS:
+ f = []
+ for (dirpath, dirnames, filenames) in walk(dirDetailedBySpecies):
+ f.extend(dirnames)
+ break
+ for species in f:
+ print(species)
+ dirSource = os.path.abspath(os.path.join(dirDetailedBySpecies,species))
+ # dirYOLO = os.path.abspath(os.path.join('yolov5','detect.py'))
+ dirWeights = os.path.abspath(os.path.join('YOLOv5','yolov5',ANNO,VERSION,'weights','best.pt'))
+ dirProject = os.path.abspath(os.path.join(dirOutBase,PROJECT,SET))
+ run(weights=dirWeights,source=dirSource,project=dirProject,name=species,imgsz=(1280, 1280),nosave=nosave,anno_type=ANNO_TYPE)
+
+ else:
+ # f = []
+ # for (dirpath, dirnames, filenames) in walk(dirDetailedBySpecies):
+ # f.extend(filenames)
+ # break
+ print(SET)
+ dirSource = dirDetailedBySpecies
+ species = os.path.basename(dirSource)
+ # dirYOLO = os.path.abspath(os.path.join('yolov5','detect.py'))
+ # dirWeights = os.path.join('yolov5','runs','train',ANNO,VERSION,'weights','best.pt')
+ dirWeights = os.path.join(dirOutBase,'runs','train','Archival_Detector','FieldPrism_Initial','FieldPrism_Initial7','weights','last.pt')
+ dirProject = os.path.join(dirOutBase,'runs','detect',PROJECT,SET)
+ run(weights=dirWeights,source=dirSource,project=dirProject,name=species,imgsz=(1280, 1280),nosave=nosave,anno_type=ANNO_TYPE)
+
+
+def runYOLOforDirOfFolders_PLANT_Botany(dirDetailedBySpecies,dirOutBase,nosave,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS):
+ if INCLUDE_SUBDIRS:
+ f = []
+ for (dirpath, dirnames, filenames) in walk(dirDetailedBySpecies):
+ f.extend(dirnames)
+ break
+ for species in f:
+ print(species)
+ dirSource = os.path.abspath(os.path.join(dirDetailedBySpecies,species))
+ # dirYOLO = os.path.abspath(os.path.join('yolov5','detect.py'))
+ dirWeights = os.path.abspath(os.path.join('yolov5','runs','train',ANNO,VERSION,'weights','best.pt'))
+ dirProject = os.path.abspath(os.path.join(dirOutBase,PROJECT,SET))
+ run(weights=dirWeights,source=dirSource,project=dirProject,name=species,imgsz=(1280, 1280),nosave=nosave)
+
+ else:
+ # f = []
+ # for (dirpath, dirnames, filenames) in walk(dirDetailedBySpecies):
+ # f.extend(filenames)
+ # break
+ print(SET)
+ dirSource = dirDetailedBySpecies
+ species = os.path.basename(dirSource)
+ # dirYOLO = os.path.abspath(os.path.join('yolov5','detect.py'))
+ # dirWeights = os.path.join('yolov5','runs','train',ANNO,VERSION,'weights','best.pt')
+ # dirWeights = os.path.abspath(os.path.join('YOLOv5','yolov5',ANNO,VERSION,'weights','best.pt'))
+ dirWeights = os.path.abspath(os.path.join('YOLOv5','yolov5',ANNO,VERSION,'weights','best.pt'))
+ dirProject = os.path.abspath(os.path.join(dirOutBase,PROJECT,SET))
+ run(weights=dirWeights,source=dirSource,project=dirProject,name=species,imgsz=(1280, 1280),nosave=nosave)
+
+ # for species in f:
+ # print(species)
+ # dirSource = os.path.abspath(os.path.join(dirDetailedBySpecies,species))
+ # dirYOLO = os.path.abspath(os.path.join('yolov5','detect.py'))
+ # dirWeights = os.path.join('yolov5','runs','train',ANNO,VERSION,'weights','best.pt')
+ # dirProject = os.path.join(PROJECT,SET)
+ # if INCLUDE_SUBDIRS:
+ # run(weights=dirWeights,source=dirSource,project=dirProject,name=species,imgsz=(1280, 1280))
+ # else:
+ # run(weights=dirWeights,source=dirSource,project=dirProject,name=SET,imgsz=(1280, 1280))
+
+
+
+### Parse the aggregated DetailedSample folder that has 2,500 images
+# fileList, nameList = readDetailedSampleJPGs(dirDetailed)
+# saveDetailedBySpecies(fileList, nameList)
+
+### Run detect for every folder in dirDetailedBySpecies
+# PROJECT = 'MAL'
+# SET = 'Detailed'
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_DetailedSample_50Spp_Ind'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,PROJECT,SET,,1)
+
+# dirDetailed = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_DetailedSample_50Spp'))
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_DetailedSample_50Spp_Ind'))
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_TargetedSample_Fraxinus'))
+
+
+### Run detect for TargetedSample
+# PREFIX = 'DT_MAL_'
+# INCLUDE_SUBDIRS = 0
+# PROJECT = 'Botany'#'Cannon'#'MAL_PLANT'
+# SET = 'Demo'#'Test_Sheets_PREP'#'Targeted'
+# ANNO = 'PREPfull'#'PLANTfull'
+# VERSION = 'baseline_all_hypEvolve'#'baseline'
+
+# dirDetailedBySpecies = os.path.abspath(os.path.join('Image_Datasets','Botany_Test_Images'))
+# dirOutBase = os.path.abspath(os.path.join('YOLOv5'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,dirOutBase,False,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS)
+# dirDetailedBySpecies = os.path.abspath(os.path.join('Image_Datasets','Cannon','Test_Sheets'))
+# dirOutBase = os.path.abspath(os.path.join('YOLOv5'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,dirOutBase,False,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS)
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_TargetedSample_Quercus_havardii'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS)
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_TargetedSample_Fraxinus'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS)
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_TargetedSample_Juglandaceae'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS)
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_TargetedSample_Lonicera'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS)
+# dirDetailedBySpecies = os.path.abspath(os.path.join(os.pardir,'Image_Datasets','GBIF_TargetedSample_Rhus'))
+# runYOLOforDirOfFolders(dirDetailedBySpecies,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS)
+
+INCLUDE_SUBDIRS = 0
+PROJECT = 'MAL_FP' #'MAL_PLANT'#'Botany'#'Cannon'#'MAL_PLANT'
+SET = 'FieldPrism_Initial' #'Detailed'#'Demo_Plant'#'Test_Sheets_PREP'#'Targeted'
+ANNO = 'PREPfull'#'PLANT_Botany'#'PLANTfull'#'PREPfull
+VERSION = 'baseline'#'Small_Adoxaceae'#'baseline'
+ANNO_TYPE = 'PREP'
+
+# # dirDetailedBySpecies = os.path.abspath(os.path.join('Image_Datasets','FieldPrism_Training_Images','FieldPrism_Training_FS-Poor'))
+# dirDetailedBySpecies = 'D:/Dropbox/LM2_Env/Image_Datasets/FieldPrism_Training_Images/FieldPrism_Training_Sheets'
+# # dirOutBase = os.path.abspath(os.path.join('YOLOv5'))
+# dirOutBase = os.path.dirname(__file__)
+# # dirOutBase > PROJECT > SET
+# # ML network: ANNO > VERSION
+# make_file_names_valid(dirDetailedBySpecies)
+# runYOLOforDirOfFolders(dirDetailedBySpecies,dirOutBase,False,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS,ANNO_TYPE)
+
+# dirDetailedBySpecies = 'D:/Dropbox/LM2_Env/Image_Datasets/FieldPrism_Training_Images/FieldPrism_Training_Outside'
+# dirOutBase = os.path.dirname(__file__)
+# make_file_names_valid(dirDetailedBySpecies)
+# runYOLOforDirOfFolders(dirDetailedBySpecies,dirOutBase,False,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS,ANNO_TYPE)
+
+dirDetailedBySpecies = 'D:/Dropbox/LM2_Env/Image_Datasets/FieldPrism_Training_Images/FieldPrism_Training_FS-Poor'
+dirOutBase = os.path.dirname(__file__)
+make_file_names_valid(dirDetailedBySpecies)
+runYOLOforDirOfFolders(dirDetailedBySpecies,dirOutBase,False,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS,ANNO_TYPE)
+
+# dirDetailedBySpecies = 'D:/Dropbox/LM2_Env/Image_Datasets/FieldPrism_Training_Images/REU_Field_QR-Code-Images'
+# dirOutBase = os.path.dirname(__file__)
+# make_file_names_valid(dirDetailedBySpecies)
+# runYOLOforDirOfFolders(dirDetailedBySpecies,dirOutBase,False,PROJECT,SET,ANNO,VERSION,INCLUDE_SUBDIRS,ANNO_TYPE)
\ No newline at end of file
diff --git a/vouchervision/component_detector/train.py b/vouchervision/component_detector/train.py
new file mode 100644
index 0000000000000000000000000000000000000000..bf74cc119e3cd709361f680a83498c6017a30953
--- /dev/null
+++ b/vouchervision/component_detector/train.py
@@ -0,0 +1,936 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+venv requirements:
+pip install torch==1.11.0+cu113 torchvision==0.12.0+cu113 torchaudio==0.11.0 --extra-index-url https://download.pytorch.org/whl/cu113
+pip3 install cython
+pip3 install opencv-python
+
+Train a YOLOv5 model on a custom dataset.
+
+Models and datasets download automatically from the latest YOLOv5 release.
+Models: https://github.com/ultralytics/yolov5/tree/master/models
+Datasets: https://github.com/ultralytics/yolov5/tree/master/data
+Tutorial: https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data
+
+Usage:
+ $ python path/to/train.py --data coco128.yaml --weights yolov5s.pt --img 640 # from pretrained (RECOMMENDED)
+ $ python path/to/train.py --data coco128.yaml --weights '' --cfg_model yolov5s.yaml --img 640 # from scratch
+"""
+
+import os,argparse, math, random, sys, time, yaml, wandb, inspect
+from dataclasses import dataclass, field
+from copy import deepcopy
+from datetime import datetime, timedelta
+from pathlib import Path
+import numpy as np
+import torch
+import torch.distributed as dist
+import torch.nn as nn
+from torch.cuda import amp
+from torch.nn.parallel import DistributedDataParallel as DDP
+from torch.optim import SGD, Adam, AdamW, lr_scheduler
+from tqdm.auto import tqdm
+import collections
+import collections.abc
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[0] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+
+import val # for end-of-epoch mAP
+from models.experimental import attempt_load
+from models.yolo import Model
+from utils.autoanchor import check_anchors
+from utils.autobatch import check_train_batch_size
+from utils.callbacks import Callbacks
+from utils.datasets import create_dataloader
+from utils.downloads import attempt_download
+from utils.general import (LOGGER, check_dataset, check_file, check_git_status, check_img_size, check_requirements,
+ check_suffix, check_yaml, colorstr, get_latest_run, increment_path, init_seeds,
+ intersect_dicts, is_ascii, labels_to_class_weights, labels_to_image_weights, methods,
+ one_cycle, print_args, print_mutation, strip_optimizer)
+from utils.loggers import Loggers
+from utils.loggers.wandb.wandb_utils import check_wandb_resume
+from utils.loss import ComputeLoss
+from utils.metrics import fitness
+from utils.plots import plot_evolve, plot_labels
+from utils.torch_utils import EarlyStopping, ModelEMA, de_parallel, select_device, torch_distributed_zero_first
+
+LOCAL_RANK = int(os.getenv('LOCAL_RANK', -1)) # https://pytorch.org/docs/stable/elastic/run.html
+RANK = int(os.getenv('RANK', -1))
+WORLD_SIZE = int(os.getenv('WORLD_SIZE', 1))
+
+currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+from machine.general_utils import get_datetime, load_cfg, get_cfg_from_full_path
+
+def train(hyp, opt, device, callbacks): # hyp is path/to/hyp.yaml or hyp dictionary
+ save_dir, epochs, batch_size, weights, single_cls, evolve, data, cfg_model, resume, noval, nosave, workers, freeze = \
+ Path(opt.save_dir), opt.epochs, opt.batch_size, opt.weights, opt.single_cls, opt.evolve, opt.data, opt.cfg_model, \
+ opt.resume, opt.noval, opt.nosave, opt.workers, opt.freeze
+ callbacks.run('on_pretrain_routine_start')
+
+ # Directories
+ w = save_dir / 'weights' # weights dir
+ (w.parent if evolve else w).mkdir(parents=True, exist_ok=True) # make dir
+ last, best = w / 'last.pt', w / 'best.pt'
+
+ # Hyperparameters
+ if isinstance(hyp, str):
+ with open(hyp, errors='ignore') as f:
+ hyp = yaml.safe_load(f) # load hyps dict
+ LOGGER.info(colorstr('hyperparameters: ') + ', '.join(f'{k}={v}' for k, v in hyp.items()))
+
+ # Save run settings
+ if not evolve:
+ with open(save_dir / 'hyp.yaml', 'w') as f:
+ yaml.safe_dump(hyp, f, sort_keys=False)
+ with open(save_dir / 'opt.yaml', 'w') as f:
+ yaml.safe_dump(vars(opt), f, sort_keys=False)
+
+ # Loggers
+ data_dict = None
+ if RANK in [-1, 0]:
+ loggers = Loggers(save_dir, weights, opt, hyp, LOGGER) # loggers instance
+ if loggers.wandb:
+ data_dict = loggers.wandb.data_dict
+ if resume:
+ weights, epochs, hyp, batch_size = opt.weights, opt.epochs, opt.hyp, opt.batch_size
+
+ # Register actions
+ for k in methods(loggers):
+ callbacks.register_action(k, callback=getattr(loggers, k))
+
+ # Config
+ plots = not evolve and not opt.noplots # create plots
+ cuda = device.type != 'cpu'
+ init_seeds(1 + RANK)
+ with torch_distributed_zero_first(LOCAL_RANK):
+ data_dict = data_dict or check_dataset(data) # check if None
+ train_path, val_path = data_dict['train'], data_dict['val']
+ nc = 1 if single_cls else int(data_dict['nc']) # number of classes
+ names = ['item'] if single_cls and len(data_dict['names']) != 1 else data_dict['names'] # class names
+ assert len(names) == nc, f'{len(names)} names found for nc={nc} dataset in {data}' # check
+ is_coco = isinstance(val_path, str) and val_path.endswith('coco/val2017.txt') # COCO dataset
+
+ # Model
+ check_suffix(weights, '.pt') # check weights
+ pretrained = weights.endswith('.pt')
+ if pretrained:
+ with torch_distributed_zero_first(LOCAL_RANK):
+ weights = attempt_download(weights) # download if not found locally
+ ckpt = torch.load(weights, map_location='cpu') # load checkpoint to CPU to avoid CUDA memory leak
+ model = Model(cfg_model or ckpt['model'].yaml, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device) # create
+ exclude = ['anchor'] if (cfg_model or hyp.get('anchors')) and not resume else [] # exclude keys
+ csd = ckpt['model'].float().state_dict() # checkpoint state_dict as FP32
+ csd = intersect_dicts(csd, model.state_dict(), exclude=exclude) # intersect
+ model.load_state_dict(csd, strict=False) # load
+ LOGGER.info(f'Transferred {len(csd)}/{len(model.state_dict())} items from {weights}') # report
+ else:
+ model = Model(cfg_model, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device) # create
+
+ # Freeze
+ freeze = [f'model.{x}.' for x in (freeze if len(freeze) > 1 else range(freeze[0]))] # layers to freeze
+ for k, v in model.named_parameters():
+ v.requires_grad = True # train all layers
+ if any(x in k for x in freeze):
+ LOGGER.info(f'freezing {k}')
+ v.requires_grad = False
+
+ # Image size
+ gs = max(int(model.stride.max()), 32) # grid size (max stride)
+ imgsz = check_img_size(opt.imgsz, gs, floor=gs * 2) # verify imgsz is gs-multiple
+
+ # Batch size
+ if RANK == -1 and batch_size == -1: # single-GPU only, estimate best batch size
+ batch_size = check_train_batch_size(model, imgsz)
+ loggers.on_params_update({"batch_size": batch_size})
+
+ # Optimizer
+ nbs = 64 # nominal batch size
+ accumulate = max(round(nbs / batch_size), 1) # accumulate loss before optimizing
+ hyp['weight_decay'] *= batch_size * accumulate / nbs # scale weight_decay
+ LOGGER.info(f"Scaled weight_decay = {hyp['weight_decay']}")
+
+ g = [], [], [] # optimizer parameter groups
+ bn = tuple(v for k, v in nn.__dict__.items() if 'Norm' in k) # normalization layers, i.e. BatchNorm2d()
+ for v in model.modules():
+ if hasattr(v, 'bias') and isinstance(v.bias, nn.Parameter): # bias
+ g[2].append(v.bias)
+ if isinstance(v, bn): # weight (no decay)
+ g[1].append(v.weight)
+ elif hasattr(v, 'weight') and isinstance(v.weight, nn.Parameter): # weight (with decay)
+ g[0].append(v.weight)
+
+ if opt.optimizer == 'Adam':
+ optimizer = Adam(g[2], lr=hyp['lr0'], betas=(hyp['momentum'], 0.999)) # adjust beta1 to momentum
+ elif opt.optimizer == 'AdamW':
+ optimizer = AdamW(g[2], lr=hyp['lr0'], betas=(hyp['momentum'], 0.999)) # adjust beta1 to momentum
+ else:
+ optimizer = SGD(g[2], lr=hyp['lr0'], momentum=hyp['momentum'], nesterov=True)
+
+ optimizer.add_param_group({'params': g[0], 'weight_decay': hyp['weight_decay']}) # add g0 with weight_decay
+ optimizer.add_param_group({'params': g[1]}) # add g1 (BatchNorm2d weights)
+ LOGGER.info(f"{colorstr('optimizer:')} {type(optimizer).__name__} with parameter groups "
+ f"{len(g[1])} weight (no decay), {len(g[0])} weight, {len(g[2])} bias")
+ del g
+
+ # Scheduler
+ if opt.cos_lr:
+ lf = one_cycle(1, hyp['lrf'], epochs) # cosine 1->hyp['lrf']
+ else:
+ lf = lambda x: (1 - x / epochs) * (1.0 - hyp['lrf']) + hyp['lrf'] # linear
+ scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf) # plot_lr_scheduler(optimizer, scheduler, epochs)
+
+ # EMA
+ ema = ModelEMA(model) if RANK in [-1, 0] else None
+
+ # Resume
+ start_epoch, best_fitness = 0, 0.0
+ if pretrained:
+ # Optimizer
+ if ckpt['optimizer'] is not None:
+ optimizer.load_state_dict(ckpt['optimizer'])
+ best_fitness = ckpt['best_fitness']
+
+ # EMA
+ if ema and ckpt.get('ema'):
+ ema.ema.load_state_dict(ckpt['ema'].float().state_dict())
+ ema.updates = ckpt['updates']
+
+ # Epochs
+ start_epoch = ckpt['epoch'] + 1
+ if resume:
+ assert start_epoch > 0, f'{weights} training to {epochs} epochs is finished, nothing to resume.'
+ if epochs < start_epoch:
+ LOGGER.info(f"{weights} has been trained for {ckpt['epoch']} epochs. Fine-tuning for {epochs} more epochs.")
+ epochs += ckpt['epoch'] # finetune additional epochs
+
+ del ckpt, csd
+
+ # DP mode
+ if cuda and RANK == -1 and torch.cuda.device_count() > 1:
+ LOGGER.warning('WARNING: DP not recommended, use torch.distributed.run for best DDP Multi-GPU results.\n'
+ 'See Multi-GPU Tutorial at https://github.com/ultralytics/yolov5/issues/475 to get started.')
+ model = torch.nn.DataParallel(model)
+
+ # SyncBatchNorm
+ if opt.sync_bn and cuda and RANK != -1:
+ model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device)
+ LOGGER.info('Using SyncBatchNorm()')
+
+ # Trainloader
+ train_loader, dataset = create_dataloader(train_path,
+ imgsz,
+ batch_size // WORLD_SIZE,
+ gs,
+ single_cls,
+ hyp=hyp,
+ augment=True,
+ cache=None if opt.cache == 'val' else opt.cache,
+ rect=opt.rect,
+ rank=LOCAL_RANK,
+ workers=opt.workers,
+ image_weights=opt.image_weights,
+ quad=opt.quad,
+ prefix=colorstr('train: '),
+ shuffle=True)
+ mlc = int(np.concatenate(dataset.labels, 0)[:, 0].max()) # max label class
+ nb = len(train_loader) # number of batches
+ assert mlc < nc, f'Label class {mlc} exceeds nc={nc} in {data}. Possible class labels are 0-{nc - 1}'
+
+ # Process 0
+ if RANK in [-1, 0]:
+ val_loader = create_dataloader(val_path,
+ imgsz,
+ batch_size // WORLD_SIZE * 2,
+ gs,
+ single_cls,
+ hyp=hyp,
+ cache=None if noval else opt.cache,
+ rect=True,
+ rank=-1,
+ workers=workers * 2,
+ pad=0.5,
+ prefix=colorstr('val: '))[0]
+
+ if not resume:
+ labels = np.concatenate(dataset.labels, 0)
+ # c = torch.tensor(labels[:, 0]) # classes
+ # cf = torch.bincount(c.long(), minlength=nc) + 1. # frequency
+ # model._initialize_biases(cf.to(device))
+ if plots:
+ plot_labels(labels, names, save_dir)
+
+ # Anchors
+ if not opt.noautoanchor:
+ check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz)
+ model.half().float() # pre-reduce anchor precision
+
+ callbacks.run('on_pretrain_routine_end')
+
+ # DDP mode
+ if cuda and RANK != -1:
+ model = DDP(model, device_ids=[LOCAL_RANK], output_device=LOCAL_RANK)
+
+ # Model attributes
+ nl = de_parallel(model).model[-1].nl # number of detection layers (to scale hyps)
+ hyp['box'] *= 3 / nl # scale to layers
+ hyp['cls'] *= nc / 80 * 3 / nl # scale to classes and layers
+ hyp['obj'] *= (imgsz / 640) ** 2 * 3 / nl # scale to image size and layers
+ hyp['label_smoothing'] = opt.label_smoothing
+ model.nc = nc # attach number of classes to model
+ model.hyp = hyp # attach hyperparameters to model
+ model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device) * nc # attach class weights
+ model.names = names
+
+ # Start training
+ t0 = time.time()
+ nw = max(round(hyp['warmup_epochs'] * nb), 100) # number of warmup iterations, max(3 epochs, 100 iterations)
+ # nw = min(nw, (epochs - start_epoch) / 2 * nb) # limit warmup to < 1/2 of training
+ last_opt_step = -1
+ maps = np.zeros(nc) # mAP per class
+ results = (0, 0, 0, 0, 0, 0, 0) # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls)
+ scheduler.last_epoch = start_epoch - 1 # do not move
+ scaler = amp.GradScaler(enabled=cuda)
+ stopper = EarlyStopping(patience=opt.patience)
+ compute_loss = ComputeLoss(model) # init loss class
+ callbacks.run('on_train_start')
+ LOGGER.info(f'Image sizes {imgsz} train, {imgsz} val\n'
+ f'Using {train_loader.num_workers * WORLD_SIZE} dataloader workers\n'
+ f"Logging results to {colorstr('bold', save_dir)}\n"
+ f'Starting training for {epochs} epochs...')
+ for epoch in range(start_epoch, epochs): # epoch ------------------------------------------------------------------
+ callbacks.run('on_train_epoch_start')
+ model.train()
+
+ # Update image weights (optional, single-GPU only)
+ if opt.image_weights:
+ cw = model.class_weights.cpu().numpy() * (1 - maps) ** 2 / nc # class weights
+ iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw) # image weights
+ dataset.indices = random.choices(range(dataset.n), weights=iw, k=dataset.n) # rand weighted idx
+
+ # Update mosaic border (optional)
+ # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs)
+ # dataset.mosaic_border = [b - imgsz, -b] # height, width borders
+
+ mloss = torch.zeros(3, device=device) # mean losses
+ if RANK != -1:
+ train_loader.sampler.set_epoch(epoch)
+ pbar = enumerate(train_loader)
+ LOGGER.info(('\n' + '%10s' * 7) % ('Epoch', 'gpu_mem', 'box', 'obj', 'cls', 'labels', 'img_size'))
+ if RANK in (-1, 0):
+ pbar = tqdm(pbar, total=nb, bar_format=' {l_bar}{bar:10}{r_bar}{bar:-10b}') # progress bar FIXED TO ALIGN WITH OTHER BAR :)
+ optimizer.zero_grad()
+ for i, (imgs, targets, paths, _) in pbar: # batch -------------------------------------------------------------
+ callbacks.run('on_train_batch_start')
+ ni = i + nb * epoch # number integrated batches (since train start)
+ imgs = imgs.to(device, non_blocking=True).float() / 255 # uint8 to float32, 0-255 to 0.0-1.0
+
+ # Warmup
+ if ni <= nw:
+ xi = [0, nw] # x interp
+ # compute_loss.gr = np.interp(ni, xi, [0.0, 1.0]) # iou loss ratio (obj_loss = 1.0 or iou)
+ accumulate = max(1, np.interp(ni, xi, [1, nbs / batch_size]).round())
+ for j, x in enumerate(optimizer.param_groups):
+ # bias lr falls from 0.1 to lr0, all other lrs rise from 0.0 to lr0
+ x['lr'] = np.interp(ni, xi, [hyp['warmup_bias_lr'] if j == 2 else 0.0, x['initial_lr'] * lf(epoch)])
+ if 'momentum' in x:
+ x['momentum'] = np.interp(ni, xi, [hyp['warmup_momentum'], hyp['momentum']])
+
+ # Multi-scale
+ if opt.multi_scale:
+ sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs # size
+ sf = sz / max(imgs.shape[2:]) # scale factor
+ if sf != 1:
+ ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:]] # new shape (stretched to gs-multiple)
+ imgs = nn.functional.interpolate(imgs, size=ns, mode='bilinear', align_corners=False)
+
+ # Forward
+ with amp.autocast(enabled=cuda):
+ pred = model(imgs) # forward
+ loss, loss_items = compute_loss(pred, targets.to(device)) # loss scaled by batch_size
+ if RANK != -1:
+ loss *= WORLD_SIZE # gradient averaged between devices in DDP mode
+ if opt.quad:
+ loss *= 4.
+
+ # Backward
+ scaler.scale(loss).backward()
+
+ # Optimize
+ if ni - last_opt_step >= accumulate:
+ scaler.step(optimizer) # optimizer.step
+ scaler.update()
+ optimizer.zero_grad()
+ if ema:
+ ema.update(model)
+ last_opt_step = ni
+
+ # Log
+ if RANK in (-1, 0):
+ mloss = (mloss * i + loss_items) / (i + 1) # update mean losses
+ mem = f'{torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0:.3g}G' # (GB)
+ pbar.set_description(('%10s' * 2 + '%10.4g' * 5) %
+ (f'{epoch}/{epochs - 1}', mem, *mloss, targets.shape[0], imgs.shape[-1]))
+ callbacks.run('on_train_batch_end', ni, model, imgs, targets, paths, plots)
+
+ if callbacks.stop_training:
+ return
+ # end batch ------------------------------------------------------------------------------------------------
+
+ # Scheduler
+ lr = [x['lr'] for x in optimizer.param_groups] # for loggers
+ scheduler.step()
+
+ if RANK in (-1, 0):
+ # mAP
+ callbacks.run('on_train_epoch_end', epoch=epoch)
+ ema.update_attr(model, include=['yaml', 'nc', 'hyp', 'names', 'stride', 'class_weights'])
+ final_epoch = (epoch + 1 == epochs) or stopper.possible_stop
+ if not noval or final_epoch: # Calculate mAP
+ results, maps, _ = val.run(data_dict,
+ batch_size=batch_size // WORLD_SIZE * 2,
+ imgsz=imgsz,
+ model=ema.ema,
+ single_cls=single_cls,
+ dataloader=val_loader,
+ save_dir=save_dir,
+ plots=False,
+ callbacks=callbacks,
+ compute_loss=compute_loss)
+
+ # Update best mAP
+ fi = fitness(np.array(results).reshape(1, -1)) # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+ if fi > best_fitness:
+ best_fitness = fi
+ log_vals = list(mloss) + list(results) + lr
+ callbacks.run('on_fit_epoch_end', log_vals, epoch, best_fitness, fi)
+
+ # Save model
+ if (not nosave) or (final_epoch and not evolve): # if save
+ ckpt = {
+ 'epoch': epoch,
+ 'best_fitness': best_fitness,
+ 'model': deepcopy(de_parallel(model)).half(),
+ 'ema': deepcopy(ema.ema).half(),
+ 'updates': ema.updates,
+ 'optimizer': optimizer.state_dict(),
+ 'wandb_id': loggers.wandb.wandb_run.id if loggers.wandb else None,
+ 'date': datetime.now().isoformat()}
+
+ # Save last, best and delete
+ torch.save(ckpt, last)
+ if best_fitness == fi:
+ torch.save(ckpt, best)
+ if (epoch > 0) and (opt.save_period > 0) and (epoch % opt.save_period == 0):
+ torch.save(ckpt, w / f'epoch{epoch}.pt')
+ del ckpt
+ callbacks.run('on_model_save', last, epoch, final_epoch, best_fitness, fi)
+
+ # Stop Single-GPU
+ if RANK == -1 and stopper(epoch=epoch, fitness=fi):
+ break
+
+ # Stop DDP TODO: known issues shttps://github.com/ultralytics/yolov5/pull/4576
+ # stop = stopper(epoch=epoch, fitness=fi)
+ # if RANK == 0:
+ # dist.broadcast_object_list([stop], 0) # broadcast 'stop' to all ranks
+
+ # Stop DPP
+ # with torch_distributed_zero_first(RANK):
+ # if stop:
+ # break # must break all DDP ranks
+
+ # end epoch ----------------------------------------------------------------------------------------------------
+ # end training -----------------------------------------------------------------------------------------------------
+ if RANK in (-1, 0):
+ LOGGER.info(f'\n{epoch - start_epoch + 1} epochs completed in {(time.time() - t0) / 3600:.3f} hours.')
+ for f in last, best:
+ if f.exists():
+ strip_optimizer(f) # strip optimizers
+ if f is best:
+ LOGGER.info(f'\nValidating {f}...')
+ results, _, _ = val.run(
+ data_dict,
+ batch_size=batch_size // WORLD_SIZE * 2,
+ imgsz=imgsz,
+ model=attempt_load(f, device).half(),
+ iou_thres=0.65 if is_coco else 0.60, # best pycocotools results at 0.65
+ single_cls=single_cls,
+ dataloader=val_loader,
+ save_dir=save_dir,
+ save_json=is_coco,
+ verbose=True,
+ plots=plots,
+ callbacks=callbacks,
+ compute_loss=compute_loss) # val best model with plots
+ if is_coco:
+ callbacks.run('on_fit_epoch_end', list(mloss) + list(results) + lr, epoch, best_fitness, fi)
+
+ callbacks.run('on_train_end', last, best, plots, epoch, results)
+ LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}")
+
+ torch.cuda.empty_cache()
+ return results
+
+
+@dataclass
+class TrainOptions:
+ '''
+ Default training values
+ Increase or decrease batch_size, n_gpu, n_workers according to machine specs
+ '''
+ '''
+ python /home/brlab/Dropbox/LM2_Env/YOLOv5/yolov5/train.py
+ --data /home/brlab/Dropbox/LM2_Env/YOLOv5/datasets/PLANT_Botany_Small/PLANT_Botany_Small.yaml
+ --project PLANT_Botany
+ --name Small_Adoxaceae
+ --weights yolov5x6.pt
+ --img 1280
+ --batch 14
+ --epochs 300
+ --cache
+ '''
+ # Parameters
+ data: str = '' #', type=str, default=ROOT / 'data/coco128.yaml', help='dataset.yaml path')
+ project: str = 'runs/train' #', default=ROOT / 'runs/train', help='save to project/name')
+ name: str = 'exp' #', default='exp', help='save to project/name')
+ weights: str = 'yolov5x6.pt' #', type=str, default=ROOT / , help='initial weights path')
+ optimizer: str = 'SGD' #', type=str, choices=['SGD', 'Adam', 'AdamW'], default='SGD', help='optimizer')
+ epochs: int = 300 #', type=int, default=300)
+ batch_size: int = 2 #
+ patience: int = 100 #', type=int, default=100, help='EarlyStopping patience (epochs without improvement)')
+ save_period: int = 10 #', type=int, default=-1, help='Save checkpoint every x epochs (disabled if < 1)')
+ imgsz: int = 1024 #', '--img', '--img-size', type=int, default=640, help='train, val image size (pixels)')
+ workers: int = 8 #', type=int, default=8, help='max dataloader workers (per RANK in DDP mode)')
+ hyp: str = 'data/hyps/hyp.scratch-low.yaml' #', type=str, default=ROOT / 'data/hyps/hyp.scratch-low.yaml', help='hyperparameters path')
+ resume: bool = False # ', nargs='?', const=True, default=False, help='resume most recent training')
+ rect: bool = False # ', action='store_true', help='rectangular training')
+ cache: bool = True #', type=str, nargs='?', const='ram', help='--cache images in "ram" (default) or "disk"')
+ freeze: list = field(init=True,default_factory=list) #', nargs='+', type=int, default=[0], help='Freeze layers: backbone=10, first3=0 1 2')
+ evolve: int = 0 # 100 or 300 ish #', type=int, nargs='?', const=300, help='evolve hyperparameters for x generations')
+ device: int = '' #', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ upload_dataset: bool = False
+ bbox_interval: int = -1
+ artifact_alias: str = 'latest'
+
+ # Recommend using defaults for below:
+ nosave: bool = False # ', action='store_true', help='only save final checkpoint')
+ noval: bool = False # ', action='store_true', help='only validate final epoch')
+ cfg_model: str = '' #', type=str, default='', help='model.yaml path')
+ noautoanchor: bool = False #', action='store_true', help='disable AutoAnchor')
+ noplots: bool = False #', action='store_true', help='save no plot files')
+ bucket: str = '' #', type=str, default='', help='gsutil bucket')
+ image_weights: bool = False #', action='store_true', help='use weighted image selection for training')
+ multi_scale: bool = False #' #', action='store_true', help='vary img-size +/- 50%%')
+ single_cls: bool = False #', action='store_true', help='train multi-class data as single-class')
+ sync_bn: bool = True #', action='store_true', help='use SyncBatchNorm, only available in DDP mode')
+ exist_ok: bool = False #', action='store_true', help='existing project/name ok, do not increment')
+ quad: bool = False #', action='store_true', help='quad dataloader')
+ cos_lr: bool = False #', action='store_true', help='cosine LR scheduler')
+ label_smoothing: float = 0.0 #', type=float, default=0.0, help='Label smoothing epsilon')
+ local_rank: bool = -1 #', type=int, default=-1, help='DDP parameter, do not modify')
+
+ new_time: str = field(init=False)
+
+ path_to_config: str = field(init=False)
+ # path_to_model: str = field(init=False)
+ # path_to_ruler_class_names: str = field(init=False)
+ # path_to_train_json: str = field(init=False)
+ # path_to_val_json: str = field(init=False)
+
+ # dir_images_train: str = field(init=False)
+ # dir_images_val: str = field(init=False)
+ # dir_images_test: str = field(init=False)
+ # dir_out: str = field(init=False)
+
+ base_architecture: str = field(init=False)
+
+ cfg: str = field(init=False)
+
+ w_and_b_key: str = field(init=False)
+ w_and_b_project: str = field(init=False)
+ entity: str = field(init=False)
+
+ def __post_init__(self) -> None:
+ '''
+ Setup
+ '''
+ self.new_time = get_datetime()
+ self.base_architecture = self.weights.split('.')[0]
+ self.freeze = [0]
+ '''
+ Configure names
+ '''
+ dir_home = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
+ path_cfg_private = os.path.join(dir_home,'PRIVATE_DATA.yaml')
+ self.cfg_private = get_cfg_from_full_path(path_cfg_private)
+
+ self.path_to_config = dir_home
+ self.cfg = load_cfg(self.path_to_config)
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['name'] is not None:
+ self.name = self.cfg['leafmachine']['component_detector_train']['model_options']['name']
+ self.name = "__".join([self.name,self.base_architecture,self.new_time])
+ # if self.cfg['leafmachine']['component_detector_train']['dir_out'] is None:
+ # self.dir_out = "__".join([['leafmachine']['component_detector_train']['plant_or_archival'], self.new_time, self.name])
+ # else:
+ # self.dir_out = self.cfg['leafmachine']['component_detector_train']['dir_out']
+ else:
+ self.name = "DEFAULT_NAME"
+ self.name = "__".join([self.name,self.base_architecture,self.new_time])
+ # if self.cfg['leafmachine']['component_detector_train']['dir_out'] is None:
+ # self.dir_out = "__".join(["leaf_seg", self.new_time, self.name])
+ # self.dir_out = os.path.join('models', self.dir_out)
+ # else:
+ # self.dir_out = self.cfg['leafmachine']['component_detector_train']['dir_out']
+
+ '''
+ Weights and Biases Info
+ https://wandb.ai/site
+ '''
+ if self.cfg_private['w_and_b']['w_and_b_key'] is not None:
+ self.w_and_b_key = self.cfg_private['w_and_b']['w_and_b_key']
+
+ if self.cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'PLANT':
+ if self.cfg_private['w_and_b']['plant_component_detector_project'] is not None:
+ self.w_and_b_project = self.cfg_private['w_and_b']['plant_component_detector_project']
+ elif self.cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'ARCHIVAL':
+ if self.cfg_private['w_and_b']['archival_component_detector_project'] is not None:
+ self.w_and_b_project = self.cfg_private['w_and_b']['archival_component_detector_project']
+ elif self.cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'LANDMARK':
+ if self.cfg_private['w_and_b']['landmark_component_detector_project'] is not None:
+ self.w_and_b_project = self.cfg_private['w_and_b']['landmark_component_detector_project']
+ elif self.cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'ARM':
+ if self.cfg_private['w_and_b']['arm_component_detector_project'] is not None:
+ self.w_and_b_project = self.cfg_private['w_and_b']['arm_component_detector_project']
+
+ if self.cfg_private['w_and_b']['entity'] is not None:
+ self.entity = self.cfg_private['w_and_b']['entity']
+
+ '''
+ Model Options
+ '''
+ # Parameters
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['data'] is not None:
+ self.data = self.cfg['leafmachine']['component_detector_train']['model_options']['data']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['project'] is not None:
+ self.project = self.cfg['leafmachine']['component_detector_train']['model_options']['project']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['name'] is not None:
+ self.name = self.cfg['leafmachine']['component_detector_train']['model_options']['name']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['weights'] is not None:
+ self.weights = self.cfg['leafmachine']['component_detector_train']['model_options']['weights']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['optimizer'] is not None:
+ self.optimizer = self.cfg['leafmachine']['component_detector_train']['model_options']['optimizer']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['batch_size'] is not None:
+ self.batch_size = self.cfg['leafmachine']['component_detector_train']['model_options']['batch_size']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['epochs'] is not None:
+ self.epochs = self.cfg['leafmachine']['component_detector_train']['model_options']['epochs']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['patience'] is not None:
+ self.patience = self.cfg['leafmachine']['component_detector_train']['model_options']['patience']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['save_period'] is not None:
+ self.save_period = self.cfg['leafmachine']['component_detector_train']['model_options']['save_period']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['imgsz'] is not None:
+ self.imgsz = self.cfg['leafmachine']['component_detector_train']['model_options']['imgsz']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['workers'] is not None:
+ self.workers = self.cfg['leafmachine']['component_detector_train']['model_options']['workers']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['hyp'] is not None:
+ self.hyp = self.cfg['leafmachine']['component_detector_train']['model_options']['hyp']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['resume'] is not None:
+ self.resume = self.cfg['leafmachine']['component_detector_train']['model_options']['resume']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['rect'] is not None:
+ self.rect = self.cfg['leafmachine']['component_detector_train']['model_options']['rect']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['cache'] is not None:
+ self.cache = self.cfg['leafmachine']['component_detector_train']['model_options']['cache']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['freeze'] is not None:
+ self.freeze = self.cfg['leafmachine']['component_detector_train']['model_options']['freeze']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['evolve'] is not None:
+ self.evolve = self.cfg['leafmachine']['component_detector_train']['model_options']['evolve']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['device'] is not None:
+ self.device = self.cfg['leafmachine']['component_detector_train']['model_options']['device']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['upload_dataset'] is not None:
+ self.upload_dataset = self.cfg['leafmachine']['component_detector_train']['model_options']['upload_dataset']
+ # Recommend using defaults for below:
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['nosave'] is not None:
+ self.nosave = self.cfg['leafmachine']['component_detector_train']['model_options']['nosave']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['noval'] is not None:
+ self.noval = self.cfg['leafmachine']['component_detector_train']['model_options']['noval']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['cfg_model'] is not None:
+ self.cfg_model = self.cfg['leafmachine']['component_detector_train']['model_options']['cfg_model']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['noautoanchor'] is not None:
+ self.noautoanchor = self.cfg['leafmachine']['component_detector_train']['model_options']['noautoanchor']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['noplots'] is not None:
+ self.noplots = self.cfg['leafmachine']['component_detector_train']['model_options']['noplots']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['bucket'] is not None:
+ self.bucket = self.cfg['leafmachine']['component_detector_train']['model_options']['bucket']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['image_weights'] is not None:
+ self.image_weights = self.cfg['leafmachine']['component_detector_train']['model_options']['image_weights']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['multi_scale'] is not None:
+ self.multi_scale = self.cfg['leafmachine']['component_detector_train']['model_options']['multi_scale']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['single_cls'] is not None:
+ self.single_cls = self.cfg['leafmachine']['component_detector_train']['model_options']['single_cls']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['sync_bn'] is not None:
+ self.sync_bn = self.cfg['leafmachine']['component_detector_train']['model_options']['sync_bn']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['exist_ok'] is not None:
+ self.exist_ok = self.cfg['leafmachine']['component_detector_train']['model_options']['exist_ok']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['quad'] is not None:
+ self.quad = self.cfg['leafmachine']['component_detector_train']['model_options']['quad']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['cos_lr'] is not None:
+ self.cos_lr = self.cfg['leafmachine']['component_detector_train']['model_options']['cos_lr']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['label_smoothing'] is not None:
+ self.label_smoothing = self.cfg['leafmachine']['component_detector_train']['model_options']['label_smoothing']
+ if self.cfg['leafmachine']['component_detector_train']['model_options']['local_rank'] is not None:
+ self.local_rank = self.cfg['leafmachine']['component_detector_train']['model_options']['local_rank']
+
+ self.project = os.path.join(ROOT,self.project,self.w_and_b_project,self.name) #', default=ROOT / 'runs/train', help='save to project/name')
+ self.weights = os.path.join(ROOT,self.weights)
+ self.hyp = os.path.join(ROOT,self.hyp)
+
+
+
+
+
+def parse_opt(known=False):
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--weights', type=str, default=ROOT / 'yolov5s.pt', help='initial weights path')
+ parser.add_argument('--cfg-model', type=str, default='', help='model.yaml path')
+ parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='dataset.yaml path')
+ parser.add_argument('--hyp', type=str, default=ROOT / 'data/hyps/hyp.scratch-low.yaml', help='hyperparameters path')
+ parser.add_argument('--epochs', type=int, default=300)
+ parser.add_argument('--batch-size', type=int, default=16, help='total batch size for all GPUs, -1 for autobatch')
+ parser.add_argument('--imgsz', '--img', '--img-size', type=int, default=640, help='train, val image size (pixels)')
+ parser.add_argument('--rect', action='store_true', help='rectangular training')
+ parser.add_argument('--resume', nargs='?', const=True, default=False, help='resume most recent training')
+ parser.add_argument('--nosave', action='store_true', help='only save final checkpoint')
+ parser.add_argument('--noval', action='store_true', help='only validate final epoch')
+ parser.add_argument('--noautoanchor', action='store_true', help='disable AutoAnchor')
+ parser.add_argument('--noplots', action='store_true', help='save no plot files')
+ parser.add_argument('--evolve', type=int, nargs='?', const=300, help='evolve hyperparameters for x generations')
+ parser.add_argument('--bucket', type=str, default='', help='gsutil bucket')
+ parser.add_argument('--cache', type=str, nargs='?', const='ram', help='--cache images in "ram" (default) or "disk"')
+ parser.add_argument('--image-weights', action='store_true', help='use weighted image selection for training')
+ parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--multi-scale', action='store_true', help='vary img-size +/- 50%%')
+ parser.add_argument('--single-cls', action='store_true', help='train multi-class data as single-class')
+ parser.add_argument('--optimizer', type=str, choices=['SGD', 'Adam', 'AdamW'], default='SGD', help='optimizer')
+ parser.add_argument('--sync-bn', action='store_true', help='use SyncBatchNorm, only available in DDP mode')
+ parser.add_argument('--workers', type=int, default=8, help='max dataloader workers (per RANK in DDP mode)')
+ parser.add_argument('--project', default=ROOT / 'runs/train', help='save to project/name')
+ parser.add_argument('--name', default='exp', help='save to project/name')
+ parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+ parser.add_argument('--quad', action='store_true', help='quad dataloader')
+ parser.add_argument('--cos-lr', action='store_true', help='cosine LR scheduler')
+ parser.add_argument('--label-smoothing', type=float, default=0.0, help='Label smoothing epsilon')
+ parser.add_argument('--patience', type=int, default=100, help='EarlyStopping patience (epochs without improvement)')
+ parser.add_argument('--freeze', nargs='+', type=int, default=[0], help='Freeze layers: backbone=10, first3=0 1 2')
+ parser.add_argument('--save-period', type=int, default=-1, help='Save checkpoint every x epochs (disabled if < 1)')
+ parser.add_argument('--local_rank', type=int, default=-1, help='DDP parameter, do not modify')
+
+ # Weights & Biases arguments
+ parser.add_argument('--entity', default=None, help='W&B: Entity')
+ parser.add_argument('--upload_dataset', nargs='?', const=True, default=False, help='W&B: Upload data, "val" option')
+ parser.add_argument('--bbox_interval', type=int, default=-1, help='W&B: Set bounding-box image logging interval')
+ parser.add_argument('--artifact_alias', type=str, default='latest', help='W&B: Version of dataset artifact to use')
+
+ opt = parser.parse_known_args()[0] if known else parser.parse_args()
+ return opt
+
+
+def main(opt, callbacks=Callbacks()):
+
+ # Checks
+ # if RANK in (-1, 0):
+ # print_args(vars(opt))
+ # check_git_status()
+ # check_requirements(exclude=['thop'])
+
+ # Resume
+ if opt.resume and not check_wandb_resume(opt) and not opt.evolve: # resume an interrupted run
+ ckpt = opt.resume if isinstance(opt.resume, str) else get_latest_run() # specified or most recent path
+ assert os.path.isfile(ckpt), 'ERROR: --resume checkpoint does not exist'
+ with open(Path(ckpt).parent.parent / 'opt.yaml', errors='ignore') as f:
+ opt = argparse.Namespace(**yaml.safe_load(f)) # replace
+ opt.cfg_model, opt.weights, opt.resume = '', ckpt, True # reinstate
+ LOGGER.info(f'Resuming training from {ckpt}')
+ else:
+ opt.data, opt.cfg_model, opt.hyp, opt.weights, opt.project = \
+ check_file(opt.data), check_yaml(opt.cfg_model), check_yaml(opt.hyp), str(opt.weights), str(opt.project) # checks
+ assert len(opt.cfg_model) or len(opt.weights), 'either --cfg or --weights must be specified'
+ if opt.evolve:
+ if opt.project == str(ROOT / 'runs/train'): # if default project name, rename to runs/evolve
+ opt.project = str(ROOT / 'runs/evolve')
+ opt.exist_ok, opt.resume = opt.resume, False # pass resume to exist_ok and disable resume
+ if opt.name == 'cfg_model':
+ opt.name = Path(opt.cfg_model).stem # use model.yaml as name
+ opt.save_dir = str(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok))
+
+ # DDP mode
+ device = select_device(opt.device, batch_size=opt.batch_size)
+ if LOCAL_RANK != -1:
+ msg = 'is not compatible with YOLOv5 Multi-GPU DDP training'
+ assert not opt.image_weights, f'--image-weights {msg}'
+ assert not opt.evolve, f'--evolve {msg}'
+ assert opt.batch_size != -1, f'AutoBatch with --batch-size -1 {msg}, please pass a valid --batch-size'
+ assert opt.batch_size % WORLD_SIZE == 0, f'--batch-size {opt.batch_size} must be multiple of WORLD_SIZE'
+ assert torch.cuda.device_count() > LOCAL_RANK, 'insufficient CUDA devices for DDP command'
+ torch.cuda.set_device(LOCAL_RANK)
+ device = torch.device('cuda', LOCAL_RANK)
+ dist.init_process_group(backend="nccl" if dist.is_nccl_available() else "gloo")
+
+ # Train
+ if not opt.evolve:
+ train(opt.hyp, opt, device, callbacks)
+ if WORLD_SIZE > 1 and RANK == 0:
+ LOGGER.info('Destroying process group... ')
+ dist.destroy_process_group()
+
+ # Evolve hyperparameters (optional)
+ else:
+ # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit)
+ meta = {
+ 'lr0': (1, 1e-5, 1e-1), # initial learning rate (SGD=1E-2, Adam=1E-3)
+ 'lrf': (1, 0.01, 1.0), # final OneCycleLR learning rate (lr0 * lrf)
+ 'momentum': (0.3, 0.6, 0.98), # SGD momentum/Adam beta1
+ 'weight_decay': (1, 0.0, 0.001), # optimizer weight decay
+ 'warmup_epochs': (1, 0.0, 5.0), # warmup epochs (fractions ok)
+ 'warmup_momentum': (1, 0.0, 0.95), # warmup initial momentum
+ 'warmup_bias_lr': (1, 0.0, 0.2), # warmup initial bias lr
+ 'box': (1, 0.02, 0.2), # box loss gain
+ 'cls': (1, 0.2, 4.0), # cls loss gain
+ 'cls_pw': (1, 0.5, 2.0), # cls BCELoss positive_weight
+ 'obj': (1, 0.2, 4.0), # obj loss gain (scale with pixels)
+ 'obj_pw': (1, 0.5, 2.0), # obj BCELoss positive_weight
+ 'iou_t': (0, 0.1, 0.7), # IoU training threshold
+ 'anchor_t': (1, 2.0, 8.0), # anchor-multiple threshold
+ 'anchors': (2, 2.0, 10.0), # anchors per output grid (0 to ignore)
+ 'fl_gamma': (0, 0.0, 2.0), # focal loss gamma (efficientDet default gamma=1.5)
+ 'hsv_h': (1, 0.0, 0.1), # image HSV-Hue augmentation (fraction)
+ 'hsv_s': (1, 0.0, 0.9), # image HSV-Saturation augmentation (fraction)
+ 'hsv_v': (1, 0.0, 0.9), # image HSV-Value augmentation (fraction)
+ 'degrees': (1, 0.0, 45.0), # image rotation (+/- deg)
+ 'translate': (1, 0.0, 0.9), # image translation (+/- fraction)
+ 'scale': (1, 0.0, 0.9), # image scale (+/- gain)
+ 'shear': (1, 0.0, 10.0), # image shear (+/- deg)
+ 'perspective': (0, 0.0, 0.001), # image perspective (+/- fraction), range 0-0.001
+ 'flipud': (1, 0.0, 1.0), # image flip up-down (probability)
+ 'fliplr': (0, 0.0, 1.0), # image flip left-right (probability)
+ 'mosaic': (1, 0.0, 1.0), # image mixup (probability)
+ 'mixup': (1, 0.0, 1.0), # image mixup (probability)
+ 'copy_paste': (1, 0.0, 1.0)} # segment copy-paste (probability)
+
+ with open(opt.hyp, errors='ignore') as f:
+ hyp = yaml.safe_load(f) # load hyps dict
+ if 'anchors' not in hyp: # anchors commented in hyp.yaml
+ hyp['anchors'] = 3
+ opt.noval, opt.nosave, save_dir = True, True, Path(opt.save_dir) # only val/save final epoch
+ # ei = [isinstance(x, (int, float)) for x in hyp.values()] # evolvable indices
+ evolve_yaml, evolve_csv = save_dir / 'hyp_evolve.yaml', save_dir / 'evolve.csv'
+ if opt.bucket:
+ os.system(f'gsutil cp gs://{opt.bucket}/evolve.csv {evolve_csv}') # download evolve.csv if exists
+
+ for _ in range(opt.evolve): # generations to evolve
+ if evolve_csv.exists(): # if evolve.csv exists: select best hyps and mutate
+ # Select parent(s)
+ parent = 'single' # parent selection method: 'single' or 'weighted'
+ x = np.loadtxt(evolve_csv, ndmin=2, delimiter=',', skiprows=1)
+ n = min(5, len(x)) # number of previous results to consider
+ x = x[np.argsort(-fitness(x))][:n] # top n mutations
+ w = fitness(x) - fitness(x).min() + 1E-6 # weights (sum > 0)
+ if parent == 'single' or len(x) == 1:
+ # x = x[random.randint(0, n - 1)] # random selection
+ x = x[random.choices(range(n), weights=w)[0]] # weighted selection
+ elif parent == 'weighted':
+ x = (x * w.reshape(n, 1)).sum(0) / w.sum() # weighted combination
+
+ # Mutate
+ mp, s = 0.8, 0.2 # mutation probability, sigma
+ npr = np.random
+ npr.seed(int(time.time()))
+ g = np.array([meta[k][0] for k in hyp.keys()]) # gains 0-1
+ ng = len(meta)
+ v = np.ones(ng)
+ while all(v == 1): # mutate until a change occurs (prevent duplicates)
+ v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0)
+ for i, k in enumerate(hyp.keys()): # plt.hist(v.ravel(), 300)
+ hyp[k] = float(x[i + 7] * v[i]) # mutate
+
+ # Constrain to limits
+ for k, v in meta.items():
+ hyp[k] = max(hyp[k], v[1]) # lower limit
+ hyp[k] = min(hyp[k], v[2]) # upper limit
+ hyp[k] = round(hyp[k], 5) # significant digits
+
+ # Train mutation
+ results = train(hyp.copy(), opt, device, callbacks)
+ callbacks = Callbacks()
+ # Write mutation results
+ print_mutation(results, hyp.copy(), save_dir, opt.bucket)
+
+ # Plot results
+ plot_evolve(evolve_csv)
+ LOGGER.info(f'Hyperparameter evolution finished {opt.evolve} generations\n'
+ f"Results saved to {colorstr('bold', save_dir)}\n"
+ f'Usage example: $ python train.py --hyp {evolve_yaml}')
+
+
+def run(**kwargs):
+ # Usage: import train; train.run(data='coco128.yaml', imgsz=320, weights='yolov5m.pt')
+ try:
+ opts = parse_opt(True)
+ except:
+ opts = TrainOptions()
+
+ for k, v in kwargs.items():
+ setattr(opts, k, v)
+ main(opts)
+ return opts
+
+
+if __name__ == "__main__":
+ dir_home = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
+ path_cfg_private = os.path.join(dir_home,'PRIVATE_DATA.yaml')
+ cfg_private = get_cfg_from_full_path(path_cfg_private)
+ if cfg_private['w_and_b']['w_and_b_key'] is not None:
+ w_and_b_key = cfg_private['w_and_b']['w_and_b_key']
+
+ path_to_config = dir_home
+ cfg = load_cfg(path_to_config)
+
+
+ if cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'PLANT':
+ if cfg_private['w_and_b']['plant_component_detector_project'] is not None:
+ w_and_b_project = cfg_private['w_and_b']['plant_component_detector_project']
+ elif cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'ARCHIVAL':
+ if cfg_private['w_and_b']['archival_component_detector_project'] is not None:
+ w_and_b_project = cfg_private['w_and_b']['archival_component_detector_project']
+ elif cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'LANDMARK':
+ if cfg_private['w_and_b']['landmark_component_detector_project'] is not None:
+ w_and_b_project = cfg_private['w_and_b']['landmark_component_detector_project']
+ elif cfg['leafmachine']['component_detector_train']['plant_or_archival'] == 'ARM':
+ if cfg_private['w_and_b']['arm_component_detector_project'] is not None:
+ w_and_b_project = cfg_private['w_and_b']['arm_component_detector_project']
+
+ if cfg_private['w_and_b']['entity'] is not None:
+ entity = cfg_private['w_and_b']['entity']
+ if cfg['leafmachine']['component_detector_train']['model_options']['name'] is not None:
+ name = cfg['leafmachine']['component_detector_train']['model_options']['name']
+ project = os.path.join(ROOT,w_and_b_project,w_and_b_project,name)
+
+ wandb.init(project=w_and_b_project,name=name, entity=entity)
+ wandb.config = {
+ "learning_rate": 0.001,
+ "epochs": 300,
+ "batch_size": 14,
+ # "upload_dataset": opts.upload_dataset,
+ # "bbox_interval": opts.bbox_interval,
+ # "artifact_alias": opts.artifact_alias
+ }
+ # try:
+ # opts = parse_opt()
+ # except:
+ opts = TrainOptions()
+
+ main(opts)
diff --git a/vouchervision/component_detector/utils/__init__.py b/vouchervision/component_detector/utils/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..d77d75f0be08138e2556d40d83ee650e921379f9
--- /dev/null
+++ b/vouchervision/component_detector/utils/__init__.py
@@ -0,0 +1,123 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+utils/initialization
+"""
+
+
+# def notebook_init(verbose=True):
+# # Check system software and hardware
+# print('Checking setup...')
+
+# import os
+# import shutil
+
+# from utils.general import check_requirements, emojis, is_colab
+# from utils.torch_utils import select_device # imports
+
+# check_requirements(('psutil', 'IPython'))
+# import psutil
+# from IPython import display # to display images and clear console output
+
+# if is_colab():
+# shutil.rmtree('/content/sample_data', ignore_errors=True) # remove colab /sample_data directory
+
+# # System info
+# if verbose:
+# gb = 1 << 30 # bytes to GiB (1024 ** 3)
+# ram = psutil.virtual_memory().total
+# total, used, free = shutil.disk_usage("/")
+# display.clear_output()
+# s = f'({os.cpu_count()} CPUs, {ram / gb:.1f} GB RAM, {(total - free) / gb:.1f}/{total / gb:.1f} GB disk)'
+# else:
+# s = ''
+
+# select_device(newline=False)
+# print(emojis(f'Setup complete ✅ {s}'))
+# return display
+
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+utils/initialization
+"""
+
+import contextlib
+import platform
+import threading
+
+
+def emojis(str=''):
+ # Return platform-dependent emoji-safe version of string
+ return str.encode().decode('ascii', 'ignore') if platform.system() == 'Windows' else str
+
+
+class TryExcept(contextlib.ContextDecorator):
+ # YOLOv5 TryExcept class. Usage: @TryExcept() decorator or 'with TryExcept():' context manager
+ def __init__(self, msg=''):
+ self.msg = msg
+
+ def __enter__(self):
+ pass
+
+ def __exit__(self, exc_type, value, traceback):
+ if value:
+ print(emojis(f"{self.msg}{': ' if self.msg else ''}{value}"))
+ return True
+
+
+def threaded(func):
+ # Multi-threads a target function and returns thread. Usage: @threaded decorator
+ def wrapper(*args, **kwargs):
+ thread = threading.Thread(target=func, args=args, kwargs=kwargs, daemon=True)
+ thread.start()
+ return thread
+
+ return wrapper
+
+
+def join_threads(verbose=False):
+ # Join all daemon threads, i.e. atexit.register(lambda: join_threads())
+ main_thread = threading.current_thread()
+ for t in threading.enumerate():
+ if t is not main_thread:
+ if verbose:
+ print(f'Joining thread {t.name}')
+ t.join()
+
+
+def notebook_init(verbose=True):
+ # Check system software and hardware
+ print('Checking setup...')
+
+ import os
+ import shutil
+
+ from ultralytics.utils.checks import check_requirements
+
+ from utils.general import check_font, is_colab
+ from utils.torch_utils import select_device # imports
+
+ check_font()
+
+ import psutil
+
+ if check_requirements('wandb', install=False):
+ os.system('pip uninstall -y wandb') # eliminate unexpected account creation prompt with infinite hang
+ if is_colab():
+ shutil.rmtree('/content/sample_data', ignore_errors=True) # remove colab /sample_data directory
+
+ # System info
+ display = None
+ if verbose:
+ gb = 1 << 30 # bytes to GiB (1024 ** 3)
+ ram = psutil.virtual_memory().total
+ total, used, free = shutil.disk_usage('/')
+ with contextlib.suppress(Exception): # clear display if ipython is installed
+ from IPython import display
+ display.clear_output()
+ s = f'({os.cpu_count()} CPUs, {ram / gb:.1f} GB RAM, {(total - free) / gb:.1f}/{total / gb:.1f} GB disk)'
+ else:
+ s = ''
+
+ select_device(newline=False)
+ print(emojis(f'Setup complete ✅ {s}'))
+ return display
diff --git a/vouchervision/component_detector/utils/activations.py b/vouchervision/component_detector/utils/activations.py
new file mode 100644
index 0000000000000000000000000000000000000000..084ce8c41230dcde25f0c01311a4c0abcd4584e7
--- /dev/null
+++ b/vouchervision/component_detector/utils/activations.py
@@ -0,0 +1,103 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Activation functions
+"""
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class SiLU(nn.Module):
+ # SiLU activation https://arxiv.org/pdf/1606.08415.pdf
+ @staticmethod
+ def forward(x):
+ return x * torch.sigmoid(x)
+
+
+class Hardswish(nn.Module):
+ # Hard-SiLU activation
+ @staticmethod
+ def forward(x):
+ # return x * F.hardsigmoid(x) # for TorchScript and CoreML
+ return x * F.hardtanh(x + 3, 0.0, 6.0) / 6.0 # for TorchScript, CoreML and ONNX
+
+
+class Mish(nn.Module):
+ # Mish activation https://github.com/digantamisra98/Mish
+ @staticmethod
+ def forward(x):
+ return x * F.softplus(x).tanh()
+
+
+class MemoryEfficientMish(nn.Module):
+ # Mish activation memory-efficient
+ class F(torch.autograd.Function):
+
+ @staticmethod
+ def forward(ctx, x):
+ ctx.save_for_backward(x)
+ return x.mul(torch.tanh(F.softplus(x))) # x * tanh(ln(1 + exp(x)))
+
+ @staticmethod
+ def backward(ctx, grad_output):
+ x = ctx.saved_tensors[0]
+ sx = torch.sigmoid(x)
+ fx = F.softplus(x).tanh()
+ return grad_output * (fx + x * sx * (1 - fx * fx))
+
+ def forward(self, x):
+ return self.F.apply(x)
+
+
+class FReLU(nn.Module):
+ # FReLU activation https://arxiv.org/abs/2007.11824
+ def __init__(self, c1, k=3): # ch_in, kernel
+ super().__init__()
+ self.conv = nn.Conv2d(c1, c1, k, 1, 1, groups=c1, bias=False)
+ self.bn = nn.BatchNorm2d(c1)
+
+ def forward(self, x):
+ return torch.max(x, self.bn(self.conv(x)))
+
+
+class AconC(nn.Module):
+ r""" ACON activation (activate or not)
+ AconC: (p1*x-p2*x) * sigmoid(beta*(p1*x-p2*x)) + p2*x, beta is a learnable parameter
+ according to "Activate or Not: Learning Customized Activation" .
+ """
+
+ def __init__(self, c1):
+ super().__init__()
+ self.p1 = nn.Parameter(torch.randn(1, c1, 1, 1))
+ self.p2 = nn.Parameter(torch.randn(1, c1, 1, 1))
+ self.beta = nn.Parameter(torch.ones(1, c1, 1, 1))
+
+ def forward(self, x):
+ dpx = (self.p1 - self.p2) * x
+ return dpx * torch.sigmoid(self.beta * dpx) + self.p2 * x
+
+
+class MetaAconC(nn.Module):
+ r""" ACON activation (activate or not)
+ MetaAconC: (p1*x-p2*x) * sigmoid(beta*(p1*x-p2*x)) + p2*x, beta is generated by a small network
+ according to "Activate or Not: Learning Customized Activation" .
+ """
+
+ def __init__(self, c1, k=1, s=1, r=16): # ch_in, kernel, stride, r
+ super().__init__()
+ c2 = max(r, c1 // r)
+ self.p1 = nn.Parameter(torch.randn(1, c1, 1, 1))
+ self.p2 = nn.Parameter(torch.randn(1, c1, 1, 1))
+ self.fc1 = nn.Conv2d(c1, c2, k, s, bias=True)
+ self.fc2 = nn.Conv2d(c2, c1, k, s, bias=True)
+ # self.bn1 = nn.BatchNorm2d(c2)
+ # self.bn2 = nn.BatchNorm2d(c1)
+
+ def forward(self, x):
+ y = x.mean(dim=2, keepdims=True).mean(dim=3, keepdims=True)
+ # batch-size 1 bug/instabilities https://github.com/ultralytics/yolov5/issues/2891
+ # beta = torch.sigmoid(self.bn2(self.fc2(self.bn1(self.fc1(y))))) # bug/unstable
+ beta = torch.sigmoid(self.fc2(self.fc1(y))) # bug patch BN layers removed
+ dpx = (self.p1 - self.p2) * x
+ return dpx * torch.sigmoid(beta * dpx) + self.p2 * x
diff --git a/vouchervision/component_detector/utils/augmentations.py b/vouchervision/component_detector/utils/augmentations.py
new file mode 100644
index 0000000000000000000000000000000000000000..3f764c06ae3b366496230bcba63c5e8621ce1c95
--- /dev/null
+++ b/vouchervision/component_detector/utils/augmentations.py
@@ -0,0 +1,284 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Image augmentation functions
+"""
+
+import math
+import random
+
+import cv2
+import numpy as np
+
+from utils.general import LOGGER, check_version, colorstr, resample_segments, segment2box
+from utils.metrics import bbox_ioa
+
+
+class Albumentations:
+ # YOLOv5 Albumentations class (optional, only used if package is installed)
+ def __init__(self):
+ self.transform = None
+ try:
+ import albumentations as A
+ check_version(A.__version__, '1.0.3', hard=True) # version requirement
+
+ T = [
+ A.Blur(p=0.01),
+ A.MedianBlur(p=0.01),
+ A.ToGray(p=0.01),
+ A.CLAHE(p=0.01),
+ A.RandomBrightnessContrast(p=0.0),
+ A.RandomGamma(p=0.0),
+ A.ImageCompression(quality_lower=75, p=0.0)] # transforms
+ self.transform = A.Compose(T, bbox_params=A.BboxParams(format='yolo', label_fields=['class_labels']))
+
+ LOGGER.info(colorstr('albumentations: ') + ', '.join(f'{x}' for x in self.transform.transforms if x.p))
+ except ImportError: # package not installed, skip
+ pass
+ except Exception as e:
+ LOGGER.info(colorstr('albumentations: ') + f'{e}')
+
+ def __call__(self, im, labels, p=1.0):
+ if self.transform and random.random() < p:
+ new = self.transform(image=im, bboxes=labels[:, 1:], class_labels=labels[:, 0]) # transformed
+ im, labels = new['image'], np.array([[c, *b] for c, b in zip(new['class_labels'], new['bboxes'])])
+ return im, labels
+
+
+def augment_hsv(im, hgain=0.5, sgain=0.5, vgain=0.5):
+ # HSV color-space augmentation
+ if hgain or sgain or vgain:
+ r = np.random.uniform(-1, 1, 3) * [hgain, sgain, vgain] + 1 # random gains
+ hue, sat, val = cv2.split(cv2.cvtColor(im, cv2.COLOR_BGR2HSV))
+ dtype = im.dtype # uint8
+
+ x = np.arange(0, 256, dtype=r.dtype)
+ lut_hue = ((x * r[0]) % 180).astype(dtype)
+ lut_sat = np.clip(x * r[1], 0, 255).astype(dtype)
+ lut_val = np.clip(x * r[2], 0, 255).astype(dtype)
+
+ im_hsv = cv2.merge((cv2.LUT(hue, lut_hue), cv2.LUT(sat, lut_sat), cv2.LUT(val, lut_val)))
+ cv2.cvtColor(im_hsv, cv2.COLOR_HSV2BGR, dst=im) # no return needed
+
+
+def hist_equalize(im, clahe=True, bgr=False):
+ # Equalize histogram on BGR image 'im' with im.shape(n,m,3) and range 0-255
+ yuv = cv2.cvtColor(im, cv2.COLOR_BGR2YUV if bgr else cv2.COLOR_RGB2YUV)
+ if clahe:
+ c = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
+ yuv[:, :, 0] = c.apply(yuv[:, :, 0])
+ else:
+ yuv[:, :, 0] = cv2.equalizeHist(yuv[:, :, 0]) # equalize Y channel histogram
+ return cv2.cvtColor(yuv, cv2.COLOR_YUV2BGR if bgr else cv2.COLOR_YUV2RGB) # convert YUV image to RGB
+
+
+def replicate(im, labels):
+ # Replicate labels
+ h, w = im.shape[:2]
+ boxes = labels[:, 1:].astype(int)
+ x1, y1, x2, y2 = boxes.T
+ s = ((x2 - x1) + (y2 - y1)) / 2 # side length (pixels)
+ for i in s.argsort()[:round(s.size * 0.5)]: # smallest indices
+ x1b, y1b, x2b, y2b = boxes[i]
+ bh, bw = y2b - y1b, x2b - x1b
+ yc, xc = int(random.uniform(0, h - bh)), int(random.uniform(0, w - bw)) # offset x, y
+ x1a, y1a, x2a, y2a = [xc, yc, xc + bw, yc + bh]
+ im[y1a:y2a, x1a:x2a] = im[y1b:y2b, x1b:x2b] # im4[ymin:ymax, xmin:xmax]
+ labels = np.append(labels, [[labels[i, 0], x1a, y1a, x2a, y2a]], axis=0)
+
+ return im, labels
+
+
+def letterbox(im, new_shape=(640, 640), color=(114, 114, 114), auto=True, scaleFill=False, scaleup=True, stride=32):
+ # Resize and pad image while meeting stride-multiple constraints
+ shape = im.shape[:2] # current shape [height, width]
+ if isinstance(new_shape, int):
+ new_shape = (new_shape, new_shape)
+
+ # Scale ratio (new / old)
+ r = min(new_shape[0] / shape[0], new_shape[1] / shape[1])
+ if not scaleup: # only scale down, do not scale up (for better val mAP)
+ r = min(r, 1.0)
+
+ # Compute padding
+ ratio = r, r # width, height ratios
+ new_unpad = int(round(shape[1] * r)), int(round(shape[0] * r))
+ dw, dh = new_shape[1] - new_unpad[0], new_shape[0] - new_unpad[1] # wh padding
+ if auto: # minimum rectangle
+ dw, dh = np.mod(dw, stride), np.mod(dh, stride) # wh padding
+ elif scaleFill: # stretch
+ dw, dh = 0.0, 0.0
+ new_unpad = (new_shape[1], new_shape[0])
+ ratio = new_shape[1] / shape[1], new_shape[0] / shape[0] # width, height ratios
+
+ dw /= 2 # divide padding into 2 sides
+ dh /= 2
+
+ if shape[::-1] != new_unpad: # resize
+ im = cv2.resize(im, new_unpad, interpolation=cv2.INTER_LINEAR)
+ top, bottom = int(round(dh - 0.1)), int(round(dh + 0.1))
+ left, right = int(round(dw - 0.1)), int(round(dw + 0.1))
+ im = cv2.copyMakeBorder(im, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color) # add border
+ return im, ratio, (dw, dh)
+
+
+def random_perspective(im,
+ targets=(),
+ segments=(),
+ degrees=10,
+ translate=.1,
+ scale=.1,
+ shear=10,
+ perspective=0.0,
+ border=(0, 0)):
+ # torchvision.transforms.RandomAffine(degrees=(-10, 10), translate=(0.1, 0.1), scale=(0.9, 1.1), shear=(-10, 10))
+ # targets = [cls, xyxy]
+
+ height = im.shape[0] + border[0] * 2 # shape(h,w,c)
+ width = im.shape[1] + border[1] * 2
+
+ # Center
+ C = np.eye(3)
+ C[0, 2] = -im.shape[1] / 2 # x translation (pixels)
+ C[1, 2] = -im.shape[0] / 2 # y translation (pixels)
+
+ # Perspective
+ P = np.eye(3)
+ P[2, 0] = random.uniform(-perspective, perspective) # x perspective (about y)
+ P[2, 1] = random.uniform(-perspective, perspective) # y perspective (about x)
+
+ # Rotation and Scale
+ R = np.eye(3)
+ a = random.uniform(-degrees, degrees)
+ # a += random.choice([-180, -90, 0, 90]) # add 90deg rotations to small rotations
+ s = random.uniform(1 - scale, 1 + scale)
+ # s = 2 ** random.uniform(-scale, scale)
+ R[:2] = cv2.getRotationMatrix2D(angle=a, center=(0, 0), scale=s)
+
+ # Shear
+ S = np.eye(3)
+ S[0, 1] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # x shear (deg)
+ S[1, 0] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # y shear (deg)
+
+ # Translation
+ T = np.eye(3)
+ T[0, 2] = random.uniform(0.5 - translate, 0.5 + translate) * width # x translation (pixels)
+ T[1, 2] = random.uniform(0.5 - translate, 0.5 + translate) * height # y translation (pixels)
+
+ # Combined rotation matrix
+ M = T @ S @ R @ P @ C # order of operations (right to left) is IMPORTANT
+ if (border[0] != 0) or (border[1] != 0) or (M != np.eye(3)).any(): # image changed
+ if perspective:
+ im = cv2.warpPerspective(im, M, dsize=(width, height), borderValue=(114, 114, 114))
+ else: # affine
+ im = cv2.warpAffine(im, M[:2], dsize=(width, height), borderValue=(114, 114, 114))
+
+ # Visualize
+ # import matplotlib.pyplot as plt
+ # ax = plt.subplots(1, 2, figsize=(12, 6))[1].ravel()
+ # ax[0].imshow(im[:, :, ::-1]) # base
+ # ax[1].imshow(im2[:, :, ::-1]) # warped
+
+ # Transform label coordinates
+ n = len(targets)
+ if n:
+ use_segments = any(x.any() for x in segments)
+ new = np.zeros((n, 4))
+ if use_segments: # warp segments
+ segments = resample_segments(segments) # upsample
+ for i, segment in enumerate(segments):
+ xy = np.ones((len(segment), 3))
+ xy[:, :2] = segment
+ xy = xy @ M.T # transform
+ xy = xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2] # perspective rescale or affine
+
+ # clip
+ new[i] = segment2box(xy, width, height)
+
+ else: # warp boxes
+ xy = np.ones((n * 4, 3))
+ xy[:, :2] = targets[:, [1, 2, 3, 4, 1, 4, 3, 2]].reshape(n * 4, 2) # x1y1, x2y2, x1y2, x2y1
+ xy = xy @ M.T # transform
+ xy = (xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2]).reshape(n, 8) # perspective rescale or affine
+
+ # create new boxes
+ x = xy[:, [0, 2, 4, 6]]
+ y = xy[:, [1, 3, 5, 7]]
+ new = np.concatenate((x.min(1), y.min(1), x.max(1), y.max(1))).reshape(4, n).T
+
+ # clip
+ new[:, [0, 2]] = new[:, [0, 2]].clip(0, width)
+ new[:, [1, 3]] = new[:, [1, 3]].clip(0, height)
+
+ # filter candidates
+ i = box_candidates(box1=targets[:, 1:5].T * s, box2=new.T, area_thr=0.01 if use_segments else 0.10)
+ targets = targets[i]
+ targets[:, 1:5] = new[i]
+
+ return im, targets
+
+
+def copy_paste(im, labels, segments, p=0.5):
+ # Implement Copy-Paste augmentation https://arxiv.org/abs/2012.07177, labels as nx5 np.array(cls, xyxy)
+ n = len(segments)
+ if p and n:
+ h, w, c = im.shape # height, width, channels
+ im_new = np.zeros(im.shape, np.uint8)
+ for j in random.sample(range(n), k=round(p * n)):
+ l, s = labels[j], segments[j]
+ box = w - l[3], l[2], w - l[1], l[4]
+ ioa = bbox_ioa(box, labels[:, 1:5]) # intersection over area
+ if (ioa < 0.30).all(): # allow 30% obscuration of existing labels
+ labels = np.concatenate((labels, [[l[0], *box]]), 0)
+ segments.append(np.concatenate((w - s[:, 0:1], s[:, 1:2]), 1))
+ cv2.drawContours(im_new, [segments[j].astype(np.int32)], -1, (255, 255, 255), cv2.FILLED)
+
+ result = cv2.bitwise_and(src1=im, src2=im_new)
+ result = cv2.flip(result, 1) # augment segments (flip left-right)
+ i = result > 0 # pixels to replace
+ # i[:, :] = result.max(2).reshape(h, w, 1) # act over ch
+ im[i] = result[i] # cv2.imwrite('debug.jpg', im) # debug
+
+ return im, labels, segments
+
+
+def cutout(im, labels, p=0.5):
+ # Applies image cutout augmentation https://arxiv.org/abs/1708.04552
+ if random.random() < p:
+ h, w = im.shape[:2]
+ scales = [0.5] * 1 + [0.25] * 2 + [0.125] * 4 + [0.0625] * 8 + [0.03125] * 16 # image size fraction
+ for s in scales:
+ mask_h = random.randint(1, int(h * s)) # create random masks
+ mask_w = random.randint(1, int(w * s))
+
+ # box
+ xmin = max(0, random.randint(0, w) - mask_w // 2)
+ ymin = max(0, random.randint(0, h) - mask_h // 2)
+ xmax = min(w, xmin + mask_w)
+ ymax = min(h, ymin + mask_h)
+
+ # apply random color mask
+ im[ymin:ymax, xmin:xmax] = [random.randint(64, 191) for _ in range(3)]
+
+ # return unobscured labels
+ if len(labels) and s > 0.03:
+ box = np.array([xmin, ymin, xmax, ymax], dtype=np.float32)
+ ioa = bbox_ioa(box, labels[:, 1:5]) # intersection over area
+ labels = labels[ioa < 0.60] # remove >60% obscured labels
+
+ return labels
+
+
+def mixup(im, labels, im2, labels2):
+ # Applies MixUp augmentation https://arxiv.org/pdf/1710.09412.pdf
+ r = np.random.beta(32.0, 32.0) # mixup ratio, alpha=beta=32.0
+ im = (im * r + im2 * (1 - r)).astype(np.uint8)
+ labels = np.concatenate((labels, labels2), 0)
+ return im, labels
+
+
+def box_candidates(box1, box2, wh_thr=2, ar_thr=100, area_thr=0.1, eps=1e-16): # box1(4,n), box2(4,n)
+ # Compute candidate boxes: box1 before augment, box2 after augment, wh_thr (pixels), aspect_ratio_thr, area_ratio
+ w1, h1 = box1[2] - box1[0], box1[3] - box1[1]
+ w2, h2 = box2[2] - box2[0], box2[3] - box2[1]
+ ar = np.maximum(w2 / (h2 + eps), h2 / (w2 + eps)) # aspect ratio
+ return (w2 > wh_thr) & (h2 > wh_thr) & (w2 * h2 / (w1 * h1 + eps) > area_thr) & (ar < ar_thr) # candidates
diff --git a/vouchervision/component_detector/utils/augmentations_torchscript.py b/vouchervision/component_detector/utils/augmentations_torchscript.py
new file mode 100644
index 0000000000000000000000000000000000000000..479599fbb62c168dcc1cd435fbd02970359e06e5
--- /dev/null
+++ b/vouchervision/component_detector/utils/augmentations_torchscript.py
@@ -0,0 +1,397 @@
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+Image augmentation functions
+"""
+
+import math
+import random
+
+import cv2
+import numpy as np
+import torch
+import torchvision.transforms as T
+import torchvision.transforms.functional as TF
+
+from utils.general import LOGGER, check_version, colorstr, resample_segments, segment2box, xywhn2xyxy
+from utils.metrics import bbox_ioa
+
+IMAGENET_MEAN = 0.485, 0.456, 0.406 # RGB mean
+IMAGENET_STD = 0.229, 0.224, 0.225 # RGB standard deviation
+
+
+class Albumentations:
+ # YOLOv5 Albumentations class (optional, only used if package is installed)
+ def __init__(self, size=640):
+ self.transform = None
+ prefix = colorstr('albumentations: ')
+ try:
+ import albumentations as A
+ check_version(A.__version__, '1.0.3', hard=True) # version requirement
+
+ T = [
+ A.RandomResizedCrop(height=size, width=size, scale=(0.8, 1.0), ratio=(0.9, 1.11), p=0.0),
+ A.Blur(p=0.01),
+ A.MedianBlur(p=0.01),
+ A.ToGray(p=0.01),
+ A.CLAHE(p=0.01),
+ A.RandomBrightnessContrast(p=0.0),
+ A.RandomGamma(p=0.0),
+ A.ImageCompression(quality_lower=75, p=0.0)] # transforms
+ self.transform = A.Compose(T, bbox_params=A.BboxParams(format='yolo', label_fields=['class_labels']))
+
+ LOGGER.info(prefix + ', '.join(f'{x}'.replace('always_apply=False, ', '') for x in T if x.p))
+ except ImportError: # package not installed, skip
+ pass
+ except Exception as e:
+ LOGGER.info(f'{prefix}{e}')
+
+ def __call__(self, im, labels, p=1.0):
+ if self.transform and random.random() < p:
+ new = self.transform(image=im, bboxes=labels[:, 1:], class_labels=labels[:, 0]) # transformed
+ im, labels = new['image'], np.array([[c, *b] for c, b in zip(new['class_labels'], new['bboxes'])])
+ return im, labels
+
+
+def normalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD, inplace=False):
+ # Denormalize RGB images x per ImageNet stats in BCHW format, i.e. = (x - mean) / std
+ return TF.normalize(x, mean, std, inplace=inplace)
+
+
+def denormalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD):
+ # Denormalize RGB images x per ImageNet stats in BCHW format, i.e. = x * std + mean
+ for i in range(3):
+ x[:, i] = x[:, i] * std[i] + mean[i]
+ return x
+
+
+def augment_hsv(im, hgain=0.5, sgain=0.5, vgain=0.5):
+ # HSV color-space augmentation
+ if hgain or sgain or vgain:
+ r = np.random.uniform(-1, 1, 3) * [hgain, sgain, vgain] + 1 # random gains
+ hue, sat, val = cv2.split(cv2.cvtColor(im, cv2.COLOR_BGR2HSV))
+ dtype = im.dtype # uint8
+
+ x = np.arange(0, 256, dtype=r.dtype)
+ lut_hue = ((x * r[0]) % 180).astype(dtype)
+ lut_sat = np.clip(x * r[1], 0, 255).astype(dtype)
+ lut_val = np.clip(x * r[2], 0, 255).astype(dtype)
+
+ im_hsv = cv2.merge((cv2.LUT(hue, lut_hue), cv2.LUT(sat, lut_sat), cv2.LUT(val, lut_val)))
+ cv2.cvtColor(im_hsv, cv2.COLOR_HSV2BGR, dst=im) # no return needed
+
+
+def hist_equalize(im, clahe=True, bgr=False):
+ # Equalize histogram on BGR image 'im' with im.shape(n,m,3) and range 0-255
+ yuv = cv2.cvtColor(im, cv2.COLOR_BGR2YUV if bgr else cv2.COLOR_RGB2YUV)
+ if clahe:
+ c = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
+ yuv[:, :, 0] = c.apply(yuv[:, :, 0])
+ else:
+ yuv[:, :, 0] = cv2.equalizeHist(yuv[:, :, 0]) # equalize Y channel histogram
+ return cv2.cvtColor(yuv, cv2.COLOR_YUV2BGR if bgr else cv2.COLOR_YUV2RGB) # convert YUV image to RGB
+
+
+def replicate(im, labels):
+ # Replicate labels
+ h, w = im.shape[:2]
+ boxes = labels[:, 1:].astype(int)
+ x1, y1, x2, y2 = boxes.T
+ s = ((x2 - x1) + (y2 - y1)) / 2 # side length (pixels)
+ for i in s.argsort()[:round(s.size * 0.5)]: # smallest indices
+ x1b, y1b, x2b, y2b = boxes[i]
+ bh, bw = y2b - y1b, x2b - x1b
+ yc, xc = int(random.uniform(0, h - bh)), int(random.uniform(0, w - bw)) # offset x, y
+ x1a, y1a, x2a, y2a = [xc, yc, xc + bw, yc + bh]
+ im[y1a:y2a, x1a:x2a] = im[y1b:y2b, x1b:x2b] # im4[ymin:ymax, xmin:xmax]
+ labels = np.append(labels, [[labels[i, 0], x1a, y1a, x2a, y2a]], axis=0)
+
+ return im, labels
+
+
+def letterbox(im, new_shape=(640, 640), color=(114, 114, 114), auto=True, scaleFill=False, scaleup=True, stride=32):
+ # Resize and pad image while meeting stride-multiple constraints
+ shape = im.shape[:2] # current shape [height, width]
+ if isinstance(new_shape, int):
+ new_shape = (new_shape, new_shape)
+
+ # Scale ratio (new / old)
+ r = min(new_shape[0] / shape[0], new_shape[1] / shape[1])
+ if not scaleup: # only scale down, do not scale up (for better val mAP)
+ r = min(r, 1.0)
+
+ # Compute padding
+ ratio = r, r # width, height ratios
+ new_unpad = int(round(shape[1] * r)), int(round(shape[0] * r))
+ dw, dh = new_shape[1] - new_unpad[0], new_shape[0] - new_unpad[1] # wh padding
+ if auto: # minimum rectangle
+ dw, dh = np.mod(dw, stride), np.mod(dh, stride) # wh padding
+ elif scaleFill: # stretch
+ dw, dh = 0.0, 0.0
+ new_unpad = (new_shape[1], new_shape[0])
+ ratio = new_shape[1] / shape[1], new_shape[0] / shape[0] # width, height ratios
+
+ dw /= 2 # divide padding into 2 sides
+ dh /= 2
+
+ if shape[::-1] != new_unpad: # resize
+ im = cv2.resize(im, new_unpad, interpolation=cv2.INTER_LINEAR)
+ top, bottom = int(round(dh - 0.1)), int(round(dh + 0.1))
+ left, right = int(round(dw - 0.1)), int(round(dw + 0.1))
+ im = cv2.copyMakeBorder(im, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color) # add border
+ return im, ratio, (dw, dh)
+
+
+def random_perspective(im,
+ targets=(),
+ segments=(),
+ degrees=10,
+ translate=.1,
+ scale=.1,
+ shear=10,
+ perspective=0.0,
+ border=(0, 0)):
+ # torchvision.transforms.RandomAffine(degrees=(-10, 10), translate=(0.1, 0.1), scale=(0.9, 1.1), shear=(-10, 10))
+ # targets = [cls, xyxy]
+
+ height = im.shape[0] + border[0] * 2 # shape(h,w,c)
+ width = im.shape[1] + border[1] * 2
+
+ # Center
+ C = np.eye(3)
+ C[0, 2] = -im.shape[1] / 2 # x translation (pixels)
+ C[1, 2] = -im.shape[0] / 2 # y translation (pixels)
+
+ # Perspective
+ P = np.eye(3)
+ P[2, 0] = random.uniform(-perspective, perspective) # x perspective (about y)
+ P[2, 1] = random.uniform(-perspective, perspective) # y perspective (about x)
+
+ # Rotation and Scale
+ R = np.eye(3)
+ a = random.uniform(-degrees, degrees)
+ # a += random.choice([-180, -90, 0, 90]) # add 90deg rotations to small rotations
+ s = random.uniform(1 - scale, 1 + scale)
+ # s = 2 ** random.uniform(-scale, scale)
+ R[:2] = cv2.getRotationMatrix2D(angle=a, center=(0, 0), scale=s)
+
+ # Shear
+ S = np.eye(3)
+ S[0, 1] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # x shear (deg)
+ S[1, 0] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # y shear (deg)
+
+ # Translation
+ T = np.eye(3)
+ T[0, 2] = random.uniform(0.5 - translate, 0.5 + translate) * width # x translation (pixels)
+ T[1, 2] = random.uniform(0.5 - translate, 0.5 + translate) * height # y translation (pixels)
+
+ # Combined rotation matrix
+ M = T @ S @ R @ P @ C # order of operations (right to left) is IMPORTANT
+ if (border[0] != 0) or (border[1] != 0) or (M != np.eye(3)).any(): # image changed
+ if perspective:
+ im = cv2.warpPerspective(im, M, dsize=(width, height), borderValue=(114, 114, 114))
+ else: # affine
+ im = cv2.warpAffine(im, M[:2], dsize=(width, height), borderValue=(114, 114, 114))
+
+ # Visualize
+ # import matplotlib.pyplot as plt
+ # ax = plt.subplots(1, 2, figsize=(12, 6))[1].ravel()
+ # ax[0].imshow(im[:, :, ::-1]) # base
+ # ax[1].imshow(im2[:, :, ::-1]) # warped
+
+ # Transform label coordinates
+ n = len(targets)
+ if n:
+ use_segments = any(x.any() for x in segments) and len(segments) == n
+ new = np.zeros((n, 4))
+ if use_segments: # warp segments
+ segments = resample_segments(segments) # upsample
+ for i, segment in enumerate(segments):
+ xy = np.ones((len(segment), 3))
+ xy[:, :2] = segment
+ xy = xy @ M.T # transform
+ xy = xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2] # perspective rescale or affine
+
+ # clip
+ new[i] = segment2box(xy, width, height)
+
+ else: # warp boxes
+ xy = np.ones((n * 4, 3))
+ xy[:, :2] = targets[:, [1, 2, 3, 4, 1, 4, 3, 2]].reshape(n * 4, 2) # x1y1, x2y2, x1y2, x2y1
+ xy = xy @ M.T # transform
+ xy = (xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2]).reshape(n, 8) # perspective rescale or affine
+
+ # create new boxes
+ x = xy[:, [0, 2, 4, 6]]
+ y = xy[:, [1, 3, 5, 7]]
+ new = np.concatenate((x.min(1), y.min(1), x.max(1), y.max(1))).reshape(4, n).T
+
+ # clip
+ new[:, [0, 2]] = new[:, [0, 2]].clip(0, width)
+ new[:, [1, 3]] = new[:, [1, 3]].clip(0, height)
+
+ # filter candidates
+ i = box_candidates(box1=targets[:, 1:5].T * s, box2=new.T, area_thr=0.01 if use_segments else 0.10)
+ targets = targets[i]
+ targets[:, 1:5] = new[i]
+
+ return im, targets
+
+
+def copy_paste(im, labels, segments, p=0.5):
+ # Implement Copy-Paste augmentation https://arxiv.org/abs/2012.07177, labels as nx5 np.array(cls, xyxy)
+ n = len(segments)
+ if p and n:
+ h, w, c = im.shape # height, width, channels
+ im_new = np.zeros(im.shape, np.uint8)
+ for j in random.sample(range(n), k=round(p * n)):
+ l, s = labels[j], segments[j]
+ box = w - l[3], l[2], w - l[1], l[4]
+ ioa = bbox_ioa(box, labels[:, 1:5]) # intersection over area
+ if (ioa < 0.30).all(): # allow 30% obscuration of existing labels
+ labels = np.concatenate((labels, [[l[0], *box]]), 0)
+ segments.append(np.concatenate((w - s[:, 0:1], s[:, 1:2]), 1))
+ cv2.drawContours(im_new, [segments[j].astype(np.int32)], -1, (1, 1, 1), cv2.FILLED)
+
+ result = cv2.flip(im, 1) # augment segments (flip left-right)
+ i = cv2.flip(im_new, 1).astype(bool)
+ im[i] = result[i] # cv2.imwrite('debug.jpg', im) # debug
+
+ return im, labels, segments
+
+
+def cutout(im, labels, p=0.5):
+ # Applies image cutout augmentation https://arxiv.org/abs/1708.04552
+ if random.random() < p:
+ h, w = im.shape[:2]
+ scales = [0.5] * 1 + [0.25] * 2 + [0.125] * 4 + [0.0625] * 8 + [0.03125] * 16 # image size fraction
+ for s in scales:
+ mask_h = random.randint(1, int(h * s)) # create random masks
+ mask_w = random.randint(1, int(w * s))
+
+ # box
+ xmin = max(0, random.randint(0, w) - mask_w // 2)
+ ymin = max(0, random.randint(0, h) - mask_h // 2)
+ xmax = min(w, xmin + mask_w)
+ ymax = min(h, ymin + mask_h)
+
+ # apply random color mask
+ im[ymin:ymax, xmin:xmax] = [random.randint(64, 191) for _ in range(3)]
+
+ # return unobscured labels
+ if len(labels) and s > 0.03:
+ box = np.array([xmin, ymin, xmax, ymax], dtype=np.float32)
+ ioa = bbox_ioa(box, xywhn2xyxy(labels[:, 1:5], w, h)) # intersection over area
+ labels = labels[ioa < 0.60] # remove >60% obscured labels
+
+ return labels
+
+
+def mixup(im, labels, im2, labels2):
+ # Applies MixUp augmentation https://arxiv.org/pdf/1710.09412.pdf
+ r = np.random.beta(32.0, 32.0) # mixup ratio, alpha=beta=32.0
+ im = (im * r + im2 * (1 - r)).astype(np.uint8)
+ labels = np.concatenate((labels, labels2), 0)
+ return im, labels
+
+
+def box_candidates(box1, box2, wh_thr=2, ar_thr=100, area_thr=0.1, eps=1e-16): # box1(4,n), box2(4,n)
+ # Compute candidate boxes: box1 before augment, box2 after augment, wh_thr (pixels), aspect_ratio_thr, area_ratio
+ w1, h1 = box1[2] - box1[0], box1[3] - box1[1]
+ w2, h2 = box2[2] - box2[0], box2[3] - box2[1]
+ ar = np.maximum(w2 / (h2 + eps), h2 / (w2 + eps)) # aspect ratio
+ return (w2 > wh_thr) & (h2 > wh_thr) & (w2 * h2 / (w1 * h1 + eps) > area_thr) & (ar < ar_thr) # candidates
+
+
+def classify_albumentations(
+ augment=True,
+ size=224,
+ scale=(0.08, 1.0),
+ ratio=(0.75, 1.0 / 0.75), # 0.75, 1.33
+ hflip=0.5,
+ vflip=0.0,
+ jitter=0.4,
+ mean=IMAGENET_MEAN,
+ std=IMAGENET_STD,
+ auto_aug=False):
+ # YOLOv5 classification Albumentations (optional, only used if package is installed)
+ prefix = colorstr('albumentations: ')
+ try:
+ import albumentations as A
+ from albumentations.pytorch import ToTensorV2
+ check_version(A.__version__, '1.0.3', hard=True) # version requirement
+ if augment: # Resize and crop
+ T = [A.RandomResizedCrop(height=size, width=size, scale=scale, ratio=ratio)]
+ if auto_aug:
+ # TODO: implement AugMix, AutoAug & RandAug in albumentation
+ LOGGER.info(f'{prefix}auto augmentations are currently not supported')
+ else:
+ if hflip > 0:
+ T += [A.HorizontalFlip(p=hflip)]
+ if vflip > 0:
+ T += [A.VerticalFlip(p=vflip)]
+ if jitter > 0:
+ color_jitter = (float(jitter), ) * 3 # repeat value for brightness, contrast, satuaration, 0 hue
+ T += [A.ColorJitter(*color_jitter, 0)]
+ else: # Use fixed crop for eval set (reproducibility)
+ T = [A.SmallestMaxSize(max_size=size), A.CenterCrop(height=size, width=size)]
+ T += [A.Normalize(mean=mean, std=std), ToTensorV2()] # Normalize and convert to Tensor
+ LOGGER.info(prefix + ', '.join(f'{x}'.replace('always_apply=False, ', '') for x in T if x.p))
+ return A.Compose(T)
+
+ except ImportError: # package not installed, skip
+ LOGGER.warning(f'{prefix}⚠️ not found, install with `pip install albumentations` (recommended)')
+ except Exception as e:
+ LOGGER.info(f'{prefix}{e}')
+
+
+def classify_transforms(size=224):
+ # Transforms to apply if albumentations not installed
+ assert isinstance(size, int), f'ERROR: classify_transforms size {size} must be integer, not (list, tuple)'
+ # T.Compose([T.ToTensor(), T.Resize(size), T.CenterCrop(size), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)])
+ return T.Compose([CenterCrop(size), ToTensor(), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)])
+
+
+class LetterBox:
+ # YOLOv5 LetterBox class for image preprocessing, i.e. T.Compose([LetterBox(size), ToTensor()])
+ def __init__(self, size=(640, 640), auto=False, stride=32):
+ super().__init__()
+ self.h, self.w = (size, size) if isinstance(size, int) else size
+ self.auto = auto # pass max size integer, automatically solve for short side using stride
+ self.stride = stride # used with auto
+
+ def __call__(self, im): # im = np.array HWC
+ imh, imw = im.shape[:2]
+ r = min(self.h / imh, self.w / imw) # ratio of new/old
+ h, w = round(imh * r), round(imw * r) # resized image
+ hs, ws = (math.ceil(x / self.stride) * self.stride for x in (h, w)) if self.auto else self.h, self.w
+ top, left = round((hs - h) / 2 - 0.1), round((ws - w) / 2 - 0.1)
+ im_out = np.full((self.h, self.w, 3), 114, dtype=im.dtype)
+ im_out[top:top + h, left:left + w] = cv2.resize(im, (w, h), interpolation=cv2.INTER_LINEAR)
+ return im_out
+
+
+class CenterCrop:
+ # YOLOv5 CenterCrop class for image preprocessing, i.e. T.Compose([CenterCrop(size), ToTensor()])
+ def __init__(self, size=640):
+ super().__init__()
+ self.h, self.w = (size, size) if isinstance(size, int) else size
+
+ def __call__(self, im): # im = np.array HWC
+ imh, imw = im.shape[:2]
+ m = min(imh, imw) # min dimension
+ top, left = (imh - m) // 2, (imw - m) // 2
+ return cv2.resize(im[top:top + m, left:left + m], (self.w, self.h), interpolation=cv2.INTER_LINEAR)
+
+
+class ToTensor:
+ # YOLOv5 ToTensor class for image preprocessing, i.e. T.Compose([LetterBox(size), ToTensor()])
+ def __init__(self, half=False):
+ super().__init__()
+ self.half = half
+
+ def __call__(self, im): # im = np.array HWC in BGR order
+ im = np.ascontiguousarray(im.transpose((2, 0, 1))[::-1]) # HWC to CHW -> BGR to RGB -> contiguous
+ im = torch.from_numpy(im) # to torch
+ im = im.half() if self.half else im.float() # uint8 to fp16/32
+ im /= 255.0 # 0-255 to 0.0-1.0
+ return im
\ No newline at end of file
diff --git a/vouchervision/component_detector/utils/autoanchor.py b/vouchervision/component_detector/utils/autoanchor.py
new file mode 100644
index 0000000000000000000000000000000000000000..cdcecd855a5103ebdc7033308dd372320f9e10e8
--- /dev/null
+++ b/vouchervision/component_detector/utils/autoanchor.py
@@ -0,0 +1,170 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+AutoAnchor utils
+"""
+
+import random
+
+import numpy as np
+import torch
+import yaml
+from tqdm.auto import tqdm
+
+from utils.general import LOGGER, colorstr, emojis
+
+PREFIX = colorstr('AutoAnchor: ')
+
+
+def check_anchor_order(m):
+ # Check anchor order against stride order for YOLOv5 Detect() module m, and correct if necessary
+ a = m.anchors.prod(-1).mean(-1).view(-1) # mean anchor area per output layer
+ da = a[-1] - a[0] # delta a
+ ds = m.stride[-1] - m.stride[0] # delta s
+ if da and (da.sign() != ds.sign()): # same order
+ LOGGER.info(f'{PREFIX}Reversing anchor order')
+ m.anchors[:] = m.anchors.flip(0)
+
+
+def check_anchors(dataset, model, thr=4.0, imgsz=640):
+ # Check anchor fit to data, recompute if necessary
+ m = model.module.model[-1] if hasattr(model, 'module') else model.model[-1] # Detect()
+ shapes = imgsz * dataset.shapes / dataset.shapes.max(1, keepdims=True)
+ scale = np.random.uniform(0.9, 1.1, size=(shapes.shape[0], 1)) # augment scale
+ wh = torch.tensor(np.concatenate([l[:, 3:5] * s for s, l in zip(shapes * scale, dataset.labels)])).float() # wh
+
+ def metric(k): # compute metric
+ r = wh[:, None] / k[None]
+ x = torch.min(r, 1 / r).min(2)[0] # ratio metric
+ best = x.max(1)[0] # best_x
+ aat = (x > 1 / thr).float().sum(1).mean() # anchors above threshold
+ bpr = (best > 1 / thr).float().mean() # best possible recall
+ return bpr, aat
+
+ stride = m.stride.to(m.anchors.device).view(-1, 1, 1) # model strides
+ anchors = m.anchors.clone() * stride # current anchors
+ bpr, aat = metric(anchors.cpu().view(-1, 2))
+ s = f'\n{PREFIX}{aat:.2f} anchors/target, {bpr:.3f} Best Possible Recall (BPR). '
+ if bpr > 0.98: # threshold to recompute
+ LOGGER.info(emojis(f'{s}Current anchors are a good fit to dataset ✅'))
+ else:
+ LOGGER.info(emojis(f'{s}Anchors are a poor fit to dataset ⚠️, attempting to improve...'))
+ na = m.anchors.numel() // 2 # number of anchors
+ try:
+ anchors = kmean_anchors(dataset, n=na, img_size=imgsz, thr=thr, gen=1000, verbose=False)
+ except Exception as e:
+ LOGGER.info(f'{PREFIX}ERROR: {e}')
+ new_bpr = metric(anchors)[0]
+ if new_bpr > bpr: # replace anchors
+ anchors = torch.tensor(anchors, device=m.anchors.device).type_as(m.anchors)
+ m.anchors[:] = anchors.clone().view_as(m.anchors)
+ check_anchor_order(m) # must be in pixel-space (not grid-space)
+ m.anchors /= stride
+ s = f'{PREFIX}Done ✅ (optional: update model *.yaml to use these anchors in the future)'
+ else:
+ s = f'{PREFIX}Done ⚠️ (original anchors better than new anchors, proceeding with original anchors)'
+ LOGGER.info(emojis(s))
+
+
+def kmean_anchors(dataset='./data/coco128.yaml', n=9, img_size=640, thr=4.0, gen=1000, verbose=True):
+ """ Creates kmeans-evolved anchors from training dataset
+
+ Arguments:
+ dataset: path to data.yaml, or a loaded dataset
+ n: number of anchors
+ img_size: image size used for training
+ thr: anchor-label wh ratio threshold hyperparameter hyp['anchor_t'] used for training, default=4.0
+ gen: generations to evolve anchors using genetic algorithm
+ verbose: print all results
+
+ Return:
+ k: kmeans evolved anchors
+
+ Usage:
+ from utils.autoanchor import *; _ = kmean_anchors()
+ """
+ from scipy.cluster.vq import kmeans
+
+ npr = np.random
+ thr = 1 / thr
+
+ def metric(k, wh): # compute metrics
+ r = wh[:, None] / k[None]
+ x = torch.min(r, 1 / r).min(2)[0] # ratio metric
+ # x = wh_iou(wh, torch.tensor(k)) # iou metric
+ return x, x.max(1)[0] # x, best_x
+
+ def anchor_fitness(k): # mutation fitness
+ _, best = metric(torch.tensor(k, dtype=torch.float32), wh)
+ return (best * (best > thr).float()).mean() # fitness
+
+ def print_results(k, verbose=True):
+ k = k[np.argsort(k.prod(1))] # sort small to large
+ x, best = metric(k, wh0)
+ bpr, aat = (best > thr).float().mean(), (x > thr).float().mean() * n # best possible recall, anch > thr
+ s = f'{PREFIX}thr={thr:.2f}: {bpr:.4f} best possible recall, {aat:.2f} anchors past thr\n' \
+ f'{PREFIX}n={n}, img_size={img_size}, metric_all={x.mean():.3f}/{best.mean():.3f}-mean/best, ' \
+ f'past_thr={x[x > thr].mean():.3f}-mean: '
+ for i, x in enumerate(k):
+ s += '%i,%i, ' % (round(x[0]), round(x[1]))
+ if verbose:
+ LOGGER.info(s[:-2])
+ return k
+
+ if isinstance(dataset, str): # *.yaml file
+ with open(dataset, errors='ignore') as f:
+ data_dict = yaml.safe_load(f) # model dict
+ from utils.datasets import LoadImagesAndLabels
+ dataset = LoadImagesAndLabels(data_dict['train'], augment=True, rect=True)
+
+ # Get label wh
+ shapes = img_size * dataset.shapes / dataset.shapes.max(1, keepdims=True)
+ wh0 = np.concatenate([l[:, 3:5] * s for s, l in zip(shapes, dataset.labels)]) # wh
+
+ # Filter
+ i = (wh0 < 3.0).any(1).sum()
+ if i:
+ LOGGER.info(f'{PREFIX}WARNING: Extremely small objects found: {i} of {len(wh0)} labels are < 3 pixels in size')
+ wh = wh0[(wh0 >= 2.0).any(1)] # filter > 2 pixels
+ # wh = wh * (npr.rand(wh.shape[0], 1) * 0.9 + 0.1) # multiply by random scale 0-1
+
+ # Kmeans init
+ try:
+ LOGGER.info(f'{PREFIX}Running kmeans for {n} anchors on {len(wh)} points...')
+ assert n <= len(wh) # apply overdetermined constraint
+ s = wh.std(0) # sigmas for whitening
+ k = kmeans(wh / s, n, iter=30)[0] * s # points
+ assert n == len(k) # kmeans may return fewer points than requested if wh is insufficient or too similar
+ except Exception:
+ LOGGER.warning(f'{PREFIX}WARNING: switching strategies from kmeans to random init')
+ k = np.sort(npr.rand(n * 2)).reshape(n, 2) * img_size # random init
+ wh, wh0 = (torch.tensor(x, dtype=torch.float32) for x in (wh, wh0))
+ k = print_results(k, verbose=False)
+
+ # Plot
+ # k, d = [None] * 20, [None] * 20
+ # for i in tqdm(range(1, 21)):
+ # k[i-1], d[i-1] = kmeans(wh / s, i) # points, mean distance
+ # fig, ax = plt.subplots(1, 2, figsize=(14, 7), tight_layout=True)
+ # ax = ax.ravel()
+ # ax[0].plot(np.arange(1, 21), np.array(d) ** 2, marker='.')
+ # fig, ax = plt.subplots(1, 2, figsize=(14, 7)) # plot wh
+ # ax[0].hist(wh[wh[:, 0]<100, 0],400)
+ # ax[1].hist(wh[wh[:, 1]<100, 1],400)
+ # fig.savefig('wh.png', dpi=200)
+
+ # Evolve
+ f, sh, mp, s = anchor_fitness(k), k.shape, 0.9, 0.1 # fitness, generations, mutation prob, sigma
+ pbar = tqdm(range(gen), bar_format='{l_bar}{bar:10}{r_bar}{bar:-10b}') # progress bar
+ for _ in pbar:
+ v = np.ones(sh)
+ while (v == 1).all(): # mutate until a change occurs (prevent duplicates)
+ v = ((npr.random(sh) < mp) * random.random() * npr.randn(*sh) * s + 1).clip(0.3, 3.0)
+ kg = (k.copy() * v).clip(min=2.0)
+ fg = anchor_fitness(kg)
+ if fg > f:
+ f, k = fg, kg.copy()
+ pbar.desc = f'{PREFIX}Evolving anchors with Genetic Algorithm: fitness = {f:.4f}'
+ if verbose:
+ print_results(k, verbose)
+
+ return print_results(k)
diff --git a/vouchervision/component_detector/utils/autobatch.py b/vouchervision/component_detector/utils/autobatch.py
new file mode 100644
index 0000000000000000000000000000000000000000..e53b4787b87df5a46b1df0eb28d8d97bc1f811fd
--- /dev/null
+++ b/vouchervision/component_detector/utils/autobatch.py
@@ -0,0 +1,58 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Auto-batch utils
+"""
+
+from copy import deepcopy
+
+import numpy as np
+import torch
+from torch.cuda import amp
+
+from utils.general import LOGGER, colorstr
+from utils.torch_utils import profile
+
+
+def check_train_batch_size(model, imgsz=640):
+ # Check YOLOv5 training batch size
+ with amp.autocast():
+ return autobatch(deepcopy(model).train(), imgsz) # compute optimal batch size
+
+
+def autobatch(model, imgsz=640, fraction=0.9, batch_size=16):
+ # Automatically estimate best batch size to use `fraction` of available CUDA memory
+ # Usage:
+ # import torch
+ # from utils.autobatch import autobatch
+ # model = torch.hub.load('ultralytics/yolov5', 'yolov5s', autoshape=False)
+ # print(autobatch(model))
+
+ prefix = colorstr('AutoBatch: ')
+ LOGGER.info(f'{prefix}Computing optimal batch size for --imgsz {imgsz}')
+ device = next(model.parameters()).device # get model device
+ if device.type == 'cpu':
+ LOGGER.info(f'{prefix}CUDA not detected, using default CPU batch-size {batch_size}')
+ return batch_size
+
+ gb = 1 << 30 # bytes to GiB (1024 ** 3)
+ d = str(device).upper() # 'CUDA:0'
+ properties = torch.cuda.get_device_properties(device) # device properties
+ t = properties.total_memory / gb # (GiB)
+ r = torch.cuda.memory_reserved(device) / gb # (GiB)
+ a = torch.cuda.memory_allocated(device) / gb # (GiB)
+ f = t - (r + a) # free inside reserved
+ LOGGER.info(f'{prefix}{d} ({properties.name}) {t:.2f}G total, {r:.2f}G reserved, {a:.2f}G allocated, {f:.2f}G free')
+
+ batch_sizes = [1, 2, 4, 8, 16]
+ try:
+ img = [torch.zeros(b, 3, imgsz, imgsz) for b in batch_sizes]
+ y = profile(img, model, n=3, device=device)
+ except Exception as e:
+ LOGGER.warning(f'{prefix}{e}')
+
+ y = [x[2] for x in y if x] # memory [2]
+ batch_sizes = batch_sizes[:len(y)]
+ p = np.polyfit(batch_sizes, y, deg=1) # first degree polynomial fit
+ b = int((f * fraction - p[1]) / p[0]) # y intercept (optimal batch size)
+ LOGGER.info(f'{prefix}Using batch-size {b} for {d} {t * fraction:.2f}G/{t:.2f}G ({fraction * 100:.0f}%)')
+ return b
diff --git a/vouchervision/component_detector/utils/aws/__init__.py b/vouchervision/component_detector/utils/aws/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/vouchervision/component_detector/utils/aws/mime.sh b/vouchervision/component_detector/utils/aws/mime.sh
new file mode 100644
index 0000000000000000000000000000000000000000..c319a83cfbdf09bea634c3bd9fca737c0b1dd505
--- /dev/null
+++ b/vouchervision/component_detector/utils/aws/mime.sh
@@ -0,0 +1,26 @@
+# AWS EC2 instance startup 'MIME' script https://aws.amazon.com/premiumsupport/knowledge-center/execute-user-data-ec2/
+# This script will run on every instance restart, not only on first start
+# --- DO NOT COPY ABOVE COMMENTS WHEN PASTING INTO USERDATA ---
+
+Content-Type: multipart/mixed; boundary="//"
+MIME-Version: 1.0
+
+--//
+Content-Type: text/cloud-config; charset="us-ascii"
+MIME-Version: 1.0
+Content-Transfer-Encoding: 7bit
+Content-Disposition: attachment; filename="cloud-config.txt"
+
+#cloud-config
+cloud_final_modules:
+- [scripts-user, always]
+
+--//
+Content-Type: text/x-shellscript; charset="us-ascii"
+MIME-Version: 1.0
+Content-Transfer-Encoding: 7bit
+Content-Disposition: attachment; filename="userdata.txt"
+
+#!/bin/bash
+# --- paste contents of userdata.sh here ---
+--//
diff --git a/vouchervision/component_detector/utils/aws/resume.py b/vouchervision/component_detector/utils/aws/resume.py
new file mode 100644
index 0000000000000000000000000000000000000000..b21731c979a121ab8227280351b70d6062efd983
--- /dev/null
+++ b/vouchervision/component_detector/utils/aws/resume.py
@@ -0,0 +1,40 @@
+# Resume all interrupted trainings in yolov5/ dir including DDP trainings
+# Usage: $ python utils/aws/resume.py
+
+import os
+import sys
+from pathlib import Path
+
+import torch
+import yaml
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[2] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+
+port = 0 # --master_port
+path = Path('').resolve()
+for last in path.rglob('*/**/last.pt'):
+ ckpt = torch.load(last)
+ if ckpt['optimizer'] is None:
+ continue
+
+ # Load opt.yaml
+ with open(last.parent.parent / 'opt.yaml', errors='ignore') as f:
+ opt = yaml.safe_load(f)
+
+ # Get device count
+ d = opt['device'].split(',') # devices
+ nd = len(d) # number of devices
+ ddp = nd > 1 or (nd == 0 and torch.cuda.device_count() > 1) # distributed data parallel
+
+ if ddp: # multi-GPU
+ port += 1
+ cmd = f'python -m torch.distributed.run --nproc_per_node {nd} --master_port {port} train.py --resume {last}'
+ else: # single-GPU
+ cmd = f'python train.py --resume {last}'
+
+ cmd += ' > /dev/null 2>&1 &' # redirect output to dev/null and run in daemon thread
+ print(cmd)
+ os.system(cmd)
diff --git a/vouchervision/component_detector/utils/aws/userdata.sh b/vouchervision/component_detector/utils/aws/userdata.sh
new file mode 100644
index 0000000000000000000000000000000000000000..5fc1332ac1b0d1794cf8f8c5f6918059ae5dc381
--- /dev/null
+++ b/vouchervision/component_detector/utils/aws/userdata.sh
@@ -0,0 +1,27 @@
+#!/bin/bash
+# AWS EC2 instance startup script https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/user-data.html
+# This script will run only once on first instance start (for a re-start script see mime.sh)
+# /home/ubuntu (ubuntu) or /home/ec2-user (amazon-linux) is working dir
+# Use >300 GB SSD
+
+cd home/ubuntu
+if [ ! -d yolov5 ]; then
+ echo "Running first-time script." # install dependencies, download COCO, pull Docker
+ git clone https://github.com/ultralytics/yolov5 -b master && sudo chmod -R 777 yolov5
+ cd yolov5
+ bash data/scripts/get_coco.sh && echo "COCO done." &
+ sudo docker pull ultralytics/yolov5:latest && echo "Docker done." &
+ python -m pip install --upgrade pip && pip install -r requirements.txt && python detect.py && echo "Requirements done." &
+ wait && echo "All tasks done." # finish background tasks
+else
+ echo "Running re-start script." # resume interrupted runs
+ i=0
+ list=$(sudo docker ps -qa) # container list i.e. $'one\ntwo\nthree\nfour'
+ while IFS= read -r id; do
+ ((i++))
+ echo "restarting container $i: $id"
+ sudo docker start $id
+ # sudo docker exec -it $id python train.py --resume # single-GPU
+ sudo docker exec -d $id python utils/aws/resume.py # multi-scenario
+ done <<<"$list"
+fi
diff --git a/vouchervision/component_detector/utils/benchmarks.py b/vouchervision/component_detector/utils/benchmarks.py
new file mode 100644
index 0000000000000000000000000000000000000000..c3636b9e4df4741c73d592098dd374398a3c5df5
--- /dev/null
+++ b/vouchervision/component_detector/utils/benchmarks.py
@@ -0,0 +1,149 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Run YOLOv5 benchmarks on all supported export formats
+
+Format | `export.py --include` | Model
+--- | --- | ---
+PyTorch | - | yolov5s.pt
+TorchScript | `torchscript` | yolov5s.torchscript
+ONNX | `onnx` | yolov5s.onnx
+OpenVINO | `openvino` | yolov5s_openvino_model/
+TensorRT | `engine` | yolov5s.engine
+CoreML | `coreml` | yolov5s.mlmodel
+TensorFlow SavedModel | `saved_model` | yolov5s_saved_model/
+TensorFlow GraphDef | `pb` | yolov5s.pb
+TensorFlow Lite | `tflite` | yolov5s.tflite
+TensorFlow Edge TPU | `edgetpu` | yolov5s_edgetpu.tflite
+TensorFlow.js | `tfjs` | yolov5s_web_model/
+
+Requirements:
+ $ pip install -r requirements.txt coremltools onnx onnx-simplifier onnxruntime openvino-dev tensorflow-cpu # CPU
+ $ pip install -r requirements.txt coremltools onnx onnx-simplifier onnxruntime-gpu openvino-dev tensorflow # GPU
+ $ pip install -U nvidia-tensorrt --index-url https://pypi.ngc.nvidia.com # TensorRT
+
+Usage:
+ $ python utils/benchmarks.py --weights yolov5s.pt --img 640
+"""
+
+import argparse
+import sys
+import time
+from pathlib import Path
+
+import pandas as pd
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+# ROOT = ROOT.relative_to(Path.cwd()) # relative
+
+import export
+import val
+from utils import notebook_init
+from utils.general import LOGGER, print_args
+from utils.torch_utils import select_device
+
+
+def run(
+ weights=ROOT / 'yolov5s.pt', # weights path
+ imgsz=640, # inference size (pixels)
+ batch_size=1, # batch size
+ data=ROOT / 'data/coco128.yaml', # dataset.yaml path
+ device='', # cuda device, i.e. 0 or 0,1,2,3 or cpu
+ half=False, # use FP16 half-precision inference
+ test=False, # test exports only
+ pt_only=False, # test PyTorch only
+):
+ y, t = [], time.time()
+ formats = export.export_formats()
+ device = select_device(device)
+ for i, (name, f, suffix, gpu) in formats.iterrows(): # index, (name, file, suffix, gpu-capable)
+ try:
+ assert i != 9, 'Edge TPU not supported'
+ assert i != 10, 'TF.js not supported'
+ if device.type != 'cpu':
+ assert gpu, f'{name} inference not supported on GPU'
+
+ # Export
+ if f == '-':
+ w = weights # PyTorch format
+ else:
+ w = export.run(weights=weights, imgsz=[imgsz], include=[f], device=device, half=half)[-1] # all others
+ assert suffix in str(w), 'export failed'
+
+ # Validate
+ result = val.run(data, w, batch_size, imgsz, plots=False, device=device, task='benchmark', half=half)
+ metrics = result[0] # metrics (mp, mr, map50, map, *losses(box, obj, cls))
+ speeds = result[2] # times (preprocess, inference, postprocess)
+ y.append([name, round(metrics[3], 4), round(speeds[1], 2)]) # mAP, t_inference
+ except Exception as e:
+ LOGGER.warning(f'WARNING: Benchmark failure for {name}: {e}')
+ y.append([name, None, None]) # mAP, t_inference
+ if pt_only and i == 0:
+ break # break after PyTorch
+
+ # Print results
+ LOGGER.info('\n')
+ parse_opt()
+ notebook_init() # print system info
+ py = pd.DataFrame(y, columns=['Format', 'mAP@0.5:0.95', 'Inference time (ms)'] if map else ['Format', 'Export', ''])
+ LOGGER.info(f'\nBenchmarks complete ({time.time() - t:.2f}s)')
+ LOGGER.info(str(py if map else py.iloc[:, :2]))
+ return py
+
+
+def test(
+ weights=ROOT / 'yolov5s.pt', # weights path
+ imgsz=640, # inference size (pixels)
+ batch_size=1, # batch size
+ data=ROOT / 'data/coco128.yaml', # dataset.yaml path
+ device='', # cuda device, i.e. 0 or 0,1,2,3 or cpu
+ half=False, # use FP16 half-precision inference
+ test=False, # test exports only
+ pt_only=False, # test PyTorch only
+):
+ y, t = [], time.time()
+ formats = export.export_formats()
+ device = select_device(device)
+ for i, (name, f, suffix, gpu) in formats.iterrows(): # index, (name, file, suffix, gpu-capable)
+ try:
+ w = weights if f == '-' else \
+ export.run(weights=weights, imgsz=[imgsz], include=[f], device=device, half=half)[-1] # weights
+ assert suffix in str(w), 'export failed'
+ y.append([name, True])
+ except Exception:
+ y.append([name, False]) # mAP, t_inference
+
+ # Print results
+ LOGGER.info('\n')
+ parse_opt()
+ notebook_init() # print system info
+ py = pd.DataFrame(y, columns=['Format', 'Export'])
+ LOGGER.info(f'\nExports complete ({time.time() - t:.2f}s)')
+ LOGGER.info(str(py))
+ return py
+
+
+def parse_opt():
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--weights', type=str, default=ROOT / 'yolov5s.pt', help='weights path')
+ parser.add_argument('--imgsz', '--img', '--img-size', type=int, default=640, help='inference size (pixels)')
+ parser.add_argument('--batch-size', type=int, default=1, help='batch size')
+ parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='dataset.yaml path')
+ parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--half', action='store_true', help='use FP16 half-precision inference')
+ parser.add_argument('--test', action='store_true', help='test exports only')
+ parser.add_argument('--pt-only', action='store_true', help='test PyTorch only')
+ opt = parser.parse_args()
+ print_args(vars(opt))
+ return opt
+
+
+def main(opt):
+ test(**vars(opt)) if opt.test else run(**vars(opt))
+
+
+if __name__ == "__main__":
+ opt = parse_opt()
+ main(opt)
diff --git a/vouchervision/component_detector/utils/callbacks.py b/vouchervision/component_detector/utils/callbacks.py
new file mode 100644
index 0000000000000000000000000000000000000000..2b32df0bf1c13ffaaec2e7598bb7c16ae76ab14c
--- /dev/null
+++ b/vouchervision/component_detector/utils/callbacks.py
@@ -0,0 +1,71 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Callback utils
+"""
+
+
+class Callbacks:
+ """"
+ Handles all registered callbacks for YOLOv5 Hooks
+ """
+
+ def __init__(self):
+ # Define the available callbacks
+ self._callbacks = {
+ 'on_pretrain_routine_start': [],
+ 'on_pretrain_routine_end': [],
+ 'on_train_start': [],
+ 'on_train_epoch_start': [],
+ 'on_train_batch_start': [],
+ 'optimizer_step': [],
+ 'on_before_zero_grad': [],
+ 'on_train_batch_end': [],
+ 'on_train_epoch_end': [],
+ 'on_val_start': [],
+ 'on_val_batch_start': [],
+ 'on_val_image_end': [],
+ 'on_val_batch_end': [],
+ 'on_val_end': [],
+ 'on_fit_epoch_end': [], # fit = train + val
+ 'on_model_save': [],
+ 'on_train_end': [],
+ 'on_params_update': [],
+ 'teardown': [],}
+ self.stop_training = False # set True to interrupt training
+
+ def register_action(self, hook, name='', callback=None):
+ """
+ Register a new action to a callback hook
+
+ Args:
+ hook: The callback hook name to register the action to
+ name: The name of the action for later reference
+ callback: The callback to fire
+ """
+ assert hook in self._callbacks, f"hook '{hook}' not found in callbacks {self._callbacks}"
+ assert callable(callback), f"callback '{callback}' is not callable"
+ self._callbacks[hook].append({'name': name, 'callback': callback})
+
+ def get_registered_actions(self, hook=None):
+ """"
+ Returns all the registered actions by callback hook
+
+ Args:
+ hook: The name of the hook to check, defaults to all
+ """
+ return self._callbacks[hook] if hook else self._callbacks
+
+ def run(self, hook, *args, **kwargs):
+ """
+ Loop through the registered actions and fire all callbacks
+
+ Args:
+ hook: The name of the hook to check, defaults to all
+ args: Arguments to receive from YOLOv5
+ kwargs: Keyword Arguments to receive from YOLOv5
+ """
+
+ assert hook in self._callbacks, f"hook '{hook}' not found in callbacks {self._callbacks}"
+
+ for logger in self._callbacks[hook]:
+ logger['callback'](*args, **kwargs)
diff --git a/vouchervision/component_detector/utils/dataloaders.py b/vouchervision/component_detector/utils/dataloaders.py
new file mode 100644
index 0000000000000000000000000000000000000000..5eab6201f4f44595c2b279eff15146841a539e46
--- /dev/null
+++ b/vouchervision/component_detector/utils/dataloaders.py
@@ -0,0 +1,1222 @@
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+Dataloaders and dataset utils
+"""
+
+import contextlib
+import glob
+import hashlib
+import json
+import math
+import os
+import random
+import shutil
+import time
+from itertools import repeat
+from multiprocessing.pool import Pool, ThreadPool
+from pathlib import Path
+from threading import Thread
+from urllib.parse import urlparse
+
+import numpy as np
+import psutil
+import torch
+import torch.nn.functional as F
+import torchvision
+import yaml
+from PIL import ExifTags, Image, ImageOps
+from torch.utils.data import DataLoader, Dataset, dataloader, distributed
+from tqdm import tqdm
+
+from utils.augmentations_torchscript import (Albumentations, augment_hsv, classify_albumentations, classify_transforms, copy_paste,
+ letterbox, mixup, random_perspective)
+from utils.general_torchscript import (DATASETS_DIR, LOGGER, NUM_THREADS, TQDM_BAR_FORMAT, check_dataset, check_requirements,
+ check_yaml, clean_str, cv2, is_colab, is_kaggle, segments2boxes, unzip_file, xyn2xy,
+ xywh2xyxy, xywhn2xyxy, xyxy2xywhn)
+from utils.torch_utils import torch_distributed_zero_first
+
+# Parameters
+HELP_URL = 'See https://docs.ultralytics.com/yolov5/tutorials/train_custom_data'
+IMG_FORMATS = 'bmp', 'dng', 'jpeg', 'jpg', 'mpo', 'png', 'tif', 'tiff', 'webp', 'pfm' # include image suffixes
+VID_FORMATS = 'asf', 'avi', 'gif', 'm4v', 'mkv', 'mov', 'mp4', 'mpeg', 'mpg', 'ts', 'wmv' # include video suffixes
+LOCAL_RANK = int(os.getenv('LOCAL_RANK', -1)) # https://pytorch.org/docs/stable/elastic/run.html
+RANK = int(os.getenv('RANK', -1))
+PIN_MEMORY = str(os.getenv('PIN_MEMORY', True)).lower() == 'true' # global pin_memory for dataloaders
+
+# Get orientation exif tag
+for orientation in ExifTags.TAGS.keys():
+ if ExifTags.TAGS[orientation] == 'Orientation':
+ break
+
+
+def get_hash(paths):
+ # Returns a single hash value of a list of paths (files or dirs)
+ size = sum(os.path.getsize(p) for p in paths if os.path.exists(p)) # sizes
+ h = hashlib.sha256(str(size).encode()) # hash sizes
+ h.update(''.join(paths).encode()) # hash paths
+ return h.hexdigest() # return hash
+
+
+def exif_size(img):
+ # Returns exif-corrected PIL size
+ s = img.size # (width, height)
+ with contextlib.suppress(Exception):
+ rotation = dict(img._getexif().items())[orientation]
+ if rotation in [6, 8]: # rotation 270 or 90
+ s = (s[1], s[0])
+ return s
+
+
+def exif_transpose(image):
+ """
+ Transpose a PIL image accordingly if it has an EXIF Orientation tag.
+ Inplace version of https://github.com/python-pillow/Pillow/blob/master/src/PIL/ImageOps.py exif_transpose()
+
+ :param image: The image to transpose.
+ :return: An image.
+ """
+ exif = image.getexif()
+ orientation = exif.get(0x0112, 1) # default 1
+ if orientation > 1:
+ method = {
+ 2: Image.FLIP_LEFT_RIGHT,
+ 3: Image.ROTATE_180,
+ 4: Image.FLIP_TOP_BOTTOM,
+ 5: Image.TRANSPOSE,
+ 6: Image.ROTATE_270,
+ 7: Image.TRANSVERSE,
+ 8: Image.ROTATE_90}.get(orientation)
+ if method is not None:
+ image = image.transpose(method)
+ del exif[0x0112]
+ image.info['exif'] = exif.tobytes()
+ return image
+
+
+def seed_worker(worker_id):
+ # Set dataloader worker seed https://pytorch.org/docs/stable/notes/randomness.html#dataloader
+ worker_seed = torch.initial_seed() % 2 ** 32
+ np.random.seed(worker_seed)
+ random.seed(worker_seed)
+
+
+def create_dataloader(path,
+ imgsz,
+ batch_size,
+ stride,
+ single_cls=False,
+ hyp=None,
+ augment=False,
+ cache=False,
+ pad=0.0,
+ rect=False,
+ rank=-1,
+ workers=8,
+ image_weights=False,
+ quad=False,
+ prefix='',
+ shuffle=False,
+ seed=0):
+ if rect and shuffle:
+ LOGGER.warning('WARNING ⚠️ --rect is incompatible with DataLoader shuffle, setting shuffle=False')
+ shuffle = False
+ with torch_distributed_zero_first(rank): # init dataset *.cache only once if DDP
+ dataset = LoadImagesAndLabels(
+ path,
+ imgsz,
+ batch_size,
+ augment=augment, # augmentation
+ hyp=hyp, # hyperparameters
+ rect=rect, # rectangular batches
+ cache_images=cache,
+ single_cls=single_cls,
+ stride=int(stride),
+ pad=pad,
+ image_weights=image_weights,
+ prefix=prefix)
+
+ batch_size = min(batch_size, len(dataset))
+ nd = torch.cuda.device_count() # number of CUDA devices
+ nw = min([os.cpu_count() // max(nd, 1), batch_size if batch_size > 1 else 0, workers]) # number of workers
+ sampler = None if rank == -1 else distributed.DistributedSampler(dataset, shuffle=shuffle)
+ loader = DataLoader if image_weights else InfiniteDataLoader # only DataLoader allows for attribute updates
+ generator = torch.Generator()
+ generator.manual_seed(6148914691236517205 + seed + RANK)
+ return loader(dataset,
+ batch_size=batch_size,
+ shuffle=shuffle and sampler is None,
+ num_workers=nw,
+ sampler=sampler,
+ pin_memory=PIN_MEMORY,
+ collate_fn=LoadImagesAndLabels.collate_fn4 if quad else LoadImagesAndLabels.collate_fn,
+ worker_init_fn=seed_worker,
+ generator=generator), dataset
+
+
+class InfiniteDataLoader(dataloader.DataLoader):
+ """ Dataloader that reuses workers
+
+ Uses same syntax as vanilla DataLoader
+ """
+
+ def __init__(self, *args, **kwargs):
+ super().__init__(*args, **kwargs)
+ object.__setattr__(self, 'batch_sampler', _RepeatSampler(self.batch_sampler))
+ self.iterator = super().__iter__()
+
+ def __len__(self):
+ return len(self.batch_sampler.sampler)
+
+ def __iter__(self):
+ for _ in range(len(self)):
+ yield next(self.iterator)
+
+
+class _RepeatSampler:
+ """ Sampler that repeats forever
+
+ Args:
+ sampler (Sampler)
+ """
+
+ def __init__(self, sampler):
+ self.sampler = sampler
+
+ def __iter__(self):
+ while True:
+ yield from iter(self.sampler)
+
+
+class LoadScreenshots:
+ # YOLOv5 screenshot dataloader, i.e. `python detect.py --source "screen 0 100 100 512 256"`
+ def __init__(self, source, img_size=640, stride=32, auto=True, transforms=None):
+ # source = [screen_number left top width height] (pixels)
+ check_requirements('mss')
+ import mss
+
+ source, *params = source.split()
+ self.screen, left, top, width, height = 0, None, None, None, None # default to full screen 0
+ if len(params) == 1:
+ self.screen = int(params[0])
+ elif len(params) == 4:
+ left, top, width, height = (int(x) for x in params)
+ elif len(params) == 5:
+ self.screen, left, top, width, height = (int(x) for x in params)
+ self.img_size = img_size
+ self.stride = stride
+ self.transforms = transforms
+ self.auto = auto
+ self.mode = 'stream'
+ self.frame = 0
+ self.sct = mss.mss()
+
+ # Parse monitor shape
+ monitor = self.sct.monitors[self.screen]
+ self.top = monitor['top'] if top is None else (monitor['top'] + top)
+ self.left = monitor['left'] if left is None else (monitor['left'] + left)
+ self.width = width or monitor['width']
+ self.height = height or monitor['height']
+ self.monitor = {'left': self.left, 'top': self.top, 'width': self.width, 'height': self.height}
+
+ def __iter__(self):
+ return self
+
+ def __next__(self):
+ # mss screen capture: get raw pixels from the screen as np array
+ im0 = np.array(self.sct.grab(self.monitor))[:, :, :3] # [:, :, :3] BGRA to BGR
+ s = f'screen {self.screen} (LTWH): {self.left},{self.top},{self.width},{self.height}: '
+
+ if self.transforms:
+ im = self.transforms(im0) # transforms
+ else:
+ im = letterbox(im0, self.img_size, stride=self.stride, auto=self.auto)[0] # padded resize
+ im = im.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB
+ im = np.ascontiguousarray(im) # contiguous
+ self.frame += 1
+ return str(self.screen), im, im0, None, s # screen, img, original img, im0s, s
+
+
+class LoadImages:
+ # YOLOv5 image/video dataloader, i.e. `python detect.py --source image.jpg/vid.mp4`
+ def __init__(self, path, img_size=640, stride=32, auto=True, transforms=None, vid_stride=1):
+ if isinstance(path, str) and Path(path).suffix == '.txt': # *.txt file with img/vid/dir on each line
+ path = Path(path).read_text().rsplit()
+ files = []
+ for p in sorted(path) if isinstance(path, (list, tuple)) else [path]:
+ p = str(Path(p).resolve())
+ if '*' in p:
+ files.extend(sorted(glob.glob(p, recursive=True))) # glob
+ elif os.path.isdir(p):
+ files.extend(sorted(glob.glob(os.path.join(p, '*.*')))) # dir
+ elif os.path.isfile(p):
+ files.append(p) # files
+ else:
+ raise FileNotFoundError(f'{p} does not exist')
+
+ images = [x for x in files if x.split('.')[-1].lower() in IMG_FORMATS]
+ videos = [x for x in files if x.split('.')[-1].lower() in VID_FORMATS]
+ ni, nv = len(images), len(videos)
+
+ self.img_size = img_size
+ self.stride = stride
+ self.files = images + videos
+ self.nf = ni + nv # number of files
+ self.video_flag = [False] * ni + [True] * nv
+ self.mode = 'image'
+ self.auto = auto
+ self.transforms = transforms # optional
+ self.vid_stride = vid_stride # video frame-rate stride
+ if any(videos):
+ self._new_video(videos[0]) # new video
+ else:
+ self.cap = None
+ assert self.nf > 0, f'No images or videos found in {p}. ' \
+ f'Supported formats are:\nimages: {IMG_FORMATS}\nvideos: {VID_FORMATS}'
+
+ def __iter__(self):
+ self.count = 0
+ return self
+
+ def __next__(self):
+ if self.count == self.nf:
+ raise StopIteration
+ path = self.files[self.count]
+
+ if self.video_flag[self.count]:
+ # Read video
+ self.mode = 'video'
+ for _ in range(self.vid_stride):
+ self.cap.grab()
+ ret_val, im0 = self.cap.retrieve()
+ while not ret_val:
+ self.count += 1
+ self.cap.release()
+ if self.count == self.nf: # last video
+ raise StopIteration
+ path = self.files[self.count]
+ self._new_video(path)
+ ret_val, im0 = self.cap.read()
+
+ self.frame += 1
+ # im0 = self._cv2_rotate(im0) # for use if cv2 autorotation is False
+ s = f'video {self.count + 1}/{self.nf} ({self.frame}/{self.frames}) {path}: '
+
+ else:
+ # Read image
+ self.count += 1
+ im0 = cv2.imread(path) # BGR
+ assert im0 is not None, f'Image Not Found {path}'
+ s = f'image {self.count}/{self.nf} {path}: '
+
+ if self.transforms:
+ im = self.transforms(im0) # transforms
+ else:
+ im = letterbox(im0, self.img_size, stride=self.stride, auto=self.auto)[0] # padded resize
+ im = im.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB
+ im = np.ascontiguousarray(im) # contiguous
+
+ return path, im, im0, self.cap, s
+
+ def _new_video(self, path):
+ # Create a new video capture object
+ self.frame = 0
+ self.cap = cv2.VideoCapture(path)
+ self.frames = int(self.cap.get(cv2.CAP_PROP_FRAME_COUNT) / self.vid_stride)
+ self.orientation = int(self.cap.get(cv2.CAP_PROP_ORIENTATION_META)) # rotation degrees
+ # self.cap.set(cv2.CAP_PROP_ORIENTATION_AUTO, 0) # disable https://github.com/ultralytics/yolov5/issues/8493
+
+ def _cv2_rotate(self, im):
+ # Rotate a cv2 video manually
+ if self.orientation == 0:
+ return cv2.rotate(im, cv2.ROTATE_90_CLOCKWISE)
+ elif self.orientation == 180:
+ return cv2.rotate(im, cv2.ROTATE_90_COUNTERCLOCKWISE)
+ elif self.orientation == 90:
+ return cv2.rotate(im, cv2.ROTATE_180)
+ return im
+
+ def __len__(self):
+ return self.nf # number of files
+
+
+class LoadStreams:
+ # YOLOv5 streamloader, i.e. `python detect.py --source 'rtsp://example.com/media.mp4' # RTSP, RTMP, HTTP streams`
+ def __init__(self, sources='file.streams', img_size=640, stride=32, auto=True, transforms=None, vid_stride=1):
+ torch.backends.cudnn.benchmark = True # faster for fixed-size inference
+ self.mode = 'stream'
+ self.img_size = img_size
+ self.stride = stride
+ self.vid_stride = vid_stride # video frame-rate stride
+ sources = Path(sources).read_text().rsplit() if os.path.isfile(sources) else [sources]
+ n = len(sources)
+ self.sources = [clean_str(x) for x in sources] # clean source names for later
+ self.imgs, self.fps, self.frames, self.threads = [None] * n, [0] * n, [0] * n, [None] * n
+ for i, s in enumerate(sources): # index, source
+ # Start thread to read frames from video stream
+ st = f'{i + 1}/{n}: {s}... '
+ if urlparse(s).hostname in ('www.youtube.com', 'youtube.com', 'youtu.be'): # if source is YouTube video
+ # YouTube format i.e. 'https://www.youtube.com/watch?v=Zgi9g1ksQHc' or 'https://youtu.be/Zgi9g1ksQHc'
+ check_requirements(('pafy', 'youtube_dl==2020.12.2'))
+ import pafy
+ s = pafy.new(s).getbest(preftype='mp4').url # YouTube URL
+ s = eval(s) if s.isnumeric() else s # i.e. s = '0' local webcam
+ if s == 0:
+ assert not is_colab(), '--source 0 webcam unsupported on Colab. Rerun command in a local environment.'
+ assert not is_kaggle(), '--source 0 webcam unsupported on Kaggle. Rerun command in a local environment.'
+ cap = cv2.VideoCapture(s)
+ assert cap.isOpened(), f'{st}Failed to open {s}'
+ w = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
+ h = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
+ fps = cap.get(cv2.CAP_PROP_FPS) # warning: may return 0 or nan
+ self.frames[i] = max(int(cap.get(cv2.CAP_PROP_FRAME_COUNT)), 0) or float('inf') # infinite stream fallback
+ self.fps[i] = max((fps if math.isfinite(fps) else 0) % 100, 0) or 30 # 30 FPS fallback
+
+ _, self.imgs[i] = cap.read() # guarantee first frame
+ self.threads[i] = Thread(target=self.update, args=([i, cap, s]), daemon=True)
+ LOGGER.info(f'{st} Success ({self.frames[i]} frames {w}x{h} at {self.fps[i]:.2f} FPS)')
+ self.threads[i].start()
+ LOGGER.info('') # newline
+
+ # check for common shapes
+ s = np.stack([letterbox(x, img_size, stride=stride, auto=auto)[0].shape for x in self.imgs])
+ self.rect = np.unique(s, axis=0).shape[0] == 1 # rect inference if all shapes equal
+ self.auto = auto and self.rect
+ self.transforms = transforms # optional
+ if not self.rect:
+ LOGGER.warning('WARNING ⚠️ Stream shapes differ. For optimal performance supply similarly-shaped streams.')
+
+ def update(self, i, cap, stream):
+ # Read stream `i` frames in daemon thread
+ n, f = 0, self.frames[i] # frame number, frame array
+ while cap.isOpened() and n < f:
+ n += 1
+ cap.grab() # .read() = .grab() followed by .retrieve()
+ if n % self.vid_stride == 0:
+ success, im = cap.retrieve()
+ if success:
+ self.imgs[i] = im
+ else:
+ LOGGER.warning('WARNING ⚠️ Video stream unresponsive, please check your IP camera connection.')
+ self.imgs[i] = np.zeros_like(self.imgs[i])
+ cap.open(stream) # re-open stream if signal was lost
+ time.sleep(0.0) # wait time
+
+ def __iter__(self):
+ self.count = -1
+ return self
+
+ def __next__(self):
+ self.count += 1
+ if not all(x.is_alive() for x in self.threads) or cv2.waitKey(1) == ord('q'): # q to quit
+ cv2.destroyAllWindows()
+ raise StopIteration
+
+ im0 = self.imgs.copy()
+ if self.transforms:
+ im = np.stack([self.transforms(x) for x in im0]) # transforms
+ else:
+ im = np.stack([letterbox(x, self.img_size, stride=self.stride, auto=self.auto)[0] for x in im0]) # resize
+ im = im[..., ::-1].transpose((0, 3, 1, 2)) # BGR to RGB, BHWC to BCHW
+ im = np.ascontiguousarray(im) # contiguous
+
+ return self.sources, im, im0, None, ''
+
+ def __len__(self):
+ return len(self.sources) # 1E12 frames = 32 streams at 30 FPS for 30 years
+
+
+def img2label_paths(img_paths):
+ # Define label paths as a function of image paths
+ sa, sb = f'{os.sep}images{os.sep}', f'{os.sep}labels{os.sep}' # /images/, /labels/ substrings
+ return [sb.join(x.rsplit(sa, 1)).rsplit('.', 1)[0] + '.txt' for x in img_paths]
+
+
+class LoadImagesAndLabels(Dataset):
+ # YOLOv5 train_loader/val_loader, loads images and labels for training and validation
+ cache_version = 0.6 # dataset labels *.cache version
+ rand_interp_methods = [cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4]
+
+ def __init__(self,
+ path,
+ img_size=640,
+ batch_size=16,
+ augment=False,
+ hyp=None,
+ rect=False,
+ image_weights=False,
+ cache_images=False,
+ single_cls=False,
+ stride=32,
+ pad=0.0,
+ min_items=0,
+ prefix=''):
+ self.img_size = img_size
+ self.augment = augment
+ self.hyp = hyp
+ self.image_weights = image_weights
+ self.rect = False if image_weights else rect
+ self.mosaic = self.augment and not self.rect # load 4 images at a time into a mosaic (only during training)
+ self.mosaic_border = [-img_size // 2, -img_size // 2]
+ self.stride = stride
+ self.path = path
+ self.albumentations = Albumentations(size=img_size) if augment else None
+
+ try:
+ f = [] # image files
+ for p in path if isinstance(path, list) else [path]:
+ p = Path(p) # os-agnostic
+ if p.is_dir(): # dir
+ f += glob.glob(str(p / '**' / '*.*'), recursive=True)
+ # f = list(p.rglob('*.*')) # pathlib
+ elif p.is_file(): # file
+ with open(p) as t:
+ t = t.read().strip().splitlines()
+ parent = str(p.parent) + os.sep
+ f += [x.replace('./', parent, 1) if x.startswith('./') else x for x in t] # to global path
+ # f += [p.parent / x.lstrip(os.sep) for x in t] # to global path (pathlib)
+ else:
+ raise FileNotFoundError(f'{prefix}{p} does not exist')
+ self.im_files = sorted(x.replace('/', os.sep) for x in f if x.split('.')[-1].lower() in IMG_FORMATS)
+ # self.img_files = sorted([x for x in f if x.suffix[1:].lower() in IMG_FORMATS]) # pathlib
+ assert self.im_files, f'{prefix}No images found'
+ except Exception as e:
+ raise Exception(f'{prefix}Error loading data from {path}: {e}\n{HELP_URL}') from e
+
+ # Check cache
+ self.label_files = img2label_paths(self.im_files) # labels
+ cache_path = (p if p.is_file() else Path(self.label_files[0]).parent).with_suffix('.cache')
+ try:
+ cache, exists = np.load(cache_path, allow_pickle=True).item(), True # load dict
+ assert cache['version'] == self.cache_version # matches current version
+ assert cache['hash'] == get_hash(self.label_files + self.im_files) # identical hash
+ except Exception:
+ cache, exists = self.cache_labels(cache_path, prefix), False # run cache ops
+
+ # Display cache
+ nf, nm, ne, nc, n = cache.pop('results') # found, missing, empty, corrupt, total
+ if exists and LOCAL_RANK in {-1, 0}:
+ d = f'Scanning {cache_path}... {nf} images, {nm + ne} backgrounds, {nc} corrupt'
+ tqdm(None, desc=prefix + d, total=n, initial=n, bar_format=TQDM_BAR_FORMAT) # display cache results
+ if cache['msgs']:
+ LOGGER.info('\n'.join(cache['msgs'])) # display warnings
+ assert nf > 0 or not augment, f'{prefix}No labels found in {cache_path}, can not start training. {HELP_URL}'
+
+ # Read cache
+ [cache.pop(k) for k in ('hash', 'version', 'msgs')] # remove items
+ labels, shapes, self.segments = zip(*cache.values())
+ nl = len(np.concatenate(labels, 0)) # number of labels
+ assert nl > 0 or not augment, f'{prefix}All labels empty in {cache_path}, can not start training. {HELP_URL}'
+ self.labels = list(labels)
+ self.shapes = np.array(shapes)
+ self.im_files = list(cache.keys()) # update
+ self.label_files = img2label_paths(cache.keys()) # update
+
+ # Filter images
+ if min_items:
+ include = np.array([len(x) >= min_items for x in self.labels]).nonzero()[0].astype(int)
+ LOGGER.info(f'{prefix}{n - len(include)}/{n} images filtered from dataset')
+ self.im_files = [self.im_files[i] for i in include]
+ self.label_files = [self.label_files[i] for i in include]
+ self.labels = [self.labels[i] for i in include]
+ self.segments = [self.segments[i] for i in include]
+ self.shapes = self.shapes[include] # wh
+
+ # Create indices
+ n = len(self.shapes) # number of images
+ bi = np.floor(np.arange(n) / batch_size).astype(int) # batch index
+ nb = bi[-1] + 1 # number of batches
+ self.batch = bi # batch index of image
+ self.n = n
+ self.indices = range(n)
+
+ # Update labels
+ include_class = [] # filter labels to include only these classes (optional)
+ self.segments = list(self.segments)
+ include_class_array = np.array(include_class).reshape(1, -1)
+ for i, (label, segment) in enumerate(zip(self.labels, self.segments)):
+ if include_class:
+ j = (label[:, 0:1] == include_class_array).any(1)
+ self.labels[i] = label[j]
+ if segment:
+ self.segments[i] = [segment[idx] for idx, elem in enumerate(j) if elem]
+ if single_cls: # single-class training, merge all classes into 0
+ self.labels[i][:, 0] = 0
+
+ # Rectangular Training
+ if self.rect:
+ # Sort by aspect ratio
+ s = self.shapes # wh
+ ar = s[:, 1] / s[:, 0] # aspect ratio
+ irect = ar.argsort()
+ self.im_files = [self.im_files[i] for i in irect]
+ self.label_files = [self.label_files[i] for i in irect]
+ self.labels = [self.labels[i] for i in irect]
+ self.segments = [self.segments[i] for i in irect]
+ self.shapes = s[irect] # wh
+ ar = ar[irect]
+
+ # Set training image shapes
+ shapes = [[1, 1]] * nb
+ for i in range(nb):
+ ari = ar[bi == i]
+ mini, maxi = ari.min(), ari.max()
+ if maxi < 1:
+ shapes[i] = [maxi, 1]
+ elif mini > 1:
+ shapes[i] = [1, 1 / mini]
+
+ self.batch_shapes = np.ceil(np.array(shapes) * img_size / stride + pad).astype(int) * stride
+
+ # Cache images into RAM/disk for faster training
+ if cache_images == 'ram' and not self.check_cache_ram(prefix=prefix):
+ cache_images = False
+ self.ims = [None] * n
+ self.npy_files = [Path(f).with_suffix('.npy') for f in self.im_files]
+ if cache_images:
+ b, gb = 0, 1 << 30 # bytes of cached images, bytes per gigabytes
+ self.im_hw0, self.im_hw = [None] * n, [None] * n
+ fcn = self.cache_images_to_disk if cache_images == 'disk' else self.load_image
+ results = ThreadPool(NUM_THREADS).imap(fcn, range(n))
+ pbar = tqdm(enumerate(results), total=n, bar_format=TQDM_BAR_FORMAT, disable=LOCAL_RANK > 0)
+ for i, x in pbar:
+ if cache_images == 'disk':
+ b += self.npy_files[i].stat().st_size
+ else: # 'ram'
+ self.ims[i], self.im_hw0[i], self.im_hw[i] = x # im, hw_orig, hw_resized = load_image(self, i)
+ b += self.ims[i].nbytes
+ pbar.desc = f'{prefix}Caching images ({b / gb:.1f}GB {cache_images})'
+ pbar.close()
+
+ def check_cache_ram(self, safety_margin=0.1, prefix=''):
+ # Check image caching requirements vs available memory
+ b, gb = 0, 1 << 30 # bytes of cached images, bytes per gigabytes
+ n = min(self.n, 30) # extrapolate from 30 random images
+ for _ in range(n):
+ im = cv2.imread(random.choice(self.im_files)) # sample image
+ ratio = self.img_size / max(im.shape[0], im.shape[1]) # max(h, w) # ratio
+ b += im.nbytes * ratio ** 2
+ mem_required = b * self.n / n # GB required to cache dataset into RAM
+ mem = psutil.virtual_memory()
+ cache = mem_required * (1 + safety_margin) < mem.available # to cache or not to cache, that is the question
+ if not cache:
+ LOGGER.info(f'{prefix}{mem_required / gb:.1f}GB RAM required, '
+ f'{mem.available / gb:.1f}/{mem.total / gb:.1f}GB available, '
+ f"{'caching images ✅' if cache else 'not caching images ⚠️'}")
+ return cache
+
+ def cache_labels(self, path=Path('./labels.cache'), prefix=''):
+ # Cache dataset labels, check images and read shapes
+ x = {} # dict
+ nm, nf, ne, nc, msgs = 0, 0, 0, 0, [] # number missing, found, empty, corrupt, messages
+ desc = f'{prefix}Scanning {path.parent / path.stem}...'
+ with Pool(NUM_THREADS) as pool:
+ pbar = tqdm(pool.imap(verify_image_label, zip(self.im_files, self.label_files, repeat(prefix))),
+ desc=desc,
+ total=len(self.im_files),
+ bar_format=TQDM_BAR_FORMAT)
+ for im_file, lb, shape, segments, nm_f, nf_f, ne_f, nc_f, msg in pbar:
+ nm += nm_f
+ nf += nf_f
+ ne += ne_f
+ nc += nc_f
+ if im_file:
+ x[im_file] = [lb, shape, segments]
+ if msg:
+ msgs.append(msg)
+ pbar.desc = f'{desc} {nf} images, {nm + ne} backgrounds, {nc} corrupt'
+
+ pbar.close()
+ if msgs:
+ LOGGER.info('\n'.join(msgs))
+ if nf == 0:
+ LOGGER.warning(f'{prefix}WARNING ⚠️ No labels found in {path}. {HELP_URL}')
+ x['hash'] = get_hash(self.label_files + self.im_files)
+ x['results'] = nf, nm, ne, nc, len(self.im_files)
+ x['msgs'] = msgs # warnings
+ x['version'] = self.cache_version # cache version
+ try:
+ np.save(path, x) # save cache for next time
+ path.with_suffix('.cache.npy').rename(path) # remove .npy suffix
+ LOGGER.info(f'{prefix}New cache created: {path}')
+ except Exception as e:
+ LOGGER.warning(f'{prefix}WARNING ⚠️ Cache directory {path.parent} is not writeable: {e}') # not writeable
+ return x
+
+ def __len__(self):
+ return len(self.im_files)
+
+ # def __iter__(self):
+ # self.count = -1
+ # print('ran dataset iter')
+ # #self.shuffled_vector = np.random.permutation(self.nF) if self.augment else np.arange(self.nF)
+ # return self
+
+ def __getitem__(self, index):
+ index = self.indices[index] # linear, shuffled, or image_weights
+
+ hyp = self.hyp
+ mosaic = self.mosaic and random.random() < hyp['mosaic']
+ if mosaic:
+ # Load mosaic
+ img, labels = self.load_mosaic(index)
+ shapes = None
+
+ # MixUp augmentation
+ if random.random() < hyp['mixup']:
+ img, labels = mixup(img, labels, *self.load_mosaic(random.randint(0, self.n - 1)))
+
+ else:
+ # Load image
+ img, (h0, w0), (h, w) = self.load_image(index)
+
+ # Letterbox
+ shape = self.batch_shapes[self.batch[index]] if self.rect else self.img_size # final letterboxed shape
+ img, ratio, pad = letterbox(img, shape, auto=False, scaleup=self.augment)
+ shapes = (h0, w0), ((h / h0, w / w0), pad) # for COCO mAP rescaling
+
+ labels = self.labels[index].copy()
+ if labels.size: # normalized xywh to pixel xyxy format
+ labels[:, 1:] = xywhn2xyxy(labels[:, 1:], ratio[0] * w, ratio[1] * h, padw=pad[0], padh=pad[1])
+
+ if self.augment:
+ img, labels = random_perspective(img,
+ labels,
+ degrees=hyp['degrees'],
+ translate=hyp['translate'],
+ scale=hyp['scale'],
+ shear=hyp['shear'],
+ perspective=hyp['perspective'])
+
+ nl = len(labels) # number of labels
+ if nl:
+ labels[:, 1:5] = xyxy2xywhn(labels[:, 1:5], w=img.shape[1], h=img.shape[0], clip=True, eps=1E-3)
+
+ if self.augment:
+ # Albumentations
+ img, labels = self.albumentations(img, labels)
+ nl = len(labels) # update after albumentations
+
+ # HSV color-space
+ augment_hsv(img, hgain=hyp['hsv_h'], sgain=hyp['hsv_s'], vgain=hyp['hsv_v'])
+
+ # Flip up-down
+ if random.random() < hyp['flipud']:
+ img = np.flipud(img)
+ if nl:
+ labels[:, 2] = 1 - labels[:, 2]
+
+ # Flip left-right
+ if random.random() < hyp['fliplr']:
+ img = np.fliplr(img)
+ if nl:
+ labels[:, 1] = 1 - labels[:, 1]
+
+ # Cutouts
+ # labels = cutout(img, labels, p=0.5)
+ # nl = len(labels) # update after cutout
+
+ labels_out = torch.zeros((nl, 6))
+ if nl:
+ labels_out[:, 1:] = torch.from_numpy(labels)
+
+ # Convert
+ img = img.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB
+ img = np.ascontiguousarray(img)
+
+ return torch.from_numpy(img), labels_out, self.im_files[index], shapes
+
+ def load_image(self, i):
+ # Loads 1 image from dataset index 'i', returns (im, original hw, resized hw)
+ im, f, fn = self.ims[i], self.im_files[i], self.npy_files[i],
+ if im is None: # not cached in RAM
+ if fn.exists(): # load npy
+ im = np.load(fn)
+ else: # read image
+ im = cv2.imread(f) # BGR
+ assert im is not None, f'Image Not Found {f}'
+ h0, w0 = im.shape[:2] # orig hw
+ r = self.img_size / max(h0, w0) # ratio
+ if r != 1: # if sizes are not equal
+ interp = cv2.INTER_LINEAR if (self.augment or r > 1) else cv2.INTER_AREA
+ im = cv2.resize(im, (math.ceil(w0 * r), math.ceil(h0 * r)), interpolation=interp)
+ return im, (h0, w0), im.shape[:2] # im, hw_original, hw_resized
+ return self.ims[i], self.im_hw0[i], self.im_hw[i] # im, hw_original, hw_resized
+
+ def cache_images_to_disk(self, i):
+ # Saves an image as an *.npy file for faster loading
+ f = self.npy_files[i]
+ if not f.exists():
+ np.save(f.as_posix(), cv2.imread(self.im_files[i]))
+
+ def load_mosaic(self, index):
+ # YOLOv5 4-mosaic loader. Loads 1 image + 3 random images into a 4-image mosaic
+ labels4, segments4 = [], []
+ s = self.img_size
+ yc, xc = (int(random.uniform(-x, 2 * s + x)) for x in self.mosaic_border) # mosaic center x, y
+ indices = [index] + random.choices(self.indices, k=3) # 3 additional image indices
+ random.shuffle(indices)
+ for i, index in enumerate(indices):
+ # Load image
+ img, _, (h, w) = self.load_image(index)
+
+ # place img in img4
+ if i == 0: # top left
+ img4 = np.full((s * 2, s * 2, img.shape[2]), 114, dtype=np.uint8) # base image with 4 tiles
+ x1a, y1a, x2a, y2a = max(xc - w, 0), max(yc - h, 0), xc, yc # xmin, ymin, xmax, ymax (large image)
+ x1b, y1b, x2b, y2b = w - (x2a - x1a), h - (y2a - y1a), w, h # xmin, ymin, xmax, ymax (small image)
+ elif i == 1: # top right
+ x1a, y1a, x2a, y2a = xc, max(yc - h, 0), min(xc + w, s * 2), yc
+ x1b, y1b, x2b, y2b = 0, h - (y2a - y1a), min(w, x2a - x1a), h
+ elif i == 2: # bottom left
+ x1a, y1a, x2a, y2a = max(xc - w, 0), yc, xc, min(s * 2, yc + h)
+ x1b, y1b, x2b, y2b = w - (x2a - x1a), 0, w, min(y2a - y1a, h)
+ elif i == 3: # bottom right
+ x1a, y1a, x2a, y2a = xc, yc, min(xc + w, s * 2), min(s * 2, yc + h)
+ x1b, y1b, x2b, y2b = 0, 0, min(w, x2a - x1a), min(y2a - y1a, h)
+
+ img4[y1a:y2a, x1a:x2a] = img[y1b:y2b, x1b:x2b] # img4[ymin:ymax, xmin:xmax]
+ padw = x1a - x1b
+ padh = y1a - y1b
+
+ # Labels
+ labels, segments = self.labels[index].copy(), self.segments[index].copy()
+ if labels.size:
+ labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padw, padh) # normalized xywh to pixel xyxy format
+ segments = [xyn2xy(x, w, h, padw, padh) for x in segments]
+ labels4.append(labels)
+ segments4.extend(segments)
+
+ # Concat/clip labels
+ labels4 = np.concatenate(labels4, 0)
+ for x in (labels4[:, 1:], *segments4):
+ np.clip(x, 0, 2 * s, out=x) # clip when using random_perspective()
+ # img4, labels4 = replicate(img4, labels4) # replicate
+
+ # Augment
+ img4, labels4, segments4 = copy_paste(img4, labels4, segments4, p=self.hyp['copy_paste'])
+ img4, labels4 = random_perspective(img4,
+ labels4,
+ segments4,
+ degrees=self.hyp['degrees'],
+ translate=self.hyp['translate'],
+ scale=self.hyp['scale'],
+ shear=self.hyp['shear'],
+ perspective=self.hyp['perspective'],
+ border=self.mosaic_border) # border to remove
+
+ return img4, labels4
+
+ def load_mosaic9(self, index):
+ # YOLOv5 9-mosaic loader. Loads 1 image + 8 random images into a 9-image mosaic
+ labels9, segments9 = [], []
+ s = self.img_size
+ indices = [index] + random.choices(self.indices, k=8) # 8 additional image indices
+ random.shuffle(indices)
+ hp, wp = -1, -1 # height, width previous
+ for i, index in enumerate(indices):
+ # Load image
+ img, _, (h, w) = self.load_image(index)
+
+ # place img in img9
+ if i == 0: # center
+ img9 = np.full((s * 3, s * 3, img.shape[2]), 114, dtype=np.uint8) # base image with 4 tiles
+ h0, w0 = h, w
+ c = s, s, s + w, s + h # xmin, ymin, xmax, ymax (base) coordinates
+ elif i == 1: # top
+ c = s, s - h, s + w, s
+ elif i == 2: # top right
+ c = s + wp, s - h, s + wp + w, s
+ elif i == 3: # right
+ c = s + w0, s, s + w0 + w, s + h
+ elif i == 4: # bottom right
+ c = s + w0, s + hp, s + w0 + w, s + hp + h
+ elif i == 5: # bottom
+ c = s + w0 - w, s + h0, s + w0, s + h0 + h
+ elif i == 6: # bottom left
+ c = s + w0 - wp - w, s + h0, s + w0 - wp, s + h0 + h
+ elif i == 7: # left
+ c = s - w, s + h0 - h, s, s + h0
+ elif i == 8: # top left
+ c = s - w, s + h0 - hp - h, s, s + h0 - hp
+
+ padx, pady = c[:2]
+ x1, y1, x2, y2 = (max(x, 0) for x in c) # allocate coords
+
+ # Labels
+ labels, segments = self.labels[index].copy(), self.segments[index].copy()
+ if labels.size:
+ labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padx, pady) # normalized xywh to pixel xyxy format
+ segments = [xyn2xy(x, w, h, padx, pady) for x in segments]
+ labels9.append(labels)
+ segments9.extend(segments)
+
+ # Image
+ img9[y1:y2, x1:x2] = img[y1 - pady:, x1 - padx:] # img9[ymin:ymax, xmin:xmax]
+ hp, wp = h, w # height, width previous
+
+ # Offset
+ yc, xc = (int(random.uniform(0, s)) for _ in self.mosaic_border) # mosaic center x, y
+ img9 = img9[yc:yc + 2 * s, xc:xc + 2 * s]
+
+ # Concat/clip labels
+ labels9 = np.concatenate(labels9, 0)
+ labels9[:, [1, 3]] -= xc
+ labels9[:, [2, 4]] -= yc
+ c = np.array([xc, yc]) # centers
+ segments9 = [x - c for x in segments9]
+
+ for x in (labels9[:, 1:], *segments9):
+ np.clip(x, 0, 2 * s, out=x) # clip when using random_perspective()
+ # img9, labels9 = replicate(img9, labels9) # replicate
+
+ # Augment
+ img9, labels9, segments9 = copy_paste(img9, labels9, segments9, p=self.hyp['copy_paste'])
+ img9, labels9 = random_perspective(img9,
+ labels9,
+ segments9,
+ degrees=self.hyp['degrees'],
+ translate=self.hyp['translate'],
+ scale=self.hyp['scale'],
+ shear=self.hyp['shear'],
+ perspective=self.hyp['perspective'],
+ border=self.mosaic_border) # border to remove
+
+ return img9, labels9
+
+ @staticmethod
+ def collate_fn(batch):
+ im, label, path, shapes = zip(*batch) # transposed
+ for i, lb in enumerate(label):
+ lb[:, 0] = i # add target image index for build_targets()
+ return torch.stack(im, 0), torch.cat(label, 0), path, shapes
+
+ @staticmethod
+ def collate_fn4(batch):
+ im, label, path, shapes = zip(*batch) # transposed
+ n = len(shapes) // 4
+ im4, label4, path4, shapes4 = [], [], path[:n], shapes[:n]
+
+ ho = torch.tensor([[0.0, 0, 0, 1, 0, 0]])
+ wo = torch.tensor([[0.0, 0, 1, 0, 0, 0]])
+ s = torch.tensor([[1, 1, 0.5, 0.5, 0.5, 0.5]]) # scale
+ for i in range(n): # zidane torch.zeros(16,3,720,1280) # BCHW
+ i *= 4
+ if random.random() < 0.5:
+ im1 = F.interpolate(im[i].unsqueeze(0).float(), scale_factor=2.0, mode='bilinear',
+ align_corners=False)[0].type(im[i].type())
+ lb = label[i]
+ else:
+ im1 = torch.cat((torch.cat((im[i], im[i + 1]), 1), torch.cat((im[i + 2], im[i + 3]), 1)), 2)
+ lb = torch.cat((label[i], label[i + 1] + ho, label[i + 2] + wo, label[i + 3] + ho + wo), 0) * s
+ im4.append(im1)
+ label4.append(lb)
+
+ for i, lb in enumerate(label4):
+ lb[:, 0] = i # add target image index for build_targets()
+
+ return torch.stack(im4, 0), torch.cat(label4, 0), path4, shapes4
+
+
+# Ancillary functions --------------------------------------------------------------------------------------------------
+def flatten_recursive(path=DATASETS_DIR / 'coco128'):
+ # Flatten a recursive directory by bringing all files to top level
+ new_path = Path(f'{str(path)}_flat')
+ if os.path.exists(new_path):
+ shutil.rmtree(new_path) # delete output folder
+ os.makedirs(new_path) # make new output folder
+ for file in tqdm(glob.glob(f'{str(Path(path))}/**/*.*', recursive=True)):
+ shutil.copyfile(file, new_path / Path(file).name)
+
+
+def extract_boxes(path=DATASETS_DIR / 'coco128'): # from utils.dataloaders import *; extract_boxes()
+ # Convert detection dataset into classification dataset, with one directory per class
+ path = Path(path) # images dir
+ shutil.rmtree(path / 'classification') if (path / 'classification').is_dir() else None # remove existing
+ files = list(path.rglob('*.*'))
+ n = len(files) # number of files
+ for im_file in tqdm(files, total=n):
+ if im_file.suffix[1:] in IMG_FORMATS:
+ # image
+ im = cv2.imread(str(im_file))[..., ::-1] # BGR to RGB
+ h, w = im.shape[:2]
+
+ # labels
+ lb_file = Path(img2label_paths([str(im_file)])[0])
+ if Path(lb_file).exists():
+ with open(lb_file) as f:
+ lb = np.array([x.split() for x in f.read().strip().splitlines()], dtype=np.float32) # labels
+
+ for j, x in enumerate(lb):
+ c = int(x[0]) # class
+ f = (path / 'classifier') / f'{c}' / f'{path.stem}_{im_file.stem}_{j}.jpg' # new filename
+ if not f.parent.is_dir():
+ f.parent.mkdir(parents=True)
+
+ b = x[1:] * [w, h, w, h] # box
+ # b[2:] = b[2:].max() # rectangle to square
+ b[2:] = b[2:] * 1.2 + 3 # pad
+ b = xywh2xyxy(b.reshape(-1, 4)).ravel().astype(int)
+
+ b[[0, 2]] = np.clip(b[[0, 2]], 0, w) # clip boxes outside of image
+ b[[1, 3]] = np.clip(b[[1, 3]], 0, h)
+ assert cv2.imwrite(str(f), im[b[1]:b[3], b[0]:b[2]]), f'box failure in {f}'
+
+
+def autosplit(path=DATASETS_DIR / 'coco128/images', weights=(0.9, 0.1, 0.0), annotated_only=False):
+ """ Autosplit a dataset into train/val/test splits and save path/autosplit_*.txt files
+ Usage: from utils.dataloaders import *; autosplit()
+ Arguments
+ path: Path to images directory
+ weights: Train, val, test weights (list, tuple)
+ annotated_only: Only use images with an annotated txt file
+ """
+ path = Path(path) # images dir
+ files = sorted(x for x in path.rglob('*.*') if x.suffix[1:].lower() in IMG_FORMATS) # image files only
+ n = len(files) # number of files
+ random.seed(0) # for reproducibility
+ indices = random.choices([0, 1, 2], weights=weights, k=n) # assign each image to a split
+
+ txt = ['autosplit_train.txt', 'autosplit_val.txt', 'autosplit_test.txt'] # 3 txt files
+ for x in txt:
+ if (path.parent / x).exists():
+ (path.parent / x).unlink() # remove existing
+
+ print(f'Autosplitting images from {path}' + ', using *.txt labeled images only' * annotated_only)
+ for i, img in tqdm(zip(indices, files), total=n):
+ if not annotated_only or Path(img2label_paths([str(img)])[0]).exists(): # check label
+ with open(path.parent / txt[i], 'a') as f:
+ f.write(f'./{img.relative_to(path.parent).as_posix()}' + '\n') # add image to txt file
+
+
+def verify_image_label(args):
+ # Verify one image-label pair
+ im_file, lb_file, prefix = args
+ nm, nf, ne, nc, msg, segments = 0, 0, 0, 0, '', [] # number (missing, found, empty, corrupt), message, segments
+ try:
+ # verify images
+ im = Image.open(im_file)
+ im.verify() # PIL verify
+ shape = exif_size(im) # image size
+ assert (shape[0] > 9) & (shape[1] > 9), f'image size {shape} <10 pixels'
+ assert im.format.lower() in IMG_FORMATS, f'invalid image format {im.format}'
+ if im.format.lower() in ('jpg', 'jpeg'):
+ with open(im_file, 'rb') as f:
+ f.seek(-2, 2)
+ if f.read() != b'\xff\xd9': # corrupt JPEG
+ ImageOps.exif_transpose(Image.open(im_file)).save(im_file, 'JPEG', subsampling=0, quality=100)
+ msg = f'{prefix}WARNING ⚠️ {im_file}: corrupt JPEG restored and saved'
+
+ # verify labels
+ if os.path.isfile(lb_file):
+ nf = 1 # label found
+ with open(lb_file) as f:
+ lb = [x.split() for x in f.read().strip().splitlines() if len(x)]
+ if any(len(x) > 6 for x in lb): # is segment
+ classes = np.array([x[0] for x in lb], dtype=np.float32)
+ segments = [np.array(x[1:], dtype=np.float32).reshape(-1, 2) for x in lb] # (cls, xy1...)
+ lb = np.concatenate((classes.reshape(-1, 1), segments2boxes(segments)), 1) # (cls, xywh)
+ lb = np.array(lb, dtype=np.float32)
+ nl = len(lb)
+ if nl:
+ assert lb.shape[1] == 5, f'labels require 5 columns, {lb.shape[1]} columns detected'
+ assert (lb >= 0).all(), f'negative label values {lb[lb < 0]}'
+ assert (lb[:, 1:] <= 1).all(), f'non-normalized or out of bounds coordinates {lb[:, 1:][lb[:, 1:] > 1]}'
+ _, i = np.unique(lb, axis=0, return_index=True)
+ if len(i) < nl: # duplicate row check
+ lb = lb[i] # remove duplicates
+ if segments:
+ segments = [segments[x] for x in i]
+ msg = f'{prefix}WARNING ⚠️ {im_file}: {nl - len(i)} duplicate labels removed'
+ else:
+ ne = 1 # label empty
+ lb = np.zeros((0, 5), dtype=np.float32)
+ else:
+ nm = 1 # label missing
+ lb = np.zeros((0, 5), dtype=np.float32)
+ return im_file, lb, shape, segments, nm, nf, ne, nc, msg
+ except Exception as e:
+ nc = 1
+ msg = f'{prefix}WARNING ⚠️ {im_file}: ignoring corrupt image/label: {e}'
+ return [None, None, None, None, nm, nf, ne, nc, msg]
+
+
+class HUBDatasetStats():
+ """ Class for generating HUB dataset JSON and `-hub` dataset directory
+
+ Arguments
+ path: Path to data.yaml or data.zip (with data.yaml inside data.zip)
+ autodownload: Attempt to download dataset if not found locally
+
+ Usage
+ from utils.dataloaders import HUBDatasetStats
+ stats = HUBDatasetStats('coco128.yaml', autodownload=True) # usage 1
+ stats = HUBDatasetStats('path/to/coco128.zip') # usage 2
+ stats.get_json(save=False)
+ stats.process_images()
+ """
+
+ def __init__(self, path='coco128.yaml', autodownload=False):
+ # Initialize class
+ zipped, data_dir, yaml_path = self._unzip(Path(path))
+ try:
+ with open(check_yaml(yaml_path), errors='ignore') as f:
+ data = yaml.safe_load(f) # data dict
+ if zipped:
+ data['path'] = data_dir
+ except Exception as e:
+ raise Exception('error/HUB/dataset_stats/yaml_load') from e
+
+ check_dataset(data, autodownload) # download dataset if missing
+ self.hub_dir = Path(data['path'] + '-hub')
+ self.im_dir = self.hub_dir / 'images'
+ self.im_dir.mkdir(parents=True, exist_ok=True) # makes /images
+ self.stats = {'nc': data['nc'], 'names': list(data['names'].values())} # statistics dictionary
+ self.data = data
+
+ @staticmethod
+ def _find_yaml(dir):
+ # Return data.yaml file
+ files = list(dir.glob('*.yaml')) or list(dir.rglob('*.yaml')) # try root level first and then recursive
+ assert files, f'No *.yaml file found in {dir}'
+ if len(files) > 1:
+ files = [f for f in files if f.stem == dir.stem] # prefer *.yaml files that match dir name
+ assert files, f'Multiple *.yaml files found in {dir}, only 1 *.yaml file allowed'
+ assert len(files) == 1, f'Multiple *.yaml files found: {files}, only 1 *.yaml file allowed in {dir}'
+ return files[0]
+
+ def _unzip(self, path):
+ # Unzip data.zip
+ if not str(path).endswith('.zip'): # path is data.yaml
+ return False, None, path
+ assert Path(path).is_file(), f'Error unzipping {path}, file not found'
+ unzip_file(path, path=path.parent)
+ dir = path.with_suffix('') # dataset directory == zip name
+ assert dir.is_dir(), f'Error unzipping {path}, {dir} not found. path/to/abc.zip MUST unzip to path/to/abc/'
+ return True, str(dir), self._find_yaml(dir) # zipped, data_dir, yaml_path
+
+ def _hub_ops(self, f, max_dim=1920):
+ # HUB ops for 1 image 'f': resize and save at reduced quality in /dataset-hub for web/app viewing
+ f_new = self.im_dir / Path(f).name # dataset-hub image filename
+ try: # use PIL
+ im = Image.open(f)
+ r = max_dim / max(im.height, im.width) # ratio
+ if r < 1.0: # image too large
+ im = im.resize((int(im.width * r), int(im.height * r)))
+ im.save(f_new, 'JPEG', quality=50, optimize=True) # save
+ except Exception as e: # use OpenCV
+ LOGGER.info(f'WARNING ⚠️ HUB ops PIL failure {f}: {e}')
+ im = cv2.imread(f)
+ im_height, im_width = im.shape[:2]
+ r = max_dim / max(im_height, im_width) # ratio
+ if r < 1.0: # image too large
+ im = cv2.resize(im, (int(im_width * r), int(im_height * r)), interpolation=cv2.INTER_AREA)
+ cv2.imwrite(str(f_new), im)
+
+ def get_json(self, save=False, verbose=False):
+ # Return dataset JSON for Ultralytics HUB
+ def _round(labels):
+ # Update labels to integer class and 6 decimal place floats
+ return [[int(c), *(round(x, 4) for x in points)] for c, *points in labels]
+
+ for split in 'train', 'val', 'test':
+ if self.data.get(split) is None:
+ self.stats[split] = None # i.e. no test set
+ continue
+ dataset = LoadImagesAndLabels(self.data[split]) # load dataset
+ x = np.array([
+ np.bincount(label[:, 0].astype(int), minlength=self.data['nc'])
+ for label in tqdm(dataset.labels, total=dataset.n, desc='Statistics')]) # shape(128x80)
+ self.stats[split] = {
+ 'instance_stats': {
+ 'total': int(x.sum()),
+ 'per_class': x.sum(0).tolist()},
+ 'image_stats': {
+ 'total': dataset.n,
+ 'unlabelled': int(np.all(x == 0, 1).sum()),
+ 'per_class': (x > 0).sum(0).tolist()},
+ 'labels': [{
+ str(Path(k).name): _round(v.tolist())} for k, v in zip(dataset.im_files, dataset.labels)]}
+
+ # Save, print and return
+ if save:
+ stats_path = self.hub_dir / 'stats.json'
+ print(f'Saving {stats_path.resolve()}...')
+ with open(stats_path, 'w') as f:
+ json.dump(self.stats, f) # save stats.json
+ if verbose:
+ print(json.dumps(self.stats, indent=2, sort_keys=False))
+ return self.stats
+
+ def process_images(self):
+ # Compress images for Ultralytics HUB
+ for split in 'train', 'val', 'test':
+ if self.data.get(split) is None:
+ continue
+ dataset = LoadImagesAndLabels(self.data[split]) # load dataset
+ desc = f'{split} images'
+ for _ in tqdm(ThreadPool(NUM_THREADS).imap(self._hub_ops, dataset.im_files), total=dataset.n, desc=desc):
+ pass
+ print(f'Done. All images saved to {self.im_dir}')
+ return self.im_dir
+
+
+# Classification dataloaders -------------------------------------------------------------------------------------------
+class ClassificationDataset(torchvision.datasets.ImageFolder):
+ """
+ YOLOv5 Classification Dataset.
+ Arguments
+ root: Dataset path
+ transform: torchvision transforms, used by default
+ album_transform: Albumentations transforms, used if installed
+ """
+
+ def __init__(self, root, augment, imgsz, cache=False):
+ super().__init__(root=root)
+ self.torch_transforms = classify_transforms(imgsz)
+ self.album_transforms = classify_albumentations(augment, imgsz) if augment else None
+ self.cache_ram = cache is True or cache == 'ram'
+ self.cache_disk = cache == 'disk'
+ self.samples = [list(x) + [Path(x[0]).with_suffix('.npy'), None] for x in self.samples] # file, index, npy, im
+
+ def __getitem__(self, i):
+ f, j, fn, im = self.samples[i] # filename, index, filename.with_suffix('.npy'), image
+ if self.cache_ram and im is None:
+ im = self.samples[i][3] = cv2.imread(f)
+ elif self.cache_disk:
+ if not fn.exists(): # load npy
+ np.save(fn.as_posix(), cv2.imread(f))
+ im = np.load(fn)
+ else: # read image
+ im = cv2.imread(f) # BGR
+ if self.album_transforms:
+ sample = self.album_transforms(image=cv2.cvtColor(im, cv2.COLOR_BGR2RGB))['image']
+ else:
+ sample = self.torch_transforms(im)
+ return sample, j
+
+
+def create_classification_dataloader(path,
+ imgsz=224,
+ batch_size=16,
+ augment=True,
+ cache=False,
+ rank=-1,
+ workers=8,
+ shuffle=True):
+ # Returns Dataloader object to be used with YOLOv5 Classifier
+ with torch_distributed_zero_first(rank): # init dataset *.cache only once if DDP
+ dataset = ClassificationDataset(root=path, imgsz=imgsz, augment=augment, cache=cache)
+ batch_size = min(batch_size, len(dataset))
+ nd = torch.cuda.device_count()
+ nw = min([os.cpu_count() // max(nd, 1), batch_size if batch_size > 1 else 0, workers])
+ sampler = None if rank == -1 else distributed.DistributedSampler(dataset, shuffle=shuffle)
+ generator = torch.Generator()
+ generator.manual_seed(6148914691236517205 + RANK)
+ return InfiniteDataLoader(dataset,
+ batch_size=batch_size,
+ shuffle=shuffle and sampler is None,
+ num_workers=nw,
+ sampler=sampler,
+ pin_memory=PIN_MEMORY,
+ worker_init_fn=seed_worker,
+ generator=generator) # or DataLoader(persistent_workers=True)
\ No newline at end of file
diff --git a/vouchervision/component_detector/utils/datasets.py b/vouchervision/component_detector/utils/datasets.py
new file mode 100644
index 0000000000000000000000000000000000000000..7075eaae7c739457e2062a094b5b072685d76951
--- /dev/null
+++ b/vouchervision/component_detector/utils/datasets.py
@@ -0,0 +1,1122 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Dataloaders and dataset utils
+"""
+"""
+All 'np.int' have been changed to 'int'
+"""
+import glob
+import hashlib
+import json
+import math
+import os
+import random
+import shutil
+import time
+from itertools import repeat
+from multiprocessing.pool import Pool, ThreadPool
+from pathlib import Path
+from threading import Thread
+from urllib.parse import urlparse
+from zipfile import ZipFile
+
+import numpy as np
+import torch
+import torch.nn.functional as F
+import yaml
+from PIL import ExifTags, Image, ImageOps
+from torch.utils.data import DataLoader, Dataset, dataloader, distributed
+from tqdm.auto import tqdm
+
+from utils.augmentations import Albumentations, augment_hsv, copy_paste, letterbox, mixup, random_perspective
+from utils.general import (DATASETS_DIR, LOGGER, NUM_THREADS, check_dataset, check_requirements, check_yaml, clean_str,
+ cv2, segments2boxes, xyn2xy, xywh2xyxy, xywhn2xyxy, xyxy2xywhn)
+from utils.torch_utils import torch_distributed_zero_first
+
+# Parameters
+HELP_URL = 'https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data'
+IMG_FORMATS = 'bmp', 'dng', 'jpeg', 'jpg', 'mpo', 'png', 'tif', 'tiff', 'webp' # include image suffixes
+VID_FORMATS = 'asf', 'avi', 'gif', 'm4v', 'mkv', 'mov', 'mp4', 'mpeg', 'mpg', 'ts', 'wmv' # include video suffixes
+BAR_FORMAT = '{l_bar}{bar:10}{r_bar}{bar:-10b}' # tqdm bar format
+LOCAL_RANK = int(os.getenv('LOCAL_RANK', -1)) # https://pytorch.org/docs/stable/elastic/run.html
+
+# Get orientation exif tag
+for orientation in ExifTags.TAGS.keys():
+ if ExifTags.TAGS[orientation] == 'Orientation':
+ break
+
+
+def get_hash(paths):
+ # Returns a single hash value of a list of paths (files or dirs)
+ size = sum(os.path.getsize(p) for p in paths if os.path.exists(p)) # sizes
+ h = hashlib.md5(str(size).encode()) # hash sizes
+ h.update(''.join(paths).encode()) # hash paths
+ return h.hexdigest() # return hash
+
+
+def exif_size(img):
+ # Returns exif-corrected PIL size
+ s = img.size # (width, height)
+ try:
+ rotation = dict(img._getexif().items())[orientation]
+ if rotation == 6: # rotation 270
+ s = (s[1], s[0])
+ elif rotation == 8: # rotation 90
+ s = (s[1], s[0])
+ except Exception:
+ pass
+
+ return s
+
+
+def exif_transpose(image):
+ """
+ Transpose a PIL image accordingly if it has an EXIF Orientation tag.
+ Inplace version of https://github.com/python-pillow/Pillow/blob/master/src/PIL/ImageOps.py exif_transpose()
+
+ :param image: The image to transpose.
+ :return: An image.
+ """
+ exif = image.getexif()
+ orientation = exif.get(0x0112, 1) # default 1
+ if orientation > 1:
+ method = {
+ 2: Image.FLIP_LEFT_RIGHT,
+ 3: Image.ROTATE_180,
+ 4: Image.FLIP_TOP_BOTTOM,
+ 5: Image.TRANSPOSE,
+ 6: Image.ROTATE_270,
+ 7: Image.TRANSVERSE,
+ 8: Image.ROTATE_90,}.get(orientation)
+ if method is not None:
+ image = image.transpose(method)
+ del exif[0x0112]
+ image.info["exif"] = exif.tobytes()
+ return image
+
+
+def create_dataloader(path,
+ imgsz,
+ batch_size,
+ stride,
+ single_cls=False,
+ hyp=None,
+ augment=False,
+ cache=False,
+ pad=0.0,
+ rect=False,
+ rank=-1,
+ workers=8,
+ image_weights=False,
+ quad=False,
+ prefix='',
+ shuffle=False):
+ if rect and shuffle:
+ LOGGER.warning('WARNING: --rect is incompatible with DataLoader shuffle, setting shuffle=False')
+ shuffle = False
+ with torch_distributed_zero_first(rank): # init dataset *.cache only once if DDP
+ dataset = LoadImagesAndLabels(
+ path,
+ imgsz,
+ batch_size,
+ augment=augment, # augmentation
+ hyp=hyp, # hyperparameters
+ rect=rect, # rectangular batches
+ cache_images=cache,
+ single_cls=single_cls,
+ stride=int(stride),
+ pad=pad,
+ image_weights=image_weights,
+ prefix=prefix)
+
+ batch_size = min(batch_size, len(dataset))
+ nd = torch.cuda.device_count() # number of CUDA devices
+ nw = min([os.cpu_count() // max(nd, 1), batch_size if batch_size > 1 else 0, workers]) # number of workers
+ sampler = None if rank == -1 else distributed.DistributedSampler(dataset, shuffle=shuffle)
+ loader = DataLoader if image_weights else InfiniteDataLoader # only DataLoader allows for attribute updates
+ return loader(dataset,
+ batch_size=batch_size,
+ shuffle=shuffle and sampler is None,
+ num_workers=nw,
+ sampler=sampler,
+ pin_memory=True,
+ collate_fn=LoadImagesAndLabels.collate_fn4 if quad else LoadImagesAndLabels.collate_fn), dataset
+
+
+class InfiniteDataLoader(dataloader.DataLoader):
+ """ Dataloader that reuses workers
+
+ Uses same syntax as vanilla DataLoader
+ """
+
+ def __init__(self, *args, **kwargs):
+ super().__init__(*args, **kwargs)
+ object.__setattr__(self, 'batch_sampler', _RepeatSampler(self.batch_sampler))
+ self.iterator = super().__iter__()
+
+ def __len__(self):
+ return len(self.batch_sampler.sampler)
+
+ def __iter__(self):
+ for i in range(len(self)):
+ yield next(self.iterator)
+
+
+class _RepeatSampler:
+ """ Sampler that repeats forever
+
+ Args:
+ sampler (Sampler)
+ """
+
+ def __init__(self, sampler):
+ self.sampler = sampler
+
+ def __iter__(self):
+ while True:
+ yield from iter(self.sampler)
+
+
+class LoadImages:
+ # YOLOv5 image/video dataloader, i.e. `python detect.py --source image.jpg/vid.mp4`
+ '''def __init__(self, path, img_size=640, stride=32, auto=True):
+ p = str(Path(path).resolve()) # os-agnostic absolute path
+ if '*' in p:
+ files = sorted(glob.glob(p, recursive=True)) # glob
+ elif os.path.isdir(p):
+ files = sorted(glob.glob(os.path.join(p, '*.*'))) # dir
+ elif os.path.isfile(p):
+ files = [p] # files
+ else:
+ raise Exception(f'ERROR: {p} does not exist')
+
+ images = [x for x in files if x.split('.')[-1].lower() in IMG_FORMATS]
+ videos = [x for x in files if x.split('.')[-1].lower() in VID_FORMATS]
+ ni, nv = len(images), len(videos)
+
+ self.img_size = img_size
+ self.stride = stride
+ self.files = images + videos
+ self.nf = ni + nv # number of files
+ self.video_flag = [False] * ni + [True] * nv
+ self.mode = 'image'
+ self.auto = auto
+ if any(videos):
+ self.new_video(videos[0]) # new video
+ else:
+ self.cap = None
+ assert self.nf > 0, f'No images or videos found in {p}. ' \
+ f'Supported formats are:\nimages: {IMG_FORMATS}\nvideos: {VID_FORMATS}'''
+ def __init__(self, path, img_size=640, stride=32, auto=True):
+ if isinstance(path, list):
+ files = []
+ for p in path:
+ p = str(Path(p).resolve())
+ if '*' in p:
+ files.extend(sorted(glob.glob(p, recursive=True)))
+ elif os.path.isdir(p):
+ files.extend(sorted(glob.glob(os.path.join(p, '*.*'))))
+ elif os.path.isfile(p):
+ files.append(p)
+ else:
+ raise Exception(f'ERROR: {p} does not exist')
+ else:
+ p = str(Path(path).resolve())
+ if '*' in p:
+ files = sorted(glob.glob(p, recursive=True)) # glob
+ elif os.path.isdir(p):
+ files = sorted(glob.glob(os.path.join(p, '*.*'))) # dir
+ elif os.path.isfile(p):
+ files = [p] # files
+ else:
+ raise Exception(f'ERROR: {p} does not exist')
+
+ images = [x for x in files if x.split('.')[-1].lower() in IMG_FORMATS]
+ videos = [x for x in files if x.split('.')[-1].lower() in VID_FORMATS]
+ ni, nv = len(images), len(videos)
+
+ self.img_size = img_size
+ self.stride = stride
+ self.files = images + videos
+ self.nf = ni + nv # number of files
+ self.video_flag = [False] * ni + [True] * nv
+ self.mode = 'image'
+ self.auto = auto
+ if any(videos):
+ self.new_video(videos[0]) # new video
+ else:
+ self.cap = None
+ assert self.nf > 0, f'No images or videos found in {path}. ' \
+ f'Supported formats are:\nimages: {IMG_FORMATS}\nvideos: {VID_FORMATS}'
+
+
+ def __iter__(self):
+ self.count = 0
+ return self
+
+ def __next__(self):
+ if self.count == self.nf:
+ raise StopIteration
+ path = self.files[self.count]
+
+ if self.video_flag[self.count]:
+ # Read video
+ self.mode = 'video'
+ ret_val, img0 = self.cap.read()
+ while not ret_val:
+ self.count += 1
+ self.cap.release()
+ if self.count == self.nf: # last video
+ raise StopIteration
+ else:
+ path = self.files[self.count]
+ self.new_video(path)
+ ret_val, img0 = self.cap.read()
+
+ self.frame += 1
+ s = f'video {self.count + 1}/{self.nf} ({self.frame}/{self.frames}) {path}: '
+
+ else:
+ # Read image
+ self.count += 1
+ img0 = cv2.imread(path) # BGR
+ assert img0 is not None, f'Image Not Found {path}'
+ s = f'Image {self.count}/{self.nf} {os.path.basename(path)}: '
+
+ # Padded resize
+ img = letterbox(img0, self.img_size, stride=self.stride, auto=self.auto)[0]
+
+ # Convert
+ img = img.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB
+ img = np.ascontiguousarray(img)
+
+ return path, img, img0, self.cap, s
+
+ def new_video(self, path):
+ self.frame = 0
+ self.cap = cv2.VideoCapture(path)
+ self.frames = int(self.cap.get(cv2.CAP_PROP_FRAME_COUNT))
+
+ def __len__(self):
+ return self.nf # number of files
+
+
+class LoadWebcam: # for inference
+ # YOLOv5 local webcam dataloader, i.e. `python detect.py --source 0`
+ def __init__(self, pipe='0', img_size=640, stride=32):
+ self.img_size = img_size
+ self.stride = stride
+ self.pipe = eval(pipe) if pipe.isnumeric() else pipe
+ self.cap = cv2.VideoCapture(self.pipe) # video capture object
+ self.cap.set(cv2.CAP_PROP_BUFFERSIZE, 3) # set buffer size
+
+ def __iter__(self):
+ self.count = -1
+ return self
+
+ def __next__(self):
+ self.count += 1
+ if cv2.waitKey(1) == ord('q'): # q to quit
+ self.cap.release()
+ cv2.destroyAllWindows()
+ raise StopIteration
+
+ # Read frame
+ ret_val, img0 = self.cap.read()
+ img0 = cv2.flip(img0, 1) # flip left-right
+
+ # Print
+ assert ret_val, f'Camera Error {self.pipe}'
+ img_path = 'webcam.jpg'
+ s = f'webcam {self.count}: '
+
+ # Padded resize
+ img = letterbox(img0, self.img_size, stride=self.stride)[0]
+
+ # Convert
+ img = img.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB
+ img = np.ascontiguousarray(img)
+
+ return img_path, img, img0, None, s
+
+ def __len__(self):
+ return 0
+
+
+class LoadStreams:
+ # YOLOv5 streamloader, i.e. `python detect.py --source 'rtsp://example.com/media.mp4' # RTSP, RTMP, HTTP streams`
+ def __init__(self, sources='streams.txt', img_size=640, stride=32, auto=True):
+ self.mode = 'stream'
+ self.img_size = img_size
+ self.stride = stride
+
+ if os.path.isfile(sources):
+ with open(sources) as f:
+ sources = [x.strip() for x in f.read().strip().splitlines() if len(x.strip())]
+ else:
+ sources = [sources]
+
+ n = len(sources)
+ self.imgs, self.fps, self.frames, self.threads = [None] * n, [0] * n, [0] * n, [None] * n
+ self.sources = [clean_str(x) for x in sources] # clean source names for later
+ self.auto = auto
+ for i, s in enumerate(sources): # index, source
+ # Start thread to read frames from video stream
+ st = f'{i + 1}/{n}: {s}... '
+ if urlparse(s).hostname in ('www.youtube.com', 'youtube.com', 'youtu.be'): # if source is YouTube video
+ check_requirements(('pafy', 'youtube_dl==2020.12.2'))
+ import pafy
+ s = pafy.new(s).getbest(preftype="mp4").url # YouTube URL
+ s = eval(s) if s.isnumeric() else s # i.e. s = '0' local webcam
+ cap = cv2.VideoCapture(s)
+ assert cap.isOpened(), f'{st}Failed to open {s}'
+ w = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
+ h = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
+ fps = cap.get(cv2.CAP_PROP_FPS) # warning: may return 0 or nan
+ self.frames[i] = max(int(cap.get(cv2.CAP_PROP_FRAME_COUNT)), 0) or float('inf') # infinite stream fallback
+ self.fps[i] = max((fps if math.isfinite(fps) else 0) % 100, 0) or 30 # 30 FPS fallback
+
+ _, self.imgs[i] = cap.read() # guarantee first frame
+ self.threads[i] = Thread(target=self.update, args=([i, cap, s]), daemon=True)
+ LOGGER.info(f"{st} Success ({self.frames[i]} frames {w}x{h} at {self.fps[i]:.2f} FPS)")
+ self.threads[i].start()
+ LOGGER.info('') # newline
+
+ # check for common shapes
+ s = np.stack([letterbox(x, self.img_size, stride=self.stride, auto=self.auto)[0].shape for x in self.imgs])
+ self.rect = np.unique(s, axis=0).shape[0] == 1 # rect inference if all shapes equal
+ if not self.rect:
+ LOGGER.warning('WARNING: Stream shapes differ. For optimal performance supply similarly-shaped streams.')
+
+ def update(self, i, cap, stream):
+ # Read stream `i` frames in daemon thread
+ n, f, read = 0, self.frames[i], 1 # frame number, frame array, inference every 'read' frame
+ while cap.isOpened() and n < f:
+ n += 1
+ # _, self.imgs[index] = cap.read()
+ cap.grab()
+ if n % read == 0:
+ success, im = cap.retrieve()
+ if success:
+ self.imgs[i] = im
+ else:
+ LOGGER.warning('WARNING: Video stream unresponsive, please check your IP camera connection.')
+ self.imgs[i] = np.zeros_like(self.imgs[i])
+ cap.open(stream) # re-open stream if signal was lost
+ time.sleep(1 / self.fps[i]) # wait time
+
+ def __iter__(self):
+ self.count = -1
+ return self
+
+ def __next__(self):
+ self.count += 1
+ if not all(x.is_alive() for x in self.threads) or cv2.waitKey(1) == ord('q'): # q to quit
+ cv2.destroyAllWindows()
+ raise StopIteration
+
+ # Letterbox
+ img0 = self.imgs.copy()
+ img = [letterbox(x, self.img_size, stride=self.stride, auto=self.rect and self.auto)[0] for x in img0]
+
+ # Stack
+ img = np.stack(img, 0)
+
+ # Convert
+ img = img[..., ::-1].transpose((0, 3, 1, 2)) # BGR to RGB, BHWC to BCHW
+ img = np.ascontiguousarray(img)
+
+ return self.sources, img, img0, None, ''
+
+ def __len__(self):
+ return len(self.sources) # 1E12 frames = 32 streams at 30 FPS for 30 years
+
+
+def img2label_paths(img_paths):
+ # Define label paths as a function of image paths
+ sa, sb = os.sep + 'images' + os.sep, os.sep + 'labels' + os.sep # /images/, /labels/ substrings
+ return [sb.join(x.rsplit(sa, 1)).rsplit('.', 1)[0] + '.txt' for x in img_paths]
+
+
+class LoadImagesAndLabels(Dataset):
+ # YOLOv5 train_loader/val_loader, loads images and labels for training and validation
+ cache_version = 0.6 # dataset labels *.cache version
+
+ def __init__(self,
+ path,
+ img_size=640,
+ batch_size=16,
+ augment=False,
+ hyp=None,
+ rect=False,
+ image_weights=False,
+ cache_images=False,
+ single_cls=False,
+ stride=32,
+ pad=0.0,
+ prefix=''):
+ self.img_size = img_size
+ self.augment = augment
+ self.hyp = hyp
+ self.image_weights = image_weights
+ self.rect = False if image_weights else rect
+ self.mosaic = self.augment and not self.rect # load 4 images at a time into a mosaic (only during training)
+ self.mosaic_border = [-img_size // 2, -img_size // 2]
+ self.stride = stride
+ self.path = path
+ self.albumentations = Albumentations() if augment else None
+
+ try:
+ f = [] # image files
+ for p in path if isinstance(path, list) else [path]:
+ p = Path(p) # os-agnostic
+ if p.is_dir(): # dir
+ f += glob.glob(str(p / '**' / '*.*'), recursive=True)
+ # f = list(p.rglob('*.*')) # pathlib
+ elif p.is_file(): # file
+ with open(p) as t:
+ t = t.read().strip().splitlines()
+ parent = str(p.parent) + os.sep
+ f += [x.replace('./', parent) if x.startswith('./') else x for x in t] # local to global path
+ # f += [p.parent / x.lstrip(os.sep) for x in t] # local to global path (pathlib)
+ else:
+ raise Exception(f'{prefix}{p} does not exist')
+ self.im_files = sorted(x.replace('/', os.sep) for x in f if x.split('.')[-1].lower() in IMG_FORMATS)
+ # self.img_files = sorted([x for x in f if x.suffix[1:].lower() in IMG_FORMATS]) # pathlib
+ assert self.im_files, f'{prefix}No images found'
+ except Exception as e:
+ raise Exception(f'{prefix}Error loading data from {path}: {e}\nSee {HELP_URL}')
+
+ # Check cache
+ self.label_files = img2label_paths(self.im_files) # labels
+ cache_path = (p if p.is_file() else Path(self.label_files[0]).parent).with_suffix('.cache')
+ try:
+ cache, exists = np.load(cache_path, allow_pickle=True).item(), True # load dict
+ assert cache['version'] == self.cache_version # same version
+ assert cache['hash'] == get_hash(self.label_files + self.im_files) # same hash
+ except Exception:
+ cache, exists = self.cache_labels(cache_path, prefix), False # cache
+
+ # Display cache
+ nf, nm, ne, nc, n = cache.pop('results') # found, missing, empty, corrupt, total
+ if exists and LOCAL_RANK in (-1, 0):
+ d = f"Scanning '{cache_path}' images and labels... {nf} found, {nm} missing, {ne} empty, {nc} corrupt"
+ tqdm(None, desc=prefix + d, total=n, initial=n, bar_format=BAR_FORMAT) # display cache results
+ if cache['msgs']:
+ LOGGER.info('\n'.join(cache['msgs'])) # display warnings
+ assert nf > 0 or not augment, f'{prefix}No labels in {cache_path}. Can not train without labels. See {HELP_URL}'
+
+ # Read cache
+ [cache.pop(k) for k in ('hash', 'version', 'msgs')] # remove items
+ labels, shapes, self.segments = zip(*cache.values())
+ self.labels = list(labels)
+ self.shapes = np.array(shapes, dtype=np.float64)
+ self.im_files = list(cache.keys()) # update
+ self.label_files = img2label_paths(cache.keys()) # update
+ n = len(shapes) # number of images
+ bi = np.floor(np.arange(n) / batch_size).astype(int) # batch index # WW changed 'int' to 'int8'
+ nb = bi[-1] + 1 # number of batches
+ self.batch = bi # batch index of image
+ self.n = n
+ self.indices = range(n)
+
+ # Update labels
+ include_class = [] # filter labels to include only these classes (optional)
+ include_class_array = np.array(include_class).reshape(1, -1)
+ for i, (label, segment) in enumerate(zip(self.labels, self.segments)):
+ if include_class:
+ j = (label[:, 0:1] == include_class_array).any(1)
+ self.labels[i] = label[j]
+ if segment:
+ self.segments[i] = segment[j]
+ if single_cls: # single-class training, merge all classes into 0
+ self.labels[i][:, 0] = 0
+ if segment:
+ self.segments[i][:, 0] = 0
+
+ # Rectangular Training
+ if self.rect:
+ # Sort by aspect ratio
+ s = self.shapes # wh
+ ar = s[:, 1] / s[:, 0] # aspect ratio
+ irect = ar.argsort()
+ self.im_files = [self.im_files[i] for i in irect]
+ self.label_files = [self.label_files[i] for i in irect]
+ self.labels = [self.labels[i] for i in irect]
+ self.shapes = s[irect] # wh
+ ar = ar[irect]
+
+ # Set training image shapes
+ shapes = [[1, 1]] * nb
+ for i in range(nb):
+ ari = ar[bi == i]
+ mini, maxi = ari.min(), ari.max()
+ if maxi < 1:
+ shapes[i] = [maxi, 1]
+ elif mini > 1:
+ shapes[i] = [1, 1 / mini]
+
+ self.batch_shapes = np.ceil(np.array(shapes) * img_size / stride + pad).astype(int) * stride
+
+ # Cache images into RAM/disk for faster training (WARNING: large datasets may exceed system resources)
+ self.ims = [None] * n
+ self.npy_files = [Path(f).with_suffix('.npy') for f in self.im_files]
+ if cache_images:
+ gb = 0 # Gigabytes of cached images
+ self.im_hw0, self.im_hw = [None] * n, [None] * n
+ fcn = self.cache_images_to_disk if cache_images == 'disk' else self.load_image
+ results = ThreadPool(NUM_THREADS).imap(fcn, range(n))
+ pbar = tqdm(enumerate(results), total=n, bar_format=BAR_FORMAT, disable=LOCAL_RANK > 0)
+ for i, x in pbar:
+ if cache_images == 'disk':
+ gb += self.npy_files[i].stat().st_size
+ else: # 'ram'
+ self.ims[i], self.im_hw0[i], self.im_hw[i] = x # im, hw_orig, hw_resized = load_image(self, i)
+ gb += self.ims[i].nbytes
+ pbar.desc = f'{prefix}Caching images ({gb / 1E9:.1f}GB {cache_images})'
+ pbar.close()
+
+ def cache_labels(self, path=Path('./labels.cache'), prefix=''):
+ # Cache dataset labels, check images and read shapes
+ x = {} # dict
+ nm, nf, ne, nc, msgs = 0, 0, 0, 0, [] # number missing, found, empty, corrupt, messages
+ desc = f"{prefix}Scanning '{path.parent / path.stem}' images and labels..."
+ with Pool(NUM_THREADS) as pool:
+ pbar = tqdm(pool.imap(verify_image_label, zip(self.im_files, self.label_files, repeat(prefix))),
+ desc=desc,
+ total=len(self.im_files),
+ bar_format=BAR_FORMAT)
+ for im_file, lb, shape, segments, nm_f, nf_f, ne_f, nc_f, msg in pbar:
+ nm += nm_f
+ nf += nf_f
+ ne += ne_f
+ nc += nc_f
+ if im_file:
+ x[im_file] = [lb, shape, segments]
+ if msg:
+ msgs.append(msg)
+ pbar.desc = f"{desc}{nf} found, {nm} missing, {ne} empty, {nc} corrupt"
+
+ pbar.close()
+ if msgs:
+ LOGGER.info('\n'.join(msgs))
+ if nf == 0:
+ LOGGER.warning(f'{prefix}WARNING: No labels found in {path}. See {HELP_URL}')
+ x['hash'] = get_hash(self.label_files + self.im_files)
+ x['results'] = nf, nm, ne, nc, len(self.im_files)
+ x['msgs'] = msgs # warnings
+ x['version'] = self.cache_version # cache version
+ try:
+ np.save(path, x) # save cache for next time
+ path.with_suffix('.cache.npy').rename(path) # remove .npy suffix
+ LOGGER.info(f'{prefix}New cache created: {path}')
+ except Exception as e:
+ LOGGER.warning(f'{prefix}WARNING: Cache directory {path.parent} is not writeable: {e}') # not writeable
+ return x
+
+ def __len__(self):
+ return len(self.im_files)
+
+ # def __iter__(self):
+ # self.count = -1
+ # print('ran dataset iter')
+ # #self.shuffled_vector = np.random.permutation(self.nF) if self.augment else np.arange(self.nF)
+ # return self
+
+ def __getitem__(self, index):
+ index = self.indices[index] # linear, shuffled, or image_weights
+
+ hyp = self.hyp
+ mosaic = self.mosaic and random.random() < hyp['mosaic']
+ if mosaic:
+ # Load mosaic
+ img, labels = self.load_mosaic(index)
+ shapes = None
+
+ # MixUp augmentation
+ if random.random() < hyp['mixup']:
+ img, labels = mixup(img, labels, *self.load_mosaic(random.randint(0, self.n - 1)))
+
+ else:
+ # Load image
+ img, (h0, w0), (h, w) = self.load_image(index)
+
+ # Letterbox
+ shape = self.batch_shapes[self.batch[index]] if self.rect else self.img_size # final letterboxed shape
+ img, ratio, pad = letterbox(img, shape, auto=False, scaleup=self.augment)
+ shapes = (h0, w0), ((h / h0, w / w0), pad) # for COCO mAP rescaling
+
+ labels = self.labels[index].copy()
+ if labels.size: # normalized xywh to pixel xyxy format
+ labels[:, 1:] = xywhn2xyxy(labels[:, 1:], ratio[0] * w, ratio[1] * h, padw=pad[0], padh=pad[1])
+
+ if self.augment:
+ img, labels = random_perspective(img,
+ labels,
+ degrees=hyp['degrees'],
+ translate=hyp['translate'],
+ scale=hyp['scale'],
+ shear=hyp['shear'],
+ perspective=hyp['perspective'])
+
+ nl = len(labels) # number of labels
+ if nl:
+ labels[:, 1:5] = xyxy2xywhn(labels[:, 1:5], w=img.shape[1], h=img.shape[0], clip=True, eps=1E-3)
+
+ if self.augment:
+ # Albumentations
+ img, labels = self.albumentations(img, labels)
+ nl = len(labels) # update after albumentations
+
+ # HSV color-space
+ augment_hsv(img, hgain=hyp['hsv_h'], sgain=hyp['hsv_s'], vgain=hyp['hsv_v'])
+
+ # Flip up-down
+ if random.random() < hyp['flipud']:
+ img = np.flipud(img)
+ if nl:
+ labels[:, 2] = 1 - labels[:, 2]
+
+ # Flip left-right
+ if random.random() < hyp['fliplr']:
+ img = np.fliplr(img)
+ if nl:
+ labels[:, 1] = 1 - labels[:, 1]
+
+ # Cutouts
+ # labels = cutout(img, labels, p=0.5)
+ # nl = len(labels) # update after cutout
+
+ labels_out = torch.zeros((nl, 6))
+ if nl:
+ labels_out[:, 1:] = torch.from_numpy(labels)
+
+ # Convert
+ img = img.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB
+ img = np.ascontiguousarray(img)
+
+ return torch.from_numpy(img), labels_out, self.im_files[index], shapes
+
+ def load_image(self, i):
+ # Loads 1 image from dataset index 'i', returns (im, original hw, resized hw)
+ im, f, fn = self.ims[i], self.im_files[i], self.npy_files[i],
+ if im is None: # not cached in RAM
+ if fn.exists(): # load npy
+ im = np.load(fn)
+ else: # read image
+ im = cv2.imread(f) # BGR
+ assert im is not None, f'Image Not Found {f}'
+ h0, w0 = im.shape[:2] # orig hw
+ r = self.img_size / max(h0, w0) # ratio
+ if r != 1: # if sizes are not equal
+ im = cv2.resize(im, (int(w0 * r), int(h0 * r)),
+ interpolation=cv2.INTER_LINEAR if (self.augment or r > 1) else cv2.INTER_AREA)
+ return im, (h0, w0), im.shape[:2] # im, hw_original, hw_resized
+ else:
+ return self.ims[i], self.im_hw0[i], self.im_hw[i] # im, hw_original, hw_resized
+
+ def cache_images_to_disk(self, i):
+ # Saves an image as an *.npy file for faster loading
+ f = self.npy_files[i]
+ if not f.exists():
+ np.save(f.as_posix(), cv2.imread(self.im_files[i]))
+
+ def load_mosaic(self, index):
+ # YOLOv5 4-mosaic loader. Loads 1 image + 3 random images into a 4-image mosaic
+ labels4, segments4 = [], []
+ s = self.img_size
+ yc, xc = (int(random.uniform(-x, 2 * s + x)) for x in self.mosaic_border) # mosaic center x, y
+ indices = [index] + random.choices(self.indices, k=3) # 3 additional image indices
+ random.shuffle(indices)
+ for i, index in enumerate(indices):
+ # Load image
+ img, _, (h, w) = self.load_image(index)
+
+ # place img in img4
+ if i == 0: # top left
+ img4 = np.full((s * 2, s * 2, img.shape[2]), 114, dtype=np.uint8) # base image with 4 tiles
+ x1a, y1a, x2a, y2a = max(xc - w, 0), max(yc - h, 0), xc, yc # xmin, ymin, xmax, ymax (large image)
+ x1b, y1b, x2b, y2b = w - (x2a - x1a), h - (y2a - y1a), w, h # xmin, ymin, xmax, ymax (small image)
+ elif i == 1: # top right
+ x1a, y1a, x2a, y2a = xc, max(yc - h, 0), min(xc + w, s * 2), yc
+ x1b, y1b, x2b, y2b = 0, h - (y2a - y1a), min(w, x2a - x1a), h
+ elif i == 2: # bottom left
+ x1a, y1a, x2a, y2a = max(xc - w, 0), yc, xc, min(s * 2, yc + h)
+ x1b, y1b, x2b, y2b = w - (x2a - x1a), 0, w, min(y2a - y1a, h)
+ elif i == 3: # bottom right
+ x1a, y1a, x2a, y2a = xc, yc, min(xc + w, s * 2), min(s * 2, yc + h)
+ x1b, y1b, x2b, y2b = 0, 0, min(w, x2a - x1a), min(y2a - y1a, h)
+
+ img4[y1a:y2a, x1a:x2a] = img[y1b:y2b, x1b:x2b] # img4[ymin:ymax, xmin:xmax]
+ padw = x1a - x1b
+ padh = y1a - y1b
+
+ # Labels
+ labels, segments = self.labels[index].copy(), self.segments[index].copy()
+ if labels.size:
+ labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padw, padh) # normalized xywh to pixel xyxy format
+ segments = [xyn2xy(x, w, h, padw, padh) for x in segments]
+ labels4.append(labels)
+ segments4.extend(segments)
+
+ # Concat/clip labels
+ labels4 = np.concatenate(labels4, 0)
+ for x in (labels4[:, 1:], *segments4):
+ np.clip(x, 0, 2 * s, out=x) # clip when using random_perspective()
+ # img4, labels4 = replicate(img4, labels4) # replicate
+
+ # Augment
+ img4, labels4, segments4 = copy_paste(img4, labels4, segments4, p=self.hyp['copy_paste'])
+ img4, labels4 = random_perspective(img4,
+ labels4,
+ segments4,
+ degrees=self.hyp['degrees'],
+ translate=self.hyp['translate'],
+ scale=self.hyp['scale'],
+ shear=self.hyp['shear'],
+ perspective=self.hyp['perspective'],
+ border=self.mosaic_border) # border to remove
+
+ return img4, labels4
+
+ def load_mosaic9(self, index):
+ # YOLOv5 9-mosaic loader. Loads 1 image + 8 random images into a 9-image mosaic
+ labels9, segments9 = [], []
+ s = self.img_size
+ indices = [index] + random.choices(self.indices, k=8) # 8 additional image indices
+ random.shuffle(indices)
+ hp, wp = -1, -1 # height, width previous
+ for i, index in enumerate(indices):
+ # Load image
+ img, _, (h, w) = self.load_image(index)
+
+ # place img in img9
+ if i == 0: # center
+ img9 = np.full((s * 3, s * 3, img.shape[2]), 114, dtype=np.uint8) # base image with 4 tiles
+ h0, w0 = h, w
+ c = s, s, s + w, s + h # xmin, ymin, xmax, ymax (base) coordinates
+ elif i == 1: # top
+ c = s, s - h, s + w, s
+ elif i == 2: # top right
+ c = s + wp, s - h, s + wp + w, s
+ elif i == 3: # right
+ c = s + w0, s, s + w0 + w, s + h
+ elif i == 4: # bottom right
+ c = s + w0, s + hp, s + w0 + w, s + hp + h
+ elif i == 5: # bottom
+ c = s + w0 - w, s + h0, s + w0, s + h0 + h
+ elif i == 6: # bottom left
+ c = s + w0 - wp - w, s + h0, s + w0 - wp, s + h0 + h
+ elif i == 7: # left
+ c = s - w, s + h0 - h, s, s + h0
+ elif i == 8: # top left
+ c = s - w, s + h0 - hp - h, s, s + h0 - hp
+
+ padx, pady = c[:2]
+ x1, y1, x2, y2 = (max(x, 0) for x in c) # allocate coords
+
+ # Labels
+ labels, segments = self.labels[index].copy(), self.segments[index].copy()
+ if labels.size:
+ labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padx, pady) # normalized xywh to pixel xyxy format
+ segments = [xyn2xy(x, w, h, padx, pady) for x in segments]
+ labels9.append(labels)
+ segments9.extend(segments)
+
+ # Image
+ img9[y1:y2, x1:x2] = img[y1 - pady:, x1 - padx:] # img9[ymin:ymax, xmin:xmax]
+ hp, wp = h, w # height, width previous
+
+ # Offset
+ yc, xc = (int(random.uniform(0, s)) for _ in self.mosaic_border) # mosaic center x, y
+ img9 = img9[yc:yc + 2 * s, xc:xc + 2 * s]
+
+ # Concat/clip labels
+ labels9 = np.concatenate(labels9, 0)
+ labels9[:, [1, 3]] -= xc
+ labels9[:, [2, 4]] -= yc
+ c = np.array([xc, yc]) # centers
+ segments9 = [x - c for x in segments9]
+
+ for x in (labels9[:, 1:], *segments9):
+ np.clip(x, 0, 2 * s, out=x) # clip when using random_perspective()
+ # img9, labels9 = replicate(img9, labels9) # replicate
+
+ # Augment
+ img9, labels9 = random_perspective(img9,
+ labels9,
+ segments9,
+ degrees=self.hyp['degrees'],
+ translate=self.hyp['translate'],
+ scale=self.hyp['scale'],
+ shear=self.hyp['shear'],
+ perspective=self.hyp['perspective'],
+ border=self.mosaic_border) # border to remove
+
+ return img9, labels9
+
+ @staticmethod
+ def collate_fn(batch):
+ im, label, path, shapes = zip(*batch) # transposed
+ for i, lb in enumerate(label):
+ lb[:, 0] = i # add target image index for build_targets()
+ return torch.stack(im, 0), torch.cat(label, 0), path, shapes
+
+ @staticmethod
+ def collate_fn4(batch):
+ img, label, path, shapes = zip(*batch) # transposed
+ n = len(shapes) // 4
+ im4, label4, path4, shapes4 = [], [], path[:n], shapes[:n]
+
+ ho = torch.tensor([[0.0, 0, 0, 1, 0, 0]])
+ wo = torch.tensor([[0.0, 0, 1, 0, 0, 0]])
+ s = torch.tensor([[1, 1, 0.5, 0.5, 0.5, 0.5]]) # scale
+ for i in range(n): # zidane torch.zeros(16,3,720,1280) # BCHW
+ i *= 4
+ if random.random() < 0.5:
+ im = F.interpolate(img[i].unsqueeze(0).float(), scale_factor=2.0, mode='bilinear',
+ align_corners=False)[0].type(img[i].type())
+ lb = label[i]
+ else:
+ im = torch.cat((torch.cat((img[i], img[i + 1]), 1), torch.cat((img[i + 2], img[i + 3]), 1)), 2)
+ lb = torch.cat((label[i], label[i + 1] + ho, label[i + 2] + wo, label[i + 3] + ho + wo), 0) * s
+ im4.append(im)
+ label4.append(lb)
+
+ for i, lb in enumerate(label4):
+ lb[:, 0] = i # add target image index for build_targets()
+
+ return torch.stack(im4, 0), torch.cat(label4, 0), path4, shapes4
+
+
+# Ancillary functions --------------------------------------------------------------------------------------------------
+def create_folder(path='./new'):
+ # Create folder
+ if os.path.exists(path):
+ shutil.rmtree(path) # delete output folder
+ os.makedirs(path) # make new output folder
+
+
+def flatten_recursive(path=DATASETS_DIR / 'coco128'):
+ # Flatten a recursive directory by bringing all files to top level
+ new_path = Path(str(path) + '_flat')
+ create_folder(new_path)
+ for file in tqdm(glob.glob(str(Path(path)) + '/**/*.*', recursive=True)):
+ shutil.copyfile(file, new_path / Path(file).name)
+
+
+def extract_boxes(path=DATASETS_DIR / 'coco128'): # from utils.datasets import *; extract_boxes()
+ # Convert detection dataset into classification dataset, with one directory per class
+ path = Path(path) # images dir
+ shutil.rmtree(path / 'classifier') if (path / 'classifier').is_dir() else None # remove existing
+ files = list(path.rglob('*.*'))
+ n = len(files) # number of files
+ for im_file in tqdm(files, total=n):
+ if im_file.suffix[1:] in IMG_FORMATS:
+ # image
+ im = cv2.imread(str(im_file))[..., ::-1] # BGR to RGB
+ h, w = im.shape[:2]
+
+ # labels
+ lb_file = Path(img2label_paths([str(im_file)])[0])
+ if Path(lb_file).exists():
+ with open(lb_file) as f:
+ lb = np.array([x.split() for x in f.read().strip().splitlines()], dtype=np.float32) # labels
+
+ for j, x in enumerate(lb):
+ c = int(x[0]) # class
+ f = (path / 'classifier') / f'{c}' / f'{path.stem}_{im_file.stem}_{j}.jpg' # new filename
+ if not f.parent.is_dir():
+ f.parent.mkdir(parents=True)
+
+ b = x[1:] * [w, h, w, h] # box
+ # b[2:] = b[2:].max() # rectangle to square
+ b[2:] = b[2:] * 1.2 + 3 # pad
+ b = xywh2xyxy(b.reshape(-1, 4)).ravel().astype(int)
+
+ b[[0, 2]] = np.clip(b[[0, 2]], 0, w) # clip boxes outside of image
+ b[[1, 3]] = np.clip(b[[1, 3]], 0, h)
+ assert cv2.imwrite(str(f), im[b[1]:b[3], b[0]:b[2]]), f'box failure in {f}'
+
+
+def autosplit(path=DATASETS_DIR / 'coco128/images', weights=(0.9, 0.1, 0.0), annotated_only=False):
+ """ Autosplit a dataset into train/val/test splits and save path/autosplit_*.txt files
+ Usage: from utils.datasets import *; autosplit()
+ Arguments
+ path: Path to images directory
+ weights: Train, val, test weights (list, tuple)
+ annotated_only: Only use images with an annotated txt file
+ """
+ path = Path(path) # images dir
+ files = sorted(x for x in path.rglob('*.*') if x.suffix[1:].lower() in IMG_FORMATS) # image files only
+ n = len(files) # number of files
+ random.seed(0) # for reproducibility
+ indices = random.choices([0, 1, 2], weights=weights, k=n) # assign each image to a split
+
+ txt = ['autosplit_train.txt', 'autosplit_val.txt', 'autosplit_test.txt'] # 3 txt files
+ [(path.parent / x).unlink(missing_ok=True) for x in txt] # remove existing
+
+ print(f'Autosplitting images from {path}' + ', using *.txt labeled images only' * annotated_only)
+ for i, img in tqdm(zip(indices, files), total=n):
+ if not annotated_only or Path(img2label_paths([str(img)])[0]).exists(): # check label
+ with open(path.parent / txt[i], 'a') as f:
+ f.write('./' + img.relative_to(path.parent).as_posix() + '\n') # add image to txt file
+
+
+def verify_image_label(args):
+ # Verify one image-label pair
+ im_file, lb_file, prefix = args
+ nm, nf, ne, nc, msg, segments = 0, 0, 0, 0, '', [] # number (missing, found, empty, corrupt), message, segments
+ try:
+ # verify images
+ im = Image.open(im_file)
+ im.verify() # PIL verify
+ shape = exif_size(im) # image size
+ assert (shape[0] > 9) & (shape[1] > 9), f'image size {shape} <10 pixels'
+ assert im.format.lower() in IMG_FORMATS, f'invalid image format {im.format}'
+ if im.format.lower() in ('jpg', 'jpeg'):
+ with open(im_file, 'rb') as f:
+ f.seek(-2, 2)
+ if f.read() != b'\xff\xd9': # corrupt JPEG
+ ImageOps.exif_transpose(Image.open(im_file)).save(im_file, 'JPEG', subsampling=0, quality=100)
+ msg = f'{prefix}WARNING: {im_file}: corrupt JPEG restored and saved'
+
+ # verify labels
+ if os.path.isfile(lb_file):
+ nf = 1 # label found
+ with open(lb_file) as f:
+ lb = [x.split() for x in f.read().strip().splitlines() if len(x)]
+ if any(len(x) > 6 for x in lb): # is segment
+ classes = np.array([x[0] for x in lb], dtype=np.float32)
+ segments = [np.array(x[1:], dtype=np.float32).reshape(-1, 2) for x in lb] # (cls, xy1...)
+ lb = np.concatenate((classes.reshape(-1, 1), segments2boxes(segments)), 1) # (cls, xywh)
+ lb = np.array(lb, dtype=np.float32)
+ nl = len(lb)
+ if nl:
+ assert lb.shape[1] == 5, f'labels require 5 columns, {lb.shape[1]} columns detected'
+ assert (lb >= 0).all(), f'negative label values {lb[lb < 0]}'
+ assert (lb[:, 1:] <= 1).all(), f'non-normalized or out of bounds coordinates {lb[:, 1:][lb[:, 1:] > 1]}'
+ _, i = np.unique(lb, axis=0, return_index=True)
+ if len(i) < nl: # duplicate row check
+ lb = lb[i] # remove duplicates
+ if segments:
+ segments = segments[i]
+ msg = f'{prefix}WARNING: {im_file}: {nl - len(i)} duplicate labels removed'
+ else:
+ ne = 1 # label empty
+ lb = np.zeros((0, 5), dtype=np.float32)
+ else:
+ nm = 1 # label missing
+ lb = np.zeros((0, 5), dtype=np.float32)
+ return im_file, lb, shape, segments, nm, nf, ne, nc, msg
+ except Exception as e:
+ nc = 1
+ msg = f'{prefix}WARNING: {im_file}: ignoring corrupt image/label: {e}'
+ return [None, None, None, None, nm, nf, ne, nc, msg]
+
+
+def dataset_stats(path='coco128.yaml', autodownload=False, verbose=False, profile=False, hub=False):
+ """ Return dataset statistics dictionary with images and instances counts per split per class
+ To run in parent directory: export PYTHONPATH="$PWD/yolov5"
+ Usage1: from utils.datasets import *; dataset_stats('coco128.yaml', autodownload=True)
+ Usage2: from utils.datasets import *; dataset_stats('path/to/coco128_with_yaml.zip')
+ Arguments
+ path: Path to data.yaml or data.zip (with data.yaml inside data.zip)
+ autodownload: Attempt to download dataset if not found locally
+ verbose: Print stats dictionary
+ """
+
+ def round_labels(labels):
+ # Update labels to integer class and 6 decimal place floats
+ return [[int(c), *(round(x, 4) for x in points)] for c, *points in labels]
+
+ def unzip(path):
+ # Unzip data.zip TODO: CONSTRAINT: path/to/abc.zip MUST unzip to 'path/to/abc/'
+ if str(path).endswith('.zip'): # path is data.zip
+ assert Path(path).is_file(), f'Error unzipping {path}, file not found'
+ ZipFile(path).extractall(path=path.parent) # unzip
+ dir = path.with_suffix('') # dataset directory == zip name
+ return True, str(dir), next(dir.rglob('*.yaml')) # zipped, data_dir, yaml_path
+ else: # path is data.yaml
+ return False, None, path
+
+ def hub_ops(f, max_dim=1920):
+ # HUB ops for 1 image 'f': resize and save at reduced quality in /dataset-hub for web/app viewing
+ f_new = im_dir / Path(f).name # dataset-hub image filename
+ try: # use PIL
+ im = Image.open(f)
+ r = max_dim / max(im.height, im.width) # ratio
+ if r < 1.0: # image too large
+ im = im.resize((int(im.width * r), int(im.height * r)))
+ im.save(f_new, 'JPEG', quality=75, optimize=True) # save
+ except Exception as e: # use OpenCV
+ print(f'WARNING: HUB ops PIL failure {f}: {e}')
+ im = cv2.imread(f)
+ im_height, im_width = im.shape[:2]
+ r = max_dim / max(im_height, im_width) # ratio
+ if r < 1.0: # image too large
+ im = cv2.resize(im, (int(im_width * r), int(im_height * r)), interpolation=cv2.INTER_AREA)
+ cv2.imwrite(str(f_new), im)
+
+ zipped, data_dir, yaml_path = unzip(Path(path))
+ with open(check_yaml(yaml_path), errors='ignore') as f:
+ data = yaml.safe_load(f) # data dict
+ if zipped:
+ data['path'] = data_dir # TODO: should this be dir.resolve()?
+ check_dataset(data, autodownload) # download dataset if missing
+ hub_dir = Path(data['path'] + ('-hub' if hub else ''))
+ stats = {'nc': data['nc'], 'names': data['names']} # statistics dictionary
+ for split in 'train', 'val', 'test':
+ if data.get(split) is None:
+ stats[split] = None # i.e. no test set
+ continue
+ x = []
+ dataset = LoadImagesAndLabels(data[split]) # load dataset
+ for label in tqdm(dataset.labels, total=dataset.n, desc='Statistics'):
+ x.append(np.bincount(label[:, 0].astype(int), minlength=data['nc']))
+ x = np.array(x) # shape(128x80)
+ stats[split] = {
+ 'instance_stats': {
+ 'total': int(x.sum()),
+ 'per_class': x.sum(0).tolist()},
+ 'image_stats': {
+ 'total': dataset.n,
+ 'unlabelled': int(np.all(x == 0, 1).sum()),
+ 'per_class': (x > 0).sum(0).tolist()},
+ 'labels': [{
+ str(Path(k).name): round_labels(v.tolist())} for k, v in zip(dataset.im_files, dataset.labels)]}
+
+ if hub:
+ im_dir = hub_dir / 'images'
+ im_dir.mkdir(parents=True, exist_ok=True)
+ for _ in tqdm(ThreadPool(NUM_THREADS).imap(hub_ops, dataset.im_files), total=dataset.n, desc='HUB Ops'):
+ pass
+
+ # Profile
+ stats_path = hub_dir / 'stats.json'
+ if profile:
+ for _ in range(1):
+ file = stats_path.with_suffix('.npy')
+ t1 = time.time()
+ np.save(file, stats)
+ t2 = time.time()
+ x = np.load(file, allow_pickle=True)
+ print(f'stats.npy times: {time.time() - t2:.3f}s read, {t2 - t1:.3f}s write')
+
+ file = stats_path.with_suffix('.json')
+ t1 = time.time()
+ with open(file, 'w') as f:
+ json.dump(stats, f) # save stats *.json
+ t2 = time.time()
+ with open(file) as f:
+ x = json.load(f) # load hyps dict
+ print(f'stats.json times: {time.time() - t2:.3f}s read, {t2 - t1:.3f}s write')
+
+ # Save, print and return
+ if hub:
+ print(f'Saving {stats_path.resolve()}...')
+ with open(stats_path, 'w') as f:
+ json.dump(stats, f) # save stats.json
+ if verbose:
+ print(json.dumps(stats, indent=2, sort_keys=False))
+ return stats
diff --git a/vouchervision/component_detector/utils/docker/.dockerignore b/vouchervision/component_detector/utils/docker/.dockerignore
new file mode 100644
index 0000000000000000000000000000000000000000..af51ccc3d8df7681ca03ea6f5b669bac37e6baa6
--- /dev/null
+++ b/vouchervision/component_detector/utils/docker/.dockerignore
@@ -0,0 +1,222 @@
+# Repo-specific DockerIgnore -------------------------------------------------------------------------------------------
+#.git
+.cache
+.idea
+runs
+output
+coco
+storage.googleapis.com
+
+data/samples/*
+**/results*.csv
+*.jpg
+
+# Neural Network weights -----------------------------------------------------------------------------------------------
+**/*.pt
+**/*.pth
+**/*.onnx
+**/*.engine
+**/*.mlmodel
+**/*.torchscript
+**/*.torchscript.pt
+**/*.tflite
+**/*.h5
+**/*.pb
+*_saved_model/
+*_web_model/
+*_openvino_model/
+
+# Below Copied From .gitignore -----------------------------------------------------------------------------------------
+# Below Copied From .gitignore -----------------------------------------------------------------------------------------
+
+
+# GitHub Python GitIgnore ----------------------------------------------------------------------------------------------
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+env/
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+wandb/
+.installed.cfg
+*.egg
+
+# PyInstaller
+# Usually these files are written by a python script from a template
+# before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+.hypothesis/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# pyenv
+.python-version
+
+# celery beat schedule file
+celerybeat-schedule
+
+# SageMath parsed files
+*.sage.py
+
+# dotenv
+.env
+
+# virtualenv
+.venv*
+venv*/
+ENV*/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+
+
+# https://github.com/github/gitignore/blob/master/Global/macOS.gitignore -----------------------------------------------
+
+# General
+.DS_Store
+.AppleDouble
+.LSOverride
+
+# Icon must end with two \r
+Icon
+Icon?
+
+# Thumbnails
+._*
+
+# Files that might appear in the root of a volume
+.DocumentRevisions-V100
+.fseventsd
+.Spotlight-V100
+.TemporaryItems
+.Trashes
+.VolumeIcon.icns
+.com.apple.timemachine.donotpresent
+
+# Directories potentially created on remote AFP share
+.AppleDB
+.AppleDesktop
+Network Trash Folder
+Temporary Items
+.apdisk
+
+
+# https://github.com/github/gitignore/blob/master/Global/JetBrains.gitignore
+# Covers JetBrains IDEs: IntelliJ, RubyMine, PhpStorm, AppCode, PyCharm, CLion, Android Studio and WebStorm
+# Reference: https://intellij-support.jetbrains.com/hc/en-us/articles/206544839
+
+# User-specific stuff:
+.idea/*
+.idea/**/workspace.xml
+.idea/**/tasks.xml
+.idea/dictionaries
+.html # Bokeh Plots
+.pg # TensorFlow Frozen Graphs
+.avi # videos
+
+# Sensitive or high-churn files:
+.idea/**/dataSources/
+.idea/**/dataSources.ids
+.idea/**/dataSources.local.xml
+.idea/**/sqlDataSources.xml
+.idea/**/dynamic.xml
+.idea/**/uiDesigner.xml
+
+# Gradle:
+.idea/**/gradle.xml
+.idea/**/libraries
+
+# CMake
+cmake-build-debug/
+cmake-build-release/
+
+# Mongo Explorer plugin:
+.idea/**/mongoSettings.xml
+
+## File-based project format:
+*.iws
+
+## Plugin-specific files:
+
+# IntelliJ
+out/
+
+# mpeltonen/sbt-idea plugin
+.idea_modules/
+
+# JIRA plugin
+atlassian-ide-plugin.xml
+
+# Cursive Clojure plugin
+.idea/replstate.xml
+
+# Crashlytics plugin (for Android Studio and IntelliJ)
+com_crashlytics_export_strings.xml
+crashlytics.properties
+crashlytics-build.properties
+fabric.properties
diff --git a/vouchervision/component_detector/utils/docker/Dockerfile b/vouchervision/component_detector/utils/docker/Dockerfile
new file mode 100644
index 0000000000000000000000000000000000000000..9bb24bb6bf3e6fcfe3e32ce157d017729e8f0390
--- /dev/null
+++ b/vouchervision/component_detector/utils/docker/Dockerfile
@@ -0,0 +1,65 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Start FROM Nvidia PyTorch image https://ngc.nvidia.com/catalog/containers/nvidia:pytorch
+FROM nvcr.io/nvidia/pytorch:21.10-py3
+
+# Install linux packages
+RUN apt update && apt install -y zip htop screen libgl1-mesa-glx
+
+# Install python dependencies
+COPY requirements.txt .
+RUN python -m pip install --upgrade pip
+RUN pip uninstall -y torch torchvision torchtext
+RUN pip install --no-cache -r requirements.txt albumentations wandb gsutil notebook \
+ torch==1.11.0+cu113 torchvision==0.12.0+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
+# RUN pip install --no-cache -U torch torchvision
+
+# Create working directory
+RUN mkdir -p /usr/src/app
+WORKDIR /usr/src/app
+
+# Copy contents
+COPY . /usr/src/app
+RUN git clone https://github.com/ultralytics/yolov5 /usr/src/yolov5
+
+# Downloads to user config dir
+ADD https://ultralytics.com/assets/Arial.ttf https://ultralytics.com/assets/Arial.Unicode.ttf /root/.config/Ultralytics/
+
+# Set environment variables
+ENV OMP_NUM_THREADS=8
+
+
+# Usage Examples -------------------------------------------------------------------------------------------------------
+
+# Build and Push
+# t=ultralytics/yolov5:latest && sudo docker build -f utils/docker/Dockerfile -t $t . && sudo docker push $t
+
+# Pull and Run
+# t=ultralytics/yolov5:latest && sudo docker pull $t && sudo docker run -it --ipc=host --gpus all $t
+
+# Pull and Run with local directory access
+# t=ultralytics/yolov5:latest && sudo docker pull $t && sudo docker run -it --ipc=host --gpus all -v "$(pwd)"/datasets:/usr/src/datasets $t
+
+# Kill all
+# sudo docker kill $(sudo docker ps -q)
+
+# Kill all image-based
+# sudo docker kill $(sudo docker ps -qa --filter ancestor=ultralytics/yolov5:latest)
+
+# Bash into running container
+# sudo docker exec -it 5a9b5863d93d bash
+
+# Bash into stopped container
+# id=$(sudo docker ps -qa) && sudo docker start $id && sudo docker exec -it $id bash
+
+# Clean up
+# docker system prune -a --volumes
+
+# Update Ubuntu drivers
+# https://www.maketecheasier.com/install-nvidia-drivers-ubuntu/
+
+# DDP test
+# python -m torch.distributed.run --nproc_per_node 2 --master_port 1 train.py --epochs 3
+
+# GCP VM from Image
+# docker.io/ultralytics/yolov5:latest
diff --git a/vouchervision/component_detector/utils/docker/Dockerfile-cpu b/vouchervision/component_detector/utils/docker/Dockerfile-cpu
new file mode 100644
index 0000000000000000000000000000000000000000..d30c07e8117286da61dd45659135f35549918b7a
--- /dev/null
+++ b/vouchervision/component_detector/utils/docker/Dockerfile-cpu
@@ -0,0 +1,37 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Start FROM Ubuntu image https://hub.docker.com/_/ubuntu
+FROM ubuntu:latest
+
+# Install linux packages
+RUN apt update
+RUN DEBIAN_FRONTEND=noninteractive TZ=Etc/UTC apt install -y tzdata
+RUN apt install -y python3-pip git zip curl htop screen libgl1-mesa-glx libglib2.0-0
+RUN alias python=python3
+
+# Install python dependencies
+COPY requirements.txt .
+RUN python3 -m pip install --upgrade pip
+RUN pip install --no-cache -r requirements.txt albumentations gsutil notebook \
+ coremltools onnx onnx-simplifier onnxruntime openvino-dev tensorflow-cpu tensorflowjs \
+ torch==1.11.0+cpu torchvision==0.12.0+cpu -f https://download.pytorch.org/whl/cpu/torch_stable.html
+
+# Create working directory
+RUN mkdir -p /usr/src/app
+WORKDIR /usr/src/app
+
+# Copy contents
+COPY . /usr/src/app
+RUN git clone https://github.com/ultralytics/yolov5 /usr/src/yolov5
+
+# Downloads to user config dir
+ADD https://ultralytics.com/assets/Arial.ttf https://ultralytics.com/assets/Arial.Unicode.ttf /root/.config/Ultralytics/
+
+
+# Usage Examples -------------------------------------------------------------------------------------------------------
+
+# Build and Push
+# t=ultralytics/yolov5:latest-cpu && sudo docker build -f utils/docker/Dockerfile-cpu -t $t . && sudo docker push $t
+
+# Pull and Run
+# t=ultralytics/yolov5:latest-cpu && sudo docker pull $t && sudo docker run -it --ipc=host -v "$(pwd)"/datasets:/usr/src/datasets $t
diff --git a/vouchervision/component_detector/utils/downloads.py b/vouchervision/component_detector/utils/downloads.py
new file mode 100644
index 0000000000000000000000000000000000000000..776a8bba1755d471f3052cd3434d707137a10e54
--- /dev/null
+++ b/vouchervision/component_detector/utils/downloads.py
@@ -0,0 +1,161 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Download utils
+"""
+
+import logging
+import os
+import platform
+import subprocess
+import time
+import urllib
+from pathlib import Path
+from zipfile import ZipFile
+
+import requests
+import torch
+
+
+def gsutil_getsize(url=''):
+ # gs://bucket/file size https://cloud.google.com/storage/docs/gsutil/commands/du
+ s = subprocess.check_output(f'gsutil du {url}', shell=True).decode('utf-8')
+ return eval(s.split(' ')[0]) if len(s) else 0 # bytes
+
+
+def safe_download(file, url, url2=None, min_bytes=1E0, error_msg=''):
+ # Attempts to download file from url or url2, checks and removes incomplete downloads < min_bytes
+ from utils.general import LOGGER
+
+ file = Path(file)
+ assert_msg = f"Downloaded file '{file}' does not exist or size is < min_bytes={min_bytes}"
+ try: # url1
+ LOGGER.info(f'Downloading {url} to {file}...')
+ torch.hub.download_url_to_file(url, str(file), progress=LOGGER.level <= logging.INFO)
+ assert file.exists() and file.stat().st_size > min_bytes, assert_msg # check
+ except Exception as e: # url2
+ file.unlink(missing_ok=True) # remove partial downloads
+ LOGGER.info(f'ERROR: {e}\nRe-attempting {url2 or url} to {file}...')
+ os.system(f"curl -L '{url2 or url}' -o '{file}' --retry 3 -C -") # curl download, retry and resume on fail
+ finally:
+ if not file.exists() or file.stat().st_size < min_bytes: # check
+ file.unlink(missing_ok=True) # remove partial downloads
+ LOGGER.info(f"ERROR: {assert_msg}\n{error_msg}")
+ LOGGER.info('')
+
+
+def attempt_download(file, repo='ultralytics/yolov5'): # from utils.downloads import *; attempt_download()
+ # Attempt file download if does not exist
+ from utils.general import LOGGER
+
+ file = Path(str(file).strip().replace("'", ''))
+ if not file.exists():
+ # URL specified
+ name = Path(urllib.parse.unquote(str(file))).name # decode '%2F' to '/' etc.
+ if str(file).startswith(('http:/', 'https:/')): # download
+ url = str(file).replace(':/', '://') # Pathlib turns :// -> :/
+ file = name.split('?')[0] # parse authentication https://url.com/file.txt?auth...
+ if Path(file).is_file():
+ LOGGER.info(f'Found {url} locally at {file}') # file already exists
+ else:
+ safe_download(file=file, url=url, min_bytes=1E5)
+ return file
+
+ # GitHub assets
+ file.parent.mkdir(parents=True, exist_ok=True) # make parent dir (if required)
+ try:
+ response = requests.get(f'https://api.github.com/repos/{repo}/releases/latest').json() # github api
+ assets = [x['name'] for x in response['assets']] # release assets, i.e. ['yolov5s.pt', 'yolov5m.pt', ...]
+ tag = response['tag_name'] # i.e. 'v1.0'
+ except Exception: # fallback plan
+ assets = [
+ 'yolov5n.pt', 'yolov5s.pt', 'yolov5m.pt', 'yolov5l.pt', 'yolov5x.pt', 'yolov5n6.pt', 'yolov5s6.pt',
+ 'yolov5m6.pt', 'yolov5l6.pt', 'yolov5x6.pt']
+ try:
+ tag = subprocess.check_output('git tag', shell=True, stderr=subprocess.STDOUT).decode().split()[-1]
+ except Exception:
+ tag = 'v6.1' # current release
+
+ if name in assets:
+ url3 = 'https://drive.google.com/drive/folders/1EFQTEUeXWSFww0luse2jB9M1QNZQGwNl' # backup gdrive mirror
+ safe_download(
+ file,
+ url=f'https://github.com/{repo}/releases/download/{tag}/{name}',
+ url2=f'https://storage.googleapis.com/{repo}/{tag}/{name}', # backup url (optional)
+ min_bytes=1E5,
+ error_msg=f'{file} missing, try downloading from https://github.com/{repo}/releases/{tag} or {url3}')
+
+ return str(file)
+
+
+def gdrive_download(id='16TiPfZj7htmTyhntwcZyEEAejOUxuT6m', file='tmp.zip'):
+ # Downloads a file from Google Drive. from yolov5.utils.downloads import *; gdrive_download()
+ t = time.time()
+ file = Path(file)
+ cookie = Path('cookie') # gdrive cookie
+ print(f'Downloading https://drive.google.com/uc?export=download&id={id} as {file}... ', end='')
+ file.unlink(missing_ok=True) # remove existing file
+ cookie.unlink(missing_ok=True) # remove existing cookie
+
+ # Attempt file download
+ out = "NUL" if platform.system() == "Windows" else "/dev/null"
+ os.system(f'curl -c ./cookie -s -L "drive.google.com/uc?export=download&id={id}" > {out}')
+ if os.path.exists('cookie'): # large file
+ s = f'curl -Lb ./cookie "drive.google.com/uc?export=download&confirm={get_token()}&id={id}" -o {file}'
+ else: # small file
+ s = f'curl -s -L -o {file} "drive.google.com/uc?export=download&id={id}"'
+ r = os.system(s) # execute, capture return
+ cookie.unlink(missing_ok=True) # remove existing cookie
+
+ # Error check
+ if r != 0:
+ file.unlink(missing_ok=True) # remove partial
+ print('Download error ') # raise Exception('Download error')
+ return r
+
+ # Unzip if archive
+ if file.suffix == '.zip':
+ print('unzipping... ', end='')
+ ZipFile(file).extractall(path=file.parent) # unzip
+ file.unlink() # remove zip
+
+ print(f'Done ({time.time() - t:.1f}s)')
+ return r
+
+
+def get_token(cookie="./cookie"):
+ with open(cookie) as f:
+ for line in f:
+ if "download" in line:
+ return line.split()[-1]
+ return ""
+
+
+# Google utils: https://cloud.google.com/storage/docs/reference/libraries ----------------------------------------------
+#
+#
+# def upload_blob(bucket_name, source_file_name, destination_blob_name):
+# # Uploads a file to a bucket
+# # https://cloud.google.com/storage/docs/uploading-objects#storage-upload-object-python
+#
+# storage_client = storage.Client()
+# bucket = storage_client.get_bucket(bucket_name)
+# blob = bucket.blob(destination_blob_name)
+#
+# blob.upload_from_filename(source_file_name)
+#
+# print('File {} uploaded to {}.'.format(
+# source_file_name,
+# destination_blob_name))
+#
+#
+# def download_blob(bucket_name, source_blob_name, destination_file_name):
+# # Uploads a blob from a bucket
+# storage_client = storage.Client()
+# bucket = storage_client.get_bucket(bucket_name)
+# blob = bucket.blob(source_blob_name)
+#
+# blob.download_to_filename(destination_file_name)
+#
+# print('Blob {} downloaded to {}.'.format(
+# source_blob_name,
+# destination_file_name))
diff --git a/vouchervision/component_detector/utils/downloads_torchscript.py b/vouchervision/component_detector/utils/downloads_torchscript.py
new file mode 100644
index 0000000000000000000000000000000000000000..1955fbe859767abce5dbe916cea944173cb83c2b
--- /dev/null
+++ b/vouchervision/component_detector/utils/downloads_torchscript.py
@@ -0,0 +1,127 @@
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+Download utils
+"""
+
+import logging
+import subprocess
+import urllib
+from pathlib import Path
+
+import requests
+import torch
+
+
+def is_url(url, check=True):
+ # Check if string is URL and check if URL exists
+ try:
+ url = str(url)
+ result = urllib.parse.urlparse(url)
+ assert all([result.scheme, result.netloc]) # check if is url
+ return (urllib.request.urlopen(url).getcode() == 200) if check else True # check if exists online
+ except (AssertionError, urllib.request.HTTPError):
+ return False
+
+
+def gsutil_getsize(url=''):
+ # gs://bucket/file size https://cloud.google.com/storage/docs/gsutil/commands/du
+ output = subprocess.check_output(['gsutil', 'du', url], shell=True, encoding='utf-8')
+ if output:
+ return int(output.split()[0])
+ return 0
+
+
+def url_getsize(url='https://ultralytics.com/images/bus.jpg'):
+ # Return downloadable file size in bytes
+ response = requests.head(url, allow_redirects=True)
+ return int(response.headers.get('content-length', -1))
+
+
+def curl_download(url, filename, *, silent: bool = False) -> bool:
+ """
+ Download a file from a url to a filename using curl.
+ """
+ silent_option = 'sS' if silent else '' # silent
+ proc = subprocess.run([
+ 'curl',
+ '-#',
+ f'-{silent_option}L',
+ url,
+ '--output',
+ filename,
+ '--retry',
+ '9',
+ '-C',
+ '-', ])
+ return proc.returncode == 0
+
+
+def safe_download(file, url, url2=None, min_bytes=1E0, error_msg=''):
+ # Attempts to download file from url or url2, checks and removes incomplete downloads < min_bytes
+ from utils.general import LOGGER
+
+ file = Path(file)
+ assert_msg = f"Downloaded file '{file}' does not exist or size is < min_bytes={min_bytes}"
+ try: # url1
+ LOGGER.info(f'Downloading {url} to {file}...')
+ torch.hub.download_url_to_file(url, str(file), progress=LOGGER.level <= logging.INFO)
+ assert file.exists() and file.stat().st_size > min_bytes, assert_msg # check
+ except Exception as e: # url2
+ if file.exists():
+ file.unlink() # remove partial downloads
+ LOGGER.info(f'ERROR: {e}\nRe-attempting {url2 or url} to {file}...')
+ # curl download, retry and resume on fail
+ curl_download(url2 or url, file)
+ finally:
+ if not file.exists() or file.stat().st_size < min_bytes: # check
+ if file.exists():
+ file.unlink() # remove partial downloads
+ LOGGER.info(f'ERROR: {assert_msg}\n{error_msg}')
+ LOGGER.info('')
+
+
+def attempt_download(file, repo='ultralytics/yolov5', release='v7.0'):
+ # Attempt file download from GitHub release assets if not found locally. release = 'latest', 'v7.0', etc.
+ from utils.general import LOGGER
+
+ def github_assets(repository, version='latest'):
+ # Return GitHub repo tag (i.e. 'v7.0') and assets (i.e. ['yolov5s.pt', 'yolov5m.pt', ...])
+ if version != 'latest':
+ version = f'tags/{version}' # i.e. tags/v7.0
+ response = requests.get(f'https://api.github.com/repos/{repository}/releases/{version}').json() # github api
+ return response['tag_name'], [x['name'] for x in response['assets']] # tag, assets
+
+ file = Path(str(file).strip().replace("'", ''))
+ if not file.exists():
+ # URL specified
+ name = Path(urllib.parse.unquote(str(file))).name # decode '%2F' to '/' etc.
+ if str(file).startswith(('http:/', 'https:/')): # download
+ url = str(file).replace(':/', '://') # Pathlib turns :// -> :/
+ file = name.split('?')[0] # parse authentication https://url.com/file.txt?auth...
+ if Path(file).is_file():
+ LOGGER.info(f'Found {url} locally at {file}') # file already exists
+ else:
+ safe_download(file=file, url=url, min_bytes=1E5)
+ return file
+
+ # GitHub assets
+ assets = [f'yolov5{size}{suffix}.pt' for size in 'nsmlx' for suffix in ('', '6', '-cls', '-seg')] # default
+ try:
+ tag, assets = github_assets(repo, release)
+ except Exception:
+ try:
+ tag, assets = github_assets(repo) # latest release
+ except Exception:
+ try:
+ tag = subprocess.check_output('git tag', shell=True, stderr=subprocess.STDOUT).decode().split()[-1]
+ except Exception:
+ tag = release
+
+ if name in assets:
+ file.parent.mkdir(parents=True, exist_ok=True) # make parent dir (if required)
+ safe_download(file,
+ url=f'https://github.com/{repo}/releases/download/{tag}/{name}',
+ min_bytes=1E5,
+ error_msg=f'{file} missing, try downloading from https://github.com/{repo}/releases/{tag}')
+
+ return str(file)
\ No newline at end of file
diff --git a/vouchervision/component_detector/utils/flask_rest_api/README.md b/vouchervision/component_detector/utils/flask_rest_api/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..a726acbd92043458311dd949cc09c0195cd35400
--- /dev/null
+++ b/vouchervision/component_detector/utils/flask_rest_api/README.md
@@ -0,0 +1,73 @@
+# Flask REST API
+
+[REST](https://en.wikipedia.org/wiki/Representational_state_transfer) [API](https://en.wikipedia.org/wiki/API)s are
+commonly used to expose Machine Learning (ML) models to other services. This folder contains an example REST API
+created using Flask to expose the YOLOv5s model from [PyTorch Hub](https://pytorch.org/hub/ultralytics_yolov5/).
+
+## Requirements
+
+[Flask](https://palletsprojects.com/p/flask/) is required. Install with:
+
+```shell
+$ pip install Flask
+```
+
+## Run
+
+After Flask installation run:
+
+```shell
+$ python3 restapi.py --port 5000
+```
+
+Then use [curl](https://curl.se/) to perform a request:
+
+```shell
+$ curl -X POST -F image=@zidane.jpg 'http://localhost:5000/v1/object-detection/yolov5s'
+```
+
+The model inference results are returned as a JSON response:
+
+```json
+[
+ {
+ "class": 0,
+ "confidence": 0.8900438547,
+ "height": 0.9318675399,
+ "name": "person",
+ "width": 0.3264600933,
+ "xcenter": 0.7438579798,
+ "ycenter": 0.5207948685
+ },
+ {
+ "class": 0,
+ "confidence": 0.8440024257,
+ "height": 0.7155083418,
+ "name": "person",
+ "width": 0.6546785235,
+ "xcenter": 0.427829951,
+ "ycenter": 0.6334488392
+ },
+ {
+ "class": 27,
+ "confidence": 0.3771208823,
+ "height": 0.3902671337,
+ "name": "tie",
+ "width": 0.0696444362,
+ "xcenter": 0.3675483763,
+ "ycenter": 0.7991207838
+ },
+ {
+ "class": 27,
+ "confidence": 0.3527112305,
+ "height": 0.1540903747,
+ "name": "tie",
+ "width": 0.0336618312,
+ "xcenter": 0.7814827561,
+ "ycenter": 0.5065554976
+ }
+]
+```
+
+An example python script to perform inference using [requests](https://docs.python-requests.org/en/master/) is given
+in `example_request.py`
diff --git a/vouchervision/component_detector/utils/flask_rest_api/example_request.py b/vouchervision/component_detector/utils/flask_rest_api/example_request.py
new file mode 100644
index 0000000000000000000000000000000000000000..773ad893296750992789a77a59e0f5ad657d0e35
--- /dev/null
+++ b/vouchervision/component_detector/utils/flask_rest_api/example_request.py
@@ -0,0 +1,19 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Perform test request
+"""
+
+import pprint
+
+import requests
+
+DETECTION_URL = "http://localhost:5000/v1/object-detection/yolov5s"
+IMAGE = "zidane.jpg"
+
+# Read image
+with open(IMAGE, "rb") as f:
+ image_data = f.read()
+
+response = requests.post(DETECTION_URL, files={"image": image_data}).json()
+
+pprint.pprint(response)
diff --git a/vouchervision/component_detector/utils/flask_rest_api/restapi.py b/vouchervision/component_detector/utils/flask_rest_api/restapi.py
new file mode 100644
index 0000000000000000000000000000000000000000..7e7b900107b5055e1e94d4a02748e55e6bdc4827
--- /dev/null
+++ b/vouchervision/component_detector/utils/flask_rest_api/restapi.py
@@ -0,0 +1,46 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Run a Flask REST API exposing a YOLOv5s model
+"""
+
+import argparse
+import io
+
+import torch
+from flask import Flask, request
+from PIL import Image
+
+app = Flask(__name__)
+
+DETECTION_URL = "/v1/object-detection/yolov5s"
+
+
+@app.route(DETECTION_URL, methods=["POST"])
+def predict():
+ if not request.method == "POST":
+ return
+
+ if request.files.get("image"):
+ # Method 1
+ # with request.files["image"] as f:
+ # im = Image.open(io.BytesIO(f.read()))
+
+ # Method 2
+ im_file = request.files["image"]
+ im_bytes = im_file.read()
+ im = Image.open(io.BytesIO(im_bytes))
+
+ results = model(im, size=640) # reduce size=320 for faster inference
+ return results.pandas().xyxy[0].to_json(orient="records")
+
+
+if __name__ == "__main__":
+ parser = argparse.ArgumentParser(description="Flask API exposing YOLOv5 model")
+ parser.add_argument("--port", default=5000, type=int, help="port number")
+ opt = parser.parse_args()
+
+ # Fix known issue urllib.error.HTTPError 403: rate limit exceeded https://github.com/ultralytics/yolov5/pull/7210
+ torch.hub._validate_not_a_forked_repo = lambda a, b, c: True
+
+ model = torch.hub.load("ultralytics/yolov5", "yolov5s", force_reload=True) # force_reload to recache
+ app.run(host="0.0.0.0", port=opt.port) # debug=True causes Restarting with stat
diff --git a/vouchervision/component_detector/utils/general.py b/vouchervision/component_detector/utils/general.py
new file mode 100644
index 0000000000000000000000000000000000000000..f08c01467d392150bf53c3bd7362153b8d9621aa
--- /dev/null
+++ b/vouchervision/component_detector/utils/general.py
@@ -0,0 +1,1019 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+General utils
+"""
+
+import contextlib
+import glob
+import inspect
+import logging
+import math
+import os
+import platform
+import random
+import re
+import shutil
+import signal
+import time
+import urllib
+from datetime import datetime
+from itertools import repeat
+from multiprocessing.pool import ThreadPool
+from pathlib import Path
+from subprocess import check_output
+from typing import Optional
+from zipfile import ZipFile
+
+import cv2
+import numpy as np
+import pandas as pd
+import pkg_resources as pkg
+import torch
+import torchvision
+import yaml
+
+# from utils.downloads import gsutil_getsize
+# from utils.metrics import box_iou, fitness
+
+# Settings
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1] # YOLOv5 root directory
+DATASETS_DIR = ROOT.parent / 'datasets' # YOLOv5 datasets directory
+NUM_THREADS = min(8, max(1, os.cpu_count() - 1)) # number of YOLOv5 multiprocessing threads
+AUTOINSTALL = str(os.getenv('YOLOv5_AUTOINSTALL', True)).lower() == 'true' # global auto-install mode
+VERBOSE = str(os.getenv('YOLOv5_VERBOSE', True)).lower() == 'true' # global verbose mode
+FONT = 'Arial.ttf' # https://ultralytics.com/assets/Arial.ttf
+
+torch.set_printoptions(linewidth=320, precision=5, profile='long')
+np.set_printoptions(linewidth=320, formatter={'float_kind': '{:11.5g}'.format}) # format short g, %precision=5
+pd.options.display.max_columns = 10
+cv2.setNumThreads(0) # prevent OpenCV from multithreading (incompatible with PyTorch DataLoader)
+os.environ['NUMEXPR_MAX_THREADS'] = str(NUM_THREADS) # NumExpr max threads
+os.environ['OMP_NUM_THREADS'] = str(NUM_THREADS) # OpenMP max threads (PyTorch and SciPy)
+
+def box_iou(box1, box2):
+ # https://github.com/pytorch/vision/blob/master/torchvision/ops/boxes.py
+ """
+ Return intersection-over-union (Jaccard index) of boxes.
+ Both sets of boxes are expected to be in (x1, y1, x2, y2) format.
+ Arguments:
+ box1 (Tensor[N, 4])
+ box2 (Tensor[M, 4])
+ Returns:
+ iou (Tensor[N, M]): the NxM matrix containing the pairwise
+ IoU values for every element in boxes1 and boxes2
+ """
+
+ # inter(N,M) = (rb(N,M,2) - lt(N,M,2)).clamp(0).prod(2)
+ (a1, a2), (b1, b2) = box1[:, None].chunk(2, 2), box2.chunk(2, 1)
+ inter = (torch.min(a2, b2) - torch.max(a1, b1)).clamp(0).prod(2)
+
+ # IoU = inter / (area1 + area2 - inter)
+ return inter / (box_area(box1.T)[:, None] + box_area(box2.T) - inter)
+
+def fitness(x):
+ # Model fitness as a weighted combination of metrics
+ w = [0.0, 0.0, 0.1, 0.9] # weights for [P, R, mAP@0.5, mAP@0.5:0.95]
+ return (x[:, :4] * w).sum(1)
+
+def gsutil_getsize(url=''):
+ # gs://bucket/file size https://cloud.google.com/storage/docs/gsutil/commands/du
+ s = subprocess.check_output(f'gsutil du {url}', shell=True).decode('utf-8')
+ return eval(s.split(' ')[0]) if len(s) else 0 # bytes
+
+
+def safe_download(file, url, url2=None, min_bytes=1E0, error_msg=''):
+ # Attempts to download file from url or url2, checks and removes incomplete downloads < min_bytes
+ from utils.general import LOGGER
+
+ file = Path(file)
+ assert_msg = f"Downloaded file '{file}' does not exist or size is < min_bytes={min_bytes}"
+ try: # url1
+ LOGGER.info(f'Downloading {url} to {file}...')
+ torch.hub.download_url_to_file(url, str(file), progress=LOGGER.level <= logging.INFO)
+ assert file.exists() and file.stat().st_size > min_bytes, assert_msg # check
+ except Exception as e: # url2
+ file.unlink(missing_ok=True) # remove partial downloads
+ LOGGER.info(f'ERROR: {e}\nRe-attempting {url2 or url} to {file}...')
+ os.system(f"curl -L '{url2 or url}' -o '{file}' --retry 3 -C -") # curl download, retry and resume on fail
+ finally:
+ if not file.exists() or file.stat().st_size < min_bytes: # check
+ file.unlink(missing_ok=True) # remove partial downloads
+ LOGGER.info(f"ERROR: {assert_msg}\n{error_msg}")
+ LOGGER.info('')
+
+def is_kaggle():
+ # Is environment a Kaggle Notebook?
+ try:
+ assert os.environ.get('PWD') == '/kaggle/working'
+ assert os.environ.get('KAGGLE_URL_BASE') == 'https://www.kaggle.com'
+ return True
+ except AssertionError:
+ return False
+
+
+def is_writeable(dir, test=False):
+ # Return True if directory has write permissions, test opening a file with write permissions if test=True
+ if test: # method 1
+ file = Path(dir) / 'tmp.txt'
+ try:
+ with open(file, 'w'): # open file with write permissions
+ pass
+ file.unlink() # remove file
+ return True
+ except OSError:
+ return False
+ else: # method 2
+ return os.access(dir, os.R_OK) # possible issues on Windows
+
+
+def set_logging(name=None, verbose=VERBOSE):
+ # Sets level and returns logger
+ if is_kaggle():
+ for h in logging.root.handlers:
+ logging.root.removeHandler(h) # remove all handlers associated with the root logger object
+ rank = int(os.getenv('RANK', -1)) # rank in world for Multi-GPU trainings
+ level = logging.INFO if (verbose and rank in (-1, 0)) else logging.WARNING
+ log = logging.getLogger(name)
+ log.setLevel(level)
+ handler = logging.StreamHandler()
+ handler.setFormatter(logging.Formatter("%(message)s"))
+ handler.setLevel(level)
+ log.addHandler(handler)
+
+
+set_logging() # run before defining LOGGER
+LOGGER = logging.getLogger("yolov5") # define globally (used in train.py, val.py, detect.py, etc.)
+
+
+def user_config_dir(dir='Ultralytics', env_var='YOLOV5_CONFIG_DIR'):
+ # Return path of user configuration directory. Prefer environment variable if exists. Make dir if required.
+ env = os.getenv(env_var)
+ if env:
+ path = Path(env) # use environment variable
+ else:
+ cfg = {'Windows': 'AppData/Roaming', 'Linux': '.config', 'Darwin': 'Library/Application Support'} # 3 OS dirs
+ path = Path.home() / cfg.get(platform.system(), '') # OS-specific config dir
+ path = (path if is_writeable(path) else Path('/tmp')) / dir # GCP and AWS lambda fix, only /tmp is writeable
+ path.mkdir(exist_ok=True) # make if required
+ return path
+
+
+CONFIG_DIR = user_config_dir() # Ultralytics settings dir
+
+
+class Profile(contextlib.ContextDecorator):
+ # Usage: @Profile() decorator or 'with Profile():' context manager
+ def __enter__(self):
+ self.start = time.time()
+
+ def __exit__(self, type, value, traceback):
+ print(f'Profile results: {time.time() - self.start:.5f}s')
+
+
+class Timeout(contextlib.ContextDecorator):
+ # Usage: @Timeout(seconds) decorator or 'with Timeout(seconds):' context manager
+ def __init__(self, seconds, *, timeout_msg='', suppress_timeout_errors=True):
+ self.seconds = int(seconds)
+ self.timeout_message = timeout_msg
+ self.suppress = bool(suppress_timeout_errors)
+
+ def _timeout_handler(self, signum, frame):
+ raise TimeoutError(self.timeout_message)
+
+ def __enter__(self):
+ if platform.system() != 'Windows': # not supported on Windows
+ signal.signal(signal.SIGALRM, self._timeout_handler) # Set handler for SIGALRM
+ signal.alarm(self.seconds) # start countdown for SIGALRM to be raised
+
+ def __exit__(self, exc_type, exc_val, exc_tb):
+ if platform.system() != 'Windows':
+ signal.alarm(0) # Cancel SIGALRM if it's scheduled
+ if self.suppress and exc_type is TimeoutError: # Suppress TimeoutError
+ return True
+
+
+class WorkingDirectory(contextlib.ContextDecorator):
+ # Usage: @WorkingDirectory(dir) decorator or 'with WorkingDirectory(dir):' context manager
+ def __init__(self, new_dir):
+ self.dir = new_dir # new dir
+ self.cwd = Path.cwd().resolve() # current dir
+
+ def __enter__(self):
+ os.chdir(self.dir)
+
+ def __exit__(self, exc_type, exc_val, exc_tb):
+ os.chdir(self.cwd)
+
+
+def try_except(func):
+ # try-except function. Usage: @try_except decorator
+ def handler(*args, **kwargs):
+ try:
+ func(*args, **kwargs)
+ except Exception as e:
+ print(e)
+
+ return handler
+
+
+def methods(instance):
+ # Get class/instance methods
+ return [f for f in dir(instance) if callable(getattr(instance, f)) and not f.startswith("__")]
+
+
+def print_args(args: Optional[dict] = None, show_file=True, show_fcn=False):
+ # Print function arguments (optional args dict)
+ x = inspect.currentframe().f_back # previous frame
+ file, _, fcn, _, _ = inspect.getframeinfo(x)
+ if args is None: # get args automatically
+ args, _, _, frm = inspect.getargvalues(x)
+ args = {k: v for k, v in frm.items() if k in args}
+ s = (f'{Path(file).stem}: ' if show_file else '') + (f'{fcn}: ' if show_fcn else '')
+ LOGGER.info(colorstr(s) + ', '.join(f'{k}={v}' for k, v in args.items()))
+
+
+def init_seeds(seed=0):
+ # Initialize random number generator (RNG) seeds https://pytorch.org/docs/stable/notes/randomness.html
+ # cudnn seed 0 settings are slower and more reproducible, else faster and less reproducible
+ import torch.backends.cudnn as cudnn
+ random.seed(seed)
+ np.random.seed(seed)
+ torch.manual_seed(seed)
+ cudnn.benchmark, cudnn.deterministic = (False, True) if seed == 0 else (True, False)
+
+
+def intersect_dicts(da, db, exclude=()):
+ # Dictionary intersection of matching keys and shapes, omitting 'exclude' keys, using da values
+ return {k: v for k, v in da.items() if k in db and not any(x in k for x in exclude) and v.shape == db[k].shape}
+
+
+def get_latest_run(search_dir='.'):
+ # Return path to most recent 'last.pt' in /runs (i.e. to --resume from)
+ last_list = glob.glob(f'{search_dir}/**/last*.pt', recursive=True)
+ return max(last_list, key=os.path.getctime) if last_list else ''
+
+
+def is_docker():
+ # Is environment a Docker container?
+ return Path('/workspace').exists() # or Path('/.dockerenv').exists()
+
+
+def is_colab():
+ # Is environment a Google Colab instance?
+ try:
+ import google.colab
+ return True
+ except ImportError:
+ return False
+
+
+def is_pip():
+ # Is file in a pip package?
+ return 'site-packages' in Path(__file__).resolve().parts
+
+
+def is_ascii(s=''):
+ # Is string composed of all ASCII (no UTF) characters? (note str().isascii() introduced in python 3.7)
+ s = str(s) # convert list, tuple, None, etc. to str
+ return len(s.encode().decode('ascii', 'ignore')) == len(s)
+
+
+def is_chinese(s='人工智能'):
+ # Is string composed of any Chinese characters?
+ return True if re.search('[\u4e00-\u9fff]', str(s)) else False
+
+
+def emojis(str=''):
+ # Return platform-dependent emoji-safe version of string
+ return str.encode().decode('ascii', 'ignore') if platform.system() == 'Windows' else str
+
+
+def file_age(path=__file__):
+ # Return days since last file update
+ dt = (datetime.now() - datetime.fromtimestamp(Path(path).stat().st_mtime)) # delta
+ return dt.days # + dt.seconds / 86400 # fractional days
+
+
+def file_update_date(path=__file__):
+ # Return human-readable file modification date, i.e. '2021-3-26'
+ t = datetime.fromtimestamp(Path(path).stat().st_mtime)
+ return f'{t.year}-{t.month}-{t.day}'
+
+
+def file_size(path):
+ # Return file/dir size (MB)
+ mb = 1 << 20 # bytes to MiB (1024 ** 2)
+ path = Path(path)
+ if path.is_file():
+ return path.stat().st_size / mb
+ elif path.is_dir():
+ return sum(f.stat().st_size for f in path.glob('**/*') if f.is_file()) / mb
+ else:
+ return 0.0
+
+
+def check_online():
+ # Check internet connectivity
+ import socket
+ try:
+ socket.create_connection(("1.1.1.1", 443), 5) # check host accessibility
+ return True
+ except OSError:
+ return False
+
+
+def git_describe(path=ROOT): # path must be a directory
+ # Return human-readable git description, i.e. v5.0-5-g3e25f1e https://git-scm.com/docs/git-describe
+ try:
+ assert (Path(path) / '.git').is_dir()
+ return check_output(f'git -C {path} describe --tags --long --always', shell=True).decode()[:-1]
+ except Exception:
+ return ''
+
+
+@try_except
+@WorkingDirectory(ROOT)
+def check_git_status():
+ # Recommend 'git pull' if code is out of date
+ msg = ', for updates see https://github.com/ultralytics/yolov5'
+ s = colorstr('github: ') # string
+ assert Path('.git').exists(), s + 'skipping check (not a git repository)' + msg
+ assert not is_docker(), s + 'skipping check (Docker image)' + msg
+ assert check_online(), s + 'skipping check (offline)' + msg
+
+ cmd = 'git fetch && git config --get remote.origin.url'
+ url = check_output(cmd, shell=True, timeout=5).decode().strip().rstrip('.git') # git fetch
+ branch = check_output('git rev-parse --abbrev-ref HEAD', shell=True).decode().strip() # checked out
+ n = int(check_output(f'git rev-list {branch}..origin/master --count', shell=True)) # commits behind
+ if n > 0:
+ s += f"⚠️ YOLOv5 is out of date by {n} commit{'s' * (n > 1)}. Use `git pull` or `git clone {url}` to update."
+ else:
+ s += f'up to date with {url} ✅'
+ LOGGER.info(emojis(s)) # emoji-safe
+
+
+def check_python(minimum='3.7.0'):
+ # Check current python version vs. required python version
+ check_version(platform.python_version(), minimum, name='Python ', hard=True)
+
+
+def check_version(current='0.0.0', minimum='0.0.0', name='version ', pinned=False, hard=False, verbose=False):
+ # Check version vs. required version
+ current, minimum = (pkg.parse_version(x) for x in (current, minimum))
+ result = (current == minimum) if pinned else (current >= minimum) # bool
+ s = f'{name}{minimum} required by YOLOv5, but {name}{current} is currently installed' # string
+ if hard:
+ assert result, s # assert min requirements met
+ if verbose and not result:
+ LOGGER.warning(s)
+ return result
+
+
+@try_except
+def check_requirements(requirements=ROOT / 'requirements.txt', exclude=(), install=True, cmds=()):
+ # Check installed dependencies meet requirements (pass *.txt file or list of packages)
+ prefix = colorstr('red', 'bold', 'requirements:')
+ check_python() # check python version
+ if isinstance(requirements, (str, Path)): # requirements.txt file
+ file = Path(requirements)
+ assert file.exists(), f"{prefix} {file.resolve()} not found, check failed."
+ with file.open() as f:
+ requirements = [f'{x.name}{x.specifier}' for x in pkg.parse_requirements(f) if x.name not in exclude]
+ else: # list or tuple of packages
+ requirements = [x for x in requirements if x not in exclude]
+
+ n = 0 # number of packages updates
+ for i, r in enumerate(requirements):
+ try:
+ pkg.require(r)
+ except Exception: # DistributionNotFound or VersionConflict if requirements not met
+ s = f"{prefix} {r} not found and is required by YOLOv5"
+ if install and AUTOINSTALL: # check environment variable
+ LOGGER.info(f"{s}, attempting auto-update...")
+ try:
+ assert check_online(), f"'pip install {r}' skipped (offline)"
+ LOGGER.info(check_output(f"pip install '{r}' {cmds[i] if cmds else ''}", shell=True).decode())
+ n += 1
+ except Exception as e:
+ LOGGER.warning(f'{prefix} {e}')
+ else:
+ LOGGER.info(f'{s}. Please install and rerun your command.')
+
+ if n: # if packages updated
+ source = file.resolve() if 'file' in locals() else requirements
+ s = f"{prefix} {n} package{'s' * (n > 1)} updated per {source}\n" \
+ f"{prefix} ⚠️ {colorstr('bold', 'Restart runtime or rerun command for updates to take effect')}\n"
+ LOGGER.info(emojis(s))
+
+
+def check_img_size(imgsz, s=32, floor=0):
+ # Verify image size is a multiple of stride s in each dimension
+ if isinstance(imgsz, int): # integer i.e. img_size=640
+ new_size = max(make_divisible(imgsz, int(s)), floor)
+ else: # list i.e. img_size=[640, 480]
+ imgsz = list(imgsz) # convert to list if tuple
+ new_size = [max(make_divisible(x, int(s)), floor) for x in imgsz]
+ if new_size != imgsz:
+ LOGGER.warning(f'WARNING: --img-size {imgsz} must be multiple of max stride {s}, updating to {new_size}')
+ return new_size
+
+
+def check_imshow():
+ # Check if environment supports image displays
+ try:
+ assert not is_docker(), 'cv2.imshow() is disabled in Docker environments'
+ assert not is_colab(), 'cv2.imshow() is disabled in Google Colab environments'
+ cv2.imshow('test', np.zeros((1, 1, 3)))
+ cv2.waitKey(1)
+ cv2.destroyAllWindows()
+ cv2.waitKey(1)
+ return True
+ except Exception as e:
+ LOGGER.warning(f'WARNING: Environment does not support cv2.imshow() or PIL Image.show() image displays\n{e}')
+ return False
+
+
+def check_suffix(file='yolov5s.pt', suffix=('.pt',), msg=''):
+ # Check file(s) for acceptable suffix
+ if file and suffix:
+ if isinstance(suffix, str):
+ suffix = [suffix]
+ for f in file if isinstance(file, (list, tuple)) else [file]:
+ s = Path(f).suffix.lower() # file suffix
+ if len(s):
+ assert s in suffix, f"{msg}{f} acceptable suffix is {suffix}"
+
+
+def check_yaml(file, suffix=('.yaml', '.yml')):
+ # Search/download YAML file (if necessary) and return path, checking suffix
+ return check_file(file, suffix)
+
+
+def check_file(file, suffix=''):
+ # Search/download file (if necessary) and return path
+ check_suffix(file, suffix) # optional
+ file = str(file) # convert to str()
+ if Path(file).is_file() or file == '': # exists
+ return file
+ elif file.startswith(('http:/', 'https:/')): # download
+ url = str(Path(file)).replace(':/', '://') # Pathlib turns :// -> :/
+ file = Path(urllib.parse.unquote(file).split('?')[0]).name # '%2F' to '/', split https://url.com/file.txt?auth
+ if Path(file).is_file():
+ LOGGER.info(f'Found {url} locally at {file}') # file already exists
+ else:
+ LOGGER.info(f'Downloading {url} to {file}...')
+ torch.hub.download_url_to_file(url, file)
+ assert Path(file).exists() and Path(file).stat().st_size > 0, f'File download failed: {url}' # check
+ return file
+ else: # search
+ files = []
+ for d in 'data', 'models', 'utils': # search directories
+ files.extend(glob.glob(str(ROOT / d / '**' / file), recursive=True)) # find file
+ assert len(files), f'File not found: {file}' # assert file was found
+ assert len(files) == 1, f"Multiple files match '{file}', specify exact path: {files}" # assert unique
+ return files[0] # return file
+
+
+def check_font(font=FONT, progress=False):
+ # Download font to CONFIG_DIR if necessary
+ font = Path(font)
+ file = CONFIG_DIR / font.name
+ if not font.exists() and not file.exists():
+ url = "https://ultralytics.com/assets/" + font.name
+ LOGGER.info(f'Downloading {url} to {file}...')
+ torch.hub.download_url_to_file(url, str(file), progress=progress)
+
+
+def check_dataset(data, autodownload=True):
+ # Download and/or unzip dataset if not found locally
+ # Usage: https://github.com/ultralytics/yolov5/releases/download/v1.0/coco128_with_yaml.zip
+
+ # Download (optional)
+ extract_dir = ''
+ if isinstance(data, (str, Path)) and str(data).endswith('.zip'): # i.e. gs://bucket/dir/coco128.zip
+ download(data, dir=DATASETS_DIR, unzip=True, delete=False, curl=False, threads=1)
+ data = next((DATASETS_DIR / Path(data).stem).rglob('*.yaml'))
+ extract_dir, autodownload = data.parent, False
+
+ # Read yaml (optional)
+ if isinstance(data, (str, Path)):
+ with open(data, errors='ignore') as f:
+ data = yaml.safe_load(f) # dictionary
+
+ # Resolve paths
+ path = Path(extract_dir or data.get('path') or '') # optional 'path' default to '.'
+ if not path.is_absolute():
+ path = (ROOT / path).resolve()
+ for k in 'train', 'val', 'test':
+ if data.get(k): # prepend path
+ data[k] = str(path / data[k]) if isinstance(data[k], str) else [str(path / x) for x in data[k]]
+
+ # Parse yaml
+ assert 'nc' in data, "Dataset 'nc' key missing."
+ if 'names' not in data:
+ data['names'] = [f'class{i}' for i in range(data['nc'])] # assign class names if missing
+ train, val, test, s = (data.get(x) for x in ('train', 'val', 'test', 'download'))
+ if val:
+ val = [Path(x).resolve() for x in (val if isinstance(val, list) else [val])] # val path
+ if not all(x.exists() for x in val):
+ LOGGER.info(emojis('\nDataset not found ⚠️, missing paths %s' % [str(x) for x in val if not x.exists()]))
+ if s and autodownload: # download script
+ t = time.time()
+ root = path.parent if 'path' in data else '..' # unzip directory i.e. '../'
+ if s.startswith('http') and s.endswith('.zip'): # URL
+ f = Path(s).name # filename
+ LOGGER.info(f'Downloading {s} to {f}...')
+ torch.hub.download_url_to_file(s, f)
+ Path(root).mkdir(parents=True, exist_ok=True) # create root
+ ZipFile(f).extractall(path=root) # unzip
+ Path(f).unlink() # remove zip
+ r = None # success
+ elif s.startswith('bash '): # bash script
+ LOGGER.info(f'Running {s} ...')
+ r = os.system(s)
+ else: # python script
+ r = exec(s, {'yaml': data}) # return None
+ dt = f'({round(time.time() - t, 1)}s)'
+ s = f"success ✅ {dt}, saved to {colorstr('bold', root)}" if r in (0, None) else f"failure {dt} ❌"
+ LOGGER.info(emojis(f"Dataset download {s}"))
+ else:
+ raise Exception(emojis('Dataset not found ❌'))
+
+ check_font('Arial.ttf' if is_ascii(data['names']) else 'Arial.Unicode.ttf', progress=True) # download fonts
+ return data # dictionary
+
+
+def url2file(url):
+ # Convert URL to filename, i.e. https://url.com/file.txt?auth -> file.txt
+ url = str(Path(url)).replace(':/', '://') # Pathlib turns :// -> :/
+ file = Path(urllib.parse.unquote(url)).name.split('?')[0] # '%2F' to '/', split https://url.com/file.txt?auth
+ return file
+
+
+def download(url, dir='.', unzip=True, delete=True, curl=False, threads=1, retry=3):
+ # Multi-threaded file download and unzip function, used in data.yaml for autodownload
+ def download_one(url, dir):
+ # Download 1 file
+ success = True
+ f = dir / Path(url).name # filename
+ if Path(url).is_file(): # exists in current path
+ Path(url).rename(f) # move to dir
+ elif not f.exists():
+ LOGGER.info(f'Downloading {url} to {f}...')
+ for i in range(retry + 1):
+ if curl:
+ s = 'sS' if threads > 1 else '' # silent
+ r = os.system(f"curl -{s}L '{url}' -o '{f}' --retry 9 -C -") # curl download
+ success = r == 0
+ else:
+ torch.hub.download_url_to_file(url, f, progress=threads == 1) # torch download
+ success = f.is_file()
+ if success:
+ break
+ elif i < retry:
+ LOGGER.warning(f'Download failure, retrying {i + 1}/{retry} {url}...')
+ else:
+ LOGGER.warning(f'Failed to download {url}...')
+
+ if unzip and success and f.suffix in ('.zip', '.gz'):
+ LOGGER.info(f'Unzipping {f}...')
+ if f.suffix == '.zip':
+ ZipFile(f).extractall(path=dir) # unzip
+ elif f.suffix == '.gz':
+ os.system(f'tar xfz {f} --directory {f.parent}') # unzip
+ if delete:
+ f.unlink() # remove zip
+
+ dir = Path(dir)
+ dir.mkdir(parents=True, exist_ok=True) # make directory
+ if threads > 1:
+ pool = ThreadPool(threads)
+ pool.imap(lambda x: download_one(*x), zip(url, repeat(dir))) # multi-threaded
+ pool.close()
+ pool.join()
+ else:
+ for u in [url] if isinstance(url, (str, Path)) else url:
+ download_one(u, dir)
+
+
+def make_divisible(x, divisor):
+ # Returns nearest x divisible by divisor
+ if isinstance(divisor, torch.Tensor):
+ divisor = int(divisor.max()) # to int
+ return math.ceil(x / divisor) * divisor
+
+
+def clean_str(s):
+ # Cleans a string by replacing special characters with underscore _
+ return re.sub(pattern="[|@#!¡·$€%&()=?¿^*;:,¨´><+]", repl="_", string=s)
+
+
+def one_cycle(y1=0.0, y2=1.0, steps=100):
+ # lambda function for sinusoidal ramp from y1 to y2 https://arxiv.org/pdf/1812.01187.pdf
+ return lambda x: ((1 - math.cos(x * math.pi / steps)) / 2) * (y2 - y1) + y1
+
+
+def colorstr(*input):
+ # Colors a string https://en.wikipedia.org/wiki/ANSI_escape_code, i.e. colorstr('blue', 'hello world')
+ *args, string = input if len(input) > 1 else ('blue', 'bold', input[0]) # color arguments, string
+ colors = {
+ 'black': '\033[30m', # basic colors
+ 'red': '\033[31m',
+ 'green': '\033[32m',
+ 'yellow': '\033[33m',
+ 'blue': '\033[34m',
+ 'magenta': '\033[35m',
+ 'cyan': '\033[36m',
+ 'white': '\033[37m',
+ 'bright_black': '\033[90m', # bright colors
+ 'bright_red': '\033[91m',
+ 'bright_green': '\033[92m',
+ 'bright_yellow': '\033[93m',
+ 'bright_blue': '\033[94m',
+ 'bright_magenta': '\033[95m',
+ 'bright_cyan': '\033[96m',
+ 'bright_white': '\033[97m',
+ 'end': '\033[0m', # misc
+ 'bold': '\033[1m',
+ 'underline': '\033[4m'}
+ return ''.join(colors[x] for x in args) + f'{string}' + colors['end']
+
+
+def labels_to_class_weights(labels, nc=80):
+ # Get class weights (inverse frequency) from training labels
+ if labels[0] is None: # no labels loaded
+ return torch.Tensor()
+
+ labels = np.concatenate(labels, 0) # labels.shape = (866643, 5) for COCO
+ classes = labels[:, 0].astype(int) # labels = [class xywh]
+ weights = np.bincount(classes, minlength=nc) # occurrences per class
+
+ # Prepend gridpoint count (for uCE training)
+ # gpi = ((320 / 32 * np.array([1, 2, 4])) ** 2 * 3).sum() # gridpoints per image
+ # weights = np.hstack([gpi * len(labels) - weights.sum() * 9, weights * 9]) ** 0.5 # prepend gridpoints to start
+
+ weights[weights == 0] = 1 # replace empty bins with 1
+ weights = 1 / weights # number of targets per class
+ weights /= weights.sum() # normalize
+ return torch.from_numpy(weights)
+
+
+def labels_to_image_weights(labels, nc=80, class_weights=np.ones(80)):
+ # Produces image weights based on class_weights and image contents
+ class_counts = np.array([np.bincount(x[:, 0].astype(int), minlength=nc) for x in labels])
+ image_weights = (class_weights.reshape(1, nc) * class_counts).sum(1)
+ # index = random.choices(range(n), weights=image_weights, k=1) # weight image sample
+ return image_weights
+
+
+def coco80_to_coco91_class(): # converts 80-index (val2014) to 91-index (paper)
+ # https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/
+ # a = np.loadtxt('data/coco.names', dtype='str', delimiter='\n')
+ # b = np.loadtxt('data/coco_paper.names', dtype='str', delimiter='\n')
+ # x1 = [list(a[i] == b).index(True) + 1 for i in range(80)] # darknet to coco
+ # x2 = [list(b[i] == a).index(True) if any(b[i] == a) else None for i in range(91)] # coco to darknet
+ x = [
+ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 27, 28, 31, 32, 33, 34,
+ 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63,
+ 64, 65, 67, 70, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 84, 85, 86, 87, 88, 89, 90]
+ return x
+
+
+def xyxy2xywh(x):
+ # Convert nx4 boxes from [x1, y1, x2, y2] to [x, y, w, h] where xy1=top-left, xy2=bottom-right
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[:, 0] = (x[:, 0] + x[:, 2]) / 2 # x center
+ y[:, 1] = (x[:, 1] + x[:, 3]) / 2 # y center
+ y[:, 2] = x[:, 2] - x[:, 0] # width
+ y[:, 3] = x[:, 3] - x[:, 1] # height
+ return y
+
+
+def xywh2xyxy(x):
+ # Convert nx4 boxes from [x, y, w, h] to [x1, y1, x2, y2] where xy1=top-left, xy2=bottom-right
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[:, 0] = x[:, 0] - x[:, 2] / 2 # top left x
+ y[:, 1] = x[:, 1] - x[:, 3] / 2 # top left y
+ y[:, 2] = x[:, 0] + x[:, 2] / 2 # bottom right x
+ y[:, 3] = x[:, 1] + x[:, 3] / 2 # bottom right y
+ return y
+
+
+def xywhn2xyxy(x, w=640, h=640, padw=0, padh=0):
+ # Convert nx4 boxes from [x, y, w, h] normalized to [x1, y1, x2, y2] where xy1=top-left, xy2=bottom-right
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[:, 0] = w * (x[:, 0] - x[:, 2] / 2) + padw # top left x
+ y[:, 1] = h * (x[:, 1] - x[:, 3] / 2) + padh # top left y
+ y[:, 2] = w * (x[:, 0] + x[:, 2] / 2) + padw # bottom right x
+ y[:, 3] = h * (x[:, 1] + x[:, 3] / 2) + padh # bottom right y
+ return y
+
+
+def xyxy2xywhn(x, w=640, h=640, clip=False, eps=0.0):
+ # Convert nx4 boxes from [x1, y1, x2, y2] to [x, y, w, h] normalized where xy1=top-left, xy2=bottom-right
+ if clip:
+ clip_coords(x, (h - eps, w - eps)) # warning: inplace clip
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[:, 0] = ((x[:, 0] + x[:, 2]) / 2) / w # x center
+ y[:, 1] = ((x[:, 1] + x[:, 3]) / 2) / h # y center
+ y[:, 2] = (x[:, 2] - x[:, 0]) / w # width
+ y[:, 3] = (x[:, 3] - x[:, 1]) / h # height
+ return y
+
+
+def xyn2xy(x, w=640, h=640, padw=0, padh=0):
+ # Convert normalized segments into pixel segments, shape (n,2)
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[:, 0] = w * x[:, 0] + padw # top left x
+ y[:, 1] = h * x[:, 1] + padh # top left y
+ return y
+
+
+def segment2box(segment, width=640, height=640):
+ # Convert 1 segment label to 1 box label, applying inside-image constraint, i.e. (xy1, xy2, ...) to (xyxy)
+ x, y = segment.T # segment xy
+ inside = (x >= 0) & (y >= 0) & (x <= width) & (y <= height)
+ x, y, = x[inside], y[inside]
+ return np.array([x.min(), y.min(), x.max(), y.max()]) if any(x) else np.zeros((1, 4)) # xyxy
+
+
+def segments2boxes(segments):
+ # Convert segment labels to box labels, i.e. (cls, xy1, xy2, ...) to (cls, xywh)
+ boxes = []
+ for s in segments:
+ x, y = s.T # segment xy
+ boxes.append([x.min(), y.min(), x.max(), y.max()]) # cls, xyxy
+ return xyxy2xywh(np.array(boxes)) # cls, xywh
+
+
+def resample_segments(segments, n=1000):
+ # Up-sample an (n,2) segment
+ for i, s in enumerate(segments):
+ x = np.linspace(0, len(s) - 1, n)
+ xp = np.arange(len(s))
+ segments[i] = np.concatenate([np.interp(x, xp, s[:, i]) for i in range(2)]).reshape(2, -1).T # segment xy
+ return segments
+
+
+def scale_coords(img1_shape, coords, img0_shape, ratio_pad=None):
+ # Rescale coords (xyxy) from img1_shape to img0_shape
+ if ratio_pad is None: # calculate from img0_shape
+ gain = min(img1_shape[0] / img0_shape[0], img1_shape[1] / img0_shape[1]) # gain = old / new
+ pad = (img1_shape[1] - img0_shape[1] * gain) / 2, (img1_shape[0] - img0_shape[0] * gain) / 2 # wh padding
+ else:
+ gain = ratio_pad[0][0]
+ pad = ratio_pad[1]
+
+ coords[:, [0, 2]] -= pad[0] # x padding
+ coords[:, [1, 3]] -= pad[1] # y padding
+ coords[:, :4] /= gain
+ clip_coords(coords, img0_shape)
+ return coords
+
+
+def clip_coords(boxes, shape):
+ # Clip bounding xyxy bounding boxes to image shape (height, width)
+ if isinstance(boxes, torch.Tensor): # faster individually
+ boxes[:, 0].clamp_(0, shape[1]) # x1
+ boxes[:, 1].clamp_(0, shape[0]) # y1
+ boxes[:, 2].clamp_(0, shape[1]) # x2
+ boxes[:, 3].clamp_(0, shape[0]) # y2
+ else: # np.array (faster grouped)
+ boxes[:, [0, 2]] = boxes[:, [0, 2]].clip(0, shape[1]) # x1, x2
+ boxes[:, [1, 3]] = boxes[:, [1, 3]].clip(0, shape[0]) # y1, y2
+
+
+def non_max_suppression(prediction,
+ conf_thres=0.25,
+ iou_thres=0.45,
+ classes=None,
+ agnostic=False,
+ multi_label=False,
+ labels=(),
+ max_det=300):
+ """Non-Maximum Suppression (NMS) on inference results to reject overlapping bounding boxes
+
+ Returns:
+ list of detections, on (n,6) tensor per image [xyxy, conf, cls]
+ """
+
+ bs = prediction.shape[0] # batch size
+ nc = prediction.shape[2] - 5 # number of classes
+ xc = prediction[..., 4] > conf_thres # candidates
+
+ # Checks
+ assert 0 <= conf_thres <= 1, f'Invalid Confidence threshold {conf_thres}, valid values are between 0.0 and 1.0'
+ assert 0 <= iou_thres <= 1, f'Invalid IoU {iou_thres}, valid values are between 0.0 and 1.0'
+
+ # Settings
+ # min_wh = 2 # (pixels) minimum box width and height
+ max_wh = 7680 # (pixels) maximum box width and height
+ max_nms = 30000 # maximum number of boxes into torchvision.ops.nms()
+ time_limit = 0.1 + 0.03 * bs # seconds to quit after
+ redundant = True # require redundant detections
+ multi_label &= nc > 1 # multiple labels per box (adds 0.5ms/img)
+ merge = False # use merge-NMS
+
+ t = time.time()
+ output = [torch.zeros((0, 6), device=prediction.device)] * bs
+ for xi, x in enumerate(prediction): # image index, image inference
+ # Apply constraints
+ # x[((x[..., 2:4] < min_wh) | (x[..., 2:4] > max_wh)).any(1), 4] = 0 # width-height
+ x = x[xc[xi]] # confidence
+
+ # Cat apriori labels if autolabelling
+ if labels and len(labels[xi]):
+ lb = labels[xi]
+ v = torch.zeros((len(lb), nc + 5), device=x.device)
+ v[:, :4] = lb[:, 1:5] # box
+ v[:, 4] = 1.0 # conf
+ v[range(len(lb)), lb[:, 0].long() + 5] = 1.0 # cls
+ x = torch.cat((x, v), 0)
+
+ # If none remain process next image
+ if not x.shape[0]:
+ continue
+
+ # Compute conf
+ x[:, 5:] *= x[:, 4:5] # conf = obj_conf * cls_conf
+
+ # Box (center x, center y, width, height) to (x1, y1, x2, y2)
+ box = xywh2xyxy(x[:, :4])
+
+ # Detections matrix nx6 (xyxy, conf, cls)
+ if multi_label:
+ i, j = (x[:, 5:] > conf_thres).nonzero(as_tuple=False).T
+ x = torch.cat((box[i], x[i, j + 5, None], j[:, None].float()), 1)
+ else: # best class only
+ conf, j = x[:, 5:].max(1, keepdim=True)
+ x = torch.cat((box, conf, j.float()), 1)[conf.view(-1) > conf_thres]
+
+ # Filter by class
+ if classes is not None:
+ x = x[(x[:, 5:6] == torch.tensor(classes, device=x.device)).any(1)]
+
+ # Apply finite constraint
+ # if not torch.isfinite(x).all():
+ # x = x[torch.isfinite(x).all(1)]
+
+ # Check shape
+ n = x.shape[0] # number of boxes
+ if not n: # no boxes
+ continue
+ elif n > max_nms: # excess boxes
+ x = x[x[:, 4].argsort(descending=True)[:max_nms]] # sort by confidence
+
+ # Batched NMS
+ c = x[:, 5:6] * (0 if agnostic else max_wh) # classes
+ boxes, scores = x[:, :4] + c, x[:, 4] # boxes (offset by class), scores
+ i = torchvision.ops.nms(boxes, scores, iou_thres) # NMS
+ if i.shape[0] > max_det: # limit detections
+ i = i[:max_det]
+ if merge and (1 < n < 3E3): # Merge NMS (boxes merged using weighted mean)
+ # update boxes as boxes(i,4) = weights(i,n) * boxes(n,4)
+ iou = box_iou(boxes[i], boxes) > iou_thres # iou matrix
+ weights = iou * scores[None] # box weights
+ x[i, :4] = torch.mm(weights, x[:, :4]).float() / weights.sum(1, keepdim=True) # merged boxes
+ if redundant:
+ i = i[iou.sum(1) > 1] # require redundancy
+
+ output[xi] = x[i]
+ if (time.time() - t) > time_limit:
+ LOGGER.warning(f'WARNING: NMS time limit {time_limit:.3f}s exceeded')
+ break # time limit exceeded
+
+ return output
+
+
+def strip_optimizer(f='best.pt', s=''): # from utils.general import *; strip_optimizer()
+ # Strip optimizer from 'f' to finalize training, optionally save as 's'
+ x = torch.load(f, map_location=torch.device('cpu'))
+ if x.get('ema'):
+ x['model'] = x['ema'] # replace model with ema
+ for k in 'optimizer', 'best_fitness', 'wandb_id', 'ema', 'updates': # keys
+ x[k] = None
+ x['epoch'] = -1
+ x['model'].half() # to FP16
+ for p in x['model'].parameters():
+ p.requires_grad = False
+ torch.save(x, s or f)
+ mb = os.path.getsize(s or f) / 1E6 # filesize
+ LOGGER.info(f"Optimizer stripped from {f},{(' saved as %s,' % s) if s else ''} {mb:.1f}MB")
+
+
+def print_mutation(results, hyp, save_dir, bucket, prefix=colorstr('evolve: ')):
+ evolve_csv = save_dir / 'evolve.csv'
+ evolve_yaml = save_dir / 'hyp_evolve.yaml'
+ keys = ('metrics/precision', 'metrics/recall', 'metrics/mAP_0.5', 'metrics/mAP_0.5:0.95', 'val/box_loss',
+ 'val/obj_loss', 'val/cls_loss') + tuple(hyp.keys()) # [results + hyps]
+ keys = tuple(x.strip() for x in keys)
+ vals = results + tuple(hyp.values())
+ n = len(keys)
+
+ # Download (optional)
+ if bucket:
+ url = f'gs://{bucket}/evolve.csv'
+ if gsutil_getsize(url) > (evolve_csv.stat().st_size if evolve_csv.exists() else 0):
+ os.system(f'gsutil cp {url} {save_dir}') # download evolve.csv if larger than local
+
+ # Log to evolve.csv
+ s = '' if evolve_csv.exists() else (('%20s,' * n % keys).rstrip(',') + '\n') # add header
+ with open(evolve_csv, 'a') as f:
+ f.write(s + ('%20.5g,' * n % vals).rstrip(',') + '\n')
+
+ # Save yaml
+ with open(evolve_yaml, 'w') as f:
+ data = pd.read_csv(evolve_csv)
+ data = data.rename(columns=lambda x: x.strip()) # strip keys
+ i = np.argmax(fitness(data.values[:, :4])) #
+ generations = len(data)
+ f.write('# YOLOv5 Hyperparameter Evolution Results\n' + f'# Best generation: {i}\n' +
+ f'# Last generation: {generations - 1}\n' + '# ' + ', '.join(f'{x.strip():>20s}' for x in keys[:7]) +
+ '\n' + '# ' + ', '.join(f'{x:>20.5g}' for x in data.values[i, :7]) + '\n\n')
+ yaml.safe_dump(data.loc[i][7:].to_dict(), f, sort_keys=False)
+
+ # Print to screen
+ LOGGER.info(prefix + f'{generations} generations finished, current result:\n' + prefix +
+ ', '.join(f'{x.strip():>20s}' for x in keys) + '\n' + prefix + ', '.join(f'{x:20.5g}'
+ for x in vals) + '\n\n')
+
+ if bucket:
+ os.system(f'gsutil cp {evolve_csv} {evolve_yaml} gs://{bucket}') # upload
+
+
+def apply_classifier(x, model, img, im0):
+ # Apply a second stage classifier to YOLO outputs
+ # Example model = torchvision.models.__dict__['efficientnet_b0'](pretrained=True).to(device).eval()
+ im0 = [im0] if isinstance(im0, np.ndarray) else im0
+ for i, d in enumerate(x): # per image
+ if d is not None and len(d):
+ d = d.clone()
+
+ # Reshape and pad cutouts
+ b = xyxy2xywh(d[:, :4]) # boxes
+ b[:, 2:] = b[:, 2:].max(1)[0].unsqueeze(1) # rectangle to square
+ b[:, 2:] = b[:, 2:] * 1.3 + 30 # pad
+ d[:, :4] = xywh2xyxy(b).long()
+
+ # Rescale boxes from img_size to im0 size
+ scale_coords(img.shape[2:], d[:, :4], im0[i].shape)
+
+ # Classes
+ pred_cls1 = d[:, 5].long()
+ ims = []
+ for j, a in enumerate(d): # per item
+ cutout = im0[i][int(a[1]):int(a[3]), int(a[0]):int(a[2])]
+ im = cv2.resize(cutout, (224, 224)) # BGR
+ # cv2.imwrite('example%i.jpg' % j, cutout)
+
+ im = im[:, :, ::-1].transpose(2, 0, 1) # BGR to RGB, to 3x416x416
+ im = np.ascontiguousarray(im, dtype=np.float32) # uint8 to float32
+ im /= 255 # 0 - 255 to 0.0 - 1.0
+ ims.append(im)
+
+ pred_cls2 = model(torch.Tensor(ims).to(d.device)).argmax(1) # classifier prediction
+ x[i] = x[i][pred_cls1 == pred_cls2] # retain matching class detections
+
+ return x
+
+
+def increment_path(path, exist_ok=False, sep='', mkdir=False):
+ # Increment file or directory path, i.e. runs/exp --> runs/exp{sep}2, runs/exp{sep}3, ... etc.
+ path = Path(path) # os-agnostic
+ if path.exists() and not exist_ok:
+ path, suffix = (path.with_suffix(''), path.suffix) if path.is_file() else (path, '')
+ dirs = glob.glob(f"{path}{sep}*") # similar paths
+ matches = [re.search(rf"%s{sep}(\d+)" % path.stem, d) for d in dirs]
+ i = [int(m.groups()[0]) for m in matches if m] # indices
+ n = max(i) + 1 if i else 2 # increment number
+ path = Path(f"{path}{sep}{n}{suffix}") # increment path
+ if mkdir:
+ path.mkdir(parents=True, exist_ok=True) # make directory
+ return path
+
+
+# OpenCV Chinese-friendly functions ------------------------------------------------------------------------------------
+imshow_ = cv2.imshow # copy to avoid recursion errors
+
+
+def imread(path, flags=cv2.IMREAD_COLOR):
+ return cv2.imdecode(np.fromfile(path, np.uint8), flags)
+
+
+def imwrite(path, im):
+ try:
+ cv2.imencode(Path(path).suffix, im)[1].tofile(path)
+ return True
+ except Exception:
+ return False
+
+
+def imshow(path, im):
+ imshow_(path.encode('unicode_escape').decode(), im)
+
+
+cv2.imread, cv2.imwrite, cv2.imshow = imread, imwrite, imshow # redefine
+
+# Variables ------------------------------------------------------------------------------------------------------------
+NCOLS = 0 if is_docker() else shutil.get_terminal_size().columns # terminal window size for tqdm
diff --git a/vouchervision/component_detector/utils/general_torchscript.py b/vouchervision/component_detector/utils/general_torchscript.py
new file mode 100644
index 0000000000000000000000000000000000000000..0fe7644b0241c78039e63c4f04477d5b4d0bbaa8
--- /dev/null
+++ b/vouchervision/component_detector/utils/general_torchscript.py
@@ -0,0 +1,1118 @@
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+General utils
+"""
+
+import contextlib
+import glob
+import inspect
+import logging
+import logging.config
+import math
+import os
+import platform
+import random
+import re
+import signal
+import subprocess
+import sys
+import time
+import urllib
+from copy import deepcopy
+from datetime import datetime
+from itertools import repeat
+from multiprocessing.pool import ThreadPool
+from pathlib import Path
+from subprocess import check_output
+from tarfile import is_tarfile
+from typing import Optional
+from zipfile import ZipFile, is_zipfile
+
+import cv2
+import numpy as np
+import pandas as pd
+import pkg_resources as pkg
+import torch
+import torchvision
+import yaml
+
+# Import 'ultralytics' package or install if if missing
+try:
+ import ultralytics
+
+ assert hasattr(ultralytics, '__version__') # verify package is not directory
+except (ImportError, AssertionError):
+ os.system('pip install -U ultralytics')
+ import ultralytics
+
+from ultralytics.utils.checks import check_requirements
+
+from utils import TryExcept, emojis
+from utils.downloads_torchscript import curl_download, gsutil_getsize
+from utils.metrics import box_iou, fitness
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1] # YOLOv5 root directory
+RANK = int(os.getenv('RANK', -1))
+
+# Settings
+NUM_THREADS = min(8, max(1, os.cpu_count() - 1)) # number of YOLOv5 multiprocessing threads
+DATASETS_DIR = Path(os.getenv('YOLOv5_DATASETS_DIR', ROOT.parent / 'datasets')) # global datasets directory
+AUTOINSTALL = str(os.getenv('YOLOv5_AUTOINSTALL', True)).lower() == 'true' # global auto-install mode
+VERBOSE = str(os.getenv('YOLOv5_VERBOSE', True)).lower() == 'true' # global verbose mode
+TQDM_BAR_FORMAT = '{l_bar}{bar:10}{r_bar}' # tqdm bar format
+FONT = 'Arial.ttf' # https://ultralytics.com/assets/Arial.ttf
+
+torch.set_printoptions(linewidth=320, precision=5, profile='long')
+np.set_printoptions(linewidth=320, formatter={'float_kind': '{:11.5g}'.format}) # format short g, %precision=5
+pd.options.display.max_columns = 10
+cv2.setNumThreads(0) # prevent OpenCV from multithreading (incompatible with PyTorch DataLoader)
+os.environ['NUMEXPR_MAX_THREADS'] = str(NUM_THREADS) # NumExpr max threads
+os.environ['OMP_NUM_THREADS'] = '1' if platform.system() == 'darwin' else str(NUM_THREADS) # OpenMP (PyTorch and SciPy)
+os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' # suppress verbose TF compiler warnings in Colab
+
+
+def is_ascii(s=''):
+ # Is string composed of all ASCII (no UTF) characters? (note str().isascii() introduced in python 3.7)
+ s = str(s) # convert list, tuple, None, etc. to str
+ return len(s.encode().decode('ascii', 'ignore')) == len(s)
+
+
+def is_chinese(s='人工智能'):
+ # Is string composed of any Chinese characters?
+ return bool(re.search('[\u4e00-\u9fff]', str(s)))
+
+
+def is_colab():
+ # Is environment a Google Colab instance?
+ return 'google.colab' in sys.modules
+
+
+def is_jupyter():
+ """
+ Check if the current script is running inside a Jupyter Notebook.
+ Verified on Colab, Jupyterlab, Kaggle, Paperspace.
+
+ Returns:
+ bool: True if running inside a Jupyter Notebook, False otherwise.
+ """
+ with contextlib.suppress(Exception):
+ from IPython import get_ipython
+ return get_ipython() is not None
+ return False
+
+
+def is_kaggle():
+ # Is environment a Kaggle Notebook?
+ return os.environ.get('PWD') == '/kaggle/working' and os.environ.get('KAGGLE_URL_BASE') == 'https://www.kaggle.com'
+
+
+def is_docker() -> bool:
+ """Check if the process runs inside a docker container."""
+ if Path('/.dockerenv').exists():
+ return True
+ try: # check if docker is in control groups
+ with open('/proc/self/cgroup') as file:
+ return any('docker' in line for line in file)
+ except OSError:
+ return False
+
+
+def is_writeable(dir, test=False):
+ # Return True if directory has write permissions, test opening a file with write permissions if test=True
+ if not test:
+ return os.access(dir, os.W_OK) # possible issues on Windows
+ file = Path(dir) / 'tmp.txt'
+ try:
+ with open(file, 'w'): # open file with write permissions
+ pass
+ file.unlink() # remove file
+ return True
+ except OSError:
+ return False
+
+
+LOGGING_NAME = 'yolov5'
+
+
+def set_logging(name=LOGGING_NAME, verbose=True):
+ # sets up logging for the given name
+ rank = int(os.getenv('RANK', -1)) # rank in world for Multi-GPU trainings
+ level = logging.INFO if verbose and rank in {-1, 0} else logging.ERROR
+ logging.config.dictConfig({
+ 'version': 1,
+ 'disable_existing_loggers': False,
+ 'formatters': {
+ name: {
+ 'format': '%(message)s'}},
+ 'handlers': {
+ name: {
+ 'class': 'logging.StreamHandler',
+ 'formatter': name,
+ 'level': level, }},
+ 'loggers': {
+ name: {
+ 'level': level,
+ 'handlers': [name],
+ 'propagate': False, }}})
+
+
+set_logging(LOGGING_NAME) # run before defining LOGGER
+LOGGER = logging.getLogger(LOGGING_NAME) # define globally (used in train.py, val.py, detect.py, etc.)
+if platform.system() == 'Windows':
+ for fn in LOGGER.info, LOGGER.warning:
+ setattr(LOGGER, fn.__name__, lambda x: fn(emojis(x))) # emoji safe logging
+
+
+def user_config_dir(dir='Ultralytics', env_var='YOLOV5_CONFIG_DIR'):
+ # Return path of user configuration directory. Prefer environment variable if exists. Make dir if required.
+ env = os.getenv(env_var)
+ if env:
+ path = Path(env) # use environment variable
+ else:
+ cfg = {'Windows': 'AppData/Roaming', 'Linux': '.config', 'Darwin': 'Library/Application Support'} # 3 OS dirs
+ path = Path.home() / cfg.get(platform.system(), '') # OS-specific config dir
+ path = (path if is_writeable(path) else Path('/tmp')) / dir # GCP and AWS lambda fix, only /tmp is writeable
+ path.mkdir(exist_ok=True) # make if required
+ return path
+
+
+CONFIG_DIR = user_config_dir() # Ultralytics settings dir
+
+
+class Profile(contextlib.ContextDecorator):
+ # YOLOv5 Profile class. Usage: @Profile() decorator or 'with Profile():' context manager
+ def __init__(self, t=0.0):
+ self.t = t
+ self.cuda = torch.cuda.is_available()
+
+ def __enter__(self):
+ self.start = self.time()
+ return self
+
+ def __exit__(self, type, value, traceback):
+ self.dt = self.time() - self.start # delta-time
+ self.t += self.dt # accumulate dt
+
+ def time(self):
+ if self.cuda:
+ torch.cuda.synchronize()
+ return time.time()
+
+
+class Timeout(contextlib.ContextDecorator):
+ # YOLOv5 Timeout class. Usage: @Timeout(seconds) decorator or 'with Timeout(seconds):' context manager
+ def __init__(self, seconds, *, timeout_msg='', suppress_timeout_errors=True):
+ self.seconds = int(seconds)
+ self.timeout_message = timeout_msg
+ self.suppress = bool(suppress_timeout_errors)
+
+ def _timeout_handler(self, signum, frame):
+ raise TimeoutError(self.timeout_message)
+
+ def __enter__(self):
+ if platform.system() != 'Windows': # not supported on Windows
+ signal.signal(signal.SIGALRM, self._timeout_handler) # Set handler for SIGALRM
+ signal.alarm(self.seconds) # start countdown for SIGALRM to be raised
+
+ def __exit__(self, exc_type, exc_val, exc_tb):
+ if platform.system() != 'Windows':
+ signal.alarm(0) # Cancel SIGALRM if it's scheduled
+ if self.suppress and exc_type is TimeoutError: # Suppress TimeoutError
+ return True
+
+
+class WorkingDirectory(contextlib.ContextDecorator):
+ # Usage: @WorkingDirectory(dir) decorator or 'with WorkingDirectory(dir):' context manager
+ def __init__(self, new_dir):
+ self.dir = new_dir # new dir
+ self.cwd = Path.cwd().resolve() # current dir
+
+ def __enter__(self):
+ os.chdir(self.dir)
+
+ def __exit__(self, exc_type, exc_val, exc_tb):
+ os.chdir(self.cwd)
+
+
+def methods(instance):
+ # Get class/instance methods
+ return [f for f in dir(instance) if callable(getattr(instance, f)) and not f.startswith('__')]
+
+
+def print_args(args: Optional[dict] = None, show_file=True, show_func=False):
+ # Print function arguments (optional args dict)
+ x = inspect.currentframe().f_back # previous frame
+ file, _, func, _, _ = inspect.getframeinfo(x)
+ if args is None: # get args automatically
+ args, _, _, frm = inspect.getargvalues(x)
+ args = {k: v for k, v in frm.items() if k in args}
+ try:
+ file = Path(file).resolve().relative_to(ROOT).with_suffix('')
+ except ValueError:
+ file = Path(file).stem
+ s = (f'{file}: ' if show_file else '') + (f'{func}: ' if show_func else '')
+ LOGGER.info(colorstr(s) + ', '.join(f'{k}={v}' for k, v in args.items()))
+
+
+def init_seeds(seed=0, deterministic=False):
+ # Initialize random number generator (RNG) seeds https://pytorch.org/docs/stable/notes/randomness.html
+ random.seed(seed)
+ np.random.seed(seed)
+ torch.manual_seed(seed)
+ torch.cuda.manual_seed(seed)
+ torch.cuda.manual_seed_all(seed) # for Multi-GPU, exception safe
+ # torch.backends.cudnn.benchmark = True # AutoBatch problem https://github.com/ultralytics/yolov5/issues/9287
+ if deterministic and check_version(torch.__version__, '1.12.0'): # https://github.com/ultralytics/yolov5/pull/8213
+ torch.use_deterministic_algorithms(True)
+ torch.backends.cudnn.deterministic = True
+ os.environ['CUBLAS_WORKSPACE_CONFIG'] = ':4096:8'
+ os.environ['PYTHONHASHSEED'] = str(seed)
+
+
+def intersect_dicts(da, db, exclude=()):
+ # Dictionary intersection of matching keys and shapes, omitting 'exclude' keys, using da values
+ return {k: v for k, v in da.items() if k in db and all(x not in k for x in exclude) and v.shape == db[k].shape}
+
+
+def get_default_args(func):
+ # Get func() default arguments
+ signature = inspect.signature(func)
+ return {k: v.default for k, v in signature.parameters.items() if v.default is not inspect.Parameter.empty}
+
+
+def get_latest_run(search_dir='.'):
+ # Return path to most recent 'last.pt' in /runs (i.e. to --resume from)
+ last_list = glob.glob(f'{search_dir}/**/last*.pt', recursive=True)
+ return max(last_list, key=os.path.getctime) if last_list else ''
+
+
+def file_age(path=__file__):
+ # Return days since last file update
+ dt = (datetime.now() - datetime.fromtimestamp(Path(path).stat().st_mtime)) # delta
+ return dt.days # + dt.seconds / 86400 # fractional days
+
+
+def file_date(path=__file__):
+ # Return human-readable file modification date, i.e. '2021-3-26'
+ t = datetime.fromtimestamp(Path(path).stat().st_mtime)
+ return f'{t.year}-{t.month}-{t.day}'
+
+
+def file_size(path):
+ # Return file/dir size (MB)
+ mb = 1 << 20 # bytes to MiB (1024 ** 2)
+ path = Path(path)
+ if path.is_file():
+ return path.stat().st_size / mb
+ elif path.is_dir():
+ return sum(f.stat().st_size for f in path.glob('**/*') if f.is_file()) / mb
+ else:
+ return 0.0
+
+
+def check_online():
+ # Check internet connectivity
+ import socket
+
+ def run_once():
+ # Check once
+ try:
+ socket.create_connection(('1.1.1.1', 443), 5) # check host accessibility
+ return True
+ except OSError:
+ return False
+
+ return run_once() or run_once() # check twice to increase robustness to intermittent connectivity issues
+
+
+def git_describe(path=ROOT): # path must be a directory
+ # Return human-readable git description, i.e. v5.0-5-g3e25f1e https://git-scm.com/docs/git-describe
+ try:
+ assert (Path(path) / '.git').is_dir()
+ return check_output(f'git -C {path} describe --tags --long --always', shell=True).decode()[:-1]
+ except Exception:
+ return ''
+
+
+@TryExcept()
+@WorkingDirectory(ROOT)
+def check_git_status(repo='ultralytics/yolov5', branch='master'):
+ # YOLOv5 status check, recommend 'git pull' if code is out of date
+ url = f'https://github.com/{repo}'
+ msg = f', for updates see {url}'
+ s = colorstr('github: ') # string
+ assert Path('.git').exists(), s + 'skipping check (not a git repository)' + msg
+ assert check_online(), s + 'skipping check (offline)' + msg
+
+ splits = re.split(pattern=r'\s', string=check_output('git remote -v', shell=True).decode())
+ matches = [repo in s for s in splits]
+ if any(matches):
+ remote = splits[matches.index(True) - 1]
+ else:
+ remote = 'ultralytics'
+ check_output(f'git remote add {remote} {url}', shell=True)
+ check_output(f'git fetch {remote}', shell=True, timeout=5) # git fetch
+ local_branch = check_output('git rev-parse --abbrev-ref HEAD', shell=True).decode().strip() # checked out
+ n = int(check_output(f'git rev-list {local_branch}..{remote}/{branch} --count', shell=True)) # commits behind
+ if n > 0:
+ pull = 'git pull' if remote == 'origin' else f'git pull {remote} {branch}'
+ s += f"⚠️ YOLOv5 is out of date by {n} commit{'s' * (n > 1)}. Use '{pull}' or 'git clone {url}' to update."
+ else:
+ s += f'up to date with {url} ✅'
+ LOGGER.info(s)
+
+
+@WorkingDirectory(ROOT)
+def check_git_info(path='.'):
+ # YOLOv5 git info check, return {remote, branch, commit}
+ check_requirements('gitpython')
+ import git
+ try:
+ repo = git.Repo(path)
+ remote = repo.remotes.origin.url.replace('.git', '') # i.e. 'https://github.com/ultralytics/yolov5'
+ commit = repo.head.commit.hexsha # i.e. '3134699c73af83aac2a481435550b968d5792c0d'
+ try:
+ branch = repo.active_branch.name # i.e. 'main'
+ except TypeError: # not on any branch
+ branch = None # i.e. 'detached HEAD' state
+ return {'remote': remote, 'branch': branch, 'commit': commit}
+ except git.exc.InvalidGitRepositoryError: # path is not a git dir
+ return {'remote': None, 'branch': None, 'commit': None}
+
+
+def check_python(minimum='3.8.0'):
+ # Check current python version vs. required python version
+ check_version(platform.python_version(), minimum, name='Python ', hard=True)
+
+
+def check_version(current='0.0.0', minimum='0.0.0', name='version ', pinned=False, hard=False, verbose=False):
+ # Check version vs. required version
+ current, minimum = (pkg.parse_version(x) for x in (current, minimum))
+ result = (current == minimum) if pinned else (current >= minimum) # bool
+ s = f'WARNING ⚠️ {name}{minimum} is required by YOLOv5, but {name}{current} is currently installed' # string
+ if hard:
+ assert result, emojis(s) # assert min requirements met
+ if verbose and not result:
+ LOGGER.warning(s)
+ return result
+
+
+def check_img_size(imgsz, s=32, floor=0):
+ # Verify image size is a multiple of stride s in each dimension
+ if isinstance(imgsz, int): # integer i.e. img_size=640
+ new_size = max(make_divisible(imgsz, int(s)), floor)
+ else: # list i.e. img_size=[640, 480]
+ imgsz = list(imgsz) # convert to list if tuple
+ new_size = [max(make_divisible(x, int(s)), floor) for x in imgsz]
+ if new_size != imgsz:
+ LOGGER.warning(f'WARNING ⚠️ --img-size {imgsz} must be multiple of max stride {s}, updating to {new_size}')
+ return new_size
+
+
+def check_imshow(warn=False):
+ # Check if environment supports image displays
+ try:
+ assert not is_jupyter()
+ assert not is_docker()
+ cv2.imshow('test', np.zeros((1, 1, 3)))
+ cv2.waitKey(1)
+ cv2.destroyAllWindows()
+ cv2.waitKey(1)
+ return True
+ except Exception as e:
+ if warn:
+ LOGGER.warning(f'WARNING ⚠️ Environment does not support cv2.imshow() or PIL Image.show()\n{e}')
+ return False
+
+
+def check_suffix(file='yolov5s.pt', suffix=('.pt', ), msg=''):
+ # Check file(s) for acceptable suffix
+ if file and suffix:
+ if isinstance(suffix, str):
+ suffix = [suffix]
+ for f in file if isinstance(file, (list, tuple)) else [file]:
+ s = Path(f).suffix.lower() # file suffix
+ if len(s):
+ assert s in suffix, f'{msg}{f} acceptable suffix is {suffix}'
+
+
+def check_yaml(file, suffix=('.yaml', '.yml')):
+ # Search/download YAML file (if necessary) and return path, checking suffix
+ return check_file(file, suffix)
+
+
+def check_file(file, suffix=''):
+ # Search/download file (if necessary) and return path
+ check_suffix(file, suffix) # optional
+ file = str(file) # convert to str()
+ if os.path.isfile(file) or not file: # exists
+ return file
+ elif file.startswith(('http:/', 'https:/')): # download
+ url = file # warning: Pathlib turns :// -> :/
+ file = Path(urllib.parse.unquote(file).split('?')[0]).name # '%2F' to '/', split https://url.com/file.txt?auth
+ if os.path.isfile(file):
+ LOGGER.info(f'Found {url} locally at {file}') # file already exists
+ else:
+ LOGGER.info(f'Downloading {url} to {file}...')
+ torch.hub.download_url_to_file(url, file)
+ assert Path(file).exists() and Path(file).stat().st_size > 0, f'File download failed: {url}' # check
+ return file
+ elif file.startswith('clearml://'): # ClearML Dataset ID
+ assert 'clearml' in sys.modules, "ClearML is not installed, so cannot use ClearML dataset. Try running 'pip install clearml'."
+ return file
+ else: # search
+ files = []
+ for d in 'data', 'models', 'utils': # search directories
+ files.extend(glob.glob(str(ROOT / d / '**' / file), recursive=True)) # find file
+ assert len(files), f'File not found: {file}' # assert file was found
+ assert len(files) == 1, f"Multiple files match '{file}', specify exact path: {files}" # assert unique
+ return files[0] # return file
+
+
+def check_font(font=FONT, progress=False):
+ # Download font to CONFIG_DIR if necessary
+ font = Path(font)
+ file = CONFIG_DIR / font.name
+ if not font.exists() and not file.exists():
+ url = f'https://ultralytics.com/assets/{font.name}'
+ LOGGER.info(f'Downloading {url} to {file}...')
+ torch.hub.download_url_to_file(url, str(file), progress=progress)
+
+
+def check_dataset(data, autodownload=True):
+ # Download, check and/or unzip dataset if not found locally
+
+ # Download (optional)
+ extract_dir = ''
+ if isinstance(data, (str, Path)) and (is_zipfile(data) or is_tarfile(data)):
+ download(data, dir=f'{DATASETS_DIR}/{Path(data).stem}', unzip=True, delete=False, curl=False, threads=1)
+ data = next((DATASETS_DIR / Path(data).stem).rglob('*.yaml'))
+ extract_dir, autodownload = data.parent, False
+
+ # Read yaml (optional)
+ if isinstance(data, (str, Path)):
+ data = yaml_load(data) # dictionary
+
+ # Checks
+ for k in 'train', 'val', 'names':
+ assert k in data, emojis(f"data.yaml '{k}:' field missing ❌")
+ if isinstance(data['names'], (list, tuple)): # old array format
+ data['names'] = dict(enumerate(data['names'])) # convert to dict
+ assert all(isinstance(k, int) for k in data['names'].keys()), 'data.yaml names keys must be integers, i.e. 2: car'
+ data['nc'] = len(data['names'])
+
+ # Resolve paths
+ path = Path(extract_dir or data.get('path') or '') # optional 'path' default to '.'
+ if not path.is_absolute():
+ path = (ROOT / path).resolve()
+ data['path'] = path # download scripts
+ for k in 'train', 'val', 'test':
+ if data.get(k): # prepend path
+ if isinstance(data[k], str):
+ x = (path / data[k]).resolve()
+ if not x.exists() and data[k].startswith('../'):
+ x = (path / data[k][3:]).resolve()
+ data[k] = str(x)
+ else:
+ data[k] = [str((path / x).resolve()) for x in data[k]]
+
+ # Parse yaml
+ train, val, test, s = (data.get(x) for x in ('train', 'val', 'test', 'download'))
+ if val:
+ val = [Path(x).resolve() for x in (val if isinstance(val, list) else [val])] # val path
+ if not all(x.exists() for x in val):
+ LOGGER.info('\nDataset not found ⚠️, missing paths %s' % [str(x) for x in val if not x.exists()])
+ if not s or not autodownload:
+ raise Exception('Dataset not found ❌')
+ t = time.time()
+ if s.startswith('http') and s.endswith('.zip'): # URL
+ f = Path(s).name # filename
+ LOGGER.info(f'Downloading {s} to {f}...')
+ torch.hub.download_url_to_file(s, f)
+ Path(DATASETS_DIR).mkdir(parents=True, exist_ok=True) # create root
+ unzip_file(f, path=DATASETS_DIR) # unzip
+ Path(f).unlink() # remove zip
+ r = None # success
+ elif s.startswith('bash '): # bash script
+ LOGGER.info(f'Running {s} ...')
+ r = subprocess.run(s, shell=True)
+ else: # python script
+ r = exec(s, {'yaml': data}) # return None
+ dt = f'({round(time.time() - t, 1)}s)'
+ s = f"success ✅ {dt}, saved to {colorstr('bold', DATASETS_DIR)}" if r in (0, None) else f'failure {dt} ❌'
+ LOGGER.info(f'Dataset download {s}')
+ check_font('Arial.ttf' if is_ascii(data['names']) else 'Arial.Unicode.ttf', progress=True) # download fonts
+ return data # dictionary
+
+
+def check_amp(model):
+ # Check PyTorch Automatic Mixed Precision (AMP) functionality. Return True on correct operation
+ from models.common import AutoShape, DetectMultiBackend
+
+ def amp_allclose(model, im):
+ # All close FP32 vs AMP results
+ m = AutoShape(model, verbose=False) # model
+ a = m(im).xywhn[0] # FP32 inference
+ m.amp = True
+ b = m(im).xywhn[0] # AMP inference
+ return a.shape == b.shape and torch.allclose(a, b, atol=0.1) # close to 10% absolute tolerance
+
+ prefix = colorstr('AMP: ')
+ device = next(model.parameters()).device # get model device
+ if device.type in ('cpu', 'mps'):
+ return False # AMP only used on CUDA devices
+ f = ROOT / 'data' / 'images' / 'bus.jpg' # image to check
+ im = f if f.exists() else 'https://ultralytics.com/images/bus.jpg' if check_online() else np.ones((640, 640, 3))
+ try:
+ assert amp_allclose(deepcopy(model), im) or amp_allclose(DetectMultiBackend('yolov5n.pt', device), im)
+ LOGGER.info(f'{prefix}checks passed ✅')
+ return True
+ except Exception:
+ help_url = 'https://github.com/ultralytics/yolov5/issues/7908'
+ LOGGER.warning(f'{prefix}checks failed ❌, disabling Automatic Mixed Precision. See {help_url}')
+ return False
+
+
+def yaml_load(file='data.yaml'):
+ # Single-line safe yaml loading
+ with open(file, errors='ignore') as f:
+ return yaml.safe_load(f)
+
+
+def yaml_save(file='data.yaml', data={}):
+ # Single-line safe yaml saving
+ with open(file, 'w') as f:
+ yaml.safe_dump({k: str(v) if isinstance(v, Path) else v for k, v in data.items()}, f, sort_keys=False)
+
+
+def unzip_file(file, path=None, exclude=('.DS_Store', '__MACOSX')):
+ # Unzip a *.zip file to path/, excluding files containing strings in exclude list
+ if path is None:
+ path = Path(file).parent # default path
+ with ZipFile(file) as zipObj:
+ for f in zipObj.namelist(): # list all archived filenames in the zip
+ if all(x not in f for x in exclude):
+ zipObj.extract(f, path=path)
+
+
+def url2file(url):
+ # Convert URL to filename, i.e. https://url.com/file.txt?auth -> file.txt
+ url = str(Path(url)).replace(':/', '://') # Pathlib turns :// -> :/
+ return Path(urllib.parse.unquote(url)).name.split('?')[0] # '%2F' to '/', split https://url.com/file.txt?auth
+
+
+def download(url, dir='.', unzip=True, delete=True, curl=False, threads=1, retry=3):
+ # Multithreaded file download and unzip function, used in data.yaml for autodownload
+ def download_one(url, dir):
+ # Download 1 file
+ success = True
+ if os.path.isfile(url):
+ f = Path(url) # filename
+ else: # does not exist
+ f = dir / Path(url).name
+ LOGGER.info(f'Downloading {url} to {f}...')
+ for i in range(retry + 1):
+ if curl:
+ success = curl_download(url, f, silent=(threads > 1))
+ else:
+ torch.hub.download_url_to_file(url, f, progress=threads == 1) # torch download
+ success = f.is_file()
+ if success:
+ break
+ elif i < retry:
+ LOGGER.warning(f'⚠️ Download failure, retrying {i + 1}/{retry} {url}...')
+ else:
+ LOGGER.warning(f'❌ Failed to download {url}...')
+
+ if unzip and success and (f.suffix == '.gz' or is_zipfile(f) or is_tarfile(f)):
+ LOGGER.info(f'Unzipping {f}...')
+ if is_zipfile(f):
+ unzip_file(f, dir) # unzip
+ elif is_tarfile(f):
+ subprocess.run(['tar', 'xf', f, '--directory', f.parent], check=True) # unzip
+ elif f.suffix == '.gz':
+ subprocess.run(['tar', 'xfz', f, '--directory', f.parent], check=True) # unzip
+ if delete:
+ f.unlink() # remove zip
+
+ dir = Path(dir)
+ dir.mkdir(parents=True, exist_ok=True) # make directory
+ if threads > 1:
+ pool = ThreadPool(threads)
+ pool.imap(lambda x: download_one(*x), zip(url, repeat(dir))) # multithreaded
+ pool.close()
+ pool.join()
+ else:
+ for u in [url] if isinstance(url, (str, Path)) else url:
+ download_one(u, dir)
+
+
+def make_divisible(x, divisor):
+ # Returns nearest x divisible by divisor
+ if isinstance(divisor, torch.Tensor):
+ divisor = int(divisor.max()) # to int
+ return math.ceil(x / divisor) * divisor
+
+
+def clean_str(s):
+ # Cleans a string by replacing special characters with underscore _
+ return re.sub(pattern='[|@#!¡·$€%&()=?¿^*;:,¨´><+]', repl='_', string=s)
+
+
+def one_cycle(y1=0.0, y2=1.0, steps=100):
+ # lambda function for sinusoidal ramp from y1 to y2 https://arxiv.org/pdf/1812.01187.pdf
+ return lambda x: ((1 - math.cos(x * math.pi / steps)) / 2) * (y2 - y1) + y1
+
+
+def colorstr(*input):
+ # Colors a string https://en.wikipedia.org/wiki/ANSI_escape_code, i.e. colorstr('blue', 'hello world')
+ *args, string = input if len(input) > 1 else ('blue', 'bold', input[0]) # color arguments, string
+ colors = {
+ 'black': '\033[30m', # basic colors
+ 'red': '\033[31m',
+ 'green': '\033[32m',
+ 'yellow': '\033[33m',
+ 'blue': '\033[34m',
+ 'magenta': '\033[35m',
+ 'cyan': '\033[36m',
+ 'white': '\033[37m',
+ 'bright_black': '\033[90m', # bright colors
+ 'bright_red': '\033[91m',
+ 'bright_green': '\033[92m',
+ 'bright_yellow': '\033[93m',
+ 'bright_blue': '\033[94m',
+ 'bright_magenta': '\033[95m',
+ 'bright_cyan': '\033[96m',
+ 'bright_white': '\033[97m',
+ 'end': '\033[0m', # misc
+ 'bold': '\033[1m',
+ 'underline': '\033[4m'}
+ return ''.join(colors[x] for x in args) + f'{string}' + colors['end']
+
+
+def labels_to_class_weights(labels, nc=80):
+ # Get class weights (inverse frequency) from training labels
+ if labels[0] is None: # no labels loaded
+ return torch.Tensor()
+
+ labels = np.concatenate(labels, 0) # labels.shape = (866643, 5) for COCO
+ classes = labels[:, 0].astype(int) # labels = [class xywh]
+ weights = np.bincount(classes, minlength=nc) # occurrences per class
+
+ # Prepend gridpoint count (for uCE training)
+ # gpi = ((320 / 32 * np.array([1, 2, 4])) ** 2 * 3).sum() # gridpoints per image
+ # weights = np.hstack([gpi * len(labels) - weights.sum() * 9, weights * 9]) ** 0.5 # prepend gridpoints to start
+
+ weights[weights == 0] = 1 # replace empty bins with 1
+ weights = 1 / weights # number of targets per class
+ weights /= weights.sum() # normalize
+ return torch.from_numpy(weights).float()
+
+
+def labels_to_image_weights(labels, nc=80, class_weights=np.ones(80)):
+ # Produces image weights based on class_weights and image contents
+ # Usage: index = random.choices(range(n), weights=image_weights, k=1) # weighted image sample
+ class_counts = np.array([np.bincount(x[:, 0].astype(int), minlength=nc) for x in labels])
+ return (class_weights.reshape(1, nc) * class_counts).sum(1)
+
+
+def coco80_to_coco91_class(): # converts 80-index (val2014) to 91-index (paper)
+ # https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/
+ # a = np.loadtxt('data/coco.names', dtype='str', delimiter='\n')
+ # b = np.loadtxt('data/coco_paper.names', dtype='str', delimiter='\n')
+ # x1 = [list(a[i] == b).index(True) + 1 for i in range(80)] # darknet to coco
+ # x2 = [list(b[i] == a).index(True) if any(b[i] == a) else None for i in range(91)] # coco to darknet
+ return [
+ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 27, 28, 31, 32, 33, 34,
+ 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63,
+ 64, 65, 67, 70, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 84, 85, 86, 87, 88, 89, 90]
+
+
+def xyxy2xywh(x):
+ # Convert nx4 boxes from [x1, y1, x2, y2] to [x, y, w, h] where xy1=top-left, xy2=bottom-right
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[..., 0] = (x[..., 0] + x[..., 2]) / 2 # x center
+ y[..., 1] = (x[..., 1] + x[..., 3]) / 2 # y center
+ y[..., 2] = x[..., 2] - x[..., 0] # width
+ y[..., 3] = x[..., 3] - x[..., 1] # height
+ return y
+
+
+def xywh2xyxy(x):
+ # Convert nx4 boxes from [x, y, w, h] to [x1, y1, x2, y2] where xy1=top-left, xy2=bottom-right
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[..., 0] = x[..., 0] - x[..., 2] / 2 # top left x
+ y[..., 1] = x[..., 1] - x[..., 3] / 2 # top left y
+ y[..., 2] = x[..., 0] + x[..., 2] / 2 # bottom right x
+ y[..., 3] = x[..., 1] + x[..., 3] / 2 # bottom right y
+ return y
+
+
+def xywhn2xyxy(x, w=640, h=640, padw=0, padh=0):
+ # Convert nx4 boxes from [x, y, w, h] normalized to [x1, y1, x2, y2] where xy1=top-left, xy2=bottom-right
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[..., 0] = w * (x[..., 0] - x[..., 2] / 2) + padw # top left x
+ y[..., 1] = h * (x[..., 1] - x[..., 3] / 2) + padh # top left y
+ y[..., 2] = w * (x[..., 0] + x[..., 2] / 2) + padw # bottom right x
+ y[..., 3] = h * (x[..., 1] + x[..., 3] / 2) + padh # bottom right y
+ return y
+
+
+def xyxy2xywhn(x, w=640, h=640, clip=False, eps=0.0):
+ # Convert nx4 boxes from [x1, y1, x2, y2] to [x, y, w, h] normalized where xy1=top-left, xy2=bottom-right
+ if clip:
+ clip_boxes(x, (h - eps, w - eps)) # warning: inplace clip
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[..., 0] = ((x[..., 0] + x[..., 2]) / 2) / w # x center
+ y[..., 1] = ((x[..., 1] + x[..., 3]) / 2) / h # y center
+ y[..., 2] = (x[..., 2] - x[..., 0]) / w # width
+ y[..., 3] = (x[..., 3] - x[..., 1]) / h # height
+ return y
+
+
+def xyn2xy(x, w=640, h=640, padw=0, padh=0):
+ # Convert normalized segments into pixel segments, shape (n,2)
+ y = x.clone() if isinstance(x, torch.Tensor) else np.copy(x)
+ y[..., 0] = w * x[..., 0] + padw # top left x
+ y[..., 1] = h * x[..., 1] + padh # top left y
+ return y
+
+
+def segment2box(segment, width=640, height=640):
+ # Convert 1 segment label to 1 box label, applying inside-image constraint, i.e. (xy1, xy2, ...) to (xyxy)
+ x, y = segment.T # segment xy
+ inside = (x >= 0) & (y >= 0) & (x <= width) & (y <= height)
+ x, y, = x[inside], y[inside]
+ return np.array([x.min(), y.min(), x.max(), y.max()]) if any(x) else np.zeros((1, 4)) # xyxy
+
+
+def segments2boxes(segments):
+ # Convert segment labels to box labels, i.e. (cls, xy1, xy2, ...) to (cls, xywh)
+ boxes = []
+ for s in segments:
+ x, y = s.T # segment xy
+ boxes.append([x.min(), y.min(), x.max(), y.max()]) # cls, xyxy
+ return xyxy2xywh(np.array(boxes)) # cls, xywh
+
+
+def resample_segments(segments, n=1000):
+ # Up-sample an (n,2) segment
+ for i, s in enumerate(segments):
+ s = np.concatenate((s, s[0:1, :]), axis=0)
+ x = np.linspace(0, len(s) - 1, n)
+ xp = np.arange(len(s))
+ segments[i] = np.concatenate([np.interp(x, xp, s[:, i]) for i in range(2)]).reshape(2, -1).T # segment xy
+ return segments
+
+
+def scale_boxes(img1_shape, boxes, img0_shape, ratio_pad=None):
+ # Rescale boxes (xyxy) from img1_shape to img0_shape
+ if ratio_pad is None: # calculate from img0_shape
+ gain = min(img1_shape[0] / img0_shape[0], img1_shape[1] / img0_shape[1]) # gain = old / new
+ pad = (img1_shape[1] - img0_shape[1] * gain) / 2, (img1_shape[0] - img0_shape[0] * gain) / 2 # wh padding
+ else:
+ gain = ratio_pad[0][0]
+ pad = ratio_pad[1]
+
+ boxes[..., [0, 2]] -= pad[0] # x padding
+ boxes[..., [1, 3]] -= pad[1] # y padding
+ boxes[..., :4] /= gain
+ clip_boxes(boxes, img0_shape)
+ return boxes
+
+
+def scale_segments(img1_shape, segments, img0_shape, ratio_pad=None, normalize=False):
+ # Rescale coords (xyxy) from img1_shape to img0_shape
+ if ratio_pad is None: # calculate from img0_shape
+ gain = min(img1_shape[0] / img0_shape[0], img1_shape[1] / img0_shape[1]) # gain = old / new
+ pad = (img1_shape[1] - img0_shape[1] * gain) / 2, (img1_shape[0] - img0_shape[0] * gain) / 2 # wh padding
+ else:
+ gain = ratio_pad[0][0]
+ pad = ratio_pad[1]
+
+ segments[:, 0] -= pad[0] # x padding
+ segments[:, 1] -= pad[1] # y padding
+ segments /= gain
+ clip_segments(segments, img0_shape)
+ if normalize:
+ segments[:, 0] /= img0_shape[1] # width
+ segments[:, 1] /= img0_shape[0] # height
+ return segments
+
+
+def clip_boxes(boxes, shape):
+ # Clip boxes (xyxy) to image shape (height, width)
+ if isinstance(boxes, torch.Tensor): # faster individually
+ boxes[..., 0].clamp_(0, shape[1]) # x1
+ boxes[..., 1].clamp_(0, shape[0]) # y1
+ boxes[..., 2].clamp_(0, shape[1]) # x2
+ boxes[..., 3].clamp_(0, shape[0]) # y2
+ else: # np.array (faster grouped)
+ boxes[..., [0, 2]] = boxes[..., [0, 2]].clip(0, shape[1]) # x1, x2
+ boxes[..., [1, 3]] = boxes[..., [1, 3]].clip(0, shape[0]) # y1, y2
+
+
+def clip_segments(segments, shape):
+ # Clip segments (xy1,xy2,...) to image shape (height, width)
+ if isinstance(segments, torch.Tensor): # faster individually
+ segments[:, 0].clamp_(0, shape[1]) # x
+ segments[:, 1].clamp_(0, shape[0]) # y
+ else: # np.array (faster grouped)
+ segments[:, 0] = segments[:, 0].clip(0, shape[1]) # x
+ segments[:, 1] = segments[:, 1].clip(0, shape[0]) # y
+
+
+def non_max_suppression(
+ prediction,
+ conf_thres=0.25,
+ iou_thres=0.45,
+ classes=None,
+ agnostic=False,
+ multi_label=False,
+ labels=(),
+ max_det=300,
+ nm=0, # number of masks
+):
+ """Non-Maximum Suppression (NMS) on inference results to reject overlapping detections
+
+ Returns:
+ list of detections, on (n,6) tensor per image [xyxy, conf, cls]
+ """
+
+ # Checks
+ assert 0 <= conf_thres <= 1, f'Invalid Confidence threshold {conf_thres}, valid values are between 0.0 and 1.0'
+ assert 0 <= iou_thres <= 1, f'Invalid IoU {iou_thres}, valid values are between 0.0 and 1.0'
+ if isinstance(prediction, (list, tuple)): # YOLOv5 model in validation model, output = (inference_out, loss_out)
+ prediction = prediction[0] # select only inference output
+
+ device = prediction.device
+ mps = 'mps' in device.type # Apple MPS
+ if mps: # MPS not fully supported yet, convert tensors to CPU before NMS
+ prediction = prediction.cpu()
+ bs = prediction.shape[0] # batch size
+ nc = prediction.shape[2] - nm - 5 # number of classes
+ xc = prediction[..., 4] > conf_thres # candidates
+
+ # Settings
+ # min_wh = 2 # (pixels) minimum box width and height
+ max_wh = 7680 # (pixels) maximum box width and height
+ max_nms = 30000 # maximum number of boxes into torchvision.ops.nms()
+ time_limit = 0.5 + 0.05 * bs # seconds to quit after
+ redundant = True # require redundant detections
+ multi_label &= nc > 1 # multiple labels per box (adds 0.5ms/img)
+ merge = False # use merge-NMS
+
+ t = time.time()
+ mi = 5 + nc # mask start index
+ output = [torch.zeros((0, 6 + nm), device=prediction.device)] * bs
+ for xi, x in enumerate(prediction): # image index, image inference
+ # Apply constraints
+ # x[((x[..., 2:4] < min_wh) | (x[..., 2:4] > max_wh)).any(1), 4] = 0 # width-height
+ x = x[xc[xi]] # confidence
+
+ # Cat apriori labels if autolabelling
+ if labels and len(labels[xi]):
+ lb = labels[xi]
+ v = torch.zeros((len(lb), nc + nm + 5), device=x.device)
+ v[:, :4] = lb[:, 1:5] # box
+ v[:, 4] = 1.0 # conf
+ v[range(len(lb)), lb[:, 0].long() + 5] = 1.0 # cls
+ x = torch.cat((x, v), 0)
+
+ # If none remain process next image
+ if not x.shape[0]:
+ continue
+
+ # Compute conf
+ x[:, 5:] *= x[:, 4:5] # conf = obj_conf * cls_conf
+
+ # Box/Mask
+ box = xywh2xyxy(x[:, :4]) # center_x, center_y, width, height) to (x1, y1, x2, y2)
+ mask = x[:, mi:] # zero columns if no masks
+
+ # Detections matrix nx6 (xyxy, conf, cls)
+ if multi_label:
+ i, j = (x[:, 5:mi] > conf_thres).nonzero(as_tuple=False).T
+ x = torch.cat((box[i], x[i, 5 + j, None], j[:, None].float(), mask[i]), 1)
+ else: # best class only
+ conf, j = x[:, 5:mi].max(1, keepdim=True)
+ x = torch.cat((box, conf, j.float(), mask), 1)[conf.view(-1) > conf_thres]
+
+ # Filter by class
+ if classes is not None:
+ x = x[(x[:, 5:6] == torch.tensor(classes, device=x.device)).any(1)]
+
+ # Apply finite constraint
+ # if not torch.isfinite(x).all():
+ # x = x[torch.isfinite(x).all(1)]
+
+ # Check shape
+ n = x.shape[0] # number of boxes
+ if not n: # no boxes
+ continue
+ x = x[x[:, 4].argsort(descending=True)[:max_nms]] # sort by confidence and remove excess boxes
+
+ # Batched NMS
+ c = x[:, 5:6] * (0 if agnostic else max_wh) # classes
+ boxes, scores = x[:, :4] + c, x[:, 4] # boxes (offset by class), scores
+ i = torchvision.ops.nms(boxes, scores, iou_thres) # NMS
+ i = i[:max_det] # limit detections
+ if merge and (1 < n < 3E3): # Merge NMS (boxes merged using weighted mean)
+ # update boxes as boxes(i,4) = weights(i,n) * boxes(n,4)
+ iou = box_iou(boxes[i], boxes) > iou_thres # iou matrix
+ weights = iou * scores[None] # box weights
+ x[i, :4] = torch.mm(weights, x[:, :4]).float() / weights.sum(1, keepdim=True) # merged boxes
+ if redundant:
+ i = i[iou.sum(1) > 1] # require redundancy
+
+ output[xi] = x[i]
+ if mps:
+ output[xi] = output[xi].to(device)
+ if (time.time() - t) > time_limit:
+ LOGGER.warning(f'WARNING ⚠️ NMS time limit {time_limit:.3f}s exceeded')
+ break # time limit exceeded
+
+ return output
+
+
+def strip_optimizer(f='best.pt', s=''): # from utils.general import *; strip_optimizer()
+ # Strip optimizer from 'f' to finalize training, optionally save as 's'
+ x = torch.load(f, map_location=torch.device('cpu'))
+ if x.get('ema'):
+ x['model'] = x['ema'] # replace model with ema
+ for k in 'optimizer', 'best_fitness', 'ema', 'updates': # keys
+ x[k] = None
+ x['epoch'] = -1
+ x['model'].half() # to FP16
+ for p in x['model'].parameters():
+ p.requires_grad = False
+ torch.save(x, s or f)
+ mb = os.path.getsize(s or f) / 1E6 # filesize
+ LOGGER.info(f"Optimizer stripped from {f},{f' saved as {s},' if s else ''} {mb:.1f}MB")
+
+
+def print_mutation(keys, results, hyp, save_dir, bucket, prefix=colorstr('evolve: ')):
+ evolve_csv = save_dir / 'evolve.csv'
+ evolve_yaml = save_dir / 'hyp_evolve.yaml'
+ keys = tuple(keys) + tuple(hyp.keys()) # [results + hyps]
+ keys = tuple(x.strip() for x in keys)
+ vals = results + tuple(hyp.values())
+ n = len(keys)
+
+ # Download (optional)
+ if bucket:
+ url = f'gs://{bucket}/evolve.csv'
+ if gsutil_getsize(url) > (evolve_csv.stat().st_size if evolve_csv.exists() else 0):
+ subprocess.run(['gsutil', 'cp', f'{url}', f'{save_dir}']) # download evolve.csv if larger than local
+
+ # Log to evolve.csv
+ s = '' if evolve_csv.exists() else (('%20s,' * n % keys).rstrip(',') + '\n') # add header
+ with open(evolve_csv, 'a') as f:
+ f.write(s + ('%20.5g,' * n % vals).rstrip(',') + '\n')
+
+ # Save yaml
+ with open(evolve_yaml, 'w') as f:
+ data = pd.read_csv(evolve_csv, skipinitialspace=True)
+ data = data.rename(columns=lambda x: x.strip()) # strip keys
+ i = np.argmax(fitness(data.values[:, :4])) #
+ generations = len(data)
+ f.write('# YOLOv5 Hyperparameter Evolution Results\n' + f'# Best generation: {i}\n' +
+ f'# Last generation: {generations - 1}\n' + '# ' + ', '.join(f'{x.strip():>20s}' for x in keys[:7]) +
+ '\n' + '# ' + ', '.join(f'{x:>20.5g}' for x in data.values[i, :7]) + '\n\n')
+ yaml.safe_dump(data.loc[i][7:].to_dict(), f, sort_keys=False)
+
+ # Print to screen
+ LOGGER.info(prefix + f'{generations} generations finished, current result:\n' + prefix +
+ ', '.join(f'{x.strip():>20s}' for x in keys) + '\n' + prefix + ', '.join(f'{x:20.5g}'
+ for x in vals) + '\n\n')
+
+ if bucket:
+ subprocess.run(['gsutil', 'cp', f'{evolve_csv}', f'{evolve_yaml}', f'gs://{bucket}']) # upload
+
+
+def apply_classifier(x, model, img, im0):
+ # Apply a second stage classifier to YOLO outputs
+ # Example model = torchvision.models.__dict__['efficientnet_b0'](pretrained=True).to(device).eval()
+ im0 = [im0] if isinstance(im0, np.ndarray) else im0
+ for i, d in enumerate(x): # per image
+ if d is not None and len(d):
+ d = d.clone()
+
+ # Reshape and pad cutouts
+ b = xyxy2xywh(d[:, :4]) # boxes
+ b[:, 2:] = b[:, 2:].max(1)[0].unsqueeze(1) # rectangle to square
+ b[:, 2:] = b[:, 2:] * 1.3 + 30 # pad
+ d[:, :4] = xywh2xyxy(b).long()
+
+ # Rescale boxes from img_size to im0 size
+ scale_boxes(img.shape[2:], d[:, :4], im0[i].shape)
+
+ # Classes
+ pred_cls1 = d[:, 5].long()
+ ims = []
+ for a in d:
+ cutout = im0[i][int(a[1]):int(a[3]), int(a[0]):int(a[2])]
+ im = cv2.resize(cutout, (224, 224)) # BGR
+
+ im = im[:, :, ::-1].transpose(2, 0, 1) # BGR to RGB, to 3x416x416
+ im = np.ascontiguousarray(im, dtype=np.float32) # uint8 to float32
+ im /= 255 # 0 - 255 to 0.0 - 1.0
+ ims.append(im)
+
+ pred_cls2 = model(torch.Tensor(ims).to(d.device)).argmax(1) # classifier prediction
+ x[i] = x[i][pred_cls1 == pred_cls2] # retain matching class detections
+
+ return x
+
+
+def increment_path(path, exist_ok=False, sep='', mkdir=False):
+ # Increment file or directory path, i.e. runs/exp --> runs/exp{sep}2, runs/exp{sep}3, ... etc.
+ path = Path(path) # os-agnostic
+ if path.exists() and not exist_ok:
+ path, suffix = (path.with_suffix(''), path.suffix) if path.is_file() else (path, '')
+
+ # Method 1
+ for n in range(2, 9999):
+ p = f'{path}{sep}{n}{suffix}' # increment path
+ if not os.path.exists(p): #
+ break
+ path = Path(p)
+
+ # Method 2 (deprecated)
+ # dirs = glob.glob(f"{path}{sep}*") # similar paths
+ # matches = [re.search(rf"{path.stem}{sep}(\d+)", d) for d in dirs]
+ # i = [int(m.groups()[0]) for m in matches if m] # indices
+ # n = max(i) + 1 if i else 2 # increment number
+ # path = Path(f"{path}{sep}{n}{suffix}") # increment path
+
+ if mkdir:
+ path.mkdir(parents=True, exist_ok=True) # make directory
+
+ return path
+
+
+# OpenCV Multilanguage-friendly functions ------------------------------------------------------------------------------------
+imshow_ = cv2.imshow # copy to avoid recursion errors
+
+
+def imread(filename, flags=cv2.IMREAD_COLOR):
+ return cv2.imdecode(np.fromfile(filename, np.uint8), flags)
+
+
+def imwrite(filename, img):
+ try:
+ cv2.imencode(Path(filename).suffix, img)[1].tofile(filename)
+ return True
+ except Exception:
+ return False
+
+
+def imshow(path, im):
+ imshow_(path.encode('unicode_escape').decode(), im)
+
+
+if Path(inspect.stack()[0].filename).parent.parent.as_posix() in inspect.stack()[-1].filename:
+ cv2.imread, cv2.imwrite, cv2.imshow = imread, imwrite, imshow # redefine
+
+# Variables ------------------------------------------------------------------------------------------------------------
\ No newline at end of file
diff --git a/vouchervision/component_detector/utils/google_app_engine/Dockerfile b/vouchervision/component_detector/utils/google_app_engine/Dockerfile
new file mode 100644
index 0000000000000000000000000000000000000000..0155618f475104e9858b81470339558156c94e13
--- /dev/null
+++ b/vouchervision/component_detector/utils/google_app_engine/Dockerfile
@@ -0,0 +1,25 @@
+FROM gcr.io/google-appengine/python
+
+# Create a virtualenv for dependencies. This isolates these packages from
+# system-level packages.
+# Use -p python3 or -p python3.7 to select python version. Default is version 2.
+RUN virtualenv /env -p python3
+
+# Setting these environment variables are the same as running
+# source /env/bin/activate.
+ENV VIRTUAL_ENV /env
+ENV PATH /env/bin:$PATH
+
+RUN apt-get update && apt-get install -y python-opencv
+
+# Copy the application's requirements.txt and run pip to install all
+# dependencies into the virtualenv.
+ADD requirements.txt /app/requirements.txt
+RUN pip install -r /app/requirements.txt
+
+# Add the application source code.
+ADD . /app
+
+# Run a WSGI server to serve the application. gunicorn must be declared as
+# a dependency in requirements.txt.
+CMD gunicorn -b :$PORT main:app
diff --git a/vouchervision/component_detector/utils/google_app_engine/additional_requirements.txt b/vouchervision/component_detector/utils/google_app_engine/additional_requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..42d7ffc0eed83e62f67adde186a711ebeef0be5a
--- /dev/null
+++ b/vouchervision/component_detector/utils/google_app_engine/additional_requirements.txt
@@ -0,0 +1,4 @@
+# add these requirements in your app on top of the existing ones
+pip==21.1
+Flask==1.0.2
+gunicorn==19.9.0
diff --git a/vouchervision/component_detector/utils/google_app_engine/app.yaml b/vouchervision/component_detector/utils/google_app_engine/app.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..5056b7c1186d6ad278957bbd6e976c3a0f169a30
--- /dev/null
+++ b/vouchervision/component_detector/utils/google_app_engine/app.yaml
@@ -0,0 +1,14 @@
+runtime: custom
+env: flex
+
+service: yolov5app
+
+liveness_check:
+ initial_delay_sec: 600
+
+manual_scaling:
+ instances: 1
+resources:
+ cpu: 1
+ memory_gb: 4
+ disk_size_gb: 20
diff --git a/vouchervision/component_detector/utils/loggers/__init__.py b/vouchervision/component_detector/utils/loggers/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..b33b04e467525e469dba496d4f96d7f93f164fb7
--- /dev/null
+++ b/vouchervision/component_detector/utils/loggers/__init__.py
@@ -0,0 +1,187 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Logging utils
+"""
+
+import os
+import warnings
+from threading import Thread
+
+import pkg_resources as pkg
+import torch
+from torch.utils.tensorboard import SummaryWriter
+
+from utils.general import colorstr, cv2, emojis
+from utils.loggers.wandb.wandb_utils import WandbLogger
+from utils.plots import plot_images, plot_results
+from utils.torch_utils import de_parallel
+
+LOGGERS = ('csv', 'tb', 'wandb') # text-file, TensorBoard, Weights & Biases
+RANK = int(os.getenv('RANK', -1))
+
+try:
+ import wandb
+
+ assert hasattr(wandb, '__version__') # verify package import not local dir
+ if pkg.parse_version(wandb.__version__) >= pkg.parse_version('0.12.2') and RANK in [0, -1]:
+ try:
+ wandb_login_success = wandb.login(timeout=30)
+ except wandb.errors.UsageError: # known non-TTY terminal issue
+ wandb_login_success = False
+ if not wandb_login_success:
+ wandb = None
+except (ImportError, AssertionError):
+ wandb = None
+
+
+class Loggers():
+ # YOLOv5 Loggers class
+ def __init__(self, save_dir=None, weights=None, opt=None, hyp=None, logger=None, include=LOGGERS):
+ self.save_dir = save_dir
+ self.weights = weights
+ self.opt = opt
+ self.hyp = hyp
+ self.logger = logger # for printing results to console
+ self.include = include
+ self.keys = [
+ 'train/box_loss',
+ 'train/obj_loss',
+ 'train/cls_loss', # train loss
+ 'metrics/precision',
+ 'metrics/recall',
+ 'metrics/mAP_0.5',
+ 'metrics/mAP_0.5:0.95', # metrics
+ 'val/box_loss',
+ 'val/obj_loss',
+ 'val/cls_loss', # val loss
+ 'x/lr0',
+ 'x/lr1',
+ 'x/lr2'] # params
+ self.best_keys = ['best/epoch', 'best/precision', 'best/recall', 'best/mAP_0.5', 'best/mAP_0.5:0.95']
+ for k in LOGGERS:
+ setattr(self, k, None) # init empty logger dictionary
+ self.csv = True # always log to csv
+
+ # Message
+ if not wandb:
+ prefix = colorstr('Weights & Biases: ')
+ s = f"{prefix}run 'pip install wandb' to automatically track and visualize YOLOv5 🚀 runs (RECOMMENDED)"
+ self.logger.info(emojis(s))
+
+ # TensorBoard
+ s = self.save_dir
+ if 'tb' in self.include and not self.opt.evolve:
+ prefix = colorstr('TensorBoard: ')
+ self.logger.info(f"{prefix}Start with 'tensorboard --logdir {s.parent}', view at http://localhost:6006/")
+ self.tb = SummaryWriter(str(s))
+
+ # W&B
+ if wandb and 'wandb' in self.include:
+ wandb_artifact_resume = isinstance(self.opt.resume, str) and self.opt.resume.startswith('wandb-artifact://')
+ run_id = torch.load(self.weights).get('wandb_id') if self.opt.resume and not wandb_artifact_resume else None
+ self.opt.hyp = self.hyp # add hyperparameters
+ self.wandb = WandbLogger(self.opt, run_id)
+ # temp warn. because nested artifacts not supported after 0.12.10
+ if pkg.parse_version(wandb.__version__) >= pkg.parse_version('0.12.11'):
+ self.logger.warning(
+ "YOLOv5 temporarily requires wandb version 0.12.10 or below. Some features may not work as expected."
+ )
+ else:
+ self.wandb = None
+
+ def on_train_start(self):
+ # Callback runs on train start
+ pass
+
+ def on_pretrain_routine_end(self):
+ # Callback runs on pre-train routine end
+ paths = self.save_dir.glob('*labels*.jpg') # training labels
+ if self.wandb:
+ self.wandb.log({"Labels": [wandb.Image(str(x), caption=x.name) for x in paths]})
+
+ def on_train_batch_end(self, ni, model, imgs, targets, paths, plots):
+ # Callback runs on train batch end
+ if plots:
+ if ni == 0:
+ if not self.opt.sync_bn: # --sync known issue https://github.com/ultralytics/yolov5/issues/3754
+ with warnings.catch_warnings():
+ warnings.simplefilter('ignore') # suppress jit trace warning
+ self.tb.add_graph(torch.jit.trace(de_parallel(model), imgs[0:1], strict=False), [])
+ if ni < 3:
+ f = self.save_dir / f'train_batch{ni}.jpg' # filename
+ Thread(target=plot_images, args=(imgs, targets, paths, f), daemon=True).start()
+ if self.wandb and ni == 10:
+ files = sorted(self.save_dir.glob('train*.jpg'))
+ self.wandb.log({'Mosaics': [wandb.Image(str(f), caption=f.name) for f in files if f.exists()]})
+
+ def on_train_epoch_end(self, epoch):
+ # Callback runs on train epoch end
+ if self.wandb:
+ self.wandb.current_epoch = epoch + 1
+
+ def on_val_image_end(self, pred, predn, path, names, im):
+ # Callback runs on val image end
+ if self.wandb:
+ self.wandb.val_one_image(pred, predn, path, names, im)
+
+ def on_val_end(self):
+ # Callback runs on val end
+ if self.wandb:
+ files = sorted(self.save_dir.glob('val*.jpg'))
+ self.wandb.log({"Validation": [wandb.Image(str(f), caption=f.name) for f in files]})
+
+ def on_fit_epoch_end(self, vals, epoch, best_fitness, fi):
+ # Callback runs at the end of each fit (train+val) epoch
+ x = {k: v for k, v in zip(self.keys, vals)} # dict
+ if self.csv:
+ file = self.save_dir / 'results.csv'
+ n = len(x) + 1 # number of cols
+ s = '' if file.exists() else (('%20s,' * n % tuple(['epoch'] + self.keys)).rstrip(',') + '\n') # add header
+ with open(file, 'a') as f:
+ f.write(s + ('%20.5g,' * n % tuple([epoch] + vals)).rstrip(',') + '\n')
+
+ if self.tb:
+ for k, v in x.items():
+ self.tb.add_scalar(k, v, epoch)
+
+ if self.wandb:
+ if best_fitness == fi:
+ best_results = [epoch] + vals[3:7]
+ for i, name in enumerate(self.best_keys):
+ self.wandb.wandb_run.summary[name] = best_results[i] # log best results in the summary
+ self.wandb.log(x)
+ self.wandb.end_epoch(best_result=best_fitness == fi)
+
+ def on_model_save(self, last, epoch, final_epoch, best_fitness, fi):
+ # Callback runs on model save event
+ if self.wandb:
+ if ((epoch + 1) % self.opt.save_period == 0 and not final_epoch) and self.opt.save_period != -1:
+ self.wandb.log_model(last.parent, self.opt, epoch, fi, best_model=best_fitness == fi)
+
+ def on_train_end(self, last, best, plots, epoch, results):
+ # Callback runs on training end
+ if plots:
+ plot_results(file=self.save_dir / 'results.csv') # save results.png
+ files = ['results.png', 'confusion_matrix.png', *(f'{x}_curve.png' for x in ('F1', 'PR', 'P', 'R'))]
+ files = [(self.save_dir / f) for f in files if (self.save_dir / f).exists()] # filter
+
+ if self.tb:
+ for f in files:
+ self.tb.add_image(f.stem, cv2.imread(str(f))[..., ::-1], epoch, dataformats='HWC')
+
+ if self.wandb:
+ self.wandb.log({k: v for k, v in zip(self.keys[3:10], results)}) # log best.pt val results
+ self.wandb.log({"Results": [wandb.Image(str(f), caption=f.name) for f in files]})
+ # Calling wandb.log. TODO: Refactor this into WandbLogger.log_model
+ if not self.opt.evolve:
+ wandb.log_artifact(str(best if best.exists() else last),
+ type='model',
+ name='run_' + self.wandb.wandb_run.id + '_model',
+ aliases=['latest', 'best', 'stripped'])
+ self.wandb.finish_run()
+
+ def on_params_update(self, params):
+ # Update hyperparams or configs of the experiment
+ # params: A dict containing {param: value} pairs
+ if self.wandb:
+ self.wandb.wandb_run.config.update(params, allow_val_change=True)
diff --git a/vouchervision/component_detector/utils/loss.py b/vouchervision/component_detector/utils/loss.py
new file mode 100644
index 0000000000000000000000000000000000000000..a32cf110dd443fe8b39ce49e6b9555836c5de267
--- /dev/null
+++ b/vouchervision/component_detector/utils/loss.py
@@ -0,0 +1,293 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Loss functions
+"""
+
+import torch
+import torch.nn as nn
+
+from utils.metrics import bbox_iou
+from utils.torch_utils import de_parallel
+
+
+def smooth_BCE(eps=0.1): # https://github.com/ultralytics/yolov3/issues/238#issuecomment-598028441
+ # return positive, negative label smoothing BCE targets
+ return 1.0 - 0.5 * eps, 0.5 * eps
+
+
+class BCEBlurWithLogitsLoss(nn.Module):
+ # BCEwithLogitLoss() with reduced missing label effects.
+ def __init__(self, alpha=0.05):
+ super().__init__()
+ self.loss_fcn = nn.BCEWithLogitsLoss(reduction='none') # must be nn.BCEWithLogitsLoss()
+ self.alpha = alpha
+
+ def forward(self, pred, true):
+ loss = self.loss_fcn(pred, true)
+ pred = torch.sigmoid(pred) # prob from logits
+ dx = pred - true # reduce only missing label effects
+ # dx = (pred - true).abs() # reduce missing label and false label effects
+ alpha_factor = 1 - torch.exp((dx - 1) / (self.alpha + 1e-4))
+ loss *= alpha_factor
+ return loss.mean()
+
+
+class FocalLoss(nn.Module):
+ # Wraps focal loss around existing loss_fcn(), i.e. criteria = FocalLoss(nn.BCEWithLogitsLoss(), gamma=1.5)
+ def __init__(self, loss_fcn, gamma=1.5, alpha=0.25):
+ super().__init__()
+ self.loss_fcn = loss_fcn # must be nn.BCEWithLogitsLoss()
+ self.gamma = gamma
+ self.alpha = alpha
+ self.reduction = loss_fcn.reduction
+ self.loss_fcn.reduction = 'none' # required to apply FL to each element
+
+ def forward(self, pred, true):
+ loss = self.loss_fcn(pred, true)
+ # p_t = torch.exp(-loss)
+ # loss *= self.alpha * (1.000001 - p_t) ** self.gamma # non-zero power for gradient stability
+
+ # TF implementation https://github.com/tensorflow/addons/blob/v0.7.1/tensorflow_addons/losses/focal_loss.py
+ pred_prob = torch.sigmoid(pred) # prob from logits
+ p_t = true * pred_prob + (1 - true) * (1 - pred_prob)
+ alpha_factor = true * self.alpha + (1 - true) * (1 - self.alpha)
+ modulating_factor = (1.0 - p_t) ** self.gamma
+ loss *= alpha_factor * modulating_factor
+
+ if self.reduction == 'mean':
+ return loss.mean()
+ elif self.reduction == 'sum':
+ return loss.sum()
+ else: # 'none'
+ return loss
+
+
+class QFocalLoss(nn.Module):
+ # Wraps Quality focal loss around existing loss_fcn(), i.e. criteria = FocalLoss(nn.BCEWithLogitsLoss(), gamma=1.5)
+ def __init__(self, loss_fcn, gamma=1.5, alpha=0.25):
+ super().__init__()
+ self.loss_fcn = loss_fcn # must be nn.BCEWithLogitsLoss()
+ self.gamma = gamma
+ self.alpha = alpha
+ self.reduction = loss_fcn.reduction
+ self.loss_fcn.reduction = 'none' # required to apply FL to each element
+
+ def forward(self, pred, true):
+ loss = self.loss_fcn(pred, true)
+
+ pred_prob = torch.sigmoid(pred) # prob from logits
+ alpha_factor = true * self.alpha + (1 - true) * (1 - self.alpha)
+ modulating_factor = torch.abs(true - pred_prob) ** self.gamma
+ loss *= alpha_factor * modulating_factor
+
+ if self.reduction == 'mean':
+ return loss.mean()
+ elif self.reduction == 'sum':
+ return loss.sum()
+ else: # 'none'
+ return loss
+
+
+class ComputeLoss:
+ sort_obj_iou = False
+
+ # Compute losses
+ def __init__(self, model, autobalance=False):
+ device = next(model.parameters()).device # get model device
+ h = model.hyp # hyperparameters
+
+ # Define criteria
+ BCEcls = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h['cls_pw']], device=device))
+ BCEobj = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h['obj_pw']], device=device))
+
+ # Class label smoothing https://arxiv.org/pdf/1902.04103.pdf eqn 3
+ self.cp, self.cn = smooth_BCE(eps=h.get('label_smoothing', 0.0)) # positive, negative BCE targets
+
+ # Focal loss
+ g = h['fl_gamma'] # focal loss gamma
+ if g > 0:
+ BCEcls, BCEobj = FocalLoss(BCEcls, g), FocalLoss(BCEobj, g)
+
+ m = de_parallel(model).model[-1] # Detect() module
+ self.balance = {3: [4.0, 1.0, 0.4]}.get(m.nl, [4.0, 1.0, 0.25, 0.06, 0.02]) # P3-P7
+ self.ssi = list(m.stride).index(16) if autobalance else 0 # stride 16 index
+ self.BCEcls, self.BCEobj, self.gr, self.hyp, self.autobalance = BCEcls, BCEobj, 1.0, h, autobalance
+ self.na = m.na # number of anchors
+ self.nc = m.nc # number of classes
+ self.nl = m.nl # number of layers
+ self.anchors = m.anchors
+ self.device = device
+
+ def __call__(self, p, targets): # predictions, targets
+ lcls = torch.zeros(1, device=self.device) # class loss
+ lbox = torch.zeros(1, device=self.device) # box loss
+ lobj = torch.zeros(1, device=self.device) # object loss
+ tcls, tbox, indices, anchors = self.build_targets(p, targets) # targets
+
+ # Losses
+ for i, pi in enumerate(p): # layer index, layer predictions
+ b, a, gj, gi = indices[i] # image, anchor, gridy, gridx
+ tobj = torch.zeros(pi.shape[:4], dtype=pi.dtype, device=self.device) # target obj
+
+ n = b.shape[0] # number of targets
+ if n:
+ # pxy, pwh, _, pcls = pi[b, a, gj, gi].tensor_split((2, 4, 5), dim=1) # faster, requires torch 1.8.0
+ pxy, pwh, _, pcls = pi[b, a, gj, gi].split((2, 2, 1, self.nc), 1) # target-subset of predictions
+
+ # Regression
+ pxy = pxy.sigmoid() * 2 - 0.5
+ pwh = (pwh.sigmoid() * 2) ** 2 * anchors[i]
+ pbox = torch.cat((pxy, pwh), 1) # predicted box
+ iou = bbox_iou(pbox, tbox[i], CIoU=True).squeeze() # iou(prediction, target)
+ lbox += (1.0 - iou).mean() # iou loss
+
+ # Objectness
+ iou = iou.detach().clamp(0).type(tobj.dtype)
+ if self.sort_obj_iou:
+ j = iou.argsort()
+ b, a, gj, gi, iou = b[j], a[j], gj[j], gi[j], iou[j]
+ if self.gr < 1:
+ iou = (1.0 - self.gr) + self.gr * iou
+ tobj[b, a, gj, gi] = iou # iou ratio
+
+ # Classification
+ if self.nc > 1: # cls loss (only if multiple classes)
+ t = torch.full_like(pcls, self.cn, device=self.device) # targets
+ t[range(n), tcls[i]] = self.cp
+ lcls += self.BCEcls(pcls, t) # BCE
+
+ # Append targets to text file
+ # with open('targets.txt', 'a') as file:
+ # [file.write('%11.5g ' * 4 % tuple(x) + '\n') for x in torch.cat((txy[i], twh[i]), 1)]
+
+ obji = self.BCEobj(pi[..., 4], tobj)
+ lobj += obji * self.balance[i] # obj loss
+ if self.autobalance:
+ self.balance[i] = self.balance[i] * 0.9999 + 0.0001 / obji.detach().item()
+
+ if self.autobalance:
+ self.balance = [x / self.balance[self.ssi] for x in self.balance]
+ lbox *= self.hyp['box']
+ lobj *= self.hyp['obj']
+ lcls *= self.hyp['cls']
+ bs = tobj.shape[0] # batch size
+
+ return (lbox + lobj + lcls) * bs, torch.cat((lbox, lobj, lcls)).detach()
+
+ # def build_targets(self, p, targets):
+ # # Build targets for compute_loss(), input targets(image,class,x,y,w,h)
+ # na, nt = self.na, targets.shape[0] # number of anchors, targets
+ # tcls, tbox, indices, anch = [], [], [], []
+ # gain = torch.ones(7, device=self.device) # normalized to gridspace gain
+ # ai = torch.arange(na, device=self.device).float().view(na, 1).repeat(1, nt) # same as .repeat_interleave(nt)
+ # targets = torch.cat((targets.repeat(na, 1, 1), ai[..., None]), 2) # append anchor indices
+
+ # g = 0.5 # bias
+ # off = torch.tensor(
+ # [
+ # [0, 0],
+ # [1, 0],
+ # [0, 1],
+ # [-1, 0],
+ # [0, -1], # j,k,l,m
+ # # [1, 1], [1, -1], [-1, 1], [-1, -1], # jk,jm,lk,lm
+ # ],
+ # device=self.device).float() * g # offsets
+
+ # for i in range(self.nl):
+ # anchors = self.anchors[i]
+ # gain[2:6] = torch.tensor(p[i].shape)[[3, 2, 3, 2]] # xyxy gain
+
+ # # Match targets to anchors
+ # t = targets * gain # shape(3,n,7)
+ # if nt:
+ # # Matches
+ # r = t[..., 4:6] / anchors[:, None] # wh ratio
+ # j = torch.max(r, 1 / r).max(2)[0] < self.hyp['anchor_t'] # compare
+ # # j = wh_iou(anchors, t[:, 4:6]) > model.hyp['iou_t'] # iou(3,n)=wh_iou(anchors(3,2), gwh(n,2))
+ # t = t[j] # filter
+
+ # # Offsets
+ # gxy = t[:, 2:4] # grid xy
+ # gxi = gain[[2, 3]] - gxy # inverse
+ # j, k = ((gxy % 1 < g) & (gxy > 1)).T
+ # l, m = ((gxi % 1 < g) & (gxi > 1)).T
+ # j = torch.stack((torch.ones_like(j), j, k, l, m))
+ # t = t.repeat((5, 1, 1))[j]
+ # offsets = (torch.zeros_like(gxy)[None] + off[:, None])[j]
+ # else:
+ # t = targets[0]
+ # offsets = 0
+
+ # # Define
+ # bc, gxy, gwh, a = t.chunk(4, 1) # (image, class), grid xy, grid wh, anchors
+ # a, (b, c) = a.long().view(-1), bc.long().T # anchors, image, class
+ # gij = (gxy - offsets).long()
+ # gi, gj = gij.T # grid indices
+
+ # # Append
+ # # indices.append((b, a, gj.clamp_(0, gain[3] - 1), gi.clamp_(0, gain[2] - 1))) # image, anchor, grid indices
+ # indices.append((b, a, gj.clamp_(0, gain[3] - 1).long(), gi.clamp_(0, gain[2] - 1).long()))
+ # tbox.append(torch.cat((gxy - gij, gwh), 1)) # box
+ # anch.append(anchors[a]) # anchors
+ # tcls.append(c) # class
+
+ # return tcls, tbox, indices, anch
+ def build_targets(self, p, targets):
+ # Build targets for compute_loss(), input targets(image,class,x,y,w,h)
+ na, nt = self.na, targets.shape[0] # number of anchors, targets
+ tcls, tbox, indices, anch = [], [], [], []
+ gain = torch.ones(7, device=self.device) # normalized to gridspace gain
+ ai = torch.arange(na, device=self.device).float().view(na, 1).repeat(1, nt) # same as .repeat_interleave(nt)
+ targets = torch.cat((targets.repeat(na, 1, 1), ai[..., None]), 2) # append anchor indices
+
+ g = 0.5 # bias
+ off = torch.tensor(
+ [
+ [0, 0],
+ [1, 0],
+ [0, 1],
+ [-1, 0],
+ [0, -1], # j,k,l,m
+ # [1, 1], [1, -1], [-1, 1], [-1, -1], # jk,jm,lk,lm
+ ],
+ device=self.device).float() * g # offsets
+
+ for i in range(self.nl):
+ anchors, shape = self.anchors[i], p[i].shape
+ gain[2:6] = torch.tensor(shape)[[3, 2, 3, 2]] # xyxy gain
+
+ # Match targets to anchors
+ t = targets * gain # shape(3,n,7)
+ if nt:
+ # Matches
+ r = t[..., 4:6] / anchors[:, None] # wh ratio
+ j = torch.max(r, 1 / r).max(2)[0] < self.hyp['anchor_t'] # compare
+ # j = wh_iou(anchors, t[:, 4:6]) > model.hyp['iou_t'] # iou(3,n)=wh_iou(anchors(3,2), gwh(n,2))
+ t = t[j] # filter
+
+ # Offsets
+ gxy = t[:, 2:4] # grid xy
+ gxi = gain[[2, 3]] - gxy # inverse
+ j, k = ((gxy % 1 < g) & (gxy > 1)).T
+ l, m = ((gxi % 1 < g) & (gxi > 1)).T
+ j = torch.stack((torch.ones_like(j), j, k, l, m))
+ t = t.repeat((5, 1, 1))[j]
+ offsets = (torch.zeros_like(gxy)[None] + off[:, None])[j]
+ else:
+ t = targets[0]
+ offsets = 0
+
+ # Define
+ bc, gxy, gwh, a = t.chunk(4, 1) # (image, class), grid xy, grid wh, anchors
+ a, (b, c) = a.long().view(-1), bc.long().T # anchors, image, class
+ gij = (gxy - offsets).long()
+ gi, gj = gij.T # grid indices
+
+ # Append
+ indices.append((b, a, gj.clamp_(0, shape[2] - 1), gi.clamp_(0, shape[3] - 1))) # image, anchor, grid
+ tbox.append(torch.cat((gxy - gij, gwh), 1)) # box
+ anch.append(anchors[a]) # anchors
+ tcls.append(c) # class
+
+ return tcls, tbox, indices, anch
diff --git a/vouchervision/component_detector/utils/metrics.py b/vouchervision/component_detector/utils/metrics.py
new file mode 100644
index 0000000000000000000000000000000000000000..ff43a3073062962a5f1882826538d9a30a21855f
--- /dev/null
+++ b/vouchervision/component_detector/utils/metrics.py
@@ -0,0 +1,348 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Model validation metrics
+"""
+
+import math
+import warnings
+from pathlib import Path
+
+import matplotlib.pyplot as plt
+import numpy as np
+import torch
+
+
+def fitness(x):
+ # Model fitness as a weighted combination of metrics
+ w = [0.0, 0.0, 0.1, 0.9] # weights for [P, R, mAP@0.5, mAP@0.5:0.95]
+ return (x[:, :4] * w).sum(1)
+
+
+def ap_per_class(tp, conf, pred_cls, target_cls, plot=False, save_dir='.', names=(), eps=1e-16):
+ """ Compute the average precision, given the recall and precision curves.
+ Source: https://github.com/rafaelpadilla/Object-Detection-Metrics.
+ # Arguments
+ tp: True positives (nparray, nx1 or nx10).
+ conf: Objectness value from 0-1 (nparray).
+ pred_cls: Predicted object classes (nparray).
+ target_cls: True object classes (nparray).
+ plot: Plot precision-recall curve at mAP@0.5
+ save_dir: Plot save directory
+ # Returns
+ The average precision as computed in py-faster-rcnn.
+ """
+
+ # Sort by objectness
+ i = np.argsort(-conf)
+ tp, conf, pred_cls = tp[i], conf[i], pred_cls[i]
+
+ # Find unique classes
+ unique_classes, nt = np.unique(target_cls, return_counts=True)
+ nc = unique_classes.shape[0] # number of classes, number of detections
+
+ # Create Precision-Recall curve and compute AP for each class
+ px, py = np.linspace(0, 1, 1000), [] # for plotting
+ ap, p, r = np.zeros((nc, tp.shape[1])), np.zeros((nc, 1000)), np.zeros((nc, 1000))
+ for ci, c in enumerate(unique_classes):
+ i = pred_cls == c
+ n_l = nt[ci] # number of labels
+ n_p = i.sum() # number of predictions
+
+ if n_p == 0 or n_l == 0:
+ continue
+ else:
+ # Accumulate FPs and TPs
+ fpc = (1 - tp[i]).cumsum(0)
+ tpc = tp[i].cumsum(0)
+
+ # Recall
+ recall = tpc / (n_l + eps) # recall curve
+ r[ci] = np.interp(-px, -conf[i], recall[:, 0], left=0) # negative x, xp because xp decreases
+
+ # Precision
+ precision = tpc / (tpc + fpc) # precision curve
+ p[ci] = np.interp(-px, -conf[i], precision[:, 0], left=1) # p at pr_score
+
+ # AP from recall-precision curve
+ for j in range(tp.shape[1]):
+ ap[ci, j], mpre, mrec = compute_ap(recall[:, j], precision[:, j])
+ if plot and j == 0:
+ py.append(np.interp(px, mrec, mpre)) # precision at mAP@0.5
+
+ # Compute F1 (harmonic mean of precision and recall)
+ f1 = 2 * p * r / (p + r + eps)
+ names = [v for k, v in names.items() if k in unique_classes] # list: only classes that have data
+ names = {i: v for i, v in enumerate(names)} # to dict
+ if plot:
+ plot_pr_curve(px, py, ap, Path(save_dir) / 'PR_curve.png', names)
+ plot_mc_curve(px, f1, Path(save_dir) / 'F1_curve.png', names, ylabel='F1')
+ plot_mc_curve(px, p, Path(save_dir) / 'P_curve.png', names, ylabel='Precision')
+ plot_mc_curve(px, r, Path(save_dir) / 'R_curve.png', names, ylabel='Recall')
+
+ i = f1.mean(0).argmax() # max F1 index
+ p, r, f1 = p[:, i], r[:, i], f1[:, i]
+ tp = (r * nt).round() # true positives
+ fp = (tp / (p + eps) - tp).round() # false positives
+ return tp, fp, p, r, f1, ap, unique_classes.astype('int32')
+
+
+def compute_ap(recall, precision):
+ """ Compute the average precision, given the recall and precision curves
+ # Arguments
+ recall: The recall curve (list)
+ precision: The precision curve (list)
+ # Returns
+ Average precision, precision curve, recall curve
+ """
+
+ # Append sentinel values to beginning and end
+ mrec = np.concatenate(([0.0], recall, [1.0]))
+ mpre = np.concatenate(([1.0], precision, [0.0]))
+
+ # Compute the precision envelope
+ mpre = np.flip(np.maximum.accumulate(np.flip(mpre)))
+
+ # Integrate area under curve
+ method = 'interp' # methods: 'continuous', 'interp'
+ if method == 'interp':
+ x = np.linspace(0, 1, 101) # 101-point interp (COCO)
+ ap = np.trapz(np.interp(x, mrec, mpre), x) # integrate
+ else: # 'continuous'
+ i = np.where(mrec[1:] != mrec[:-1])[0] # points where x axis (recall) changes
+ ap = np.sum((mrec[i + 1] - mrec[i]) * mpre[i + 1]) # area under curve
+
+ return ap, mpre, mrec
+
+
+class ConfusionMatrix:
+ # Updated version of https://github.com/kaanakan/object_detection_confusion_matrix
+ def __init__(self, nc, conf=0.25, iou_thres=0.45):
+ self.matrix = np.zeros((nc + 1, nc + 1))
+ self.nc = nc # number of classes
+ self.conf = conf
+ self.iou_thres = iou_thres
+
+ def process_batch(self, detections, labels):
+ """
+ Return intersection-over-union (Jaccard index) of boxes.
+ Both sets of boxes are expected to be in (x1, y1, x2, y2) format.
+ Arguments:
+ detections (Array[N, 6]), x1, y1, x2, y2, conf, class
+ labels (Array[M, 5]), class, x1, y1, x2, y2
+ Returns:
+ None, updates confusion matrix accordingly
+ """
+ detections = detections[detections[:, 4] > self.conf]
+ gt_classes = labels[:, 0].int()
+ detection_classes = detections[:, 5].int()
+ iou = box_iou(labels[:, 1:], detections[:, :4])
+
+ x = torch.where(iou > self.iou_thres)
+ if x[0].shape[0]:
+ matches = torch.cat((torch.stack(x, 1), iou[x[0], x[1]][:, None]), 1).cpu().numpy()
+ if x[0].shape[0] > 1:
+ matches = matches[matches[:, 2].argsort()[::-1]]
+ matches = matches[np.unique(matches[:, 1], return_index=True)[1]]
+ matches = matches[matches[:, 2].argsort()[::-1]]
+ matches = matches[np.unique(matches[:, 0], return_index=True)[1]]
+ else:
+ matches = np.zeros((0, 3))
+
+ n = matches.shape[0] > 0
+ m0, m1, _ = matches.transpose().astype(np.int16)
+ for i, gc in enumerate(gt_classes):
+ j = m0 == i
+ if n and sum(j) == 1:
+ self.matrix[detection_classes[m1[j]], gc] += 1 # correct
+ else:
+ self.matrix[self.nc, gc] += 1 # background FP
+
+ if n:
+ for i, dc in enumerate(detection_classes):
+ if not any(m1 == i):
+ self.matrix[dc, self.nc] += 1 # background FN
+
+ def matrix(self):
+ return self.matrix
+
+ def tp_fp(self):
+ tp = self.matrix.diagonal() # true positives
+ fp = self.matrix.sum(1) - tp # false positives
+ # fn = self.matrix.sum(0) - tp # false negatives (missed detections)
+ return tp[:-1], fp[:-1] # remove background class
+
+ def plot(self, normalize=True, save_dir='', names=()):
+ try:
+ import seaborn as sn
+
+ array = self.matrix / ((self.matrix.sum(0).reshape(1, -1) + 1E-9) if normalize else 1) # normalize columns
+ array[array < 0.005] = np.nan # don't annotate (would appear as 0.00)
+
+ fig = plt.figure(figsize=(12, 9), tight_layout=True)
+ nc, nn = self.nc, len(names) # number of classes, names
+ sn.set(font_scale=1.0 if nc < 50 else 0.8) # for label size
+ labels = (0 < nn < 99) and (nn == nc) # apply names to ticklabels
+ with warnings.catch_warnings():
+ warnings.simplefilter('ignore') # suppress empty matrix RuntimeWarning: All-NaN slice encountered
+ sn.heatmap(array,
+ annot=nc < 30,
+ annot_kws={
+ "size": 8},
+ cmap='Blues',
+ fmt='.2f',
+ square=True,
+ vmin=0.0,
+ xticklabels=names + ['background FP'] if labels else "auto",
+ yticklabels=names + ['background FN'] if labels else "auto").set_facecolor((1, 1, 1))
+ fig.axes[0].set_xlabel('True')
+ fig.axes[0].set_ylabel('Predicted')
+ fig.savefig(Path(save_dir) / 'confusion_matrix.png', dpi=250)
+ plt.close()
+ except Exception as e:
+ print(f'WARNING: ConfusionMatrix plot failure: {e}')
+
+ def print(self):
+ for i in range(self.nc + 1):
+ print(' '.join(map(str, self.matrix[i])))
+
+
+def bbox_iou(box1, box2, xywh=True, GIoU=False, DIoU=False, CIoU=False, eps=1e-7):
+ # Returns Intersection over Union (IoU) of box1(1,4) to box2(n,4)
+
+ # Get the coordinates of bounding boxes
+ if xywh: # transform from xywh to xyxy
+ (x1, y1, w1, h1), (x2, y2, w2, h2) = box1.chunk(4, 1), box2.chunk(4, 1)
+ w1_, h1_, w2_, h2_ = w1 / 2, h1 / 2, w2 / 2, h2 / 2
+ b1_x1, b1_x2, b1_y1, b1_y2 = x1 - w1_, x1 + w1_, y1 - h1_, y1 + h1_
+ b2_x1, b2_x2, b2_y1, b2_y2 = x2 - w2_, x2 + w2_, y2 - h2_, y2 + h2_
+ else: # x1, y1, x2, y2 = box1
+ b1_x1, b1_y1, b1_x2, b1_y2 = box1.chunk(4, 1)
+ b2_x1, b2_y1, b2_x2, b2_y2 = box2.chunk(4, 1)
+ w1, h1 = b1_x2 - b1_x1, b1_y2 - b1_y1 + eps
+ w2, h2 = b2_x2 - b2_x1, b2_y2 - b2_y1 + eps
+
+ # Intersection area
+ inter = (torch.min(b1_x2, b2_x2) - torch.max(b1_x1, b2_x1)).clamp(0) * \
+ (torch.min(b1_y2, b2_y2) - torch.max(b1_y1, b2_y1)).clamp(0)
+
+ # Union Area
+ union = w1 * h1 + w2 * h2 - inter + eps
+
+ # IoU
+ iou = inter / union
+ if CIoU or DIoU or GIoU:
+ cw = torch.max(b1_x2, b2_x2) - torch.min(b1_x1, b2_x1) # convex (smallest enclosing box) width
+ ch = torch.max(b1_y2, b2_y2) - torch.min(b1_y1, b2_y1) # convex height
+ if CIoU or DIoU: # Distance or Complete IoU https://arxiv.org/abs/1911.08287v1
+ c2 = cw ** 2 + ch ** 2 + eps # convex diagonal squared
+ rho2 = ((b2_x1 + b2_x2 - b1_x1 - b1_x2) ** 2 + (b2_y1 + b2_y2 - b1_y1 - b1_y2) ** 2) / 4 # center dist ** 2
+ if CIoU: # https://github.com/Zzh-tju/DIoU-SSD-pytorch/blob/master/utils/box/box_utils.py#L47
+ v = (4 / math.pi ** 2) * torch.pow(torch.atan(w2 / h2) - torch.atan(w1 / h1), 2)
+ with torch.no_grad():
+ alpha = v / (v - iou + (1 + eps))
+ return iou - (rho2 / c2 + v * alpha) # CIoU
+ return iou - rho2 / c2 # DIoU
+ c_area = cw * ch + eps # convex area
+ return iou - (c_area - union) / c_area # GIoU https://arxiv.org/pdf/1902.09630.pdf
+ return iou # IoU
+
+
+def box_area(box):
+ # box = xyxy(4,n)
+ return (box[2] - box[0]) * (box[3] - box[1])
+
+
+def box_iou(box1, box2):
+ # https://github.com/pytorch/vision/blob/master/torchvision/ops/boxes.py
+ """
+ Return intersection-over-union (Jaccard index) of boxes.
+ Both sets of boxes are expected to be in (x1, y1, x2, y2) format.
+ Arguments:
+ box1 (Tensor[N, 4])
+ box2 (Tensor[M, 4])
+ Returns:
+ iou (Tensor[N, M]): the NxM matrix containing the pairwise
+ IoU values for every element in boxes1 and boxes2
+ """
+
+ # inter(N,M) = (rb(N,M,2) - lt(N,M,2)).clamp(0).prod(2)
+ (a1, a2), (b1, b2) = box1[:, None].chunk(2, 2), box2.chunk(2, 1)
+ inter = (torch.min(a2, b2) - torch.max(a1, b1)).clamp(0).prod(2)
+
+ # IoU = inter / (area1 + area2 - inter)
+ return inter / (box_area(box1.T)[:, None] + box_area(box2.T) - inter)
+
+
+def bbox_ioa(box1, box2, eps=1E-7):
+ """ Returns the intersection over box2 area given box1, box2. Boxes are x1y1x2y2
+ box1: np.array of shape(4)
+ box2: np.array of shape(nx4)
+ returns: np.array of shape(n)
+ """
+
+ # Get the coordinates of bounding boxes
+ b1_x1, b1_y1, b1_x2, b1_y2 = box1
+ b2_x1, b2_y1, b2_x2, b2_y2 = box2.T
+
+ # Intersection area
+ inter_area = (np.minimum(b1_x2, b2_x2) - np.maximum(b1_x1, b2_x1)).clip(0) * \
+ (np.minimum(b1_y2, b2_y2) - np.maximum(b1_y1, b2_y1)).clip(0)
+
+ # box2 area
+ box2_area = (b2_x2 - b2_x1) * (b2_y2 - b2_y1) + eps
+
+ # Intersection over box2 area
+ return inter_area / box2_area
+
+
+def wh_iou(wh1, wh2):
+ # Returns the nxm IoU matrix. wh1 is nx2, wh2 is mx2
+ wh1 = wh1[:, None] # [N,1,2]
+ wh2 = wh2[None] # [1,M,2]
+ inter = torch.min(wh1, wh2).prod(2) # [N,M]
+ return inter / (wh1.prod(2) + wh2.prod(2) - inter) # iou = inter / (area1 + area2 - inter)
+
+
+# Plots ----------------------------------------------------------------------------------------------------------------
+
+
+def plot_pr_curve(px, py, ap, save_dir='pr_curve.png', names=()):
+ # Precision-recall curve
+ fig, ax = plt.subplots(1, 1, figsize=(9, 6), tight_layout=True)
+ py = np.stack(py, axis=1)
+
+ if 0 < len(names) < 21: # display per-class legend if < 21 classes
+ for i, y in enumerate(py.T):
+ ax.plot(px, y, linewidth=1, label=f'{names[i]} {ap[i, 0]:.3f}') # plot(recall, precision)
+ else:
+ ax.plot(px, py, linewidth=1, color='grey') # plot(recall, precision)
+
+ ax.plot(px, py.mean(1), linewidth=3, color='blue', label='all classes %.3f mAP@0.5' % ap[:, 0].mean())
+ ax.set_xlabel('Recall')
+ ax.set_ylabel('Precision')
+ ax.set_xlim(0, 1)
+ ax.set_ylim(0, 1)
+ plt.legend(bbox_to_anchor=(1.04, 1), loc="upper left")
+ fig.savefig(Path(save_dir), dpi=250)
+ plt.close()
+
+
+def plot_mc_curve(px, py, save_dir='mc_curve.png', names=(), xlabel='Confidence', ylabel='Metric'):
+ # Metric-confidence curve
+ fig, ax = plt.subplots(1, 1, figsize=(9, 6), tight_layout=True)
+
+ if 0 < len(names) < 21: # display per-class legend if < 21 classes
+ for i, y in enumerate(py):
+ ax.plot(px, y, linewidth=1, label=f'{names[i]}') # plot(confidence, metric)
+ else:
+ ax.plot(px, py.T, linewidth=1, color='grey') # plot(confidence, metric)
+
+ y = py.mean(0)
+ ax.plot(px, y, linewidth=3, color='blue', label=f'all classes {y.max():.2f} at {px[y.argmax()]:.3f}')
+ ax.set_xlabel(xlabel)
+ ax.set_ylabel(ylabel)
+ ax.set_xlim(0, 1)
+ ax.set_ylim(0, 1)
+ plt.legend(bbox_to_anchor=(1.04, 1), loc="upper left")
+ fig.savefig(Path(save_dir), dpi=250)
+ plt.close()
diff --git a/vouchervision/component_detector/utils/plots.py b/vouchervision/component_detector/utils/plots.py
new file mode 100644
index 0000000000000000000000000000000000000000..7b781d9a2db7052ad43d52b4d71c07f88a75da67
--- /dev/null
+++ b/vouchervision/component_detector/utils/plots.py
@@ -0,0 +1,560 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Plotting utils
+"""
+
+import os, math, sys, cv2
+from copy import copy
+from pathlib import Path
+from urllib.error import URLError
+import matplotlib
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+import seaborn as sn
+import torch
+from PIL import Image, ImageDraw, ImageFont
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[0] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+from general import (CONFIG_DIR, FONT, LOGGER, Timeout, check_font, check_requirements, clip_coords, increment_path, is_ascii, try_except, xywh2xyxy, xyxy2xywh)
+from metrics import fitness
+
+# Settings
+RANK = int(os.getenv('RANK', -1))
+matplotlib.rc('font', **{'size': 11})
+matplotlib.use('Agg') # for writing to files only
+
+
+class Colors:
+ # Ultralytics color palette https://ultralytics.com/
+ def __init__(self):
+ # hex = matplotlib.colors.TABLEAU_COLORS.values()
+ # if PROFILE == "PREP":
+ # hex = ('FF0046', '008941', 'F2FF00', '0000FF', '00FBFF',
+ # 'A30059', 'FFCDDC', 'FFAC28', '8C8C8C', '00D4BB',
+ # '2C99A8', '00C2FF', '344593', '6473FF', '0018EC', '8438FF', '520085', 'CB38FF', 'FF95C8', 'FF37C7')
+ # elif PROFILE == "PLANT":
+ hex = ('00ff37', '69fffc', 'ffcb00', 'fcff00', '000000',
+ 'ff34ff', '9a00ff', 'ff0009', 'ceffc4', 'ff8600',
+ '901616', '00C2FF', '344593', '6473FF', '0018EC', '8438FF', '520085', 'CB38FF', 'FF95C8', 'FF37C7')
+ # else:
+ # hex = ('FF3838', 'FF9D97', 'FF701F', 'FFB21D', 'CFD231', '48F90A', '92CC17', '3DDB86', '1A9334', '00D4BB',
+ # '2C99A8', '00C2FF', '344593', '6473FF', '0018EC', '8438FF', '520085', 'CB38FF', 'FF95C8', 'FF37C7')
+ self.palette = [self.hex2rgb('#' + c) for c in hex]
+ self.n = len(self.palette)
+
+ def __call__(self, i, bgr=False):
+ c = self.palette[int(i) % self.n]
+ return (c[2], c[1], c[0]) if bgr else c
+
+ @staticmethod
+ def hex2rgb(h): # rgb order (PIL)
+ return tuple(int(h[1 + i:1 + i + 2], 16) for i in (0, 2, 4))
+
+
+colors = Colors() # create instance for 'from utils.plots import colors'
+
+
+def check_pil_font(font=FONT, size=10):
+ # Return a PIL TrueType Font, downloading to CONFIG_DIR if necessary
+ font = Path(font)
+ font = font if font.exists() else (CONFIG_DIR / font.name)
+ try:
+ return ImageFont.truetype(str(font) if font.exists() else font.name, size)
+ except Exception: # download if missing
+ try:
+ check_font(font)
+ return ImageFont.truetype(str(font), size)
+ except TypeError:
+ check_requirements('Pillow>=8.4.0') # known issue https://github.com/ultralytics/yolov5/issues/5374
+ except URLError: # not online
+ return ImageFont.load_default()
+
+class AnnotatorLandmark:
+ # YOLOv5 Annotator for train/val mosaics and jpgs and detect/hub inference annotations
+ def __init__(self, im, line_width=None, font_size=None, font='Arial.ttf', pil=False, example='abc'):
+ assert im.data.contiguous, 'Image not contiguous. Apply np.ascontiguousarray(im) to Annotator() input images.'
+ non_ascii = not is_ascii(example) # non-latin labels, i.e. asian, arabic, cyrillic
+ self.pil = pil or non_ascii
+ if self.pil: # use PIL
+ self.im = im if isinstance(im, Image.Image) else Image.fromarray(im)
+ self.draw = ImageDraw.Draw(self.im)
+ self.font = check_pil_font(font='Arial.Unicode.ttf' if non_ascii else font,
+ size=font_size or max(round(sum(self.im.size) / 2 * 0.035), 12))
+ else: # use cv2
+ self.im = im
+ self.lw = line_width #or max(round(sum(im.shape) / 2 * 0.003), 2) # line width
+
+ def box_label(self, box, label='', color=(128, 128, 128), txt_color=(255, 255, 255)):
+ # Add one xyxy box to image with label
+ if self.pil or not is_ascii(label):
+ self.draw.rectangle(box, width=self.lw, outline=color) # box
+ # if label:
+ # w, h = self.font.getsize(label) # text width, height
+ # outside = box[1] - h >= 0 # label fits outside box
+ # self.draw.rectangle(
+ # (box[0], box[1] - h if outside else box[1], box[0] + w + 1,
+ # box[1] + 1 if outside else box[1] + h + 1),
+ # fill=color,
+ # )
+ # # self.draw.text((box[0], box[1]), label, fill=txt_color, font=self.font, anchor='ls') # for PIL>8.0
+ # self.draw.text((box[0], box[1] - h if outside else box[1]), label, fill=txt_color, font=self.font)
+ else: # cv2
+ p1, p2 = (int(box[0]), int(box[1])), (int(box[2]), int(box[3]))
+ cv2.circle(self.im, (int((p1[0]+p2[0])/2), int((p1[1]+p2[1])/2)), radius=self.lw, color=color, thickness=-1)
+ # cv2.rectangle(self.im, p1, p2, color, thickness=self.lw, lineType=cv2.LINE_AA)
+ # if label:
+ # tf = 6 #max(self.lw - 1, 1) # font thickness
+ # w, h = cv2.getTextSize(label, 0, fontScale=self.lw / 3, thickness=tf)[0] # text width, height
+ # outside = p1[1] - h - 3 >= 0 # label fits outside box
+ # p2 = p1[0] + w, p1[1] - h - 3 if outside else p1[1] + h + 3
+ # cv2.rectangle(self.im, p1, p2, color, -1, cv2.LINE_AA) # filled
+ # cv2.putText(self.im,
+ # label, (p1[0], p1[1] - 2 if outside else p1[1] + h + 2),
+ # 0,
+ # self.lw,
+ # txt_color,
+ # thickness=tf,
+ # lineType=cv2.LINE_AA)
+
+ def rectangle(self, xy, fill=None, outline=None, width=1):
+ # Add rectangle to image (PIL-only)
+ self.draw.rectangle(xy, fill, outline, width)
+
+ def text(self, xy, text, txt_color=(255, 255, 255)):
+ # Add text to image (PIL-only)
+ w, h = self.font.getsize(text) # text width, height
+ self.draw.text((xy[0], xy[1] - h + 1), text, fill=txt_color, font=self.font)
+
+ def result(self):
+ # Return annotated image as array
+ return np.asarray(self.im)
+
+class Annotator:
+ # YOLOv5 Annotator for train/val mosaics and jpgs and detect/hub inference annotations
+ def __init__(self, im, line_width=None, font_size=None, font='Arial.ttf', pil=False, example='abc'):
+ assert im.data.contiguous, 'Image not contiguous. Apply np.ascontiguousarray(im) to Annotator() input images.'
+ non_ascii = not is_ascii(example) # non-latin labels, i.e. asian, arabic, cyrillic
+ self.pil = pil or non_ascii
+ if self.pil: # use PIL
+ self.im = im if isinstance(im, Image.Image) else Image.fromarray(im)
+ self.draw = ImageDraw.Draw(self.im)
+ self.font = check_pil_font(font='Arial.Unicode.ttf' if non_ascii else font,
+ size=font_size or max(round(sum(self.im.size) / 2 * 0.035), 12))
+ else: # use cv2
+ self.im = im
+ self.lw = line_width or max(round(sum(im.shape) / 2 * 0.003), 2) # line width
+
+ def box_label(self, box, label='', color=(128, 128, 128), txt_color=(255, 255, 255)):
+ # Add one xyxy box to image with label
+ if self.pil or not is_ascii(label):
+ self.draw.rectangle(box, width=self.lw, outline=color) # box
+ if label:
+ # w, h = self.font.getsize(label) # text width, height
+ w = 38
+ h = 38
+ outside = box[1] - h >= 0 # label fits outside box
+ self.draw.rectangle(
+ (box[0], box[1] - h if outside else box[1], box[0] + w + 1,
+ box[1] + 1 if outside else box[1] + h + 1),
+ fill=color,
+ )
+ # self.draw.text((box[0], box[1]), label, fill=txt_color, font=self.font, anchor='ls') # for PIL>8.0
+ self.draw.text((box[0], box[1] - h if outside else box[1]), label, fill=txt_color, font=self.font)
+ else: # cv2
+ p1, p2 = (int(box[0]), int(box[1])), (int(box[2]), int(box[3]))
+ cv2.rectangle(self.im, p1, p2, color, thickness=self.lw, lineType=cv2.LINE_AA)
+ if label:
+ tf = max(self.lw - 1, 1) # font thickness
+ w, h = cv2.getTextSize(label, 0, fontScale=self.lw / 3, thickness=tf)[0] # text width, height
+ outside = p1[1] - h - 3 >= 0 # label fits outside box
+ p2 = p1[0] + w, p1[1] - h - 3 if outside else p1[1] + h + 3
+ cv2.rectangle(self.im, p1, p2, color, -1, cv2.LINE_AA) # filled
+ cv2.putText(self.im,
+ label, (p1[0], p1[1] - 2 if outside else p1[1] + h + 2),
+ 0,
+ self.lw / 3,
+ txt_color,
+ thickness=tf,
+ lineType=cv2.LINE_AA)
+
+ def rectangle(self, xy, fill=None, outline=None, width=1):
+ # Add rectangle to image (PIL-only)
+ self.draw.rectangle(xy, fill, outline, width)
+
+ def text(self, xy, text, txt_color=(255, 255, 255)):
+ # Add text to image (PIL-only)
+ # w, h = self.font.getsize(text) # text width, height
+ w = 38
+ h = 38
+ self.draw.text((xy[0], xy[1] - h + 1), text, fill=txt_color, font=self.font)
+
+ def result(self):
+ # Return annotated image as array
+ return np.asarray(self.im)
+
+
+def feature_visualization(x, module_type, stage, n=32, save_dir=Path('runs/detect/exp')):
+ """
+ x: Features to be visualized
+ module_type: Module type
+ stage: Module stage within model
+ n: Maximum number of feature maps to plot
+ save_dir: Directory to save results
+ """
+ if 'Detect' not in module_type:
+ batch, channels, height, width = x.shape # batch, channels, height, width
+ if height > 1 and width > 1:
+ f = save_dir / f"stage{stage}_{module_type.split('.')[-1]}_features.png" # filename
+
+ blocks = torch.chunk(x[0].cpu(), channels, dim=0) # select batch index 0, block by channels
+ n = min(n, channels) # number of plots
+ fig, ax = plt.subplots(math.ceil(n / 8), 8, tight_layout=True) # 8 rows x n/8 cols
+ ax = ax.ravel()
+ plt.subplots_adjust(wspace=0.05, hspace=0.05)
+ for i in range(n):
+ ax[i].imshow(blocks[i].squeeze()) # cmap='gray'
+ ax[i].axis('off')
+
+ LOGGER.info(f'Saving {f}... ({n}/{channels})')
+ plt.savefig(f, dpi=300, bbox_inches='tight')
+ plt.close()
+ np.save(str(f.with_suffix('.npy')), x[0].cpu().numpy()) # npy save
+
+
+def hist2d(x, y, n=100):
+ # 2d histogram used in labels.png and evolve.png
+ xedges, yedges = np.linspace(x.min(), x.max(), n), np.linspace(y.min(), y.max(), n)
+ hist, xedges, yedges = np.histogram2d(x, y, (xedges, yedges))
+ xidx = np.clip(np.digitize(x, xedges) - 1, 0, hist.shape[0] - 1)
+ yidx = np.clip(np.digitize(y, yedges) - 1, 0, hist.shape[1] - 1)
+ return np.log(hist[xidx, yidx])
+
+
+def butter_lowpass_filtfilt(data, cutoff=1500, fs=50000, order=5):
+ from scipy.signal import butter, filtfilt
+
+ # https://stackoverflow.com/questions/28536191/how-to-filter-smooth-with-scipy-numpy
+ def butter_lowpass(cutoff, fs, order):
+ nyq = 0.5 * fs
+ normal_cutoff = cutoff / nyq
+ return butter(order, normal_cutoff, btype='low', analog=False)
+
+ b, a = butter_lowpass(cutoff, fs, order=order)
+ return filtfilt(b, a, data) # forward-backward filter
+
+
+def output_to_target(output):
+ # Convert model output to target format [batch_id, class_id, x, y, w, h, conf]
+ targets = []
+ for i, o in enumerate(output):
+ for *box, conf, cls in o.cpu().numpy():
+ targets.append([i, cls, *list(*xyxy2xywh(np.array(box)[None])), conf])
+ return np.array(targets)
+
+
+def plot_images(images, targets, paths=None, fname='images.jpg', names=None, max_size=1920, max_subplots=16):
+ # Plot image grid with labels
+ if isinstance(images, torch.Tensor):
+ images = images.cpu().float().numpy()
+ if isinstance(targets, torch.Tensor):
+ targets = targets.cpu().numpy()
+ if np.max(images[0]) <= 1:
+ images *= 255 # de-normalise (optional)
+ bs, _, h, w = images.shape # batch size, _, height, width
+ bs = min(bs, max_subplots) # limit plot images
+ ns = np.ceil(bs ** 0.5) # number of subplots (square)
+
+ # Build Image
+ mosaic = np.full((int(ns * h), int(ns * w), 3), 255, dtype=np.uint8) # init
+ for i, im in enumerate(images):
+ if i == max_subplots: # if last batch has fewer images than we expect
+ break
+ x, y = int(w * (i // ns)), int(h * (i % ns)) # block origin
+ im = im.transpose(1, 2, 0)
+ mosaic[y:y + h, x:x + w, :] = im
+
+ # Resize (optional)
+ scale = max_size / ns / max(h, w)
+ if scale < 1:
+ h = math.ceil(scale * h)
+ w = math.ceil(scale * w)
+ mosaic = cv2.resize(mosaic, tuple(int(x * ns) for x in (w, h)))
+
+ # Annotate
+ fs = int((h + w) * ns * 0.01) # font size
+ annotator = Annotator(mosaic, line_width=round(fs / 10), font_size=fs, pil=True, example=names)
+ for i in range(i + 1):
+ x, y = int(w * (i // ns)), int(h * (i % ns)) # block origin
+ annotator.rectangle([x, y, x + w, y + h], None, (255, 255, 255), width=2) # borders
+ if paths:
+ annotator.text((x + 5, y + 5 + h), text=Path(paths[i]).name[:40], txt_color=(220, 220, 220)) # filenames
+ if len(targets) > 0:
+ ti = targets[targets[:, 0] == i] # image targets
+ boxes = xywh2xyxy(ti[:, 2:6]).T
+ classes = ti[:, 1].astype('int')
+ labels = ti.shape[1] == 6 # labels if no conf column
+ conf = None if labels else ti[:, 6] # check for confidence presence (label vs pred)
+
+ if boxes.shape[1]:
+ if boxes.max() <= 1.01: # if normalized with tolerance 0.01
+ boxes[[0, 2]] *= w # scale to pixels
+ boxes[[1, 3]] *= h
+ elif scale < 1: # absolute coords need scale if image scales
+ boxes *= scale
+ boxes[[0, 2]] += x
+ boxes[[1, 3]] += y
+ for j, box in enumerate(boxes.T.tolist()):
+ cls = classes[j]
+ color = colors(cls)
+ cls = names[cls] if names else cls
+ if labels or conf[j] > 0.25: # 0.25 conf thresh
+ label = f'{cls}' if labels else f'{cls} {conf[j]:.1f}'
+ annotator.box_label(box, label, color=color)
+ annotator.im.save(fname) # save
+
+
+def plot_lr_scheduler(optimizer, scheduler, epochs=300, save_dir=''):
+ # Plot LR simulating training for full epochs
+ optimizer, scheduler = copy(optimizer), copy(scheduler) # do not modify originals
+ y = []
+ for _ in range(epochs):
+ scheduler.step()
+ y.append(optimizer.param_groups[0]['lr'])
+ plt.plot(y, '.-', label='LR')
+ plt.xlabel('epoch')
+ plt.ylabel('LR')
+ plt.grid()
+ plt.xlim(0, epochs)
+ plt.ylim(0)
+ plt.savefig(Path(save_dir) / 'LR.png', dpi=200)
+ plt.close()
+
+
+def plot_val_txt(): # from utils.plots import *; plot_val()
+ # Plot val.txt histograms
+ x = np.loadtxt('val.txt', dtype=np.float32)
+ box = xyxy2xywh(x[:, :4])
+ cx, cy = box[:, 0], box[:, 1]
+
+ fig, ax = plt.subplots(1, 1, figsize=(6, 6), tight_layout=True)
+ ax.hist2d(cx, cy, bins=600, cmax=10, cmin=0)
+ ax.set_aspect('equal')
+ plt.savefig('hist2d.png', dpi=300)
+
+ fig, ax = plt.subplots(1, 2, figsize=(12, 6), tight_layout=True)
+ ax[0].hist(cx, bins=600)
+ ax[1].hist(cy, bins=600)
+ plt.savefig('hist1d.png', dpi=200)
+
+
+def plot_targets_txt(): # from utils.plots import *; plot_targets_txt()
+ # Plot targets.txt histograms
+ x = np.loadtxt('targets.txt', dtype=np.float32).T
+ s = ['x targets', 'y targets', 'width targets', 'height targets']
+ fig, ax = plt.subplots(2, 2, figsize=(8, 8), tight_layout=True)
+ ax = ax.ravel()
+ for i in range(4):
+ ax[i].hist(x[i], bins=100, label=f'{x[i].mean():.3g} +/- {x[i].std():.3g}')
+ ax[i].legend()
+ ax[i].set_title(s[i])
+ plt.savefig('targets.jpg', dpi=200)
+
+
+def plot_val_study(file='', dir='', x=None): # from utils.plots import *; plot_val_study()
+ # Plot file=study.txt generated by val.py (or plot all study*.txt in dir)
+ save_dir = Path(file).parent if file else Path(dir)
+ plot2 = False # plot additional results
+ if plot2:
+ ax = plt.subplots(2, 4, figsize=(10, 6), tight_layout=True)[1].ravel()
+
+ fig2, ax2 = plt.subplots(1, 1, figsize=(8, 4), tight_layout=True)
+ # for f in [save_dir / f'study_coco_{x}.txt' for x in ['yolov5n6', 'yolov5s6', 'yolov5m6', 'yolov5l6', 'yolov5x6']]:
+ for f in sorted(save_dir.glob('study*.txt')):
+ y = np.loadtxt(f, dtype=np.float32, usecols=[0, 1, 2, 3, 7, 8, 9], ndmin=2).T
+ x = np.arange(y.shape[1]) if x is None else np.array(x)
+ if plot2:
+ s = ['P', 'R', 'mAP@.5', 'mAP@.5:.95', 't_preprocess (ms/img)', 't_inference (ms/img)', 't_NMS (ms/img)']
+ for i in range(7):
+ ax[i].plot(x, y[i], '.-', linewidth=2, markersize=8)
+ ax[i].set_title(s[i])
+
+ j = y[3].argmax() + 1
+ ax2.plot(y[5, 1:j],
+ y[3, 1:j] * 1E2,
+ '.-',
+ linewidth=2,
+ markersize=8,
+ label=f.stem.replace('study_coco_', '').replace('yolo', 'YOLO'))
+
+ ax2.plot(1E3 / np.array([209, 140, 97, 58, 35, 18]), [34.6, 40.5, 43.0, 47.5, 49.7, 51.5],
+ 'k.-',
+ linewidth=2,
+ markersize=8,
+ alpha=.25,
+ label='EfficientDet')
+
+ ax2.grid(alpha=0.2)
+ ax2.set_yticks(np.arange(20, 60, 5))
+ ax2.set_xlim(0, 57)
+ ax2.set_ylim(25, 55)
+ ax2.set_xlabel('GPU Speed (ms/img)')
+ ax2.set_ylabel('COCO AP val')
+ ax2.legend(loc='lower right')
+ f = save_dir / 'study.png'
+ print(f'Saving {f}...')
+ plt.savefig(f, dpi=300)
+
+
+@try_except # known issue https://github.com/ultralytics/yolov5/issues/5395
+@Timeout(30) # known issue https://github.com/ultralytics/yolov5/issues/5611
+def plot_labels(labels, names=(), save_dir=Path('')):
+ # plot dataset labels
+ LOGGER.info(f"Plotting labels to {save_dir / 'labels.jpg'}... ")
+ c, b = labels[:, 0], labels[:, 1:].transpose() # classes, boxes
+ nc = int(c.max() + 1) # number of classes
+ x = pd.DataFrame(b.transpose(), columns=['x', 'y', 'width', 'height'])
+
+ # seaborn correlogram
+ sn.pairplot(x, corner=True, diag_kind='auto', kind='hist', diag_kws=dict(bins=50), plot_kws=dict(pmax=0.9))
+ plt.savefig(save_dir / 'labels_correlogram.jpg', dpi=200)
+ plt.close()
+
+ # matplotlib labels
+ matplotlib.use('svg') # faster
+ ax = plt.subplots(2, 2, figsize=(8, 8), tight_layout=True)[1].ravel()
+ y = ax[0].hist(c, bins=np.linspace(0, nc, nc + 1) - 0.5, rwidth=0.8)
+ try: # color histogram bars by class
+ [y[2].patches[i].set_color([x / 255 for x in colors(i)]) for i in range(nc)] # known issue #3195
+ except Exception:
+ pass
+ ax[0].set_ylabel('instances')
+ if 0 < len(names) < 30:
+ ax[0].set_xticks(range(len(names)))
+ ax[0].set_xticklabels(names, rotation=90, fontsize=10)
+ else:
+ ax[0].set_xlabel('classes')
+ sn.histplot(x, x='x', y='y', ax=ax[2], bins=50, pmax=0.9)
+ sn.histplot(x, x='width', y='height', ax=ax[3], bins=50, pmax=0.9)
+
+ # rectangles
+ labels[:, 1:3] = 0.5 # center
+ labels[:, 1:] = xywh2xyxy(labels[:, 1:]) * 2000
+ img = Image.fromarray(np.ones((2000, 2000, 3), dtype=np.uint8) * 255)
+ for cls, *box in labels[:1000]:
+ ImageDraw.Draw(img).rectangle(box, width=1, outline=colors(cls)) # plot
+ ax[1].imshow(img)
+ ax[1].axis('off')
+
+ for a in [0, 1, 2, 3]:
+ for s in ['top', 'right', 'left', 'bottom']:
+ ax[a].spines[s].set_visible(False)
+
+ plt.savefig(save_dir / 'labels.jpg', dpi=200)
+ matplotlib.use('Agg')
+ plt.close()
+
+
+def plot_evolve(evolve_csv='path/to/evolve.csv'): # from utils.plots import *; plot_evolve()
+ # Plot evolve.csv hyp evolution results
+ evolve_csv = Path(evolve_csv)
+ data = pd.read_csv(evolve_csv)
+ keys = [x.strip() for x in data.columns]
+ x = data.values
+ f = fitness(x)
+ j = np.argmax(f) # max fitness index
+ plt.figure(figsize=(10, 12), tight_layout=True)
+ matplotlib.rc('font', **{'size': 8})
+ print(f'Best results from row {j} of {evolve_csv}:')
+ for i, k in enumerate(keys[7:]):
+ v = x[:, 7 + i]
+ mu = v[j] # best single result
+ plt.subplot(6, 5, i + 1)
+ plt.scatter(v, f, c=hist2d(v, f, 20), cmap='viridis', alpha=.8, edgecolors='none')
+ plt.plot(mu, f.max(), 'k+', markersize=15)
+ plt.title(f'{k} = {mu:.3g}', fontdict={'size': 9}) # limit to 40 characters
+ if i % 5 != 0:
+ plt.yticks([])
+ print(f'{k:>15}: {mu:.3g}')
+ f = evolve_csv.with_suffix('.png') # filename
+ plt.savefig(f, dpi=200)
+ plt.close()
+ print(f'Saved {f}')
+
+
+def plot_results(file='path/to/results.csv', dir=''):
+ # Plot training results.csv. Usage: from utils.plots import *; plot_results('path/to/results.csv')
+ save_dir = Path(file).parent if file else Path(dir)
+ fig, ax = plt.subplots(2, 5, figsize=(12, 6), tight_layout=True)
+ ax = ax.ravel()
+ files = list(save_dir.glob('results*.csv'))
+ assert len(files), f'No results.csv files found in {save_dir.resolve()}, nothing to plot.'
+ for fi, f in enumerate(files):
+ try:
+ data = pd.read_csv(f)
+ s = [x.strip() for x in data.columns]
+ x = data.values[:, 0]
+ for i, j in enumerate([1, 2, 3, 4, 5, 8, 9, 10, 6, 7]):
+ y = data.values[:, j]
+ # y[y == 0] = np.nan # don't show zero values
+ ax[i].plot(x, y, marker='.', label=f.stem, linewidth=2, markersize=8)
+ ax[i].set_title(s[j], fontsize=12)
+ # if j in [8, 9, 10]: # share train and val loss y axes
+ # ax[i].get_shared_y_axes().join(ax[i], ax[i - 5])
+ except Exception as e:
+ LOGGER.info(f'Warning: Plotting error for {f}: {e}')
+ ax[1].legend()
+ fig.savefig(save_dir / 'results.png', dpi=200)
+ plt.close()
+
+
+def profile_idetection(start=0, stop=0, labels=(), save_dir=''):
+ # Plot iDetection '*.txt' per-image logs. from utils.plots import *; profile_idetection()
+ ax = plt.subplots(2, 4, figsize=(12, 6), tight_layout=True)[1].ravel()
+ s = ['Images', 'Free Storage (GB)', 'RAM Usage (GB)', 'Battery', 'dt_raw (ms)', 'dt_smooth (ms)', 'real-world FPS']
+ files = list(Path(save_dir).glob('frames*.txt'))
+ for fi, f in enumerate(files):
+ try:
+ results = np.loadtxt(f, ndmin=2).T[:, 90:-30] # clip first and last rows
+ n = results.shape[1] # number of rows
+ x = np.arange(start, min(stop, n) if stop else n)
+ results = results[:, x]
+ t = (results[0] - results[0].min()) # set t0=0s
+ results[0] = x
+ for i, a in enumerate(ax):
+ if i < len(results):
+ label = labels[fi] if len(labels) else f.stem.replace('frames_', '')
+ a.plot(t, results[i], marker='.', label=label, linewidth=1, markersize=5)
+ a.set_title(s[i])
+ a.set_xlabel('time (s)')
+ # if fi == len(files) - 1:
+ # a.set_ylim(bottom=0)
+ for side in ['top', 'right']:
+ a.spines[side].set_visible(False)
+ else:
+ a.remove()
+ except Exception as e:
+ print(f'Warning: Plotting error for {f}; {e}')
+ ax[1].legend()
+ plt.savefig(Path(save_dir) / 'idetection_profile.png', dpi=200)
+
+
+def save_one_box(xyxy, im, file=Path('im.jpg'), gain=1.02, pad=10, square=False, BGR=False, save=True):
+ # Save image crop as {file} with crop size multiple {gain} and {pad} pixels. Save and/or return crop
+ xyxy = torch.tensor(xyxy).view(-1, 4)
+ b = xyxy2xywh(xyxy) # boxes
+ if square:
+ b[:, 2:] = b[:, 2:].max(1)[0].unsqueeze(1) # attempt rectangle to square
+ b[:, 2:] = b[:, 2:] * gain + pad # box wh * gain + pad
+ xyxy = xywh2xyxy(b).long()
+ clip_coords(xyxy, im.shape)
+ crop = im[int(xyxy[0, 1]):int(xyxy[0, 3]), int(xyxy[0, 0]):int(xyxy[0, 2]), ::(1 if BGR else -1)]
+ if save:
+ file.parent.mkdir(parents=True, exist_ok=True) # make directory
+ f = str(increment_path(file).with_suffix('.jpg'))
+ # cv2.imwrite(f, crop) # https://github.com/ultralytics/yolov5/issues/7007 chroma subsampling issue
+ Image.fromarray(cv2.cvtColor(crop, cv2.COLOR_BGR2RGB)).save(f, quality=95, subsampling=0)
+ return crop
diff --git a/vouchervision/component_detector/utils/torch_utils.py b/vouchervision/component_detector/utils/torch_utils.py
new file mode 100644
index 0000000000000000000000000000000000000000..72f8a0fd1659784ca6dcacc3fc061dea4e560b22
--- /dev/null
+++ b/vouchervision/component_detector/utils/torch_utils.py
@@ -0,0 +1,312 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+PyTorch utils
+"""
+
+import math
+import os
+import platform
+import subprocess
+import time
+import warnings
+from contextlib import contextmanager
+from copy import deepcopy
+from pathlib import Path
+
+import torch
+import torch.distributed as dist
+import torch.nn as nn
+import torch.nn.functional as F
+
+from utils.general import LOGGER, file_update_date, git_describe
+
+try:
+ import thop # for FLOPs computation
+except ImportError:
+ thop = None
+
+# Suppress PyTorch warnings
+warnings.filterwarnings('ignore', message='User provided device_type of \'cuda\', but CUDA is not available. Disabling')
+
+
+@contextmanager
+def torch_distributed_zero_first(local_rank: int):
+ # Decorator to make all processes in distributed training wait for each local_master to do something
+ if local_rank not in [-1, 0]:
+ dist.barrier(device_ids=[local_rank])
+ yield
+ if local_rank == 0:
+ dist.barrier(device_ids=[0])
+
+
+def device_count():
+ # Returns number of CUDA devices available. Safe version of torch.cuda.device_count(). Only works on Linux.
+ assert platform.system() == 'Linux', 'device_count() function only works on Linux'
+ try:
+ cmd = 'nvidia-smi -L | wc -l'
+ return int(subprocess.run(cmd, shell=True, capture_output=True, check=True).stdout.decode().split()[-1])
+ except Exception:
+ return 0
+
+
+def select_device(device='', batch_size=0, newline=True):
+ # device = 'cpu' or '0' or '0,1,2,3'
+ s = f'YOLOv5 🚀 {git_describe() or file_update_date()} torch {torch.__version__} ' # string
+ device = str(device).strip().lower().replace('cuda:', '') # to string, 'cuda:0' to '0'
+ cpu = device == 'cpu'
+ if cpu:
+ os.environ['CUDA_VISIBLE_DEVICES'] = '-1' # force torch.cuda.is_available() = False
+ elif device: # non-cpu device requested
+ os.environ['CUDA_VISIBLE_DEVICES'] = device # set environment variable - must be before assert is_available()
+ assert torch.cuda.is_available() and torch.cuda.device_count() >= len(device.replace(',', '')), \
+ f"Invalid CUDA '--device {device}' requested, use '--device cpu' or pass valid CUDA device(s)"
+
+ cuda = not cpu and torch.cuda.is_available()
+ if cuda:
+ devices = device.split(',') if device else '0' # range(torch.cuda.device_count()) # i.e. 0,1,6,7
+ n = len(devices) # device count
+ if n > 1 and batch_size > 0: # check batch_size is divisible by device_count
+ assert batch_size % n == 0, f'batch-size {batch_size} not multiple of GPU count {n}'
+ space = ' ' * (len(s) + 1)
+ for i, d in enumerate(devices):
+ p = torch.cuda.get_device_properties(i)
+ s += f"{'' if i == 0 else space}CUDA:{d} ({p.name}, {p.total_memory / (1 << 20):.0f}MiB)\n" # bytes to MB
+ else:
+ s += 'CPU\n'
+
+ if not newline:
+ s = s.rstrip()
+ LOGGER.info(s.encode().decode('ascii', 'ignore') if platform.system() == 'Windows' else s) # emoji-safe
+ return torch.device('cuda:0' if cuda else 'cpu')
+
+
+def time_sync():
+ # PyTorch-accurate time
+ if torch.cuda.is_available():
+ torch.cuda.synchronize()
+ return time.time()
+
+
+def profile(input, ops, n=10, device=None):
+ # YOLOv5 speed/memory/FLOPs profiler
+ #
+ # Usage:
+ # input = torch.randn(16, 3, 640, 640)
+ # m1 = lambda x: x * torch.sigmoid(x)
+ # m2 = nn.SiLU()
+ # profile(input, [m1, m2], n=100) # profile over 100 iterations
+
+ results = []
+ device = device or select_device()
+ print(f"{'Params':>12s}{'GFLOPs':>12s}{'GPU_mem (GB)':>14s}{'forward (ms)':>14s}{'backward (ms)':>14s}"
+ f"{'input':>24s}{'output':>24s}")
+
+ for x in input if isinstance(input, list) else [input]:
+ x = x.to(device)
+ x.requires_grad = True
+ for m in ops if isinstance(ops, list) else [ops]:
+ m = m.to(device) if hasattr(m, 'to') else m # device
+ m = m.half() if hasattr(m, 'half') and isinstance(x, torch.Tensor) and x.dtype is torch.float16 else m
+ tf, tb, t = 0, 0, [0, 0, 0] # dt forward, backward
+ try:
+ flops = thop.profile(m, inputs=(x,), verbose=False)[0] / 1E9 * 2 # GFLOPs
+ except Exception:
+ flops = 0
+
+ try:
+ for _ in range(n):
+ t[0] = time_sync()
+ y = m(x)
+ t[1] = time_sync()
+ try:
+ _ = (sum(yi.sum() for yi in y) if isinstance(y, list) else y).sum().backward()
+ t[2] = time_sync()
+ except Exception: # no backward method
+ # print(e) # for debug
+ t[2] = float('nan')
+ tf += (t[1] - t[0]) * 1000 / n # ms per op forward
+ tb += (t[2] - t[1]) * 1000 / n # ms per op backward
+ mem = torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0 # (GB)
+ s_in = tuple(x.shape) if isinstance(x, torch.Tensor) else 'list'
+ s_out = tuple(y.shape) if isinstance(y, torch.Tensor) else 'list'
+ p = sum(list(x.numel() for x in m.parameters())) if isinstance(m, nn.Module) else 0 # parameters
+ print(f'{p:12}{flops:12.4g}{mem:>14.3f}{tf:14.4g}{tb:14.4g}{str(s_in):>24s}{str(s_out):>24s}')
+ results.append([p, flops, mem, tf, tb, s_in, s_out])
+ except Exception as e:
+ print(e)
+ results.append(None)
+ torch.cuda.empty_cache()
+ return results
+
+
+def is_parallel(model):
+ # Returns True if model is of type DP or DDP
+ return type(model) in (nn.parallel.DataParallel, nn.parallel.DistributedDataParallel)
+
+
+def de_parallel(model):
+ # De-parallelize a model: returns single-GPU model if model is of type DP or DDP
+ return model.module if is_parallel(model) else model
+
+
+def initialize_weights(model):
+ for m in model.modules():
+ t = type(m)
+ if t is nn.Conv2d:
+ pass # nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
+ elif t is nn.BatchNorm2d:
+ m.eps = 1e-3
+ m.momentum = 0.03
+ elif t in [nn.Hardswish, nn.LeakyReLU, nn.ReLU, nn.ReLU6, nn.SiLU]:
+ m.inplace = True
+
+
+def find_modules(model, mclass=nn.Conv2d):
+ # Finds layer indices matching module class 'mclass'
+ return [i for i, m in enumerate(model.module_list) if isinstance(m, mclass)]
+
+
+def sparsity(model):
+ # Return global model sparsity
+ a, b = 0, 0
+ for p in model.parameters():
+ a += p.numel()
+ b += (p == 0).sum()
+ return b / a
+
+
+def prune(model, amount=0.3):
+ # Prune model to requested global sparsity
+ import torch.nn.utils.prune as prune
+ print('Pruning model... ', end='')
+ for name, m in model.named_modules():
+ if isinstance(m, nn.Conv2d):
+ prune.l1_unstructured(m, name='weight', amount=amount) # prune
+ prune.remove(m, 'weight') # make permanent
+ print(' %.3g global sparsity' % sparsity(model))
+
+
+def fuse_conv_and_bn(conv, bn):
+ # Fuse Conv2d() and BatchNorm2d() layers https://tehnokv.com/posts/fusing-batchnorm-and-conv/
+ fusedconv = nn.Conv2d(conv.in_channels,
+ conv.out_channels,
+ kernel_size=conv.kernel_size,
+ stride=conv.stride,
+ padding=conv.padding,
+ groups=conv.groups,
+ bias=True).requires_grad_(False).to(conv.weight.device)
+
+ # Prepare filters
+ w_conv = conv.weight.clone().view(conv.out_channels, -1)
+ w_bn = torch.diag(bn.weight.div(torch.sqrt(bn.eps + bn.running_var)))
+ fusedconv.weight.copy_(torch.mm(w_bn, w_conv).view(fusedconv.weight.shape))
+
+ # Prepare spatial bias
+ b_conv = torch.zeros(conv.weight.size(0), device=conv.weight.device) if conv.bias is None else conv.bias
+ b_bn = bn.bias - bn.weight.mul(bn.running_mean).div(torch.sqrt(bn.running_var + bn.eps))
+ fusedconv.bias.copy_(torch.mm(w_bn, b_conv.reshape(-1, 1)).reshape(-1) + b_bn)
+
+ return fusedconv
+
+
+def model_info(model, verbose=False, img_size=640):
+ # Model information. img_size may be int or list, i.e. img_size=640 or img_size=[640, 320]
+ n_p = sum(x.numel() for x in model.parameters()) # number parameters
+ n_g = sum(x.numel() for x in model.parameters() if x.requires_grad) # number gradients
+ if verbose:
+ print(f"{'layer':>5} {'name':>40} {'gradient':>9} {'parameters':>12} {'shape':>20} {'mu':>10} {'sigma':>10}")
+ for i, (name, p) in enumerate(model.named_parameters()):
+ name = name.replace('module_list.', '')
+ print('%5g %40s %9s %12g %20s %10.3g %10.3g' %
+ (i, name, p.requires_grad, p.numel(), list(p.shape), p.mean(), p.std()))
+
+ try: # FLOPs
+ from thop import profile
+ stride = max(int(model.stride.max()), 32) if hasattr(model, 'stride') else 32
+ img = torch.zeros((1, model.yaml.get('ch', 3), stride, stride), device=next(model.parameters()).device) # input
+ flops = profile(deepcopy(model), inputs=(img,), verbose=False)[0] / 1E9 * 2 # stride GFLOPs
+ img_size = img_size if isinstance(img_size, list) else [img_size, img_size] # expand if int/float
+ fs = ', %.1f GFLOPs' % (flops * img_size[0] / stride * img_size[1] / stride) # 640x640 GFLOPs
+ except (ImportError, Exception):
+ fs = ''
+
+ name = Path(model.yaml_file).stem.replace('yolov5', 'YOLOv5') if hasattr(model, 'yaml_file') else 'Model'
+ LOGGER.info(f"{name} summary: {len(list(model.modules()))} layers, {n_p} parameters, {n_g} gradients{fs}")
+
+
+def scale_img(img, ratio=1.0, same_shape=False, gs=32): # img(16,3,256,416)
+ # Scales img(bs,3,y,x) by ratio constrained to gs-multiple
+ if ratio == 1.0:
+ return img
+ else:
+ h, w = img.shape[2:]
+ s = (int(h * ratio), int(w * ratio)) # new size
+ img = F.interpolate(img, size=s, mode='bilinear', align_corners=False) # resize
+ if not same_shape: # pad/crop img
+ h, w = (math.ceil(x * ratio / gs) * gs for x in (h, w))
+ return F.pad(img, [0, w - s[1], 0, h - s[0]], value=0.447) # value = imagenet mean
+
+
+def copy_attr(a, b, include=(), exclude=()):
+ # Copy attributes from b to a, options to only include [...] and to exclude [...]
+ for k, v in b.__dict__.items():
+ if (len(include) and k not in include) or k.startswith('_') or k in exclude:
+ continue
+ else:
+ setattr(a, k, v)
+
+
+class EarlyStopping:
+ # YOLOv5 simple early stopper
+ def __init__(self, patience=30):
+ self.best_fitness = 0.0 # i.e. mAP
+ self.best_epoch = 0
+ self.patience = patience or float('inf') # epochs to wait after fitness stops improving to stop
+ self.possible_stop = False # possible stop may occur next epoch
+
+ def __call__(self, epoch, fitness):
+ if fitness >= self.best_fitness: # >= 0 to allow for early zero-fitness stage of training
+ self.best_epoch = epoch
+ self.best_fitness = fitness
+ delta = epoch - self.best_epoch # epochs without improvement
+ self.possible_stop = delta >= (self.patience - 1) # possible stop may occur next epoch
+ stop = delta >= self.patience # stop training if patience exceeded
+ if stop:
+ LOGGER.info(f'Stopping training early as no improvement observed in last {self.patience} epochs. '
+ f'Best results observed at epoch {self.best_epoch}, best model saved as best.pt.\n'
+ f'To update EarlyStopping(patience={self.patience}) pass a new patience value, '
+ f'i.e. `python train.py --patience 300` or use `--patience 0` to disable EarlyStopping.')
+ return stop
+
+
+class ModelEMA:
+ """ Updated Exponential Moving Average (EMA) from https://github.com/rwightman/pytorch-image-models
+ Keeps a moving average of everything in the model state_dict (parameters and buffers)
+ For EMA details see https://www.tensorflow.org/api_docs/python/tf/train/ExponentialMovingAverage
+ """
+
+ def __init__(self, model, decay=0.9999, tau=2000, updates=0):
+ # Create EMA
+ self.ema = deepcopy(de_parallel(model)).eval() # FP32 EMA
+ # if next(model.parameters()).device.type != 'cpu':
+ # self.ema.half() # FP16 EMA
+ self.updates = updates # number of EMA updates
+ self.decay = lambda x: decay * (1 - math.exp(-x / tau)) # decay exponential ramp (to help early epochs)
+ for p in self.ema.parameters():
+ p.requires_grad_(False)
+
+ def update(self, model):
+ # Update EMA parameters
+ with torch.no_grad():
+ self.updates += 1
+ d = self.decay(self.updates)
+
+ msd = de_parallel(model).state_dict() # model state_dict
+ for k, v in self.ema.state_dict().items():
+ if v.dtype.is_floating_point:
+ v *= d
+ v += (1 - d) * msd[k].detach()
+
+ def update_attr(self, model, include=(), exclude=('process_group', 'reducer')):
+ # Update EMA attributes
+ copy_attr(self.ema, model, include, exclude)
diff --git a/vouchervision/component_detector/utils/torch_utils_torchscript.py b/vouchervision/component_detector/utils/torch_utils_torchscript.py
new file mode 100644
index 0000000000000000000000000000000000000000..ea9e9fbf5740f7da83f6db90fce6d4db4b3f743f
--- /dev/null
+++ b/vouchervision/component_detector/utils/torch_utils_torchscript.py
@@ -0,0 +1,432 @@
+# YOLOv5 🚀 by Ultralytics, AGPL-3.0 license
+"""
+PyTorch utils
+"""
+
+import math
+import os
+import platform
+import subprocess
+import time
+import warnings
+from contextlib import contextmanager
+from copy import deepcopy
+from pathlib import Path
+
+import torch
+import torch.distributed as dist
+import torch.nn as nn
+import torch.nn.functional as F
+from torch.nn.parallel import DistributedDataParallel as DDP
+
+from utils.general_torchscript import LOGGER, check_version, colorstr, file_date, git_describe
+
+LOCAL_RANK = int(os.getenv('LOCAL_RANK', -1)) # https://pytorch.org/docs/stable/elastic/run.html
+RANK = int(os.getenv('RANK', -1))
+WORLD_SIZE = int(os.getenv('WORLD_SIZE', 1))
+
+try:
+ import thop # for FLOPs computation
+except ImportError:
+ thop = None
+
+# Suppress PyTorch warnings
+warnings.filterwarnings('ignore', message='User provided device_type of \'cuda\', but CUDA is not available. Disabling')
+warnings.filterwarnings('ignore', category=UserWarning)
+
+
+def smart_inference_mode(torch_1_9=check_version(torch.__version__, '1.9.0')):
+ # Applies torch.inference_mode() decorator if torch>=1.9.0 else torch.no_grad() decorator
+ def decorate(fn):
+ return (torch.inference_mode if torch_1_9 else torch.no_grad)()(fn)
+
+ return decorate
+
+
+def smartCrossEntropyLoss(label_smoothing=0.0):
+ # Returns nn.CrossEntropyLoss with label smoothing enabled for torch>=1.10.0
+ if check_version(torch.__version__, '1.10.0'):
+ return nn.CrossEntropyLoss(label_smoothing=label_smoothing)
+ if label_smoothing > 0:
+ LOGGER.warning(f'WARNING ⚠️ label smoothing {label_smoothing} requires torch>=1.10.0')
+ return nn.CrossEntropyLoss()
+
+
+def smart_DDP(model):
+ # Model DDP creation with checks
+ assert not check_version(torch.__version__, '1.12.0', pinned=True), \
+ 'torch==1.12.0 torchvision==0.13.0 DDP training is not supported due to a known issue. ' \
+ 'Please upgrade or downgrade torch to use DDP. See https://github.com/ultralytics/yolov5/issues/8395'
+ if check_version(torch.__version__, '1.11.0'):
+ return DDP(model, device_ids=[LOCAL_RANK], output_device=LOCAL_RANK, static_graph=True)
+ else:
+ return DDP(model, device_ids=[LOCAL_RANK], output_device=LOCAL_RANK)
+
+
+def reshape_classifier_output(model, n=1000):
+ # Update a TorchVision classification model to class count 'n' if required
+ from models.common import Classify
+ name, m = list((model.model if hasattr(model, 'model') else model).named_children())[-1] # last module
+ if isinstance(m, Classify): # YOLOv5 Classify() head
+ if m.linear.out_features != n:
+ m.linear = nn.Linear(m.linear.in_features, n)
+ elif isinstance(m, nn.Linear): # ResNet, EfficientNet
+ if m.out_features != n:
+ setattr(model, name, nn.Linear(m.in_features, n))
+ elif isinstance(m, nn.Sequential):
+ types = [type(x) for x in m]
+ if nn.Linear in types:
+ i = types.index(nn.Linear) # nn.Linear index
+ if m[i].out_features != n:
+ m[i] = nn.Linear(m[i].in_features, n)
+ elif nn.Conv2d in types:
+ i = types.index(nn.Conv2d) # nn.Conv2d index
+ if m[i].out_channels != n:
+ m[i] = nn.Conv2d(m[i].in_channels, n, m[i].kernel_size, m[i].stride, bias=m[i].bias is not None)
+
+
+@contextmanager
+def torch_distributed_zero_first(local_rank: int):
+ # Decorator to make all processes in distributed training wait for each local_master to do something
+ if local_rank not in [-1, 0]:
+ dist.barrier(device_ids=[local_rank])
+ yield
+ if local_rank == 0:
+ dist.barrier(device_ids=[0])
+
+
+def device_count():
+ # Returns number of CUDA devices available. Safe version of torch.cuda.device_count(). Supports Linux and Windows
+ assert platform.system() in ('Linux', 'Windows'), 'device_count() only supported on Linux or Windows'
+ try:
+ cmd = 'nvidia-smi -L | wc -l' if platform.system() == 'Linux' else 'nvidia-smi -L | find /c /v ""' # Windows
+ return int(subprocess.run(cmd, shell=True, capture_output=True, check=True).stdout.decode().split()[-1])
+ except Exception:
+ return 0
+
+
+def select_device(device='', batch_size=0, newline=True):
+ # device = None or 'cpu' or 0 or '0' or '0,1,2,3'
+ s = f'YOLOv5 🚀 {git_describe() or file_date()} Python-{platform.python_version()} torch-{torch.__version__} '
+ device = str(device).strip().lower().replace('cuda:', '').replace('none', '') # to string, 'cuda:0' to '0'
+ cpu = device == 'cpu'
+ mps = device == 'mps' # Apple Metal Performance Shaders (MPS)
+ if cpu or mps:
+ os.environ['CUDA_VISIBLE_DEVICES'] = '-1' # force torch.cuda.is_available() = False
+ elif device: # non-cpu device requested
+ os.environ['CUDA_VISIBLE_DEVICES'] = device # set environment variable - must be before assert is_available()
+ assert torch.cuda.is_available() and torch.cuda.device_count() >= len(device.replace(',', '')), \
+ f"Invalid CUDA '--device {device}' requested, use '--device cpu' or pass valid CUDA device(s)"
+
+ if not cpu and not mps and torch.cuda.is_available(): # prefer GPU if available
+ devices = device.split(',') if device else '0' # range(torch.cuda.device_count()) # i.e. 0,1,6,7
+ n = len(devices) # device count
+ if n > 1 and batch_size > 0: # check batch_size is divisible by device_count
+ assert batch_size % n == 0, f'batch-size {batch_size} not multiple of GPU count {n}'
+ space = ' ' * (len(s) + 1)
+ for i, d in enumerate(devices):
+ p = torch.cuda.get_device_properties(i)
+ s += f"{'' if i == 0 else space}CUDA:{d} ({p.name}, {p.total_memory / (1 << 20):.0f}MiB)\n" # bytes to MB
+ arg = 'cuda:0'
+ elif mps and getattr(torch, 'has_mps', False) and torch.backends.mps.is_available(): # prefer MPS if available
+ s += 'MPS\n'
+ arg = 'mps'
+ else: # revert to CPU
+ s += 'CPU\n'
+ arg = 'cpu'
+
+ if not newline:
+ s = s.rstrip()
+ LOGGER.info(s)
+ return torch.device(arg)
+
+
+def time_sync():
+ # PyTorch-accurate time
+ if torch.cuda.is_available():
+ torch.cuda.synchronize()
+ return time.time()
+
+
+def profile(input, ops, n=10, device=None):
+ """ YOLOv5 speed/memory/FLOPs profiler
+ Usage:
+ input = torch.randn(16, 3, 640, 640)
+ m1 = lambda x: x * torch.sigmoid(x)
+ m2 = nn.SiLU()
+ profile(input, [m1, m2], n=100) # profile over 100 iterations
+ """
+ results = []
+ if not isinstance(device, torch.device):
+ device = select_device(device)
+ print(f"{'Params':>12s}{'GFLOPs':>12s}{'GPU_mem (GB)':>14s}{'forward (ms)':>14s}{'backward (ms)':>14s}"
+ f"{'input':>24s}{'output':>24s}")
+
+ for x in input if isinstance(input, list) else [input]:
+ x = x.to(device)
+ x.requires_grad = True
+ for m in ops if isinstance(ops, list) else [ops]:
+ m = m.to(device) if hasattr(m, 'to') else m # device
+ m = m.half() if hasattr(m, 'half') and isinstance(x, torch.Tensor) and x.dtype is torch.float16 else m
+ tf, tb, t = 0, 0, [0, 0, 0] # dt forward, backward
+ try:
+ flops = thop.profile(m, inputs=(x, ), verbose=False)[0] / 1E9 * 2 # GFLOPs
+ except Exception:
+ flops = 0
+
+ try:
+ for _ in range(n):
+ t[0] = time_sync()
+ y = m(x)
+ t[1] = time_sync()
+ try:
+ _ = (sum(yi.sum() for yi in y) if isinstance(y, list) else y).sum().backward()
+ t[2] = time_sync()
+ except Exception: # no backward method
+ # print(e) # for debug
+ t[2] = float('nan')
+ tf += (t[1] - t[0]) * 1000 / n # ms per op forward
+ tb += (t[2] - t[1]) * 1000 / n # ms per op backward
+ mem = torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0 # (GB)
+ s_in, s_out = (tuple(x.shape) if isinstance(x, torch.Tensor) else 'list' for x in (x, y)) # shapes
+ p = sum(x.numel() for x in m.parameters()) if isinstance(m, nn.Module) else 0 # parameters
+ print(f'{p:12}{flops:12.4g}{mem:>14.3f}{tf:14.4g}{tb:14.4g}{str(s_in):>24s}{str(s_out):>24s}')
+ results.append([p, flops, mem, tf, tb, s_in, s_out])
+ except Exception as e:
+ print(e)
+ results.append(None)
+ torch.cuda.empty_cache()
+ return results
+
+
+def is_parallel(model):
+ # Returns True if model is of type DP or DDP
+ return type(model) in (nn.parallel.DataParallel, nn.parallel.DistributedDataParallel)
+
+
+def de_parallel(model):
+ # De-parallelize a model: returns single-GPU model if model is of type DP or DDP
+ return model.module if is_parallel(model) else model
+
+
+def initialize_weights(model):
+ for m in model.modules():
+ t = type(m)
+ if t is nn.Conv2d:
+ pass # nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
+ elif t is nn.BatchNorm2d:
+ m.eps = 1e-3
+ m.momentum = 0.03
+ elif t in [nn.Hardswish, nn.LeakyReLU, nn.ReLU, nn.ReLU6, nn.SiLU]:
+ m.inplace = True
+
+
+def find_modules(model, mclass=nn.Conv2d):
+ # Finds layer indices matching module class 'mclass'
+ return [i for i, m in enumerate(model.module_list) if isinstance(m, mclass)]
+
+
+def sparsity(model):
+ # Return global model sparsity
+ a, b = 0, 0
+ for p in model.parameters():
+ a += p.numel()
+ b += (p == 0).sum()
+ return b / a
+
+
+def prune(model, amount=0.3):
+ # Prune model to requested global sparsity
+ import torch.nn.utils.prune as prune
+ for name, m in model.named_modules():
+ if isinstance(m, nn.Conv2d):
+ prune.l1_unstructured(m, name='weight', amount=amount) # prune
+ prune.remove(m, 'weight') # make permanent
+ LOGGER.info(f'Model pruned to {sparsity(model):.3g} global sparsity')
+
+
+def fuse_conv_and_bn(conv, bn):
+ # Fuse Conv2d() and BatchNorm2d() layers https://tehnokv.com/posts/fusing-batchnorm-and-conv/
+ fusedconv = nn.Conv2d(conv.in_channels,
+ conv.out_channels,
+ kernel_size=conv.kernel_size,
+ stride=conv.stride,
+ padding=conv.padding,
+ dilation=conv.dilation,
+ groups=conv.groups,
+ bias=True).requires_grad_(False).to(conv.weight.device)
+
+ # Prepare filters
+ w_conv = conv.weight.clone().view(conv.out_channels, -1)
+ w_bn = torch.diag(bn.weight.div(torch.sqrt(bn.eps + bn.running_var)))
+ fusedconv.weight.copy_(torch.mm(w_bn, w_conv).view(fusedconv.weight.shape))
+
+ # Prepare spatial bias
+ b_conv = torch.zeros(conv.weight.size(0), device=conv.weight.device) if conv.bias is None else conv.bias
+ b_bn = bn.bias - bn.weight.mul(bn.running_mean).div(torch.sqrt(bn.running_var + bn.eps))
+ fusedconv.bias.copy_(torch.mm(w_bn, b_conv.reshape(-1, 1)).reshape(-1) + b_bn)
+
+ return fusedconv
+
+
+def model_info(model, verbose=False, imgsz=640):
+ # Model information. img_size may be int or list, i.e. img_size=640 or img_size=[640, 320]
+ n_p = sum(x.numel() for x in model.parameters()) # number parameters
+ n_g = sum(x.numel() for x in model.parameters() if x.requires_grad) # number gradients
+ if verbose:
+ print(f"{'layer':>5} {'name':>40} {'gradient':>9} {'parameters':>12} {'shape':>20} {'mu':>10} {'sigma':>10}")
+ for i, (name, p) in enumerate(model.named_parameters()):
+ name = name.replace('module_list.', '')
+ print('%5g %40s %9s %12g %20s %10.3g %10.3g' %
+ (i, name, p.requires_grad, p.numel(), list(p.shape), p.mean(), p.std()))
+
+ try: # FLOPs
+ p = next(model.parameters())
+ stride = max(int(model.stride.max()), 32) if hasattr(model, 'stride') else 32 # max stride
+ im = torch.empty((1, p.shape[1], stride, stride), device=p.device) # input image in BCHW format
+ flops = thop.profile(deepcopy(model), inputs=(im, ), verbose=False)[0] / 1E9 * 2 # stride GFLOPs
+ imgsz = imgsz if isinstance(imgsz, list) else [imgsz, imgsz] # expand if int/float
+ fs = f', {flops * imgsz[0] / stride * imgsz[1] / stride:.1f} GFLOPs' # 640x640 GFLOPs
+ except Exception:
+ fs = ''
+
+ name = Path(model.yaml_file).stem.replace('yolov5', 'YOLOv5') if hasattr(model, 'yaml_file') else 'Model'
+ LOGGER.info(f'{name} summary: {len(list(model.modules()))} layers, {n_p} parameters, {n_g} gradients{fs}')
+
+
+def scale_img(img, ratio=1.0, same_shape=False, gs=32): # img(16,3,256,416)
+ # Scales img(bs,3,y,x) by ratio constrained to gs-multiple
+ if ratio == 1.0:
+ return img
+ h, w = img.shape[2:]
+ s = (int(h * ratio), int(w * ratio)) # new size
+ img = F.interpolate(img, size=s, mode='bilinear', align_corners=False) # resize
+ if not same_shape: # pad/crop img
+ h, w = (math.ceil(x * ratio / gs) * gs for x in (h, w))
+ return F.pad(img, [0, w - s[1], 0, h - s[0]], value=0.447) # value = imagenet mean
+
+
+def copy_attr(a, b, include=(), exclude=()):
+ # Copy attributes from b to a, options to only include [...] and to exclude [...]
+ for k, v in b.__dict__.items():
+ if (len(include) and k not in include) or k.startswith('_') or k in exclude:
+ continue
+ else:
+ setattr(a, k, v)
+
+
+def smart_optimizer(model, name='Adam', lr=0.001, momentum=0.9, decay=1e-5):
+ # YOLOv5 3-param group optimizer: 0) weights with decay, 1) weights no decay, 2) biases no decay
+ g = [], [], [] # optimizer parameter groups
+ bn = tuple(v for k, v in nn.__dict__.items() if 'Norm' in k) # normalization layers, i.e. BatchNorm2d()
+ for v in model.modules():
+ for p_name, p in v.named_parameters(recurse=0):
+ if p_name == 'bias': # bias (no decay)
+ g[2].append(p)
+ elif p_name == 'weight' and isinstance(v, bn): # weight (no decay)
+ g[1].append(p)
+ else:
+ g[0].append(p) # weight (with decay)
+
+ if name == 'Adam':
+ optimizer = torch.optim.Adam(g[2], lr=lr, betas=(momentum, 0.999)) # adjust beta1 to momentum
+ elif name == 'AdamW':
+ optimizer = torch.optim.AdamW(g[2], lr=lr, betas=(momentum, 0.999), weight_decay=0.0)
+ elif name == 'RMSProp':
+ optimizer = torch.optim.RMSprop(g[2], lr=lr, momentum=momentum)
+ elif name == 'SGD':
+ optimizer = torch.optim.SGD(g[2], lr=lr, momentum=momentum, nesterov=True)
+ else:
+ raise NotImplementedError(f'Optimizer {name} not implemented.')
+
+ optimizer.add_param_group({'params': g[0], 'weight_decay': decay}) # add g0 with weight_decay
+ optimizer.add_param_group({'params': g[1], 'weight_decay': 0.0}) # add g1 (BatchNorm2d weights)
+ LOGGER.info(f"{colorstr('optimizer:')} {type(optimizer).__name__}(lr={lr}) with parameter groups "
+ f'{len(g[1])} weight(decay=0.0), {len(g[0])} weight(decay={decay}), {len(g[2])} bias')
+ return optimizer
+
+
+def smart_hub_load(repo='ultralytics/yolov5', model='yolov5s', **kwargs):
+ # YOLOv5 torch.hub.load() wrapper with smart error/issue handling
+ if check_version(torch.__version__, '1.9.1'):
+ kwargs['skip_validation'] = True # validation causes GitHub API rate limit errors
+ if check_version(torch.__version__, '1.12.0'):
+ kwargs['trust_repo'] = True # argument required starting in torch 0.12
+ try:
+ return torch.hub.load(repo, model, **kwargs)
+ except Exception:
+ return torch.hub.load(repo, model, force_reload=True, **kwargs)
+
+
+def smart_resume(ckpt, optimizer, ema=None, weights='yolov5s.pt', epochs=300, resume=True):
+ # Resume training from a partially trained checkpoint
+ best_fitness = 0.0
+ start_epoch = ckpt['epoch'] + 1
+ if ckpt['optimizer'] is not None:
+ optimizer.load_state_dict(ckpt['optimizer']) # optimizer
+ best_fitness = ckpt['best_fitness']
+ if ema and ckpt.get('ema'):
+ ema.ema.load_state_dict(ckpt['ema'].float().state_dict()) # EMA
+ ema.updates = ckpt['updates']
+ if resume:
+ assert start_epoch > 0, f'{weights} training to {epochs} epochs is finished, nothing to resume.\n' \
+ f"Start a new training without --resume, i.e. 'python train.py --weights {weights}'"
+ LOGGER.info(f'Resuming training from {weights} from epoch {start_epoch} to {epochs} total epochs')
+ if epochs < start_epoch:
+ LOGGER.info(f"{weights} has been trained for {ckpt['epoch']} epochs. Fine-tuning for {epochs} more epochs.")
+ epochs += ckpt['epoch'] # finetune additional epochs
+ return best_fitness, start_epoch, epochs
+
+
+class EarlyStopping:
+ # YOLOv5 simple early stopper
+ def __init__(self, patience=30):
+ self.best_fitness = 0.0 # i.e. mAP
+ self.best_epoch = 0
+ self.patience = patience or float('inf') # epochs to wait after fitness stops improving to stop
+ self.possible_stop = False # possible stop may occur next epoch
+
+ def __call__(self, epoch, fitness):
+ if fitness >= self.best_fitness: # >= 0 to allow for early zero-fitness stage of training
+ self.best_epoch = epoch
+ self.best_fitness = fitness
+ delta = epoch - self.best_epoch # epochs without improvement
+ self.possible_stop = delta >= (self.patience - 1) # possible stop may occur next epoch
+ stop = delta >= self.patience # stop training if patience exceeded
+ if stop:
+ LOGGER.info(f'Stopping training early as no improvement observed in last {self.patience} epochs. '
+ f'Best results observed at epoch {self.best_epoch}, best model saved as best.pt.\n'
+ f'To update EarlyStopping(patience={self.patience}) pass a new patience value, '
+ f'i.e. `python train.py --patience 300` or use `--patience 0` to disable EarlyStopping.')
+ return stop
+
+
+class ModelEMA:
+ """ Updated Exponential Moving Average (EMA) from https://github.com/rwightman/pytorch-image-models
+ Keeps a moving average of everything in the model state_dict (parameters and buffers)
+ For EMA details see https://www.tensorflow.org/api_docs/python/tf/train/ExponentialMovingAverage
+ """
+
+ def __init__(self, model, decay=0.9999, tau=2000, updates=0):
+ # Create EMA
+ self.ema = deepcopy(de_parallel(model)).eval() # FP32 EMA
+ self.updates = updates # number of EMA updates
+ self.decay = lambda x: decay * (1 - math.exp(-x / tau)) # decay exponential ramp (to help early epochs)
+ for p in self.ema.parameters():
+ p.requires_grad_(False)
+
+ def update(self, model):
+ # Update EMA parameters
+ self.updates += 1
+ d = self.decay(self.updates)
+
+ msd = de_parallel(model).state_dict() # model state_dict
+ for k, v in self.ema.state_dict().items():
+ if v.dtype.is_floating_point: # true for FP16 and FP32
+ v *= d
+ v += (1 - d) * msd[k].detach()
+ # assert v.dtype == msd[k].dtype == torch.float32, f'{k}: EMA {v.dtype} and model {msd[k].dtype} must be FP32'
+
+ def update_attr(self, model, include=(), exclude=('process_group', 'reducer')):
+ # Update EMA attributes
+ copy_attr(self.ema, model, include, exclude)
diff --git a/vouchervision/component_detector/utils_check_layers.py b/vouchervision/component_detector/utils_check_layers.py
new file mode 100644
index 0000000000000000000000000000000000000000..31e4d397dd53e4b007cace8efb7428f0cfac0796
--- /dev/null
+++ b/vouchervision/component_detector/utils_check_layers.py
@@ -0,0 +1,20 @@
+import torch
+import torch.nn as nn
+from torchsummary import summary
+
+# Load the model
+model = torch.load('D:/Dropbox/FieldPrism/fieldprism/yolov5/weights_nano/best.pt')
+
+summary(model['model'] , input_size=(3, 512, 512))
+
+model.load_state_dict(checkpoint['model'])
+# Create a dummy input with the same dimensions expected by the model.
+# For a YOLO model, it might be something like (batch_size, 3, height, width)
+dummy_input = torch.randn(1, 3, 512, 512)
+
+# Get a prediction to inspect the shape
+with torch.no_grad():
+ output = model(dummy_input)
+
+# Print the output shape
+print("Output shape:", output.shape)
\ No newline at end of file
diff --git a/vouchervision/component_detector/utils_convert_to_TorchScript.py b/vouchervision/component_detector/utils_convert_to_TorchScript.py
new file mode 100644
index 0000000000000000000000000000000000000000..663d59fdc49559acb1ce557acfc3b9598445d448
--- /dev/null
+++ b/vouchervision/component_detector/utils_convert_to_TorchScript.py
@@ -0,0 +1,81 @@
+import torch
+import torch.jit
+import sys
+import os
+
+def check_if_jit_model(model_path):
+ try:
+ # Try to load the model as a TorchScript model
+ torch.jit.load(model_path)
+ return True, "The model is in TorchScript format."
+ except Exception as e:
+ return False, f"The model is not in TorchScript format. Error: {str(e)}"
+
+# Append the path to your system path (Note: Don't use quotes)
+sys.path.append("D:/Dropbox/FieldPrism/fieldprism/yolov5/weights")
+
+
+# Check if the path exists and you have read permissions
+model_path = "D:/Dropbox/FieldPrism/fieldprism/yolov5/weights_nano/best.pt"
+
+
+if not os.path.exists(model_path):
+ print(f"Model path {model_path} does not exist. Please check the path.")
+ sys.exit(1)
+
+
+is_jit_model, message = check_if_jit_model(model_path)
+print(message)
+
+
+# Load your custom model
+# Load your custom model
+try:
+ loaded_dict = torch.load(model_path, map_location='cuda:0') # Adjust the device as needed
+
+ # Assuming your model architecture is defined in a class called `YourModelClass`
+ # model = YourModelClass()
+ # model.load_state_dict(loaded_dict['model']) # If the model is saved as a state dictionary
+
+ # If the model is saved entirely (architecture + weights)
+ if isinstance(loaded_dict, dict) and 'model' in loaded_dict:
+ model = loaded_dict['model']
+ else:
+ model = loaded_dict # Assuming the loaded object is a model
+
+ # Switch the model to evaluation mode
+ model.eval()
+
+ # Create a dummy input that matches the input dimensions of the model
+ dummy_input = torch.randn(1, 3, 512, 512).half().to('cuda:0')
+
+
+ # Try tracing the model
+ try:
+ scripted_module = torch.jit.trace(model, dummy_input)
+ print("Model traced successfully.")
+ except Exception as e:
+ print(f"An error occurred during tracing. Error: {str(e)}")
+
+ # Try scripting the model
+ try:
+ scripted_module = torch.jit.script(model)
+ save_path = "D:/Dropbox/FieldPrism/fieldprism/yolov5/weights/fieldprism_v_1_0.pt"
+ scripted_module.save(save_path)
+ print(f"Saved TorchScript model to {save_path}")
+ except Exception as e:
+ print(f"An error occurred during the scripting. Error: {str(e)}")
+except Exception as e:
+ print(f"Error in loading the model: {e}")
+ sys.exit(1)
+
+
+# Script the model
+try:
+ scripted_module = torch.jit.script(model)
+ # Save the TorchScript model (make sure you have write permissions for the directory)
+ save_path = "D:/Dropbox/FieldPrism/fieldprism/yolov5/weights/fieldprism_v_1_0.pt"
+ scripted_module.save(save_path)
+ print(f"Saved TorchScript model to {save_path}")
+except Exception as e:
+ print(f"An error occurred during the scripting. Error: {str(e)}")
diff --git a/vouchervision/component_detector/val.py b/vouchervision/component_detector/val.py
new file mode 100644
index 0000000000000000000000000000000000000000..58113f016a5825df59e630d5efc34f4756970362
--- /dev/null
+++ b/vouchervision/component_detector/val.py
@@ -0,0 +1,396 @@
+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Validate a trained YOLOv5 model accuracy on a custom dataset
+
+Usage:
+ $ python path/to/val.py --weights yolov5s.pt --data coco128.yaml --img 640
+
+Usage - formats:
+ $ python path/to/val.py --weights yolov5s.pt # PyTorch
+ yolov5s.torchscript # TorchScript
+ yolov5s.onnx # ONNX Runtime or OpenCV DNN with --dnn
+ yolov5s.xml # OpenVINO
+ yolov5s.engine # TensorRT
+ yolov5s.mlmodel # CoreML (macOS-only)
+ yolov5s_saved_model # TensorFlow SavedModel
+ yolov5s.pb # TensorFlow GraphDef
+ yolov5s.tflite # TensorFlow Lite
+ yolov5s_edgetpu.tflite # TensorFlow Edge TPU
+"""
+
+import argparse
+import json
+import os
+import sys
+from pathlib import Path
+from threading import Thread
+
+import numpy as np
+import torch
+from tqdm.auto import tqdm
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[0] # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+ sys.path.append(str(ROOT)) # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative
+
+from models.common import DetectMultiBackend
+from utils.callbacks import Callbacks
+from utils.datasets import create_dataloader
+from utils.general import (LOGGER, check_dataset, check_img_size, check_requirements, check_yaml,
+ coco80_to_coco91_class, colorstr, increment_path, non_max_suppression, print_args,
+ scale_coords, xywh2xyxy, xyxy2xywh)
+from utils.metrics import ConfusionMatrix, ap_per_class, box_iou
+from utils.plots import output_to_target, plot_images, plot_val_study
+from utils.torch_utils import select_device, time_sync
+
+
+def save_one_txt(predn, save_conf, shape, file):
+ # Save one txt result
+ gn = torch.tensor(shape)[[1, 0, 1, 0]] # normalization gain whwh
+ for *xyxy, conf, cls in predn.tolist():
+ xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh
+ line = (cls, *xywh, conf) if save_conf else (cls, *xywh) # label format
+ with open(file, 'a') as f:
+ f.write(('%g ' * len(line)).rstrip() % line + '\n')
+
+
+def save_one_json(predn, jdict, path, class_map):
+ # Save one JSON result {"image_id": 42, "category_id": 18, "bbox": [258.15, 41.29, 348.26, 243.78], "score": 0.236}
+ image_id = int(path.stem) if path.stem.isnumeric() else path.stem
+ box = xyxy2xywh(predn[:, :4]) # xywh
+ box[:, :2] -= box[:, 2:] / 2 # xy center to top-left corner
+ for p, b in zip(predn.tolist(), box.tolist()):
+ jdict.append({
+ 'image_id': image_id,
+ 'category_id': class_map[int(p[5])],
+ 'bbox': [round(x, 3) for x in b],
+ 'score': round(p[4], 5)})
+
+
+def process_batch(detections, labels, iouv):
+ """
+ Return correct predictions matrix. Both sets of boxes are in (x1, y1, x2, y2) format.
+ Arguments:
+ detections (Array[N, 6]), x1, y1, x2, y2, conf, class
+ labels (Array[M, 5]), class, x1, y1, x2, y2
+ Returns:
+ correct (Array[N, 10]), for 10 IoU levels
+ """
+ correct = torch.zeros(detections.shape[0], iouv.shape[0], dtype=torch.bool, device=iouv.device)
+ iou = box_iou(labels[:, 1:], detections[:, :4])
+ x = torch.where((iou >= iouv[0]) & (labels[:, 0:1] == detections[:, 5])) # IoU above threshold and classes match
+ if x[0].shape[0]:
+ matches = torch.cat((torch.stack(x, 1), iou[x[0], x[1]][:, None]), 1).cpu().numpy() # [label, detection, iou]
+ if x[0].shape[0] > 1:
+ matches = matches[matches[:, 2].argsort()[::-1]]
+ matches = matches[np.unique(matches[:, 1], return_index=True)[1]]
+ # matches = matches[matches[:, 2].argsort()[::-1]]
+ matches = matches[np.unique(matches[:, 0], return_index=True)[1]]
+ matches = torch.from_numpy(matches).to(iouv.device)
+ correct[matches[:, 1].long()] = matches[:, 2:3] >= iouv
+ return correct
+
+
+@torch.no_grad()
+def run(
+ data,
+ weights=None, # model.pt path(s)
+ batch_size=32, # batch size
+ imgsz=640, # inference size (pixels)
+ conf_thres=0.001, # confidence threshold
+ iou_thres=0.6, # NMS IoU threshold
+ task='val', # train, val, test, speed or study
+ device='', # cuda device, i.e. 0 or 0,1,2,3 or cpu
+ workers=8, # max dataloader workers (per RANK in DDP mode)
+ single_cls=False, # treat as single-class dataset
+ augment=False, # augmented inference
+ verbose=False, # verbose output
+ save_txt=False, # save results to *.txt
+ save_hybrid=False, # save label+prediction hybrid results to *.txt
+ save_conf=False, # save confidences in --save-txt labels
+ save_json=False, # save a COCO-JSON results file
+ project=ROOT / 'runs/val', # save to project/name
+ name='exp', # save to project/name
+ exist_ok=False, # existing project/name ok, do not increment
+ half=True, # use FP16 half-precision inference
+ dnn=False, # use OpenCV DNN for ONNX inference
+ model=None,
+ dataloader=None,
+ save_dir=Path(''),
+ plots=True,
+ callbacks=Callbacks(),
+ compute_loss=None,
+):
+ # Initialize/load model and set device
+ training = model is not None
+ if training: # called by train.py
+ device, pt, jit, engine = next(model.parameters()).device, True, False, False # get model device, PyTorch model
+ half &= device.type != 'cpu' # half precision only supported on CUDA
+ model.half() if half else model.float()
+ else: # called directly
+ device = select_device(device, batch_size=batch_size)
+
+ # Directories
+ save_dir = increment_path(Path(project) / name, exist_ok=exist_ok) # increment run
+ (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir
+
+ # Load model
+ model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half)
+ stride, pt, jit, engine = model.stride, model.pt, model.jit, model.engine
+ imgsz = check_img_size(imgsz, s=stride) # check image size
+ half = model.fp16 # FP16 supported on limited backends with CUDA
+ if engine:
+ batch_size = model.batch_size
+ else:
+ device = model.device
+ if not (pt or jit):
+ batch_size = 1 # export.py models default to batch-size 1
+ LOGGER.info(f'Forcing --batch-size 1 square inference (1,3,{imgsz},{imgsz}) for non-PyTorch models')
+
+ # Data
+ data = check_dataset(data) # check
+
+ # Configure
+ model.eval()
+ cuda = device.type != 'cpu'
+ is_coco = isinstance(data.get('val'), str) and data['val'].endswith(f'coco{os.sep}val2017.txt') # COCO dataset
+ nc = 1 if single_cls else int(data['nc']) # number of classes
+ iouv = torch.linspace(0.5, 0.95, 10, device=device) # iou vector for mAP@0.5:0.95
+ niou = iouv.numel()
+
+ # Dataloader
+ if not training:
+ if pt and not single_cls: # check --weights are trained on --data
+ ncm = model.model.nc
+ assert ncm == nc, f'{weights[0]} ({ncm} classes) trained on different --data than what you passed ({nc} ' \
+ f'classes). Pass correct combination of --weights and --data that are trained together.'
+ model.warmup(imgsz=(1 if pt else batch_size, 3, imgsz, imgsz)) # warmup
+ pad = 0.0 if task in ('speed', 'benchmark') else 0.5
+ rect = False if task == 'benchmark' else pt # square inference for benchmarks
+ task = task if task in ('train', 'val', 'test') else 'val' # path to train/val/test images
+ dataloader = create_dataloader(data[task],
+ imgsz,
+ batch_size,
+ stride,
+ single_cls,
+ pad=pad,
+ rect=rect,
+ workers=workers,
+ prefix=colorstr(f'{task}: '))[0]
+
+ seen = 0
+ confusion_matrix = ConfusionMatrix(nc=nc)
+ names = {k: v for k, v in enumerate(model.names if hasattr(model, 'names') else model.module.names)}
+ class_map = coco80_to_coco91_class() if is_coco else list(range(1000))
+ s = ('%20s' + '%11s' * 6) % ('Class', 'Images', 'Labels', 'P', 'R', 'mAP@.5', 'mAP@.5:.95')
+ dt, p, r, f1, mp, mr, map50, map = [0.0, 0.0, 0.0], 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0
+ loss = torch.zeros(3, device=device)
+ jdict, stats, ap, ap_class = [], [], [], []
+ callbacks.run('on_val_start')
+ pbar = tqdm(dataloader, desc=s, bar_format='{l_bar}{bar:10}{r_bar}{bar:-10b}') # progress bar
+ for batch_i, (im, targets, paths, shapes) in enumerate(pbar):
+ callbacks.run('on_val_batch_start')
+ t1 = time_sync()
+ if cuda:
+ im = im.to(device, non_blocking=True)
+ targets = targets.to(device)
+ im = im.half() if half else im.float() # uint8 to fp16/32
+ im /= 255 # 0 - 255 to 0.0 - 1.0
+ nb, _, height, width = im.shape # batch size, channels, height, width
+ t2 = time_sync()
+ dt[0] += t2 - t1
+
+ # Inference
+ out, train_out = model(im) if training else model(im, augment=augment, val=True) # inference, loss outputs
+ dt[1] += time_sync() - t2
+
+ # Loss
+ if compute_loss:
+ loss += compute_loss([x.float() for x in train_out], targets)[1] # box, obj, cls
+
+ # NMS
+ targets[:, 2:] *= torch.tensor((width, height, width, height), device=device) # to pixels
+ lb = [targets[targets[:, 0] == i, 1:] for i in range(nb)] if save_hybrid else [] # for autolabelling
+ t3 = time_sync()
+ out = non_max_suppression(out, conf_thres, iou_thres, labels=lb, multi_label=True, agnostic=single_cls)
+ dt[2] += time_sync() - t3
+
+ # Metrics
+ for si, pred in enumerate(out):
+ labels = targets[targets[:, 0] == si, 1:]
+ nl, npr = labels.shape[0], pred.shape[0] # number of labels, predictions
+ path, shape = Path(paths[si]), shapes[si][0]
+ correct = torch.zeros(npr, niou, dtype=torch.bool, device=device) # init
+ seen += 1
+
+ if npr == 0:
+ if nl:
+ stats.append((correct, *torch.zeros((3, 0), device=device)))
+ continue
+
+ # Predictions
+ if single_cls:
+ pred[:, 5] = 0
+ predn = pred.clone()
+ scale_coords(im[si].shape[1:], predn[:, :4], shape, shapes[si][1]) # native-space pred
+
+ # Evaluate
+ if nl:
+ tbox = xywh2xyxy(labels[:, 1:5]) # target boxes
+ scale_coords(im[si].shape[1:], tbox, shape, shapes[si][1]) # native-space labels
+ labelsn = torch.cat((labels[:, 0:1], tbox), 1) # native-space labels
+ correct = process_batch(predn, labelsn, iouv)
+ if plots:
+ confusion_matrix.process_batch(predn, labelsn)
+ stats.append((correct, pred[:, 4], pred[:, 5], labels[:, 0])) # (correct, conf, pcls, tcls)
+
+ # Save/log
+ if save_txt:
+ save_one_txt(predn, save_conf, shape, file=save_dir / 'labels' / (path.stem + '.txt'))
+ if save_json:
+ save_one_json(predn, jdict, path, class_map) # append to COCO-JSON dictionary
+ callbacks.run('on_val_image_end', pred, predn, path, names, im[si])
+
+ # Plot images
+ if plots and batch_i < 3:
+ f = save_dir / f'val_batch{batch_i}_labels.jpg' # labels
+ Thread(target=plot_images, args=(im, targets, paths, f, names), daemon=True).start()
+ f = save_dir / f'val_batch{batch_i}_pred.jpg' # predictions
+ Thread(target=plot_images, args=(im, output_to_target(out), paths, f, names), daemon=True).start()
+
+ callbacks.run('on_val_batch_end')
+
+ # Compute metrics
+ stats = [torch.cat(x, 0).cpu().numpy() for x in zip(*stats)] # to numpy
+ if len(stats) and stats[0].any():
+ tp, fp, p, r, f1, ap, ap_class = ap_per_class(*stats, plot=plots, save_dir=save_dir, names=names)
+ ap50, ap = ap[:, 0], ap.mean(1) # AP@0.5, AP@0.5:0.95
+ mp, mr, map50, map = p.mean(), r.mean(), ap50.mean(), ap.mean()
+ nt = np.bincount(stats[3].astype(np.int64), minlength=nc) # number of targets per class
+ else:
+ nt = torch.zeros(1)
+
+ # Print results
+ pf = '%20s' + '%11i' * 2 + '%11.3g' * 4 # print format
+ LOGGER.info(pf % ('all', seen, nt.sum(), mp, mr, map50, map))
+
+ # Print results per class
+ if (verbose or (nc < 50 and not training)) and nc > 1 and len(stats):
+ for i, c in enumerate(ap_class):
+ LOGGER.info(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))
+
+ # Print speeds
+ t = tuple(x / seen * 1E3 for x in dt) # speeds per image
+ if not training:
+ shape = (batch_size, 3, imgsz, imgsz)
+ LOGGER.info(f'Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {shape}' % t)
+
+ # Plots
+ if plots:
+ confusion_matrix.plot(save_dir=save_dir, names=list(names.values()))
+ callbacks.run('on_val_end')
+
+ # Save JSON
+ if save_json and len(jdict):
+ w = Path(weights[0] if isinstance(weights, list) else weights).stem if weights is not None else '' # weights
+ anno_json = str(Path(data.get('path', '../coco')) / 'annotations/instances_val2017.json') # annotations json
+ pred_json = str(save_dir / f"{w}_predictions.json") # predictions json
+ LOGGER.info(f'\nEvaluating pycocotools mAP... saving {pred_json}...')
+ with open(pred_json, 'w') as f:
+ json.dump(jdict, f)
+
+ try: # https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocoEvalDemo.ipynb
+ check_requirements(['pycocotools'])
+ from pycocotools.coco import COCO
+ from pycocotools.cocoeval import COCOeval
+
+ anno = COCO(anno_json) # init annotations api
+ pred = anno.loadRes(pred_json) # init predictions api
+ eval = COCOeval(anno, pred, 'bbox')
+ if is_coco:
+ eval.params.imgIds = [int(Path(x).stem) for x in dataloader.dataset.im_files] # image IDs to evaluate
+ eval.evaluate()
+ eval.accumulate()
+ eval.summarize()
+ map, map50 = eval.stats[:2] # update results (mAP@0.5:0.95, mAP@0.5)
+ except Exception as e:
+ LOGGER.info(f'pycocotools unable to run: {e}')
+
+ # Return results
+ model.float() # for training
+ if not training:
+ s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
+ LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}{s}")
+ maps = np.zeros(nc) + map
+ for i, c in enumerate(ap_class):
+ maps[c] = ap[i]
+ return (mp, mr, map50, map, *(loss.cpu() / len(dataloader)).tolist()), maps, t
+
+
+def parse_opt():
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='dataset.yaml path')
+ parser.add_argument('--weights', nargs='+', type=str, default=ROOT / 'yolov5s.pt', help='model.pt path(s)')
+ parser.add_argument('--batch-size', type=int, default=32, help='batch size')
+ parser.add_argument('--imgsz', '--img', '--img-size', type=int, default=640, help='inference size (pixels)')
+ parser.add_argument('--conf-thres', type=float, default=0.001, help='confidence threshold')
+ parser.add_argument('--iou-thres', type=float, default=0.6, help='NMS IoU threshold')
+ parser.add_argument('--task', default='val', help='train, val, test, speed or study')
+ parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+ parser.add_argument('--workers', type=int, default=8, help='max dataloader workers (per RANK in DDP mode)')
+ parser.add_argument('--single-cls', action='store_true', help='treat as single-class dataset')
+ parser.add_argument('--augment', action='store_true', help='augmented inference')
+ parser.add_argument('--verbose', action='store_true', help='report mAP by class')
+ parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
+ parser.add_argument('--save-hybrid', action='store_true', help='save label+prediction hybrid results to *.txt')
+ parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
+ parser.add_argument('--save-json', action='store_true', help='save a COCO-JSON results file')
+ parser.add_argument('--project', default=ROOT / 'runs/val', help='save to project/name')
+ parser.add_argument('--name', default='exp', help='save to project/name')
+ parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+ parser.add_argument('--half', action='store_true', help='use FP16 half-precision inference')
+ parser.add_argument('--dnn', action='store_true', help='use OpenCV DNN for ONNX inference')
+ opt = parser.parse_args()
+ opt.data = check_yaml(opt.data) # check YAML
+ opt.save_json |= opt.data.endswith('coco.yaml')
+ opt.save_txt |= opt.save_hybrid
+ print_args(vars(opt))
+ return opt
+
+
+def main(opt):
+ check_requirements(requirements=ROOT / 'requirements.txt', exclude=('tensorboard', 'thop'))
+
+ if opt.task in ('train', 'val', 'test'): # run normally
+ if opt.conf_thres > 0.001: # https://github.com/ultralytics/yolov5/issues/1466
+ LOGGER.info(f'WARNING: confidence threshold {opt.conf_thres} >> 0.001 will produce invalid mAP values.')
+ run(**vars(opt))
+
+ else:
+ weights = opt.weights if isinstance(opt.weights, list) else [opt.weights]
+ opt.half = True # FP16 for fastest results
+ if opt.task == 'speed': # speed benchmarks
+ # python val.py --task speed --data coco.yaml --batch 1 --weights yolov5n.pt yolov5s.pt...
+ opt.conf_thres, opt.iou_thres, opt.save_json = 0.25, 0.45, False
+ for opt.weights in weights:
+ run(**vars(opt), plots=False)
+
+ elif opt.task == 'study': # speed vs mAP benchmarks
+ # python val.py --task study --data coco.yaml --iou 0.7 --weights yolov5n.pt yolov5s.pt...
+ for opt.weights in weights:
+ f = f'study_{Path(opt.data).stem}_{Path(opt.weights).stem}.txt' # filename to save to
+ x, y = list(range(256, 1536 + 128, 128)), [] # x axis (image sizes), y axis
+ for opt.imgsz in x: # img-size
+ LOGGER.info(f'\nRunning {f} --imgsz {opt.imgsz}...')
+ r, _, t = run(**vars(opt), plots=False)
+ y.append(r + t) # results and times
+ np.savetxt(f, y, fmt='%10.4g') # save
+ os.system('zip -r study.zip study_*.txt')
+ plot_val_study(x=x) # plot
+
+
+if __name__ == "__main__":
+ opt = parse_opt()
+ main(opt)
diff --git a/vouchervision/data_project.py b/vouchervision/data_project.py
new file mode 100644
index 0000000000000000000000000000000000000000..1bcf1449e9c6854f8685f7f294c1e8e5c64a1c4b
--- /dev/null
+++ b/vouchervision/data_project.py
@@ -0,0 +1,326 @@
+import os, sys, inspect, shutil, warnings
+from dataclasses import dataclass, field
+import pandas as pd
+currentdir = os.path.dirname(os.path.dirname(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+sys.path.append(currentdir)
+from vouchervision.general_utils import import_csv, import_tsv, bcolors
+from vouchervision.general_utils import Print_Verbose, print_main_warn, print_main_success, make_file_names_valid, make_images_in_dir_vertical
+from vouchervision.utils_GBIF import generate_image_filename
+from vouchervision.download_from_GBIF_all_images_in_file import download_all_images_from_GBIF_LM2
+from PIL import Image
+from tqdm import tqdm
+from pathlib import Path
+
+@dataclass
+class Project_Info():
+ batch_size: int = 50
+
+ image_location: str = ''
+
+ dir_images: str = ''
+
+ project_data: object = field(init=False)
+ project_data_list: object = field(init=False)
+
+ path_csv_combined: str = ''
+ path_csv_occ: str = ''
+ path_csv_img: str = ''
+ csv_combined: str = ''
+ csv_occ: str = ''
+ csv_img: str = ''
+
+ Dirs: object = field(init=False)
+
+ has_valid_images: bool = True
+
+ def __init__(self, cfg, logger, dir_home, Dirs) -> None:
+ self.Dirs = Dirs
+ logger.name = 'Project Info'
+ logger.info("Gathering Images and Image Metadata")
+
+ self.batch_size = cfg['leafmachine']['project']['batch_size']
+
+ self.image_location = cfg['leafmachine']['project']['image_location']
+
+ self.valid_extensions = ['.jpg', '.JPG', '.jpeg', '.JPEG', '.png', '.PNG', '.bmp', '.BMP', '.tif', '.TIF', '.tiff', '.TIFF']
+
+ self.copy_images_to_project_dir(cfg['leafmachine']['project']['dir_images_local'], Dirs)
+
+ self.make_file_names_custom(Dirs.save_original, cfg, Dirs)
+
+ # If project is local, expect:
+ # dir with images
+ # path to images.csv
+ # path to occ.csv
+ # OR path to combined.csv
+ # if self.image_location in ['local','l','L','Local']:
+ self.__import_local_files(cfg, logger, Dirs)
+
+ # If project is GBIF, expect:
+ # Darwin Core Images (or multimedia.txt) and Occurrences file pair, either .txt or .csv
+ # elif self.image_location in ['GBIF','g','G','gbif']:
+ # self.__import_GBIF_files_post_download(cfg, logger, dir_home)
+
+ self.__make_project_dict(Dirs) #, self.batch_size)
+
+ # Make sure image file names are legal
+ make_file_names_valid(Dirs.save_original, cfg)
+
+ # Make all images vertical
+ make_images_in_dir_vertical(Dirs.save_original, cfg)
+
+
+
+ @property
+ def has_valid_images(self):
+ return self.check_for_images()
+
+ @property
+ def file_ext(self):
+ return f"{['.jpg', '.JPG', '.jpeg', '.JPEG', '.png', '.PNG', '.bmp', '.BMP', '.tif', '.TIF', '.tiff', '.TIFF']}"
+
+ def check_for_images(self):
+ for filename in os.listdir(self.dir_images):
+ if filename.endswith(tuple(self.valid_extensions)):
+ return True
+ return False
+
+ def remove_non_numbers(self, s):
+ return ''.join([char for char in s if char.isdigit()])
+
+ def copy_images_to_project_dir(self, dir_images, Dirs):
+ n_total = len(os.listdir(dir_images))
+ for file in tqdm(os.listdir(dir_images), desc=f'{bcolors.HEADER} Copying images to working directory{bcolors.ENDC}',colour="white",position=0,total = n_total):
+ # Copy og image to new dir
+ # Copied image will be used for all downstream applications
+ source = os.path.join(dir_images, file)
+ destination = os.path.join(Dirs.save_original, file)
+ shutil.copy(source, destination)
+
+ def make_file_names_custom(self, dir_images, cfg, Dirs):
+ n_total = len(os.listdir(dir_images))
+ for file in tqdm(os.listdir(dir_images), desc=f'{bcolors.HEADER} Creating Catalog Number from file name{bcolors.ENDC}',colour="green",position=0,total = n_total):
+ # Copy og image to new dir
+ # Copied image will be used for all downstream applications
+ # source = os.path.join(dir_images, file)
+ # destination = os.path.join(Dirs.save_original, file)
+ # shutil.copy(source, destination)
+
+ if cfg['leafmachine']['project']['catalog_numerical_only'] or cfg['leafmachine']['project']['prefix_removal'] or cfg['leafmachine']['project']['suffix_removal']:
+ name = Path(file).stem
+ ext = Path(file).suffix
+ if cfg['leafmachine']['project']['prefix_removal']:
+ name_cleaned = name.replace(cfg['leafmachine']['project']['prefix_removal'], "")
+ if cfg['leafmachine']['project']['suffix_removal']:
+ name_cleaned = name.replace(cfg['leafmachine']['project']['suffix_removal'], "")
+ if cfg['leafmachine']['project']['catalog_numerical_only']:
+ name_cleaned = self.remove_non_numbers(name)
+ name_new = ''.join([name_cleaned,ext])
+ i = 0
+ try:
+ os.rename(os.path.join(dir_images,file), os.path.join(dir_images,name_new))
+ except:
+ warnings.warn("WARNING: duplicate file names will result given the current selections for 'prefix_removal', 'suffix_removal', or 'catalog_numerical_only'. Change them before continuing.")
+ warnings.warn("The affected file name has not been changed.")
+
+
+
+ def __create_combined_csv(self):
+ self.csv_img = self.csv_img.rename(columns={"gbifID": "gbifID_images"})
+ self.csv_img = self.csv_img.rename(columns={"identifier": "url"})
+ # print(self.csv_img.head(5))
+
+ combined = pd.merge(self.csv_img, self.csv_occ, left_on='gbifID_images', right_on='gbifID')
+ # print(combined.head(5))
+ names_list = combined.apply(generate_image_filename, axis=1, result_type='expand')
+ # print(names_list.head(5))
+ # Select columns 7, 0, 1
+ selected_columns = names_list.iloc[:,[7,0,1]]
+ # Rename columns
+ selected_columns.columns = ['fullname','filename_image','filename_image_jpg']
+ # print(selected_columns.head(5))
+ self.csv_combined = pd.concat([selected_columns, combined], axis=1)
+ # print(self.csv_combined.head(5))
+ new_name = ''.join(['combined_', os.path.basename(self.path_csv_occ).split('.')[0], '_', os.path.basename(self.path_csv_img).split('.')[0], '.csv'])
+ self.path_csv_combined = os.path.join(os.path.dirname(self.path_csv_occ), new_name)
+ self.csv_combined.to_csv(self.path_csv_combined, mode='w', header=True, index=False)
+ return self.path_csv_combined
+
+ def __import_local_files(self, cfg, logger, Dirs):
+ # Images
+ if cfg['leafmachine']['project']['dir_images_local'] is None:
+ self.dir_images = None
+ else:
+ self.dir_images = Dirs.save_original
+
+ # CSV import
+ # Combined
+ try:
+ if cfg['leafmachine']['project']['path_combined_csv_local'] is None:
+ self.csv_combined = None
+ self.path_csv_combined = None
+ else:
+ self.path_csv_combined = cfg['leafmachine']['project']['path_combined_csv_local']
+ self.csv_combined = import_csv(self.path_csv_combined)
+ # Occurrence
+ if cfg['leafmachine']['project']['path_occurrence_csv_local'] is None:
+ self.csv_occ = None
+ self.path_csv_occ = None
+ else:
+ self.path_csv_occ = cfg['leafmachine']['project']['path_occurrence_csv_local']
+ self.csv_occ = import_csv(self.path_csv_occ)
+ # Images/metadata
+ if cfg['leafmachine']['project']['path_images_csv_local'] is None:
+ self.path_csv_img = None
+ self.path_csv_img = None
+ else:
+ self.path_csv_img = cfg['leafmachine']['project']['path_images_csv_local']
+ self.csv_img = import_csv(self.path_csv_img)
+
+ # Create combined if it's missing
+ if self.csv_combined is None:
+ if cfg['leafmachine']['project']['path_combined_csv_local'] is not None:
+ # Print_Verbose(cfg, 2, 'Combined CSV file not provided, creating it now...').bold()
+ logger.info('Combined CSV file not provided, creating it now...')
+ location = self.__create_combined_csv()
+ # Print_Verbose(cfg, 2, ''.join(['Combined CSV --> ',location])).green()
+ logger.info(''.join(['Combined CSV --> ',location]))
+
+ else:
+ # Print_Verbose(cfg, 2, 'Combined CSV file not available or provided. Skipped record import.').bold()
+ logger.info('Combined CSV file not available or provided. Skipped record import.')
+ else:
+ # Print_Verbose(cfg, 2, ''.join(['Combined CSV --> ',self.path_csv_combined])).green()
+ logger.info(''.join(['Combined CSV --> ',self.path_csv_combined]))
+ except:
+ pass
+
+ # Print_Verbose(cfg, 2, ''.join(['Image Directory --> ',self.dir_images])).green()
+ logger.info(''.join(['Image Directory --> ',Dirs.save_original]))
+
+
+
+ # def __import_GBIF_files_post_download(self, cfg, logger, dir_home):
+ # # Download the images from GBIF
+ # # This pulls from /LeafMachine2/configs/config_download_from_GBIF_all_images_in_file or filter
+ # print_main_warn('Downloading Images from GBIF...')
+ # logger.info('Downloading Images from GBIF...')
+ # self.cfg_images = download_all_images_from_GBIF_LM2(dir_home, cfg['leafmachine']['project']['GBIF_mode'])
+ # self.dir_images = self.cfg_images['dir_destination_images']
+ # self.path_csv = self.cfg_images['dir_destination_csv']
+ # print_main_success(''.join(['Images saved to --> ',self.dir_images]))
+ # logger.info(''.join(['Images saved to --> ',self.dir_images]))
+
+
+ # self.path_csv_combined = os.path.join(self.path_csv, self.cfg_images['filename_combined'])
+ # self.path_csv_occ = os.path.join(self.path_csv, self.cfg_images['filename_occ'])
+ # self.path_csv_img = os.path.join(self.path_csv, self.cfg_images['filename_img'])
+
+ # if 'txt' in (self.cfg_images['filename_occ'].split('.')[1] or self.cfg_images['filename_img'].split('.')[1]):
+ # self.csv_combined = import_tsv(self.path_csv_combined)
+ # # self.csv_occ = import_tsv(self.path_csv_occ)
+ # # self.csv_img = import_tsv(self.path_csv_img)
+ # else:
+ # self.csv_combined = import_csv(self.path_csv_combined)
+ # # self.csv_occ = import_csv(self.path_csv_occ)
+ # # self.csv_img = import_csv(self.path_csv_img)
+
+ def process_in_batches(self, cfg):
+ batch_size = cfg['leafmachine']['project']['batch_size']
+ self.project_data_list = []
+ keys = list(self.project_data.keys())
+ num_batches = len(keys) // batch_size + 1
+ for i in range(num_batches):
+ start = i * batch_size
+ end = (i + 1) * batch_size
+ batch_keys = keys[start:end]
+ batch = {key: self.project_data[key] for key in batch_keys}
+ self.project_data_list.append(batch)
+ return num_batches, len(self.project_data)
+
+ # Original
+ '''def __make_project_dict(self):
+ self.project_data = {}
+ for img in os.listdir(self.dir_images):
+ if (img.endswith(".jpg") or img.endswith(".jpeg")):
+ img_name = str(img.split('.')[0])
+ self.project_data[img_name] = {}
+ '''
+ # def __make_project_dict(self): # This DELETES the invalid file, not safe
+ # self.project_data = {}
+ # for img in os.listdir(self.dir_images):
+ # img_split, ext = os.path.splitext(img)
+ # if ext.lower() in self.valid_extensions:
+ # with Image.open(os.path.join(self.dir_images, img)) as im:
+ # _, ext = os.path.splitext(img)
+ # if ext != '.jpg':
+ # im = im.convert('RGB')
+ # im.save(os.path.join(self.dir_images, img_split) + '.jpg', quality=100)
+ # img += '.jpg'
+ # os.remove(os.path.join(self.dir_images, ''.join([img_split, ext])))
+ # img_name = os.path.splitext(img)[0]
+ # self.project_data[img_split] = {}
+
+ def __make_project_dict(self, Dirs):
+ self.project_data = {}
+ invalid_dir = None
+
+ for img in os.listdir(Dirs.save_original):
+ img_split, ext = os.path.splitext(img)
+ if ext in self.valid_extensions:
+ with Image.open(os.path.join(Dirs.save_original, img)) as im:
+ _, ext = os.path.splitext(img)
+ if ext not in ['.jpg']:
+ im = im.convert('RGB')
+ new_img_name = ''.join([img_split, '.jpg'])
+ im.save(os.path.join(Dirs.save_original, new_img_name), quality=100)
+ self.project_data[img_split] = {}
+
+ # move the original file to the INVALID_FILE directory
+ if invalid_dir is None:
+ invalid_dir = os.path.join(os.path.dirname(Dirs.save_original), 'INVALID_FILES')
+ os.makedirs(invalid_dir, exist_ok=True)
+
+ # skip if the file already exists in the INVALID_FILE directory
+ if not os.path.exists(os.path.join(invalid_dir, img)):
+ shutil.move(os.path.join(Dirs.save_original, img), os.path.join(invalid_dir, img))
+
+ img = new_img_name
+ img_name = os.path.splitext(img)[0]
+ self.project_data[img_split] = {}
+ else:
+ # if the file has an invalid extension, move it to the INVALID_FILE directory
+ if invalid_dir is None:
+ invalid_dir = os.path.join(os.path.dirname(Dirs.save_original), 'INVALID_FILES')
+ os.makedirs(invalid_dir, exist_ok=True)
+
+ # skip if the file already exists in the INVALID_FILE directory
+ if not os.path.exists(os.path.join(invalid_dir, img)):
+ shutil.move(os.path.join(Dirs.save_original, img), os.path.join(invalid_dir, img))
+
+ def add_records_to_project_dict(self):
+ for img in os.listdir(self.Dirs.save_original):
+ if (img.endswith(".jpg") or img.endswith(".jpeg")):
+ img_name = str(img.split('.')[0])
+ try:
+ self.project_data[img_name]['GBIF_Record'] = self.__get_data_from_combined(img_name)
+ except:
+ self.project_data[img_name]['GBIF_Record'] = None
+
+ def __get_data_from_combined(self, img_name):
+ df = pd.DataFrame(self.csv_combined)
+ row = df[df['filename_image'] == img_name].head(1).to_dict()
+ return row
+
+
+class Project_Stats():
+ specimens = 0
+
+ rulers = 0
+
+
+ def __init__(self, cfg, logger, dir_home) -> None:
+ logger.name = 'Project Info'
+ logger.info("Gathering Images and Image Metadata")
\ No newline at end of file
diff --git a/vouchervision/directory_structure_VV.py b/vouchervision/directory_structure_VV.py
new file mode 100644
index 0000000000000000000000000000000000000000..4391a2a0634b0a2bb868aaaed776dd7fe6684350
--- /dev/null
+++ b/vouchervision/directory_structure_VV.py
@@ -0,0 +1,120 @@
+import os, pathlib, sys, inspect
+from dataclasses import dataclass, field
+currentdir = os.path.dirname(os.path.dirname(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+sys.path.append(currentdir)
+from vouchervision.general_utils import validate_dir, get_datetime
+
+@dataclass
+class Dir_Structure():
+ # Home
+ run_name: str = ''
+ dir_home: str = ''
+ dir_project: str = ''
+
+ # Processing dirs
+ path_archival_components: str = ''
+ path_config_file: str = ''
+
+ ruler_info: str = ''
+ ruler_overlay: str = ''
+ ruler_processed: str = ''
+ ruler_data: str = ''
+ ruler_class_overlay: str = ''
+ ruler_validation_summary: str = ''
+ ruler_validation: str = ''
+
+ save_per_image: str = ''
+ save_per_annotation_class: str = ''
+ binarize_labels: str = ''
+
+ # logging
+ path_log: str = ''
+
+ def __init__(self, cfg) -> None:
+ # Home
+ self.run_name = cfg['leafmachine']['project']['run_name']
+ self.dir_home = cfg['leafmachine']['project']['dir_output']
+ self.dir_project = os.path.join(self.dir_home,self.run_name)
+ validate_dir(self.dir_home)
+ self.__add_time_to_existing_project_dir()
+ validate_dir(self.dir_project)
+
+ # Processing dirs
+ self.path_archival_components = os.path.join(self.dir_project,'Archival_Components')
+ validate_dir(self.path_archival_components)
+
+ self.path_config_file = os.path.join(self.dir_project,'Config_File')
+ validate_dir(self.path_config_file)
+
+ self.path_cost = os.path.join(self.dir_project,'Cost')
+ validate_dir(self.path_cost)
+
+ # Logging
+ self.path_log = os.path.join(self.dir_project,'Logs')
+ validate_dir(self.path_log)
+
+ # self.custom_overlay_pdfs = os.path.join(self.dir_project,'Summary','Custom_Overlay_PDFs')
+ # self.custom_overlay_images = os.path.join(self.dir_project,'Summary','Custom_Overlay_Images')
+
+ ###
+ # self.custom_overlay_pdfs = os.path.join(self.dir_project,'Summary','Custom_Overlay_PDFs')
+ # if cfg['leafmachine']['overlay']['save_overlay_to_pdf']:
+ # validate_dir(self.custom_overlay_pdfs)
+
+ # self.custom_overlay_images = os.path.join(self.dir_project,'Summary','Custom_Overlay_Images')
+ # if cfg['leafmachine']['overlay']['save_overlay_to_jpgs']:
+ # validate_dir(self.custom_overlay_images)
+
+ ### Rulers
+ # self.ruler_info = os.path.join(self.dir_project,'Archival_Components','Ruler_Info')
+ # self.ruler_validation_summary = os.path.join(self.dir_project,'Archival_Components','Ruler_Info', 'Ruler_Validation_Summary')
+ # self.ruler_validation = os.path.join(self.dir_project,'Archival_Components','Ruler_Info', 'Ruler_Validation')
+ # self.ruler_processed = os.path.join(self.dir_project,'Archival_Components','Ruler_Info', 'Ruler_Processed')
+ # validate_dir(self.ruler_info)
+
+
+ validate_dir(os.path.join(self.path_archival_components, 'JSON'))
+ validate_dir(os.path.join(self.path_archival_components, 'labels'))
+
+ ### Data
+ self.transcription = os.path.join(self.dir_project,'Transcription')
+ validate_dir(self.transcription)
+ self.transcription_ind = os.path.join(self.dir_project,'Transcription','Individual')
+ validate_dir(self.transcription_ind)
+ # self.transcription_ind_helper = os.path.join(self.dir_project,'Transcription','Individual_Helper_Content')
+ # validate_dir(self.transcription_ind_helper)
+ self.transcription_ind_OCR = os.path.join(self.dir_project,'Transcription','Individual_OCR')
+ validate_dir(self.transcription_ind_OCR)
+ self.transcription_ind_OCR_bounds = os.path.join(self.dir_project,'Transcription','Individual_OCR_Bounds')
+ validate_dir(self.transcription_ind_OCR_bounds)
+ self.transcription_ind_OCR_helper = os.path.join(self.dir_project,'Transcription','Individual_OCR_Helper')
+ validate_dir(self.transcription_ind_OCR_helper)
+
+ self.save_original = os.path.join(self.dir_project,'Original_Images')
+ validate_dir(self.save_original)
+
+ self.save_per_image = os.path.join(self.dir_project,'Cropped_Images', 'By_Image')
+ self.save_per_annotation_class = os.path.join(self.dir_project,'Cropped_Images', 'By_Class')
+ self.save_per_annotation_class = os.path.join(self.dir_project,'Cropped_Images', 'By_Class')
+ if cfg['leafmachine']['cropped_components']['save_per_image']:
+ validate_dir(self.save_per_image)
+ if cfg['leafmachine']['cropped_components']['save_per_annotation_class']:
+ validate_dir(self.save_per_annotation_class)
+ if cfg['leafmachine']['cropped_components']['binarize_labels']:
+ validate_dir(self.save_per_annotation_class)
+ # self.binarize_labels = os.path.join(self.dir_project,'Cropped_Images', 'By_Class','label_binary')
+ # validate_dir(self.binarize_labels)
+
+ def __add_time_to_existing_project_dir(self) -> None:
+ path = pathlib.Path(self.dir_project)
+ if path.exists():
+ now = get_datetime()
+ path = path.with_name(path.name + "_" + now)
+ self.run_name = path.name
+ path.mkdir()
+ self.dir_project = path
+ else:
+ path.mkdir()
+ self.dir_project = path
\ No newline at end of file
diff --git a/vouchervision/download_from_GBIF_all_images_in_file.py b/vouchervision/download_from_GBIF_all_images_in_file.py
new file mode 100644
index 0000000000000000000000000000000000000000..573dae2844dfdcd5c7c2be61bb42e9104ecf23d1
--- /dev/null
+++ b/vouchervision/download_from_GBIF_all_images_in_file.py
@@ -0,0 +1,60 @@
+import os, inspect, sys
+currentdir = os.path.dirname(os.path.dirname(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+sys.path.append(currentdir)
+
+from utils_GBIF import download_all_images_in_images_csv, get_cfg_from_full_path, download_all_images_in_images_csv_multiDirs
+
+
+'''
+This script attempts to download all images (every row) that are in the provided images.csv file.
+This means that you have either pruned an images.csv file or that you wish to download every possible image.
+
+*** USAGE ***
+Use the config_download_from_GBIF_all_images_in_file.yml file to set all parameters, then run this script
+
+images.csv files used here should be in a standard Darwin Core (DWC) format.
+
+This script is meant for Darwin Core files retrieved from GBIF.org
+ *** Note: not all DWC files have the same column names
+ *** Note: This script should adjust to slightly different column names, but errors may occur with non-GBIF files
+
+*** NOTES ***
+There are different scripts in the LeafMachine2/leafmachine2/machine folder to download images from:
+ 1) non-GBIF sources i.e. SEINet, SERNEC
+ 2) provide a list of species/genera/families and retrive all available images from GBIF
+ 3) provide a list of species/genera/families and retrive a custom set of images that are available on GBIF
+'''
+def download_all_images_from_GBIF_LM2(dir_LM2, mode):
+ dir_current_config = os.path.join(dir_LM2,'configs')
+ if mode in ['all','All','ALL','a','A']:
+ path_cfg = os.path.join(dir_current_config,'config_download_from_GBIF_all_images_in_file.yml')
+ elif mode in ['filter','Filter','FILTER','f','F']:
+ path_cfg = os.path.join(dir_current_config,'config_download_from_GBIF_all_images_in_filter.yml')
+ cfg = get_cfg_from_full_path(path_cfg)
+
+ # Run Download
+ download_all_images_in_images_csv(cfg)
+ return cfg
+
+if __name__ == '__main__':
+ opt = 'single'
+
+ if opt == 'single':
+ dir_LM2 = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+ dir_current_config = os.path.join(dir_LM2,'configs')
+ path_cfg = os.path.join(dir_current_config,'config_download_from_GBIF_all_images_in_file.yml')
+ cfg = get_cfg_from_full_path(path_cfg)
+
+ # Run Download
+ download_all_images_in_images_csv(cfg)
+
+ elif opt == 'multi':
+ dir_LM2 = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+ dir_current_config = os.path.join(dir_LM2,'configs')
+ path_cfg = os.path.join(dir_current_config,'config_download_from_GBIF_all_images_in_filter_multiDirs.yml')
+ cfg = get_cfg_from_full_path(path_cfg)
+
+ # Run Download
+ download_all_images_in_images_csv_multiDirs(cfg)
\ No newline at end of file
diff --git a/vouchervision/embed_occ.py b/vouchervision/embed_occ.py
new file mode 100644
index 0000000000000000000000000000000000000000..4eb07156727482c34659a41a928f78e30cb2ca23
--- /dev/null
+++ b/vouchervision/embed_occ.py
@@ -0,0 +1,178 @@
+import openai
+import os
+import sys
+import inspect
+from tqdm import tqdm
+import pandas as pd
+import numpy as np
+from sklearn.metrics.pairwise import cosine_similarity
+import json
+import gradio as gr
+
+currentdir = os.path.dirname(os.path.abspath(
+ inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+from vouchervision.general_utils import get_cfg_from_full_path
+from prompts import PROMPT_UMICH_skeleton_all_asia
+from LLM_chatGPT_3_5 import num_tokens_from_string, OCR_to_dict
+
+'''
+This generates OpenAI embedding. These are no longer used by VoucherVision.
+We have transitioned to "hkunlp/instructor-xl"
+
+Please see: https://huggingface.co/hkunlp/instructor-xl
+
+This file has some experimentation code that can be helpful to reference,
+but is no relevant to VoucherVision.
+'''
+
+class GenerateEmbeddings:
+ def __init__(self, file_occ, file_name, dir_out="D:/D_Desktop/embedding"):
+ self.file_occ = file_occ
+ self.file_name = file_name
+ self.dir_out = dir_out
+
+ self.SEP = '!!'
+
+ # Set API key
+ dir_home = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
+ path_cfg_private = os.path.join(dir_home, 'PRIVATE_DATA.yaml')
+ cfg_private = get_cfg_from_full_path(path_cfg_private)
+ openai.api_key = cfg_private['openai']['openai_api_key']
+
+ def generate(self):
+ # Read CSV file
+ df = pd.read_csv(self.file_occ, sep='\t',
+ on_bad_lines='skip', dtype=str, low_memory=False)
+
+ # Extract headers separately
+ dwc_headers = df.columns.tolist()
+
+ # Combine columns into a single string separated by commas
+ df['combined'] = df.apply(
+ lambda row: self.SEP.join(row.values.astype(str)), axis=1)
+
+ # Wrap the get_embedding function call with tqdm progress bar
+ tqdm.pandas(desc="Generating embeddings")
+ df['ada_embedding'] = df.combined.progress_apply(
+ lambda x: self.get_embedding(x, model='text-embedding-ada-002'))
+
+ # Save to output CSV
+ output_file = os.path.join(
+ self.dir_out, f'embedded_dwc__{self.file_name}.csv')
+ df[['combined', 'ada_embedding']].to_csv(output_file, index=False)
+
+ # Save headers to a separate CSV file
+ headers_file = os.path.join(
+ self.dir_out, f'dwc_headers__{self.file_name}.csv')
+ with open(headers_file, 'w') as f:
+ f.write('\n'.join(dwc_headers))
+
+ return output_file
+
+ def get_embedding(self, text, model="text-embedding-ada-002"):
+ text = text.replace("\n", " ")
+ return openai.Embedding.create(input=[text], model=model)['data'][0]['embedding']
+
+ def load_embedded_csv(self, csv_path):
+ df = pd.read_csv(csv_path)
+ df['ada_embedding'] = df.ada_embedding.apply(eval).apply(np.array)
+
+ headers_file = os.path.join(
+ self.dir_out, f'dwc_headers__{self.file_name}.csv')
+ with open(headers_file, 'r') as f:
+ dwc_headers = f.read().splitlines()
+
+ return df, dwc_headers
+
+ def search_rows(self, dwc_headers, df, query, n=3, pprint=True):
+ query_embedding = self.get_embedding(
+ query, model="text-embedding-ada-002")
+ df["similarity"] = df.ada_embedding.apply(
+ lambda x: cosine_similarity([x], [query_embedding])[0][0])
+
+ results = df.sort_values("similarity", ascending=False).head(n)
+
+ if pprint:
+ for i in range(n):
+ row = results.iloc[i]
+ df_split = pd.DataFrame(
+ [row.combined.split(self.SEP)], columns=dwc_headers)
+ df_clean = df_split.replace(
+ 'nan', np.nan).dropna(axis=1, how='any')
+ # Convert df_clean to a dictionary
+ row_dict = df_clean.to_dict(orient='records')[0]
+
+ # Convert dictionary to a long literal string
+ row_string = json.dumps(row_dict)
+
+ print(row_string)
+ # print(df_clean)
+ # print(df_clean.to_string(index=False))
+
+ nt = num_tokens_from_string(row_string, "cl100k_base")
+ print(nt)
+
+ return results
+
+
+def create_embeddings(file_occ, file_name, dir_out):
+ # Instantiate and generate embeddings
+ embedder = GenerateEmbeddings(file_occ, file_name, dir_out)
+ output_file = embedder.generate()
+
+
+def old_method(img_path):
+ set_rules = """1. Your job is to return a new dict based on the structure of the reference dict ref_dict and these are your rules.
+ 2. You must look at ref_dict and refactor the new text called OCR to match the same formatting.
+ 3. OCR contains unstructured text, use your knowledge to put the OCR text into the correct ref_dict column.
+ 4. If there is a field that does not have a direct proxy in the OCR text, you can fill it in based on your knowledge, but you cannot generate new information.
+ 5. The dict key is the column header, the value is the new text. The separator in the new text is '!!', which indicates a new element but not strictly a new column. Remove the '!!' separator before adding text to the new dict
+ 6. Never put text from the ref_dict values into the new dict, but you must use the headers from ref_dict.
+ 7. There cannot be duplicate dictionary fields.
+ 8. Only return the new dict, do not explain your answer."""
+
+ # 4. If there is a simple typo you should correct the spelling, but do not rephrase or rework the ORC text.
+ sample_text = """['gbifID', 'abstract', 'accessRights', 'accrualMethod', 'accrualPeriodicity', 'accrualPolicy', 'alternative', 'audience', 'available', 'bibliographicCitation', 'conformsTo', 'contributor', 'coverage', 'created', 'creator', 'date', 'dateAccepted', 'dateCopyrighted', 'dateSubmitted', 'description', 'educationLevel', 'extent', 'format', 'hasFormat', 'hasPart', 'hasVersion', 'identifier', 'instructionalMethod', 'isFormatOf', 'isPartOf', 'isReferencedBy', 'isReplacedBy', 'isRequiredBy', 'isVersionOf', 'issued', 'language', 'license', 'mediator', 'medium', 'modified', 'provenance', 'publisher', 'references', 'relation', 'replaces', 'requires', 'rights', 'rightsHolder', 'source', 'spatial', 'subject', 'tableOfContents', 'temporal', 'title', 'type', 'valid', 'institutionID', 'collectionID', 'datasetID', 'institutionCode', 'collectionCode', 'datasetName', 'ownerInstitutionCode', 'basisOfRecord', 'informationWithheld', 'dataGeneralizations', 'dynamicProperties', 'occurrenceID', 'catalogNumber', 'recordNumber', 'recordedBy', 'recordedByID', 'individualCount', 'organismQuantity', 'organismQuantityType', 'sex', 'lifeStage', 'reproductiveCondition', 'behavior', 'establishmentMeans', 'degreeOfEstablishment', 'pathway', 'georeferenceVerificationStatus', 'occurrenceStatus', 'preparations', 'disposition', 'associatedOccurrences', 'associatedReferences', 'associatedSequences', 'associatedTaxa', 'otherCatalogNumbers', 'occurrenceRemarks', 'organismID', 'organismName', 'organismScope', 'associatedOrganisms', 'previousIdentifications', 'organismRemarks', 'materialSampleID', 'eventID', 'parentEventID', 'fieldNumber', 'eventDate', 'eventTime', 'startDayOfYear', 'endDayOfYear', 'year', 'month', 'day', 'verbatimEventDate', 'habitat', 'samplingProtocol', 'sampleSizeValue', 'sampleSizeUnit', 'samplingEffort', 'fieldNotes', 'eventRemarks', 'locationID', 'higherGeographyID', 'higherGeography', 'continent', 'waterBody', 'islandGroup', 'island', 'countryCode', 'stateProvince', 'county', 'municipality', 'locality', 'verbatimLocality', 'verbatimElevation', 'verticalDatum', 'verbatimDepth', 'minimumDistanceAboveSurfaceInMeters', 'maximumDistanceAboveSurfaceInMeters', 'locationAccordingTo', 'locationRemarks', 'decimalLatitude', 'decimalLongitude', 'coordinateUncertaintyInMeters', 'coordinatePrecision', 'pointRadiusSpatialFit', 'verbatimCoordinateSystem', 'verbatimSRS', 'footprintWKT', 'footprintSRS', 'footprintSpatialFit', 'georeferencedBy', 'georeferencedDate', 'georeferenceProtocol', 'georeferenceSources', 'georeferenceRemarks', 'geologicalContextID', 'earliestEonOrLowestEonothem', 'latestEonOrHighestEonothem', 'earliestEraOrLowestErathem', 'latestEraOrHighestErathem', 'earliestPeriodOrLowestSystem', 'latestPeriodOrHighestSystem', 'earliestEpochOrLowestSeries', 'latestEpochOrHighestSeries', 'earliestAgeOrLowestStage', 'latestAgeOrHighestStage', 'lowestBiostratigraphicZone', 'highestBiostratigraphicZone', 'lithostratigraphicTerms', 'group', 'formation', 'member', 'bed', 'identificationID', 'verbatimIdentification', 'identificationQualifier', 'typeStatus', 'identifiedBy', 'identifiedByID', 'dateIdentified', 'identificationReferences', 'identificationVerificationStatus', 'identificationRemarks', 'taxonID', 'scientificNameID', 'acceptedNameUsageID', 'parentNameUsageID', 'originalNameUsageID', 'nameAccordingToID', 'namePublishedInID', 'taxonConceptID', 'scientificName', 'acceptedNameUsage', 'parentNameUsage', 'originalNameUsage', 'nameAccordingTo', 'namePublishedIn', 'namePublishedInYear', 'higherClassification', 'kingdom', 'phylum', 'class', 'order', 'family', 'subfamily', 'genus', 'genericName', 'subgenus', 'infragenericEpithet', 'specificEpithet', 'infraspecificEpithet', 'cultivarEpithet', 'taxonRank', 'verbatimTaxonRank', 'vernacularName', 'nomenclaturalCode', 'taxonomicStatus', 'nomenclaturalStatus', 'taxonRemarks', 'datasetKey', 'publishingCountry', 'lastInterpreted', 'elevation', 'elevationAccuracy', 'depth', 'depthAccuracy', 'distanceAboveSurface', 'distanceAboveSurfaceAccuracy', 'issue', 'mediaType', 'hasCoordinate', 'hasGeospatialIssues', 'taxonKey', 'acceptedTaxonKey', 'kingdomKey', 'phylumKey', 'classKey', 'orderKey', 'familyKey', 'genusKey', 'subgenusKey', 'speciesKey', 'species', 'acceptedScientificName', 'verbatimScientificName', 'typifiedName', 'protocol', 'lastParsed', 'lastCrawled', 'repatriated', 'relativeOrganismQuantity', 'level0Gid', 'level0Name', 'level1Gid', 'level1Name', 'level2Gid', 'level2Name', 'level3Gid', 'level3Name', 'iucnRedListCategory']\n3898509458,nan,http://rightsstatements.org/vocab/CNE/1.0/,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,2605588,nan,nan,nan,nan,nan,nan,nan,nan,nan,CC0_1_0,nan,nan,2022-08-15T08:30:45Z,nan,nan,https://portal.neherbaria.org/portal/collections/individual/index.php?occid=2605588,nan,nan,nan,nan,Mohonk Preserve,nan,nan,nan,nan,nan,nan,nan,nan,nan,745e5369-ba4e-4b80-b4b7-d64ab309e7b7,nan,Mohonk Preserve,DSRC,nan,nan,PRESERVED_SPECIMEN,nan,nan,nan,f2d1ba77-1c4d-41f6-8569-50becee5e9c3,MOH002237,nan,Dan Smiley,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,PRESENT,nan,nan,nan,nan,nan,nan,nan,The Buff,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,1971-10-21T00:00:00,nan,294,nan,1971,10,21,10/21/71,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,US,New York,nan,nan,Mohonk Lake,nan,nan,nan,nan,nan,nan,nan,nan,41.772115,-74.153723,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,nan,830036,nan,nan,nan,nan,nan,nan,nan,Populus tremuloides Michx.,nan,nan,nan,nan,nan,nan,Plantae|Charophyta|Streptophytina|Equisetopsida|Magnoliidae|Malpighiales|Salicaceae|Populus,Plantae,Tracheophyta,Magnoliopsida,Malpighiales,Salicaceae,nan,Populus,Populus,nan,nan,tremuloides,nan,nan,SPECIES,nan,nan,nan,ACCEPTED,nan,nan,ffe1030d-42d1-4bb5-8400-1123cc859a5a,US,2022-11-29T23:03:56.952Z,nan,nan,nan,nan,nan,nan,GEODETIC_DATUM_ASSUMED_WGS84;AMBIGUOUS_COLLECTION;INSTITUTION_MATCH_FUZZY,StillImage,true,false,3040215,3040215,6,7707728,220,1414,6664,3040183,nan,3040215,Populus tremuloides,Populus tremuloides Michx.,Populus tremuloides,nan,DWC_ARCHIVE,2022-11-29T23:03:56.952Z,2022-11-29T23:02:54.980Z,false,nan,USA,United States,USA.33_1,New York,USA.33.57_1,Ulster,nan,nan,LC"""
+ sample_text_headers = """['gbifID', 'abstract', 'accessRights', 'accrualMethod', 'accrualPeriodicity', 'accrualPolicy', 'alternative', 'audience', 'available', 'bibliographicCitation', 'conformsTo', 'contributor', 'coverage', 'created', 'creator', 'date', 'dateAccepted', 'dateCopyrighted', 'dateSubmitted', 'description', 'educationLevel', 'extent', 'format', 'hasFormat', 'hasPart', 'hasVersion', 'identifier', 'instructionalMethod', 'isFormatOf', 'isPartOf', 'isReferencedBy', 'isReplacedBy', 'isRequiredBy', 'isVersionOf', 'issued', 'language', 'license', 'mediator', 'medium', 'modified', 'provenance', 'publisher', 'references', 'relation', 'replaces', 'requires', 'rights', 'rightsHolder', 'source', 'spatial', 'subject', 'tableOfContents', 'temporal', 'title', 'type', 'valid', 'institutionID', 'collectionID', 'datasetID', 'institutionCode', 'collectionCode', 'datasetName', 'ownerInstitutionCode', 'basisOfRecord', 'informationWithheld', 'dataGeneralizations', 'dynamicProperties', 'occurrenceID', 'catalogNumber', 'recordNumber', 'recordedBy', 'recordedByID', 'individualCount', 'organismQuantity', 'organismQuantityType', 'sex', 'lifeStage', 'reproductiveCondition', 'behavior', 'establishmentMeans', 'degreeOfEstablishment', 'pathway', 'georeferenceVerificationStatus', 'occurrenceStatus', 'preparations', 'disposition', 'associatedOccurrences', 'associatedReferences', 'associatedSequences', 'associatedTaxa', 'otherCatalogNumbers', 'occurrenceRemarks', 'organismID', 'organismName', 'organismScope', 'associatedOrganisms', 'previousIdentifications', 'organismRemarks', 'materialSampleID', 'eventID', 'parentEventID', 'fieldNumber', 'eventDate', 'eventTime', 'startDayOfYear', 'endDayOfYear', 'year', 'month', 'day', 'verbatimEventDate', 'habitat', 'samplingProtocol', 'sampleSizeValue', 'sampleSizeUnit', 'samplingEffort', 'fieldNotes', 'eventRemarks', 'locationID', 'higherGeographyID', 'higherGeography', 'continent', 'waterBody', 'islandGroup', 'island', 'countryCode', 'stateProvince', 'county', 'municipality', 'locality', 'verbatimLocality', 'verbatimElevation', 'verticalDatum', 'verbatimDepth', 'minimumDistanceAboveSurfaceInMeters', 'maximumDistanceAboveSurfaceInMeters', 'locationAccordingTo', 'locationRemarks', 'decimalLatitude', 'decimalLongitude', 'coordinateUncertaintyInMeters', 'coordinatePrecision', 'pointRadiusSpatialFit', 'verbatimCoordinateSystem', 'verbatimSRS', 'footprintWKT', 'footprintSRS', 'footprintSpatialFit', 'georeferencedBy', 'georeferencedDate', 'georeferenceProtocol', 'georeferenceSources', 'georeferenceRemarks', 'geologicalContextID', 'earliestEonOrLowestEonothem', 'latestEonOrHighestEonothem', 'earliestEraOrLowestErathem', 'latestEraOrHighestErathem', 'earliestPeriodOrLowestSystem', 'latestPeriodOrHighestSystem', 'earliestEpochOrLowestSeries', 'latestEpochOrHighestSeries', 'earliestAgeOrLowestStage', 'latestAgeOrHighestStage', 'lowestBiostratigraphicZone', 'highestBiostratigraphicZone', 'lithostratigraphicTerms', 'group', 'formation', 'member', 'bed', 'identificationID', 'verbatimIdentification', 'identificationQualifier', 'typeStatus', 'identifiedBy', 'identifiedByID', 'dateIdentified', 'identificationReferences', 'identificationVerificationStatus', 'identificationRemarks', 'taxonID', 'scientificNameID', 'acceptedNameUsageID', 'parentNameUsageID', 'originalNameUsageID', 'nameAccordingToID', 'namePublishedInID', 'taxonConceptID', 'scientificName', 'acceptedNameUsage', 'parentNameUsage', 'originalNameUsage', 'nameAccordingTo', 'namePublishedIn', 'namePublishedInYear', 'higherClassification', 'kingdom', 'phylum', 'class', 'order', 'family', 'subfamily', 'genus', 'genericName', 'subgenus', 'infragenericEpithet', 'specificEpithet', 'infraspecificEpithet', 'cultivarEpithet', 'taxonRank', 'verbatimTaxonRank', 'vernacularName', 'nomenclaturalCode', 'taxonomicStatus', 'nomenclaturalStatus', 'taxonRemarks', 'datasetKey', 'publishingCountry', 'lastInterpreted', 'elevation', 'elevationAccuracy', 'depth', 'depthAccuracy', 'distanceAboveSurface', 'distanceAboveSurfaceAccuracy', 'issue', 'mediaType', 'hasCoordinate', 'hasGeospatialIssues', 'taxonKey', 'acceptedTaxonKey', 'kingdomKey', 'phylumKey', 'classKey', 'orderKey', 'familyKey', 'genusKey', 'subgenusKey', 'speciesKey', 'species', 'acceptedScientificName', 'verbatimScientificName', 'typifiedName', 'protocol', 'lastParsed', 'lastCrawled', 'repatriated', 'relativeOrganismQuantity', 'level0Gid', 'level0Name', 'level1Gid', 'level1Name', 'level2Gid', 'level2Name', 'level3Gid', 'level3Name', 'iucnRedListCategory']"""
+
+ sample_OCR_response = """PLANTS OF BORNEC!! Euphorbiaceae!! Chaetocarpus castanocarpus Thwaites!! Det. JH Beaman, 15 May 2010 !!Sabah: Kota Kinabalu District: Bukit Padang, by UKMS!!temporary campus. Elev. 30 m. Eroded hills and gullies.!!scattered scrubby vegetation; Crocker Formation. Shrub.!!Lat. 5°58 N. Long. 116°06 E!!John H. Beaman 83041!!August 1983!!with Willem Meijer!!HERBARIA OF UNIVERSITI KEBANGSAAN MALAYSIA (UKMS) and!!MICHIGAN STATE UNIVERSITY (MSC)!!"""
+ sample_dict = """{"gbifID": "3898509458", "accessRights": "http://rightsstatements.org/vocab/CNE/1.0/", "identifier": "2605588", "license": "CC0_1_0", "modified": "2022-08-15T08:30:45Z", "references": "https://portal.neherbaria.org/portal/collections/individual/index.php?occid=2605588", "rightsHolder": "Mohonk Preserve", "collectionID": "745e5369-ba4e-4b80-b4b7-d64ab309e7b7", "institutionCode": "Mohonk Preserve", "collectionCode": "DSRC", "basisOfRecord": "PRESERVED_SPECIMEN", "occurrenceID": "f2d1ba77-1c4d-41f6-8569-50becee5e9c3", "catalogNumber": "MOH002237", "recordedBy": "Dan Smiley", "occurrenceStatus": "PRESENT", "occurrenceRemarks": "The Buff", "eventDate": "1971-10-21T00:00:00", "startDayOfYear": "294", "year": "1971", "month": "10", "day": "21", "verbatimEventDate": "10/21/71", "countryCode": "US", "stateProvince": "New York", "locality": "Mohonk Lake", "decimalLatitude": "41.772115", "decimalLongitude": "-74.153723", "taxonID": "830036", "scientificName": "Populus tremuloides Michx.", "higherClassification": "Plantae|Charophyta|Streptophytina|Equisetopsida|Magnoliidae|Malpighiales|Salicaceae|Populus", "kingdom": "Plantae", "phylum": "Tracheophyta", "class": "Magnoliopsida", "order": "Malpighiales", "family": "Salicaceae", "genus": "Populus", "genericName": "Populus", "specificEpithet": "tremuloides", "taxonRank": "SPECIES", "taxonomicStatus": "ACCEPTED", "datasetKey": "ffe1030d-42d1-4bb5-8400-1123cc859a5a", "publishingCountry": "US", "lastInterpreted": "2022-11-29T23:03:56.952Z", "issue": "GEODETIC_DATUM_ASSUMED_WGS84;AMBIGUOUS_COLLECTION;INSTITUTION_MATCH_FUZZY", "mediaType": "StillImage", "hasCoordinate": "true", "hasGeospatialIssues": "false", "taxonKey": "3040215", "acceptedTaxonKey": "3040215", "kingdomKey": "6", "phylumKey": "7707728", "classKey": "220", "orderKey": "1414", "familyKey": "6664", "genusKey": "3040183", "speciesKey": "3040215", "species": "Populus tremuloides", "acceptedScientificName": "Populus tremuloides Michx.", "verbatimScientificName": "Populus tremuloides", "protocol": "DWC_ARCHIVE", "lastParsed": "2022-11-29T23:03:56.952Z", "lastCrawled": "2022-11-29T23:02:54.980Z", "repatriated": "false", "level0Gid": "USA", "level0Name": "United States", "level1Gid": "USA.33_1", "level1Name": "New York", "level2Gid": "USA.33.57_1", "level2Name": "Ulster", "iucnRedListCategory": "LC"}"""
+
+ nt_rules = num_tokens_from_string(set_rules, "cl100k_base")
+ nt_dict = num_tokens_from_string(sample_dict, "cl100k_base")
+ nt_ocr = num_tokens_from_string(sample_OCR_response, "cl100k_base")
+
+ print(f"nt - nt_rules {nt_rules}")
+ print(f"nt - nt_dict {nt_dict}")
+ print(f"nt - nt_new {nt_ocr}")
+
+ do_create = False
+
+ file_occ = 'D:/Dropbox/LeafMachine2/leafmachine2/transcription/test_occ/occurrence_short.txt'
+ file_name = 'test_occ'
+ dir_out = "D:/D_Desktop/embedding"
+
+ '''
+ if do_create:
+ create_embeddings(file_occ, file_name, dir_out)
+
+
+ # Load the generated embeddings
+ output_file = os.path.join(dir_out, f'embedded_dwc__{file_name}.csv')
+ embedder = GenerateEmbeddings(file_occ, file_name, dir_out)
+ embedded_df, dwc_headers = embedder.load_embedded_csv(output_file)
+
+ # Search for reviews
+ search_query = "1971 The Buff"
+ results = embedder.search_rows(dwc_headers, embedded_df, search_query, n=1)
+ print(results)
+ '''
+ GPT_response = OCR_to_dict(img_path)
+ print(GPT_response)
+
+
+if __name__ == '__main__':
+ print()
+
diff --git a/vouchervision/embeddings_db.py b/vouchervision/embeddings_db.py
new file mode 100644
index 0000000000000000000000000000000000000000..c939d6261dce92f3c48380189e146b72b311f8ea
--- /dev/null
+++ b/vouchervision/embeddings_db.py
@@ -0,0 +1,277 @@
+import json, os, time, uuid
+import pandas as pd
+import numpy as np
+from sklearn.metrics.pairwise import cosine_similarity
+from transformers import AutoTokenizer, AutoModel
+import chromadb
+from chromadb.config import Settings
+from chromadb.utils import embedding_functions
+from InstructorEmbedding import INSTRUCTOR
+from langchain.vectorstores import Chroma
+'''
+If there is a transformers install error:
+pip install transformers==4.29.2
+Python 3.8 and above will need to upgrade the transformers to 4.2x.xx
+https://github.com/huggingface/transformers/issues/11799
+
+The goal is to creat a domain knowledge database based on existing transcribed labels.
+
+I modify the domain knowledge (an xlsx file) so that each row is embedded in a way that most closely
+resembles the raw OCR output, since that is what will be used to query against the db.
+
+Once the closest row is found, I use the id to go back to the xlsx and take the whole row, converting
+it into a dictionary similar to the desired output from the LLM.
+
+This dict is then added to the prompt as a hint for the LLM.
+'''
+
+'''
+pip uninstall protobuf
+pip install protobuf==3.19.5
+'''
+class VoucherVisionEmbedding:
+ # def __init__(self, db_name, path_domain_knowledge, logger, build_new_db=False, model_name="hkunlp/instructor-xl", device="cuda"):
+ # DB_DIR = os.path.join(os.path.dirname(__file__), db_name)
+
+ # client_settings = chromadb.config.Settings(
+ # chroma_db_impl="duckdb+parquet",
+ # persist_directory=DB_DIR,
+ # anonymized_telemetry=False
+ # )
+ # embeddings = embedding_functions.InstructorEmbeddingFunction(model_name=model_name, device=device)
+
+ # self.collection = Chroma(
+ # collection_name="langchain_store",
+ # embedding_function=embeddings,
+ # client_settings=client_settings,
+ # persist_directory=DB_DIR,
+ # )
+
+ # total_rows = len(self.domain_knowledge)
+ # for index, row in self.domain_knowledge.iterrows():
+ # try:
+ # self.logger.info(f"[Creating New Embedding DB] --- Adding Row {index+1}/{total_rows}")
+ # except:
+ # print(f"Row {index+1}/{total_rows}")
+ # id = str(row[0])
+ # document = str(' '.join(row[1:][row[1:].notna()].astype(str)))
+
+ # self.collection.add_texts(document, None, id, embedding=embeddings)
+ # self.collection.persist()
+ # print(self.collection)
+
+ def __init__(self, db_name, path_domain_knowledge, logger, build_new_db=False, model_name="hkunlp/instructor-xl", device="cuda"):
+ DB_DIR = os.path.join(os.path.dirname(__file__), db_name)
+ self.logger = logger
+ self.path_domain_knowledge = path_domain_knowledge
+ self.client = chromadb.PersistentClient(path=DB_DIR,
+ settings=Settings(anonymized_telemetry=False))
+
+ ef = embedding_functions.InstructorEmbeddingFunction(model_name=model_name, device=device)
+ self.domain_knowledge = pd.read_excel(path_domain_knowledge).fillna('').astype(str)
+
+ if build_new_db:
+ self.logger.info(f"Creating new DB from {self.path_domain_knowledge}")
+ self.collection = self.client.create_collection(name=db_name, embedding_function=ef, metadata={"hnsw:space": "cosine"})
+ self.create_db_from_xlsx()
+ else:
+ try:
+ self.collection = self.client.get_collection(name=db_name, embedding_function=ef)
+ except:
+ self.logger.error(f"Embedding database not found! Creating new DB from {self.path_domain_knowledge}")
+ self.collection = self.client.create_collection(name=db_name, embedding_function=ef, metadata={"hnsw:space": "cosine"})
+ self.create_db_from_xlsx()
+
+
+ def add_document(self, document, metadata, id):
+ id = str(id)
+ existing_documents = self.collection.get()
+ if id not in existing_documents['ids']:
+ try:
+ self.collection.add(documents=[document], ids=[id])
+ except Exception as e:
+ self.logger.error(f"Error while adding document {id}: {str(e)}")
+
+
+ # try:
+ # self.collection.add(documents=[document], ids=[id])
+ # except:
+ # try:
+ # time.sleep(0.1)
+ # self.collection.add(documents=[document], ids=[id])
+ # except:
+ # try:
+ # self.logger.info(f"[Embedding Add Doc] --- Failed, skipping: {id}")
+ # except:
+ # print(f"Failed, skipping: {id}")
+ else:
+ try:
+ self.logger.info(f"[Embedding Add Doc] --- ID already exists in the collection: {id}")
+ except:
+ print(f"ID already exists in the collection: {id}")
+
+
+ def query_db(self, query_text, n_results):
+ results = self.collection.query(query_texts=[query_text], n_results=n_results)
+
+ self.similarity = round(results['distances'][0][0],3)
+ self.similarity_exact = results['distances'][0][0]
+ try:
+ self.logger.info(f"[Embedding Search] --- Similarity (close to zero is best) {self.similarity}")
+ except:
+ print(f"Similarity (close to zero is best) --- {self.similarity}")
+
+ self.domain_knowledge.iloc[:, 0] = self.domain_knowledge.iloc[:, 0].astype(str)
+
+ # Initialize an empty list to hold dictionaries
+ for id in results['ids']:
+ row_dicts = self._get_row_from_df(id)
+ if not row_dicts:
+ # try:
+ # self.logger.info(f"[Embedding Search] --- Similar Dictionary\n{row_dicts}")
+ # except:
+ # print(row_dicts)
+ # else:
+ try:
+ self.logger.info(f"[Embedding Search] --- No row found for id {id}")
+ except:
+ print(f"No row found for id {id}")
+
+ # Return the list of dictionaries if n_results > 1, else return single dictionary
+ if n_results > 1:
+ return row_dicts
+ else:
+ return row_dicts[0] if row_dicts else None
+
+ def create_db_from_xlsx(self):
+ total_rows = len(self.domain_knowledge)
+ for index, row in self.domain_knowledge.iterrows():
+ try:
+ self.logger.info(f"[Creating New Embedding DB] --- Adding Row {index+1}/{total_rows}")
+ except:
+ print(f"Row {index+1}/{total_rows}")
+ id = str(row.iloc[0])
+ document = str(' '.join(row[0:][row[0:].notna()].astype(str)))
+ self.add_document(document, None, id)
+
+ def get_similarity(self):
+ return self.similarity_exact
+
+ def _get_row_from_df(self, ids):
+ row_dicts = [] # initialize an empty list to hold dictionaries
+ for id in ids:
+ row = self.domain_knowledge[self.domain_knowledge.iloc[:, 0] == id]
+ if not row.empty:
+ row_dict = row.iloc[0].to_dict()
+ row_dict.pop('Catalog Number', None)
+ for key in row_dict:
+ if pd.isna(row_dict[key]):
+ row_dict[key] = ''
+ row_dicts.append(row_dict) # append the dictionary to the list
+ return row_dicts if row_dicts else None # return the list of dictionaries or None if it's empty
+
+ # def _get_row_from_df(self, ids):
+ # for id in ids:
+ # row = self.domain_knowledge[self.domain_knowledge.iloc[:, 0] == id]
+ # if not row.empty:
+ # row_dict = row.iloc[0].to_dict()
+ # row_dict.pop('Catalog Number', None)
+ # for key in row_dict:
+ # if pd.isna(row_dict[key]):
+ # row_dict[key] = ''
+ # return row_dict
+ # return None
+
+
+
+
+class VoucherVisionEmbeddingTest:
+ def __init__(self, ground_truth_dir, llm_output_dir, model_name="hkunlp/instructor-xl"):
+ self.ground_truth_dir = ground_truth_dir
+ self.llm_output_dir = llm_output_dir
+ self.model_name = model_name
+ self.model = INSTRUCTOR(model_name, device="cuda")
+ self.instruction = "Represent the Science json dictionary document:"
+
+ def compare_texts(self, ground_truth_text, predicted_text):
+ # Convert the texts to embeddings using the given model
+ ground_truth_embedding = self.model.encode([[self.instruction,ground_truth_text]])
+ predicted_embedding = self.model.encode([[self.instruction,predicted_text]])
+
+ # Compute the cosine similarity between the two embeddings
+ similarity = cosine_similarity(ground_truth_embedding, predicted_embedding)
+
+ return similarity[0][0]
+
+ @staticmethod
+ def json_to_text(json_dict):
+ return str(json_dict)
+
+ def get_max_difference(self, similarities):
+ differences = [abs(1 - sim) for sim in similarities]
+ return max(differences)
+
+ def evaluate(self):
+ # Get a list of all ground truth and LLM output files
+ ground_truth_files = os.listdir(self.ground_truth_dir)
+ llm_output_files = os.listdir(self.llm_output_dir)
+
+ # Ensure file lists are sorted so they match up correctly
+ ground_truth_files.sort()
+ llm_output_files.sort()
+
+ similarities = []
+ key_similarities = [] # List to store key similarity
+
+ for ground_truth_file, llm_output_file in zip(ground_truth_files, llm_output_files):
+ # Read the files and convert them to text
+ with open(os.path.join(self.ground_truth_dir, ground_truth_file), 'r') as f:
+ ground_truth_dict = json.load(f)
+ ground_truth_text = self.json_to_text(ground_truth_dict)
+ with open(os.path.join(self.llm_output_dir, llm_output_file), 'r') as ff:
+ llm_output_dict = json.load(ff)
+ llm_output_text = self.json_to_text(llm_output_dict)
+
+ # Compute the similarity between the ground truth and the LLM output
+ similarity = self.compare_texts(ground_truth_text, llm_output_text)
+
+ # Clip and round to mitigate/smudge floating-point precision limitations
+ similarity = np.clip(similarity, -1.0, 1.0)
+ similarity = np.round(similarity, 6)
+
+ similarities.append(similarity)
+
+ # Compare keys
+ ground_truth_keys = ', '.join(sorted(ground_truth_dict.keys()))
+ llm_output_keys = ', '.join(sorted(llm_output_dict.keys()))
+ key_similarity = self.compare_texts(ground_truth_keys, llm_output_keys)
+ key_similarity = np.clip(key_similarity, -1.0, 1.0)
+ key_similarity = np.round(key_similarity, 6)
+ key_similarities.append(key_similarity)
+
+ # Compute the mean similarity
+ mean_similarity = np.mean(similarities)
+ mean_key_similarity = np.mean(key_similarities)
+
+ max_diff = self.get_max_difference(similarities)
+ max_diff_key = self.get_max_difference(key_similarities)
+
+ return mean_similarity, max_diff, similarities, mean_key_similarity, max_diff_key, key_similarities
+
+
+
+
+if __name__ == '__main__':
+ # db_name = "VV_all_asia_minimal"
+ db_name = "all_asia_minimal"
+ path_domain_knowledge = 'D:/Dropbox/LeafMachine2/leafmachine2/transcription/domain_knowledge/AllAsiaMinimalasof25May2023_2__FOR-EMBEDDING.xlsx'
+ # path_domain_knowledge = 'D:/Dropbox/LeafMachine2/leafmachine2/transcription/domain_knowledge/AllAsiaMinimalasof25May2023_2__TRIMMEDtiny.xlsx'
+
+ build_new_db = False
+
+
+ VVE = VoucherVisionEmbedding(db_name, path_domain_knowledge, build_new_db)
+
+ test_query = "Golden Thread\nHerbaria of Michigan State University (MSC) and\nUniversiti Kebangsaan Malaysia, Sabah Campus (UKMS)\nUNITED STATES\n3539788\nNATIONAL HERBARIUM\nPLANTS OF BORNEO\nBrookea tomentosa Benth.\nMalaysia. Sabah. Beaufort District: Beaufort Hill. 5°22'N,\n115°45'E. Elev. 200 m. Burned logged dipterocarp forest.\nCrocker Formation. Small tree, corolla cream.\nDet. at K, 1986\n28 August 1983\nWith: Reed S. Beaman and Teofila E. Beamann\nJohn H. Beaman 6844"
+
+ domain_knowledge_example = VVE.query_db(test_query, 1)
\ No newline at end of file
diff --git a/vouchervision/emoji_rain.py b/vouchervision/emoji_rain.py
new file mode 100644
index 0000000000000000000000000000000000000000..5fb370e1da873757e42991f1d98b6036c532e7e0
--- /dev/null
+++ b/vouchervision/emoji_rain.py
@@ -0,0 +1,243 @@
+from typing import Union
+
+import streamlit as st
+
+import inspect
+from importlib import import_module
+from pathlib import Path
+from typing import Any, Callable, Optional, TypeVar, Union, overload
+
+try:
+ from streamlit.runtime.metrics_util import gather_metrics as _gather_metrics
+except ImportError:
+
+ def _gather_metrics(name, func): # type: ignore
+ return func
+
+
+F = TypeVar("F", bound=Callable[..., Any])
+
+# Typing overloads here are actually required so that you can correctly (= with correct typing) use the decorator in different ways:
+# 1) as a decorator without parameters @extra
+# 2) as a decorator with parameters (@extra(foo="bar") but this also refers to empty parameters @extra()
+# 3) as a function: extra(my_function)
+
+
+@overload
+def extra(
+ func: F,
+) -> F:
+ ...
+
+
+@overload
+def extra(
+ func: None = None,
+) -> Callable[[F], F]:
+ ...
+
+
+def extra(
+ func: Optional[F] = None,
+) -> Union[Callable[[F], F], F]:
+
+ if func:
+
+ filename = inspect.stack()[1].filename
+ submodule = Path(filename).parent.name
+ extra_name = "streamlit_extras." + submodule
+ module = import_module(extra_name)
+
+ if hasattr(module, "__funcs__"):
+ module.__funcs__ += [func] # type: ignore
+ else:
+ module.__funcs__ = [func] # type: ignore
+
+ profiling_name = f"{submodule}.{func.__name__}"
+ try:
+ return _gather_metrics(name=profiling_name, func=func)
+ except TypeError:
+ # Don't fail on streamlit==1.13.0, which only expects a callable
+ pass
+
+ def wrapper(f: F) -> F:
+ return f
+
+ return wrapper
+
+
+@extra
+def proportional_rain(
+ emoji1: str,
+ count1: int,
+ emoji2: str,
+ count2: int,
+ font_size: int = 64,
+ falling_speed: int = 5,
+ animation_length: Union[int, str] = "infinite"
+):
+ """
+ Creates a CSS animation where input emojis fall from top to bottom of the screen.
+ The proportion of emojis is based on the provided counts.
+ """
+
+ if isinstance(animation_length, int):
+ animation_length = f"{animation_length}"
+
+ # CSS Code ...
+ st.write(
+ f"""
+
+ """,
+ unsafe_allow_html=True,
+ )
+
+ # Create emoji strings based on counts
+ emoji_str1 = "".join([f'{emoji1}
' for _ in range(count1)])
+ emoji_str2 = "".join([f'{emoji2}
' for _ in range(count2)])
+
+ st.write(
+ f"""
+
+
+ {emoji_str1}
+ {emoji_str2}
+
+ """,
+ unsafe_allow_html=True,
+ )
\ No newline at end of file
diff --git a/vouchervision/fetch_data.py b/vouchervision/fetch_data.py
new file mode 100644
index 0000000000000000000000000000000000000000..a296be52bb6a922544d07a2f45d56c7b1c569080
--- /dev/null
+++ b/vouchervision/fetch_data.py
@@ -0,0 +1,256 @@
+from __future__ import annotations
+import os, yaml, shutil
+from io import BytesIO
+from urllib.request import urlopen
+from zipfile import ZipFile
+import urllib.request
+from tqdm import tqdm
+import subprocess
+
+VERSION = 'v-2-1'
+
+def fetch_data(logger, dir_home, cfg_file_path):
+ logger.name = 'Fetch Data'
+ ready_to_use = False
+ do_fetch = True
+ current = ''.join(['release_', VERSION])
+
+ # Make sure weights are present
+ if os.path.isfile(os.path.join(dir_home,'bin','version.yml')):
+ ver = load_version(dir_home)
+
+ if ver['version'] == VERSION:
+ if current in os.listdir(os.path.join(dir_home,'bin')): # The release dir is present
+ do_fetch = False
+ ready_to_use = True
+ logger.warning(f"Version file --- {os.path.join(dir_home,'bin','version.yml')}")
+ logger.warning(f"Current version --- {ver['version']}")
+ logger.warning(f"Last updated --- {ver['last_update']}")
+ else: # right version, no release dir yet
+ do_fetch = True
+ logger.warning(f"--------------------------------")
+ logger.warning(f" Downloading data files... ")
+ logger.warning(f"--------------------------------")
+ logger.warning(f"Version file --- {os.path.join(dir_home,'bin','version.yml')}")
+ logger.warning(f"Current version --- {ver['version']}")
+ logger.warning(f"Last updated --- {ver['last_update']}")
+ else:
+ do_fetch = True
+ logger.warning(f"--------------------------------")
+ logger.warning(f" Out of date... ")
+ logger.warning(f" Downloading data files... ")
+ logger.warning(f"--------------------------------")
+ logger.warning(f"Version file --- {os.path.join(dir_home,'bin','version.yml')}")
+ logger.warning(f"Current version --- {ver['version']}")
+ logger.warning(f"Last updated --- {ver['last_update']}")
+
+
+ else:
+ do_fetch = True
+ logger.warning(f"--------------------------------")
+ logger.warning(f" Missing version.yml... ")
+ logger.warning(f" Downloading data files... ")
+ logger.warning(f"--------------------------------")
+ logger.warning(f"Version file --- {os.path.join(dir_home,'bin','version.yml')}")
+ logger.warning(f"Current version --- {ver['version']}")
+ logger.warning(f"Last updated --- {ver['last_update']}")
+
+
+ if do_fetch:
+ logger.warning(f"Fetching files for version --> {ver['version']}")
+ path_release = get_weights(dir_home, current, logger)
+ if path_release is not None:
+ logger.warning(f"Data download successful. Unzipping...")
+ move_data_to_home(path_release, dir_home, logger)
+ ready_to_use = True
+ logger.warning(f"--------------------------------")
+ logger.warning(f" LeafMachine2 is up to date ")
+ logger.warning(f"--------------------------------")
+
+ else:
+ logger.warning(f"--------------------------------")
+ logger.warning(f" LeafMachine2 is up to date ")
+ logger.warning(f"--------------------------------")
+
+ return ready_to_use
+
+
+
+def get_weights(dir_home, current, logger):
+
+ try:
+ path_zip = os.path.join(dir_home,'bin',current)
+ zipurl = ''.join(['https://leafmachine.org/LM2/', current,'.zip'])
+ headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36'}
+
+ req = urllib.request.Request(url=zipurl, headers=headers)
+
+ # Get the file size from the Content-Length header
+ with urllib.request.urlopen(req) as url_response:
+ file_size = int(url_response.headers['Content-Length'])
+
+ # Download the ZIP file from the URL with progress bar
+ with tqdm(unit='B', unit_scale=True, unit_divisor=1024, total=file_size) as pbar:
+ with urllib.request.urlopen(req) as url_response:
+ with open(current + '.zip', 'wb') as file:
+ while True:
+ chunk = url_response.read(4096)
+ if not chunk:
+ break
+ file.write(chunk)
+ pbar.update(len(chunk))
+
+ # Extract the contents of the ZIP file to the current directory
+ zipfilename = current + '.zip'
+ with ZipFile(zipfilename, 'r') as zip_file:
+ zip_file.extractall(os.path.join(dir_home,'bin'))
+
+ print(f"{bcolors.CGREENBG2}Data extracted to {path_zip}{bcolors.ENDC}")
+ logger.warning(f"Data extracted to {path_zip}")
+
+ return path_zip
+ except Exception as e:
+ print(f"{bcolors.CREDBG2}ERROR --- Could not download or extract machine learning models\n{e}{bcolors.ENDC}")
+ logger.warning(f"ERROR --- Could not download or extract machine learning models")
+ logger.warning(f"ERROR --- {e}")
+ return None
+
+
+def load_version(dir_home):
+ try:
+ with open(os.path.join(dir_home,'bin',"version.yml"), "r") as ymlfile:
+ ver = yaml.full_load(ymlfile)
+ except:
+ with open(os.path.join(os.path.dirname(os.path.dirname(dir_home)),'bin',"version.yml"), "r") as ymlfile:
+ ver = yaml.full_load(ymlfile)
+ return ver
+
+def move_data_to_home(path_release, dir_home, logger):
+ path_list_file = os.path.join(path_release, 'path_list.yml')
+
+ with open(path_list_file, 'r') as file:
+ path_list = yaml.safe_load(file)
+
+ paths = {
+ 'path_ruler_classifier': os.path.join(dir_home, *path_list['path_ruler_classifier'].split('___')),
+ 'path_ruler_binary_classifier': os.path.join(dir_home, *path_list['path_ruler_binary_classifier'].split('___')),
+ 'path_ruler_classifier_binary_classes': os.path.join(dir_home, *path_list['path_ruler_classifier_binary_classes'].split('___')),
+ 'path_ruler_classifier_ruler_classes': os.path.join(dir_home, *path_list['path_ruler_classifier_ruler_classes'].split('___')),
+ 'path_DocEnTR': os.path.join(dir_home, *path_list['path_DocEnTR'].split('___')),
+ 'path_ACD': os.path.join(dir_home, *path_list['path_ACD'].split('___')),
+ 'path_PCD': os.path.join(dir_home, *path_list['path_PCD'].split('___')),
+ 'path_landmarks': os.path.join(dir_home, *path_list['path_landmarks'].split('___')),
+ 'path_YOLO': os.path.join(dir_home, *path_list['path_YOLO'].split('___')),
+ 'path_segment': os.path.join(dir_home, *path_list['path_segment'].split('___')),
+ }
+
+
+ ### Ruler classifier
+ source_file = os.path.join(path_release, 'ruler_classifier', 'ruler_classifier_38classes_v-1.pt')
+ destination_dir = paths['path_ruler_classifier']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+
+ source_file = os.path.join(path_release, 'ruler_classifier', 'model_scripted_resnet_720_withCompression.pt')
+ destination_dir = paths['path_ruler_binary_classifier']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+ source_file = os.path.join(path_release, 'ruler_classifier', 'binary_classes.txt')
+ destination_dir = paths['path_ruler_classifier_binary_classes']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+ source_file = os.path.join(path_release, 'ruler_classifier', 'ruler_classes.txt')
+ destination_dir = paths['path_ruler_classifier_ruler_classes']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+
+ ### Ruler segmentation
+ source_file = os.path.join(path_release, 'ruler_segment', 'small_256_8__epoch-10.pt')
+ destination_dir = paths['path_DocEnTR']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+
+ ### ACD
+ source_file = os.path.join(path_release, 'acd', 'best.pt')
+ destination_dir = paths['path_ACD']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+
+ ### PCD
+ source_file = os.path.join(path_release, 'pcd', 'best.pt')
+ destination_dir = paths['path_PCD']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+
+ ### Landmarks
+ source_file = os.path.join(path_release, 'landmarks', 'best.pt')
+ destination_dir = paths['path_landmarks']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+
+ ### YOLO
+ source_file = os.path.join(path_release, 'YOLO', 'yolov5x6.pt')
+ destination_dir = paths['path_YOLO']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+
+ ### Segmentation
+ source_file = os.path.join(path_release, 'segmentation', 'model_final.pth')
+ destination_dir = paths['path_segment']
+ os.makedirs(destination_dir, exist_ok=True)
+ try_move(logger, source_file, destination_dir )
+
+def git_pull_no_clean():
+ # Define the git command to run
+ git_cmd = ['git', 'pull', '--no-clean']
+
+ # Run the git command using subprocess
+ result = subprocess.run(git_cmd, capture_output=True, text=True)
+
+ # Check if the command was successful
+ if result.returncode == 0:
+ print(result.stdout)
+ else:
+ print(result.stderr)
+
+def try_move(logger, source_file, destination_dir ):
+ try:
+ # Try to move the file using shutil.move()
+ shutil.move(source_file, destination_dir, copy_function=shutil.copy2)
+ logger.warning(f"{source_file}\nmoved to\n{destination_dir}")
+ print(f"{source_file}\nmoved to\n{destination_dir}")
+ except FileExistsError:
+ # If the file already exists in the destination directory, skip it
+ logger.warning(f"Already exists in\n{destination_dir}.\nSkipping...")
+ print(f"Already exists in\n{destination_dir}.\nSkipping...")
+ pass
+ except Exception as e:
+ # Catch any other exceptions that might occur
+ logger.warning(f"[ERROR] occurred while moving:\n{source_file}:\n{str(e)}")
+ print(f"ERROR occurred while moving:\n{source_file}:\n{str(e)}")
+
+class bcolors:
+ HEADER = '\033[95m'
+ OKBLUE = '\033[94m'
+ OKCYAN = '\033[96m'
+ OKGREEN = '\033[92m'
+ WARNING = '\033[93m'
+ FAIL = '\033[91m'
+ ENDC = '\033[0m'
+ BOLD = '\033[1m'
+ UNDERLINE = '\033[4m'
+ CGREENBG2 = '\33[102m'
+ CREDBG2 = '\33[101m'
+ CWHITEBG2 = '\33[107m'
+
+if __name__ == '__main__':
+ git_pull_no_clean()
\ No newline at end of file
diff --git a/vouchervision/general_utils.py b/vouchervision/general_utils.py
new file mode 100644
index 0000000000000000000000000000000000000000..5f8974a42b03cbdd3df2fc524568ffc8839d5616
--- /dev/null
+++ b/vouchervision/general_utils.py
@@ -0,0 +1,1370 @@
+import os, yaml, datetime, argparse, re, cv2, random, shutil, tiktoken, json, csv
+from collections import Counter
+import pandas as pd
+from pathlib import Path
+from dataclasses import dataclass
+from tqdm import tqdm
+import numpy as np
+import concurrent.futures
+from time import perf_counter
+import torch
+
+'''
+TIFF --> DNG
+Install
+https://helpx.adobe.com/camera-raw/using/adobe-dng-converter.html
+Read
+https://helpx.adobe.com/content/dam/help/en/photoshop/pdf/dng_commandline.pdf
+
+'''
+
+
+# https://stackoverflow.com/questions/287871/how-do-i-print-colored-text-to-the-terminal
+
+def validate_dir(dir):
+ if not os.path.exists(dir):
+ os.makedirs(dir)
+
+def get_cfg_from_full_path(path_cfg):
+ with open(path_cfg, "r") as ymlfile:
+ cfg = yaml.full_load(ymlfile)
+ return cfg
+
+def num_tokens_from_string(string: str, encoding_name: str) -> int:
+ encoding = tiktoken.get_encoding(encoding_name)
+ if isinstance(string, dict):
+ string = json.dumps(string)
+ num_tokens = len(encoding.encode(string))
+ return num_tokens
+
+def add_to_expense_report(dir_home, data):
+ path_expense_report = os.path.join(dir_home, 'expense_report','expense_report.csv')
+
+ # Check if the file exists
+ file_exists = os.path.isfile(path_expense_report)
+
+ # Open the file in append mode if it exists, or write mode if it doesn't
+ mode = 'a' if file_exists else 'w'
+
+ with open(path_expense_report, mode=mode, newline='') as file:
+ writer = csv.writer(file)
+
+ # If the file does not exist, write the header first
+ if not file_exists:
+ writer.writerow(['run','date','api_version','total_cost', 'n_images', 'tokens_in', 'tokens_out', 'rate_in', 'rate_out', 'cost_in', 'cost_out',])
+
+ # Write the data row
+ writer.writerow(data)
+
+def save_token_info_as_csv(Dirs, LLM_version0, path_api_cost, total_tokens_in, total_tokens_out, n_images):
+ version_mapping = {
+ 'GPT 4': 'GPT_4',
+ 'GPT 3.5': 'GPT_3_5',
+ 'Azure GPT 3.5': 'GPT_3_5',
+ 'Azure GPT 4': 'GPT_4',
+ 'PaLM 2': 'PALM2'
+ }
+ LLM_version = version_mapping[LLM_version0]
+ # Define the CSV file path
+ csv_file_path = os.path.join(Dirs.path_cost, Dirs.run_name + '.csv')
+
+ cost_in, cost_out, total_cost, rate_in, rate_out = calculate_cost(LLM_version, path_api_cost, total_tokens_in, total_tokens_out)
+
+ # The data to be written to the CSV file
+ data = [Dirs.run_name, get_datetime(),LLM_version, total_cost, n_images, total_tokens_in, total_tokens_out, rate_in, rate_out, cost_in, cost_out,]
+
+ # Open the file in write mode
+ with open(csv_file_path, mode='w', newline='') as file:
+ writer = csv.writer(file)
+
+ # Write the header
+ writer.writerow(['run','date','api_version','total_cost', 'n_images', 'tokens_in', 'tokens_out', 'rate_in', 'rate_out', 'cost_in', 'cost_out',])
+
+ # Write the data
+ writer.writerow(data)
+ # Create a summary string
+ cost_summary = (f"Cost Summary for {Dirs.run_name}:\n"
+ f" API Cost In: ${rate_in} per 1000 Tokens\n"
+ f" API Cost Out: ${rate_out} per 1000 Tokens\n"
+ f" Tokens In: {total_tokens_in} - Cost: ${cost_in:.4f}\n"
+ f" Tokens In: {total_tokens_in} - Cost: ${cost_in:.4f}\n"
+ f" Tokens Out: {total_tokens_out} - Cost: ${cost_out:.4f}\n"
+ f" Images Processed: {n_images}\n"
+ f" Total Cost: ${total_cost:.4f}")
+ return cost_summary, data, total_cost
+
+def summarize_expense_report(path_expense_report):
+ # Initialize counters and sums
+ run_count = 0
+ total_cost_sum = 0
+ tokens_in_sum = 0
+ tokens_out_sum = 0
+ rate_in_sum = 0
+ rate_out_sum = 0
+ cost_in_sum = 0
+ cost_out_sum = 0
+ n_images_sum = 0
+ api_version_counts = Counter()
+
+ # Try to read the CSV file into a DataFrame
+ try:
+ df = pd.read_csv(path_expense_report)
+
+ # Process each row in the DataFrame
+ for index, row in df.iterrows():
+ run_count += 1
+ total_cost_sum += row['total_cost']
+ tokens_in_sum += row['tokens_in']
+ tokens_out_sum += row['tokens_out']
+ rate_in_sum += row['rate_in']
+ rate_out_sum += row['rate_out']
+ cost_in_sum += row['cost_in']
+ cost_out_sum += row['cost_out']
+ n_images_sum += row['n_images']
+ api_version_counts[row['api_version']] += 1
+
+ except FileNotFoundError:
+ print(f"The file {path_expense_report} does not exist.")
+ return None
+
+ # Calculate API version percentages
+ api_version_percentages = {version: (count / run_count) * 100 for version, count in api_version_counts.items()}
+
+ # Calculate cost per image for each API version
+ cost_per_image_dict = {}
+ for version, count in api_version_counts.items():
+ total_cost = df[df['api_version'] == version]['total_cost'].sum()
+ n_images = df[df['api_version'] == version]['n_images'].sum()
+ cost_per_image = total_cost / n_images if n_images > 0 else 0
+ cost_per_image_dict[version] = cost_per_image
+
+ # Return the DataFrame and all summaries
+ return {
+ 'run_count': run_count,
+ 'total_cost_sum': total_cost_sum,
+ 'tokens_in_sum': tokens_in_sum,
+ 'tokens_out_sum': tokens_out_sum,
+ 'rate_in_sum': rate_in_sum,
+ 'rate_out_sum': rate_out_sum,
+ 'cost_in_sum': cost_in_sum,
+ 'cost_out_sum': cost_out_sum,
+ 'n_images_sum':n_images_sum,
+ 'api_version_percentages': api_version_percentages,
+ 'cost_per_image': cost_per_image_dict
+ }, df
+
+def calculate_cost(LLM_version, path_api_cost, total_tokens_in, total_tokens_out):
+ # Load the rates from the YAML file
+ with open(path_api_cost, 'r') as file:
+ cost_data = yaml.safe_load(file)
+
+ # Get the rates for the specified LLM version
+ if LLM_version in cost_data:
+ rates = cost_data[LLM_version]
+ cost_in = rates['in'] * (total_tokens_in/1000)
+ cost_out = rates['out'] * (total_tokens_out/1000)
+ total_cost = cost_in + cost_out
+ else:
+ raise ValueError(f"LLM version {LLM_version} not found in the cost data")
+
+ return cost_in, cost_out, total_cost, rates['in'], rates['out']
+
+def create_google_ocr_yaml_config(output_file, dir_images_local, dir_output):
+ # Define the configuration dictionary
+ config = {
+ 'leafmachine': {
+ 'LLM_version': 'PaLM 2',
+ 'archival_component_detector': {
+ 'detector_iteration': 'PREP_final',
+ 'detector_type': 'Archival_Detector',
+ 'detector_version': 'PREP_final',
+ 'detector_weights': 'best.pt',
+ 'do_save_prediction_overlay_images': True,
+ 'ignore_objects_for_overlay': [],
+ 'minimum_confidence_threshold': 0.5
+ },
+ 'cropped_components': {
+ 'binarize_labels': False,
+ 'binarize_labels_skeletonize': False,
+ 'do_save_cropped_annotations': True,
+ 'save_cropped_annotations': ['label', 'barcode'],
+ 'save_per_annotation_class': True,
+ 'save_per_image': False
+ },
+ 'data': {
+ 'do_apply_conversion_factor': False,
+ 'include_darwin_core_data_from_combined_file': False,
+ 'save_individual_csv_files_landmarks': False,
+ 'save_individual_csv_files_measurements': False,
+ 'save_individual_csv_files_rulers': False,
+ 'save_individual_efd_files': False,
+ 'save_json_measurements': False,
+ 'save_json_rulers': False
+ },
+ 'do': {
+ 'check_for_corrupt_images_make_vertical': True,
+ 'check_for_illegal_filenames': False
+ },
+ 'logging': {
+ 'log_level': None
+ },
+ 'modules': {
+ 'specimen_crop': True
+ },
+ 'overlay': {
+ 'alpha_transparency_archival': 0.3,
+ 'alpha_transparency_plant': 0,
+ 'alpha_transparency_seg_partial_leaf': 0.3,
+ 'alpha_transparency_seg_whole_leaf': 0.4,
+ 'ignore_archival_detections_classes': [],
+ 'ignore_landmark_classes': [],
+ 'ignore_plant_detections_classes': ['leaf_whole', 'specimen'],
+ 'line_width_archival': 12,
+ 'line_width_efd': 12,
+ 'line_width_plant': 12,
+ 'line_width_seg': 12,
+ 'overlay_background_color': 'black',
+ 'overlay_dpi': 300,
+ 'save_overlay_to_jpgs': True,
+ 'save_overlay_to_pdf': False,
+ 'show_archival_detections': True,
+ 'show_landmarks': True,
+ 'show_plant_detections': True,
+ 'show_segmentations': True
+ },
+ 'print': {
+ 'optional_warnings': True,
+ 'verbose': True
+ },
+ 'project': {
+ 'batch_size': 500,
+ 'build_new_embeddings_database': False,
+ 'catalog_numerical_only': False,
+ 'continue_run_from_partial_xlsx': '',
+ 'delete_all_temps': False,
+ 'delete_temps_keep_VVE': False,
+ 'dir_images_local': dir_images_local,
+ 'dir_output': dir_output,
+ 'embeddings_database_name': 'SLTP_UM_AllAsiaMinimalInRegion',
+ 'image_location': 'local',
+ 'num_workers': 1,
+ 'path_to_domain_knowledge_xlsx': '',
+ 'prefix_removal': '',
+ 'prompt_version': 'Version 2 PaLM 2',
+ 'run_name': 'google_vision_ocr_test',
+ 'suffix_removal': '',
+ 'use_domain_knowledge': False
+ },
+ 'use_RGB_label_images': False
+ }
+ }
+ # Generate the YAML string from the data structure
+ validate_dir(os.path.dirname(output_file))
+ yaml_str = yaml.dump(config)
+
+ # Write the YAML string to a file
+ with open(output_file, 'w') as file:
+ file.write(yaml_str)
+
+def test_GPU():
+ info = []
+ success = False
+
+ if torch.cuda.is_available():
+ num_gpus = torch.cuda.device_count()
+ info.append(f"Number of GPUs: {num_gpus}")
+
+ for i in range(num_gpus):
+ gpu = torch.cuda.get_device_properties(i)
+ info.append(f"GPU {i}: {gpu.name}")
+
+ success = True
+ else:
+ info.append("No GPU found!")
+ info.append("LeafMachine2 image cropping and embedding search will be slow or not possible.")
+
+ return success, info
+
+
+# def load_cfg(pathToCfg):
+# try:
+# with open(os.path.join(pathToCfg,"LeafMachine2.yaml"), "r") as ymlfile:
+# cfg = yaml.full_load(ymlfile)
+# except:
+# with open(os.path.join(os.path.dirname(os.path.dirname(pathToCfg)),"LeafMachine2.yaml"), "r") as ymlfile:
+# cfg = yaml.full_load(ymlfile)
+# return cfg
+
+# def load_cfg_VV(pathToCfg):
+# try:
+# with open(os.path.join(pathToCfg,"VoucherVision.yaml"), "r") as ymlfile:
+# cfg = yaml.full_load(ymlfile)
+# except:
+# with open(os.path.join(os.path.dirname(os.path.dirname(pathToCfg)),"VoucherVision.yaml"), "r") as ymlfile:
+# cfg = yaml.full_load(ymlfile)
+# return cfg
+
+def load_cfg(pathToCfg, system='LeafMachine2'):
+ if system not in ['LeafMachine2', 'VoucherVision', 'SpecimenCrop']:
+ raise ValueError("Invalid system. Expected 'LeafMachine2', 'VoucherVision' or 'SpecimenCrop'.")
+
+ try:
+ with open(os.path.join(pathToCfg, f"{system}.yaml"), "r") as ymlfile:
+ cfg = yaml.full_load(ymlfile)
+ except:
+ with open(os.path.join(os.path.dirname(os.path.dirname(pathToCfg)), f"{system}.yaml"), "r") as ymlfile:
+ cfg = yaml.full_load(ymlfile)
+ return cfg
+
+
+def import_csv(full_path):
+ csv_data = pd.read_csv(full_path,sep=',',header=0, low_memory=False, dtype=str)
+ return csv_data
+
+def import_tsv(full_path):
+ csv_data = pd.read_csv(full_path,sep='\t',header=0, low_memory=False, dtype=str)
+ return csv_data
+
+def parse_cfg():
+ parser = argparse.ArgumentParser(
+ description='Parse inputs to read config file',
+ formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+
+ optional_args = parser._action_groups.pop()
+ required_args = parser.add_argument_group('MANDATORY arguments')
+ required_args.add_argument('--path-to-cfg',
+ type=str,
+ required=True,
+ help='Path to config file - LeafMachine.yaml. Do not include the file name, just the parent dir.')
+
+ parser._action_groups.append(optional_args)
+ args = parser.parse_args()
+ return args
+
+def check_for_subdirs(cfg):
+ original_in = cfg['leafmachine']['project']['dir_images_local']
+ dirs_list = []
+ run_name = []
+ has_subdirs = False
+ if os.path.isdir(original_in):
+ # list contents of the directory
+ contents = os.listdir(original_in)
+
+ # check if any of the contents is a directory
+ subdirs = [f for f in contents if os.path.isdir(os.path.join(original_in, f))]
+
+ if len(subdirs) > 0:
+ print("The directory contains subdirectories:")
+ for subdir in subdirs:
+ has_subdirs = True
+ print(os.path.join(original_in, subdir))
+ dirs_list.append(os.path.join(original_in, subdir))
+ run_name.append(subdir)
+ else:
+ print("The directory does not contain any subdirectories.")
+ dirs_list.append(original_in)
+ run_name.append(cfg['leafmachine']['project']['run_name'])
+
+ else:
+ print("The specified path is not a directory.")
+
+ return run_name, dirs_list, has_subdirs
+
+def check_for_subdirs_VV(cfg):
+ original_in = cfg['leafmachine']['project']['dir_images_local']
+ dirs_list = []
+ run_name = []
+ has_subdirs = False
+ if os.path.isdir(original_in):
+ dirs_list.append(original_in)
+ run_name.append(os.path.basename(os.path.normpath(original_in)))
+ # list contents of the directory
+ contents = os.listdir(original_in)
+
+ # check if any of the contents is a directory
+ subdirs = [f for f in contents if os.path.isdir(os.path.join(original_in, f))]
+
+ if len(subdirs) > 0:
+ print("The directory contains subdirectories:")
+ for subdir in subdirs:
+ has_subdirs = True
+ print(os.path.join(original_in, subdir))
+ dirs_list.append(os.path.join(original_in, subdir))
+ run_name.append(subdir)
+ else:
+ print("The directory does not contain any subdirectories.")
+ dirs_list.append(original_in)
+ run_name.append(cfg['leafmachine']['project']['run_name'])
+
+ else:
+ print("The specified path is not a directory.")
+
+ return run_name, dirs_list, has_subdirs
+
+def get_datetime():
+ day = "_".join([str(datetime.datetime.now().strftime("%Y")),str(datetime.datetime.now().strftime("%m")),str(datetime.datetime.now().strftime("%d"))])
+ time = "-".join([str(datetime.datetime.now().strftime("%H")),str(datetime.datetime.now().strftime("%M")),str(datetime.datetime.now().strftime("%S"))])
+ new_time = "__".join([day,time])
+ return new_time
+
+def save_config_file(cfg, logger, Dirs):
+ logger.info("Save config file")
+ name_yaml = ''.join([Dirs.run_name,'.yaml'])
+ write_yaml(cfg, os.path.join(Dirs.path_config_file, name_yaml))
+
+def write_yaml(cfg, path_cfg):
+ with open(path_cfg, 'w') as file:
+ yaml.dump(cfg, file)
+
+def split_into_batches(Project, logger, cfg):
+ logger.name = 'Creating Batches'
+ n_batches, n_images = Project.process_in_batches(cfg)
+ m = f'Created {n_batches} Batches to Process {n_images} Images'
+ logger.info(m)
+ return Project, n_batches, m
+
+def make_images_in_dir_vertical(dir_images_unprocessed, cfg):
+ if cfg['leafmachine']['do']['check_for_corrupt_images_make_vertical']:
+ n_rotate = 0
+ n_corrupt = 0
+ n_total = len(os.listdir(dir_images_unprocessed))
+ for image_name_jpg in tqdm(os.listdir(dir_images_unprocessed), desc=f'{bcolors.BOLD} Checking Image Dimensions{bcolors.ENDC}',colour="cyan",position=0,total = n_total):
+ if image_name_jpg.endswith((".jpg",".JPG",".jpeg",".JPEG")):
+ try:
+ image = cv2.imread(os.path.join(dir_images_unprocessed, image_name_jpg))
+ h, w, img_c = image.shape
+ image, img_h, img_w, did_rotate = make_image_vertical(image, h, w, do_rotate_180=False)
+ if did_rotate:
+ n_rotate += 1
+ cv2.imwrite(os.path.join(dir_images_unprocessed,image_name_jpg), image)
+ except:
+ n_corrupt +=1
+ os.remove(os.path.join(dir_images_unprocessed, image_name_jpg))
+ # TODO check that below works as intended
+ elif image_name_jpg.endswith((".tiff",".tif",".png",".PNG",".TIFF",".TIF",".jp2",".JP2",".bmp",".BMP",".dib",".DIB")):
+ try:
+ image = cv2.imread(os.path.join(dir_images_unprocessed, image_name_jpg))
+ h, w, img_c = image.shape
+ image, img_h, img_w, did_rotate = make_image_vertical(image, h, w, do_rotate_180=False)
+ if did_rotate:
+ n_rotate += 1
+ image_name_jpg = '.'.join([image_name_jpg.split('.')[0], 'jpg'])
+ cv2.imwrite(os.path.join(dir_images_unprocessed,image_name_jpg), image)
+ except:
+ n_corrupt +=1
+ os.remove(os.path.join(dir_images_unprocessed, image_name_jpg))
+ m = ''.join(['Number of Images Rotated: ', str(n_rotate)])
+ Print_Verbose(cfg, 2, m).bold()
+ m2 = ''.join(['Number of Images Corrupted: ', str(n_corrupt)])
+ if n_corrupt > 0:
+ Print_Verbose(cfg, 2, m2).warning
+ else:
+ Print_Verbose(cfg, 2, m2).bold
+
+def make_image_vertical(image, h, w, do_rotate_180):
+ did_rotate = False
+ if do_rotate_180:
+ # try:
+ image = cv2.rotate(image, cv2.ROTATE_180)
+ img_h, img_w, img_c = image.shape
+ did_rotate = True
+ # print(" Rotated 180")
+ else:
+ if h < w:
+ # try:
+ image = cv2.rotate(image, cv2.ROTATE_90_CLOCKWISE)
+ img_h, img_w, img_c = image.shape
+ did_rotate = True
+ # print(" Rotated 90 CW")
+ elif h >= w:
+ image = image
+ img_h = h
+ img_w = w
+ # print(" Not Rotated")
+ return image, img_h, img_w, did_rotate
+
+
+def make_image_horizontal(image, h, w, do_rotate_180):
+ if h > w:
+ if do_rotate_180:
+ image = cv2.rotate(image, cv2.ROTATE_180)
+ return cv2.rotate(image, cv2.ROTATE_90_COUNTERCLOCKWISE), w, h, True
+ return image, w, h, False
+
+def make_images_in_dir_horizontal(dir_images_unprocessed, cfg):
+ # if cfg['leafmachine']['do']['check_for_corrupt_images_make_horizontal']:
+ n_rotate = 0
+ n_corrupt = 0
+ n_total = len(os.listdir(dir_images_unprocessed))
+ for image_name_jpg in tqdm(os.listdir(dir_images_unprocessed), desc=f'{bcolors.BOLD} Checking Image Dimensions{bcolors.ENDC}', colour="cyan", position=0, total=n_total):
+ if image_name_jpg.endswith((".jpg",".JPG",".jpeg",".JPEG")):
+ try:
+ image = cv2.imread(os.path.join(dir_images_unprocessed, image_name_jpg))
+ h, w, img_c = image.shape
+ image, img_h, img_w, did_rotate = make_image_horizontal(image, h, w, do_rotate_180=False)
+ if did_rotate:
+ n_rotate += 1
+ cv2.imwrite(os.path.join(dir_images_unprocessed,image_name_jpg), image)
+ except:
+ n_corrupt +=1
+ os.remove(os.path.join(dir_images_unprocessed, image_name_jpg))
+ # TODO check that below works as intended
+ elif image_name_jpg.endswith((".tiff",".tif",".png",".PNG",".TIFF",".TIF",".jp2",".JP2",".bmp",".BMP",".dib",".DIB")):
+ try:
+ image = cv2.imread(os.path.join(dir_images_unprocessed, image_name_jpg))
+ h, w, img_c = image.shape
+ image, img_h, img_w, did_rotate = make_image_horizontal(image, h, w, do_rotate_180=False)
+ if did_rotate:
+ n_rotate += 1
+ image_name_jpg = '.'.join([image_name_jpg.split('.')[0], 'jpg'])
+ cv2.imwrite(os.path.join(dir_images_unprocessed,image_name_jpg), image)
+ except:
+ n_corrupt +=1
+ os.remove(os.path.join(dir_images_unprocessed, image_name_jpg))
+ m = ''.join(['Number of Images Rotated: ', str(n_rotate)])
+ print(m)
+ # Print_Verbose(cfg, 2, m).bold()
+ m2 = ''.join(['Number of Images Corrupted: ', str(n_corrupt)])
+ print(m2)
+
+
+@dataclass
+class Print_Verbose_Error():
+ cfg: str = ''
+ indent_level: int = 0
+ message: str = ''
+ error: str = ''
+
+ def __init__(self, cfg,indent_level,message,error) -> None:
+ self.cfg = cfg
+ self.indent_level = indent_level
+ self.message = message
+ self.error = error
+
+ def print_error_to_console(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['optional_warnings']:
+ print(f"{bcolors.FAIL}{white_space}{self.message} ERROR: {self.error}{bcolors.ENDC}")
+
+ def print_warning_to_console(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['optional_warnings']:
+ print(f"{bcolors.WARNING}{white_space}{self.message} ERROR: {self.error}{bcolors.ENDC}")
+
+@dataclass
+class Print_Verbose():
+ cfg: str = ''
+ indent_level: int = 0
+ message: str = ''
+
+ def __init__(self, cfg, indent_level, message) -> None:
+ self.cfg = cfg
+ self.indent_level = indent_level
+ self.message = message
+
+ def bold(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['verbose']:
+ print(f"{bcolors.BOLD}{white_space}{self.message}{bcolors.ENDC}")
+
+ def green(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['verbose']:
+ print(f"{bcolors.OKGREEN}{white_space}{self.message}{bcolors.ENDC}")
+
+ def cyan(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['verbose']:
+ print(f"{bcolors.OKCYAN}{white_space}{self.message}{bcolors.ENDC}")
+
+ def blue(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['verbose']:
+ print(f"{bcolors.OKBLUE}{white_space}{self.message}{bcolors.ENDC}")
+
+ def warning(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['verbose']:
+ print(f"{bcolors.WARNING}{white_space}{self.message}{bcolors.ENDC}")
+
+ def plain(self):
+ white_space = " " * 5 * self.indent_level
+ if self.cfg['leafmachine']['print']['verbose']:
+ print(f"{white_space}{self.message}")
+
+def print_main_start(message):
+ indent_level = 1
+ white_space = " " * 5 * indent_level
+ end = " " * int(80 - len(message) - len(white_space))
+ # end_white_space = " " * end
+ blank = " " * 80
+ print(f"{bcolors.CBLUEBG2}{blank}{bcolors.ENDC}")
+ print(f"{bcolors.CBLUEBG2}{white_space}{message}{end}{bcolors.ENDC}")
+ print(f"{bcolors.CBLUEBG2}{blank}{bcolors.ENDC}")
+
+def print_main_success(message):
+ indent_level = 1
+ white_space = " " * 5 * indent_level
+ end = " " * int(80 - len(message) - len(white_space))
+ blank = " " * 80
+ # end_white_space = " " * end
+ print(f"{bcolors.CGREENBG2}{blank}{bcolors.ENDC}")
+ print(f"{bcolors.CGREENBG2}{white_space}{message}{end}{bcolors.ENDC}")
+ print(f"{bcolors.CGREENBG2}{blank}{bcolors.ENDC}")
+
+def print_main_warn(message):
+ indent_level = 1
+ white_space = " " * 5 * indent_level
+ end = " " * int(80 - len(message) - len(white_space))
+ # end_white_space = " " * end
+ blank = " " * 80
+ print(f"{bcolors.CYELLOWBG2}{blank}{bcolors.ENDC}")
+ print(f"{bcolors.CYELLOWBG2}{white_space}{message}{end}{bcolors.ENDC}")
+ print(f"{bcolors.CYELLOWBG2}{blank}{bcolors.ENDC}")
+
+def print_main_fail(message):
+ indent_level = 1
+ white_space = " " * 5 * indent_level
+ end = " " * int(80 - len(message) - len(white_space))
+ # end_white_space = " " * end
+ blank = " " * 80
+ print(f"{bcolors.CREDBG2}{blank}{bcolors.ENDC}")
+ print(f"{bcolors.CREDBG2}{white_space}{message}{end}{bcolors.ENDC}")
+ print(f"{bcolors.CREDBG2}{blank}{bcolors.ENDC}")
+
+def print_main_info(message):
+ indent_level = 2
+ white_space = " " * 5 * indent_level
+ end = " " * int(80 - len(message) - len(white_space))
+ # end_white_space = " " * end
+ print(f"{bcolors.CGREYBG}{white_space}{message}{end}{bcolors.ENDC}")
+
+# def report_config(dir_home, cfg_file_path):
+# print_main_start("Loading Configuration File")
+# if cfg_file_path == None:
+# print_main_info(''.join([os.path.join(dir_home, 'LeafMachine2.yaml')]))
+# elif cfg_file_path == 'test_installation':
+# print_main_info(''.join([os.path.join(dir_home, 'demo','LeafMachine2_demo.yaml')]))
+# else:
+# print_main_info(cfg_file_path)
+
+# def report_config_VV(dir_home, cfg_file_path):
+# print_main_start("Loading Configuration File")
+# if cfg_file_path == None:
+# print_main_info(''.join([os.path.join(dir_home, 'VoucherVision.yaml')]))
+# elif cfg_file_path == 'test_installation':
+# print_main_info(''.join([os.path.join(dir_home, 'demo','VoucherVision_demo.yaml')]))
+# else:
+# print_main_info(cfg_file_path)
+
+def report_config(dir_home, cfg_file_path, system='VoucherVision'):
+ print_main_start("Loading Configuration File")
+
+ if system not in ['LeafMachine2', 'VoucherVision', 'SpecimenCrop']:
+ raise ValueError("Invalid system. Expected 'LeafMachine2' or 'VoucherVision' or 'SpecimenCrop'.")
+
+ if cfg_file_path == None:
+ print_main_info(''.join([os.path.join(dir_home, f'{system}.yaml')]))
+ elif cfg_file_path == 'test_installation':
+ print_main_info(''.join([os.path.join(dir_home, 'demo', f'{system}_demo.yaml')]))
+ else:
+ print_main_info(cfg_file_path)
+
+
+def make_file_names_valid(dir, cfg):
+ if cfg['leafmachine']['do']['check_for_illegal_filenames']:
+ n_total = len(os.listdir(dir))
+ for file in tqdm(os.listdir(dir), desc=f'{bcolors.HEADER} Removing illegal characters from file names{bcolors.ENDC}',colour="cyan",position=0,total = n_total):
+ name = Path(file).stem
+ ext = Path(file).suffix
+ name_cleaned = re.sub(r"[^a-zA-Z0-9_-]","-",name)
+ name_new = ''.join([name_cleaned,ext])
+ i = 0
+ try:
+ os.rename(os.path.join(dir,file), os.path.join(dir,name_new))
+ except:
+ while os.path.exists(os.path.join(dir,name_new)):
+ i += 1
+ name_new = '_'.join([name_cleaned, str(i), ext])
+ os.rename(os.path.join(dir,file), os.path.join(dir,name_new))
+
+# def load_config_file(dir_home, cfg_file_path):
+# if cfg_file_path == None: # Default path
+# return load_cfg(dir_home)
+# else:
+# if cfg_file_path == 'test_installation':
+# path_cfg = os.path.join(dir_home,'demo','LeafMachine2_demo.yaml')
+# return get_cfg_from_full_path(path_cfg)
+# else: # Custom path
+# return get_cfg_from_full_path(cfg_file_path)
+
+# def load_config_file_VV(dir_home, cfg_file_path):
+# if cfg_file_path == None: # Default path
+# return load_cfg_VV(dir_home)
+# else:
+# if cfg_file_path == 'test_installation':
+# path_cfg = os.path.join(dir_home,'demo','VoucherVision_demo.yaml')
+# return get_cfg_from_full_path(path_cfg)
+# else: # Custom path
+# return get_cfg_from_full_path(cfg_file_path)
+
+def load_config_file(dir_home, cfg_file_path, system='LeafMachine2'):
+ if system not in ['LeafMachine2', 'VoucherVision', 'SpecimenCrop']:
+ raise ValueError("Invalid system. Expected 'LeafMachine2' or 'VoucherVision' or 'SpecimenCrop'.")
+
+ if cfg_file_path is None: # Default path
+ if system == 'LeafMachine2':
+ return load_cfg(dir_home, system='LeafMachine2') # For LeafMachine2
+
+ elif system == 'VoucherVision': # VoucherVision
+ return load_cfg(dir_home, system='VoucherVision') # For VoucherVision
+
+ elif system == 'SpecimenCrop': # SpecimenCrop
+ return load_cfg(dir_home, system='SpecimenCrop') # For SpecimenCrop
+
+ else:
+ if cfg_file_path == 'test_installation':
+ path_cfg = os.path.join(dir_home, 'demo', f'{system}_demo.yaml')
+ return get_cfg_from_full_path(path_cfg)
+ else: # Custom path
+ return get_cfg_from_full_path(cfg_file_path)
+
+
+def load_config_file_testing(dir_home, cfg_file_path):
+ if cfg_file_path == None: # Default path
+ return load_cfg(dir_home)
+ else:
+ if cfg_file_path == 'test_installation':
+ path_cfg = os.path.join(dir_home,'demo','demo.yaml')
+ return get_cfg_from_full_path(path_cfg)
+ else: # Custom path
+ return get_cfg_from_full_path(cfg_file_path)
+
+def subset_dir_images(cfg, Project, Dirs):
+ if cfg['leafmachine']['project']['process_subset_of_images']:
+ dir_images_subset = cfg['leafmachine']['project']['dir_images_subset']
+ num_images_per_species = cfg['leafmachine']['project']['n_images_per_species']
+ if cfg['leafmachine']['project']['species_list'] is not None:
+ species_list = import_csv(cfg['leafmachine']['project']['species_list'])
+ species_list = species_list.iloc[:, 0].tolist()
+ else:
+ species_list = None
+
+ validate_dir(dir_images_subset)
+
+ species_counts = {}
+ filenames = os.listdir(Project.dir_images)
+ random.shuffle(filenames)
+ for filename in filenames:
+ species_name = filename.split('.')[0]
+ species_name = species_name.split('_')[2:]
+ species_name = '_'.join([species_name[0], species_name[1], species_name[2]])
+
+ if (species_list is None) or ((species_name in species_list) and (species_list is not None)):
+
+ if species_name not in species_counts:
+ species_counts[species_name] = 0
+
+ if species_counts[species_name] < num_images_per_species:
+ species_counts[species_name] += 1
+ src_path = os.path.join(Project.dir_images, filename)
+ dest_path = os.path.join(dir_images_subset, filename)
+ shutil.copy(src_path, dest_path)
+
+ Project.dir_images = dir_images_subset
+
+ subset_csv_name = os.path.join(Dirs.dir_images_subset, '.'.join([Dirs.run_name, 'csv']))
+ df = pd.DataFrame({'species_name': list(species_counts.keys()), 'count': list(species_counts.values())})
+ df.to_csv(subset_csv_name, index=False)
+ return Project
+ else:
+ return Project
+
+'''# Define function to be executed by each worker
+def worker_crop(rank, cfg, dir_home, Project, Dirs):
+ # Set worker seed based on rank
+ np.random.seed(rank)
+ # Call function for this worker
+ crop_detections_from_images(cfg, dir_home, Project, Dirs)
+
+def crop_detections_from_images(cfg, dir_home, Project, Dirs):
+ num_workers = 6
+
+ # Initialize and start worker processes
+ processes = []
+ for rank in range(num_workers):
+ p = mp.Process(target=worker_crop, args=(rank, cfg, dir_home, Project, Dirs))
+ p.start()
+ processes.append(p)
+
+ # Wait for all worker processes to finish
+ for p in processes:
+ p.join()'''
+
+def crop_detections_from_images_worker_VV(filename, analysis, Project, Dirs, save_per_image, save_per_class, save_list, binarize_labels):
+ try:
+ full_image = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename, 'jpg'])))
+ except:
+ full_image = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename, 'jpeg'])))
+
+ try:
+ archival = analysis['Detections_Archival_Components']
+ has_archival = True
+ except:
+ has_archival = False
+
+ try:
+ plant = analysis['Detections_Plant_Components']
+ has_plant = True
+ except:
+ has_plant = False
+
+ if has_archival and (save_per_image or save_per_class):
+ crop_component_from_yolo_coords_VV('ARCHIVAL', Dirs, analysis, archival, full_image, filename, save_per_image, save_per_class, save_list)
+
+def crop_detections_from_images_worker(filename, analysis, Project, Dirs, save_per_image, save_per_class, save_list, binarize_labels):
+ try:
+ full_image = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename, 'jpg'])))
+ except:
+ full_image = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename, 'jpeg'])))
+
+ try:
+ archival = analysis['Detections_Archival_Components']
+ has_archival = True
+ except:
+ has_archival = False
+
+ try:
+ plant = analysis['Detections_Plant_Components']
+ has_plant = True
+ except:
+ has_plant = False
+
+ if has_archival and (save_per_image or save_per_class):
+ crop_component_from_yolo_coords('ARCHIVAL', Dirs, analysis, archival, full_image, filename, save_per_image, save_per_class, save_list)
+ if has_plant and (save_per_image or save_per_class):
+ crop_component_from_yolo_coords('PLANT', Dirs, analysis, plant, full_image, filename, save_per_image, save_per_class, save_list)
+
+
+def crop_detections_from_images(cfg, logger, dir_home, Project, Dirs, batch_size=50):
+ t2_start = perf_counter()
+ logger.name = 'Crop Components'
+
+ if cfg['leafmachine']['cropped_components']['do_save_cropped_annotations']:
+ detections = cfg['leafmachine']['cropped_components']['save_cropped_annotations']
+ logger.info(f"Cropping {detections} components from images")
+
+ save_per_image = cfg['leafmachine']['cropped_components']['save_per_image']
+ save_per_class = cfg['leafmachine']['cropped_components']['save_per_annotation_class']
+ save_list = cfg['leafmachine']['cropped_components']['save_cropped_annotations']
+ try:
+ binarize_labels = cfg['leafmachine']['cropped_components']['binarize_labels']
+ except:
+ binarize_labels = False
+ if cfg['leafmachine']['project']['batch_size'] is None:
+ batch_size = 50
+ else:
+ batch_size = int(cfg['leafmachine']['project']['batch_size'])
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 4
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ if binarize_labels:
+ save_per_class = True
+
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = []
+ for i in range(0, len(Project.project_data), batch_size):
+ batch = list(Project.project_data.items())[i:i+batch_size]
+ # print(f'Cropping Detections from Images {i} to {i+batch_size}')
+ logger.info(f'Cropping {detections} from images {i} to {i+batch_size} [{len(Project.project_data)}]')
+ for filename, analysis in batch:
+ if len(analysis) != 0:
+ futures.append(executor.submit(crop_detections_from_images_worker, filename, analysis, Project, Dirs, save_per_image, save_per_class, save_list, binarize_labels))
+
+ for future in concurrent.futures.as_completed(futures):
+ pass
+ futures.clear()
+
+ t2_stop = perf_counter()
+ logger.info(f"Save cropped components --- elapsed time: {round(t2_stop - t2_start)} seconds")
+
+def crop_detections_from_images_VV(cfg, logger, dir_home, Project, Dirs, batch_size=50):
+ t2_start = perf_counter()
+ logger.name = 'Crop Components'
+
+ if cfg['leafmachine']['cropped_components']['do_save_cropped_annotations']:
+ detections = cfg['leafmachine']['cropped_components']['save_cropped_annotations']
+ logger.info(f"Cropping {detections} components from images")
+
+ save_per_image = cfg['leafmachine']['cropped_components']['save_per_image']
+ save_per_class = cfg['leafmachine']['cropped_components']['save_per_annotation_class']
+ save_list = cfg['leafmachine']['cropped_components']['save_cropped_annotations']
+ binarize_labels = cfg['leafmachine']['cropped_components']['binarize_labels']
+ if cfg['leafmachine']['project']['batch_size'] is None:
+ batch_size = 50
+ else:
+ batch_size = int(cfg['leafmachine']['project']['batch_size'])
+ if cfg['leafmachine']['project']['num_workers'] is None:
+ num_workers = 4
+ else:
+ num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+ if binarize_labels:
+ save_per_class = True
+
+ with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+ futures = []
+ for i in range(0, len(Project.project_data), batch_size):
+ batch = list(Project.project_data.items())[i:i+batch_size]
+ # print(f'Cropping Detections from Images {i} to {i+batch_size}')
+ logger.info(f'Cropping {detections} from images {i} to {i+batch_size} [{len(Project.project_data)}]')
+ for filename, analysis in batch:
+ if len(analysis) != 0:
+ futures.append(executor.submit(crop_detections_from_images_worker_VV, filename, analysis, Project, Dirs, save_per_image, save_per_class, save_list, binarize_labels))
+
+ for future in concurrent.futures.as_completed(futures):
+ pass
+ futures.clear()
+
+ t2_stop = perf_counter()
+ logger.info(f"Save cropped components --- elapsed time: {round(t2_stop - t2_start)} seconds")
+# def crop_detections_from_images_VV(cfg, logger, dir_home, Project, Dirs, batch_size=50):
+# t2_start = perf_counter()
+# logger.name = 'Crop Components'
+
+# if cfg['leafmachine']['cropped_components']['do_save_cropped_annotations']:
+# detections = cfg['leafmachine']['cropped_components']['save_cropped_annotations']
+# logger.info(f"Cropping {detections} components from images")
+
+# save_per_image = cfg['leafmachine']['cropped_components']['save_per_image']
+# save_per_class = cfg['leafmachine']['cropped_components']['save_per_annotation_class']
+# save_list = cfg['leafmachine']['cropped_components']['save_cropped_annotations']
+# binarize_labels = cfg['leafmachine']['cropped_components']['binarize_labels']
+# if cfg['leafmachine']['project']['batch_size'] is None:
+# batch_size = 50
+# else:
+# batch_size = int(cfg['leafmachine']['project']['batch_size'])
+
+# if binarize_labels:
+# save_per_class = True
+
+# for i in range(0, len(Project.project_data), batch_size):
+# batch = list(Project.project_data.items())[i:i+batch_size]
+# logger.info(f"Cropping {detections} from images {i} to {i+batch_size} [{len(Project.project_data)}]")
+# for filename, analysis in batch:
+# if len(analysis) != 0:
+# crop_detections_from_images_worker_VV(filename, analysis, Project, Dirs, save_per_image, save_per_class, save_list, binarize_labels)
+
+# t2_stop = perf_counter()
+# logger.info(f"Save cropped components --- elapsed time: {round(t2_stop - t2_start)} seconds")
+
+
+# def crop_detections_from_images_SpecimenCrop(cfg, logger, dir_home, Project, Dirs, original_img_dir=None, batch_size=50):
+# t2_start = perf_counter()
+# logger.name = 'Crop Components --- Specimen Crop'
+
+# if cfg['leafmachine']['modules']['specimen_crop']:
+# # save_list = ['ruler', 'barcode', 'colorcard', 'label', 'map', 'envelope', 'photo', 'attached_item', 'weights',
+# # 'leaf_whole', 'leaf_partial', 'leaflet', 'seed_fruit_one', 'seed_fruit_many', 'flower_one', 'flower_many', 'bud', 'specimen', 'roots', 'wood']
+# save_list = cfg['leafmachine']['cropped_components']['include_these_objects_in_specimen_crop']
+
+# logger.info(f"Cropping to include {save_list} components from images")
+
+# if cfg['leafmachine']['project']['batch_size'] is None:
+# batch_size = 50
+# else:
+# batch_size = int(cfg['leafmachine']['project']['batch_size'])
+# if cfg['leafmachine']['project']['num_workers'] is None:
+# num_workers = 4
+# else:
+# num_workers = int(cfg['leafmachine']['project']['num_workers'])
+
+# with concurrent.futures.ThreadPoolExecutor(max_workers=num_workers) as executor:
+# futures = []
+# for i in range(0, len(Project.project_data), batch_size):
+# batch = list(Project.project_data.items())[i:i+batch_size]
+# # print(f'Cropping Detections from Images {i} to {i+batch_size}')
+# logger.info(f'Cropping {save_list} from images {i} to {i+batch_size} [{len(Project.project_data)}]')
+# for filename, analysis in batch:
+# if len(analysis) != 0:
+# futures.append(executor.submit(crop_detections_from_images_worker_SpecimenCrop, filename, analysis, Project, Dirs, save_list, original_img_dir))
+
+# for future in concurrent.futures.as_completed(futures):
+# pass
+# futures.clear()
+
+# t2_stop = perf_counter()
+# logger.info(f"Save cropped components --- elapsed time: {round(t2_stop - t2_start)} seconds")
+
+'''
+# Single threaded
+def crop_detections_from_images(cfg, dir_home, Project, Dirs):
+ if cfg['leafmachine']['cropped_components']['do_save_cropped_annotations']:
+ save_per_image = cfg['leafmachine']['cropped_components']['save_per_image']
+ save_per_class = cfg['leafmachine']['cropped_components']['save_per_annotation_class']
+ save_list = cfg['leafmachine']['cropped_components']['save_cropped_annotations']
+ binarize_labels = cfg['leafmachine']['cropped_components']['binarize_labels']
+ if binarize_labels:
+ save_per_class = True
+
+ for filename, analysis in tqdm(Project.project_data.items(), desc=f'{bcolors.BOLD} Cropping Detections from Images{bcolors.ENDC}',colour="cyan",position=0,total = len(Project.project_data.items())):
+ if len(analysis) != 0:
+ try:
+ full_image = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename, 'jpg'])))
+ except:
+ full_image = cv2.imread(os.path.join(Project.dir_images, '.'.join([filename, 'jpeg'])))
+
+ try:
+ archival = analysis['Detections_Archival_Components']
+ has_archival = True
+ except:
+ has_archival = False
+
+ try:
+ plant = analysis['Detections_Plant_Components']
+ has_plant = True
+ except:
+ has_plant = False
+
+ if has_archival and (save_per_image or save_per_class):
+ crop_component_from_yolo_coords('ARCHIVAL', Dirs, analysis, archival, full_image, filename, save_per_image, save_per_class, save_list)
+ if has_plant and (save_per_image or save_per_class):
+ crop_component_from_yolo_coords('PLANT', Dirs, analysis, plant, full_image, filename, save_per_image, save_per_class, save_list)
+'''
+
+
+def process_detections(success, save_list, detections, detection_type, height, width, min_x, min_y, max_x, max_y):
+ for detection in detections:
+ detection_class = detection[0]
+ detection_class = set_index_for_annotation(detection_class, detection_type)
+
+ if (detection_class in save_list) or ('save_all' in save_list):
+ location = yolo_to_position_ruler(detection, height, width)
+ ruler_polygon = [
+ (location[1], location[2]),
+ (location[3], location[2]),
+ (location[3], location[4]),
+ (location[1], location[4])
+ ]
+
+ x_coords = [x for x, y in ruler_polygon]
+ y_coords = [y for x, y in ruler_polygon]
+
+ min_x = min(min_x, *x_coords)
+ min_y = min(min_y, *y_coords)
+ max_x = max(max_x, *x_coords)
+ max_y = max(max_y, *y_coords)
+ success = True
+
+ return min_x, min_y, max_x, max_y, success
+
+
+def crop_component_from_yolo_coords_VV(anno_type, Dirs, analysis, all_detections, full_image, filename, save_per_image, save_per_class, save_list):
+ height = analysis['height']
+ width = analysis['width']
+
+ # Initialize a list to hold all the cropped images
+ cropped_images = []
+
+ if len(all_detections) < 1:
+ print(' MAKE THIS HAVE AN EMPTY PLACEHOLDER') # TODO ###################################################################################
+ else:
+ for detection in all_detections:
+ detection_class = detection[0]
+ detection_class = set_index_for_annotation(detection_class, anno_type)
+
+ if (detection_class in save_list) or ('save_all' in save_list):
+
+ location = yolo_to_position_ruler(detection, height, width)
+ ruler_polygon = [(location[1], location[2]), (location[3], location[2]), (location[3], location[4]), (location[1], location[4])]
+
+ x_coords = [x for x, y in ruler_polygon]
+ y_coords = [y for x, y in ruler_polygon]
+
+ min_x, min_y = min(x_coords), min(y_coords)
+ max_x, max_y = max(x_coords), max(y_coords)
+
+ detection_cropped = full_image[min_y:max_y, min_x:max_x]
+ cropped_images.append(detection_cropped)
+ loc = '-'.join([str(min_x), str(min_y), str(max_x), str(max_y)])
+ detection_cropped_name = '.'.join(['__'.join([filename, detection_class, loc]), 'jpg'])
+ # detection_cropped_name = '.'.join([filename,'jpg'])
+
+ # save_per_image
+ if (detection_class in save_list) and save_per_image:
+ if detection_class == 'label':
+ detection_class2 = 'label_ind'
+ else:
+ detection_class2 = detection_class
+ dir_destination = os.path.join(Dirs.save_per_image, filename, detection_class2)
+ # print(os.path.join(dir_destination,detection_cropped_name))
+ validate_dir(dir_destination)
+ # cv2.imwrite(os.path.join(dir_destination,detection_cropped_name), detection_cropped)
+
+ # save_per_class
+ if (detection_class in save_list) and save_per_class:
+ if detection_class == 'label':
+ detection_class2 = 'label_ind'
+ else:
+ detection_class2 = detection_class
+ dir_destination = os.path.join(Dirs.save_per_annotation_class, detection_class2)
+ # print(os.path.join(dir_destination,detection_cropped_name))
+ validate_dir(dir_destination)
+ # cv2.imwrite(os.path.join(dir_destination,detection_cropped_name), detection_cropped)
+ else:
+ # print(f'detection_class: {detection_class} not in save_list: {save_list}')
+ pass
+
+ # Initialize a list to hold all the acceptable cropped images
+ acceptable_cropped_images = []
+
+ for img in cropped_images:
+ # Calculate the aspect ratio of the image
+ aspect_ratio = min(img.shape[0], img.shape[1]) / max(img.shape[0], img.shape[1])
+ # Only add the image to the acceptable list if the aspect ratio is more square than 1:8
+ if aspect_ratio >= 1/8:
+ acceptable_cropped_images.append(img)
+
+ # Sort acceptable_cropped_images by area (largest first)
+ acceptable_cropped_images.sort(key=lambda img: img.shape[0] * img.shape[1], reverse=True)
+
+
+ # If there are no acceptable cropped images, set combined_image to None or to a placeholder image
+ if not acceptable_cropped_images:
+ combined_image = None # Or a placeholder image here
+ else:
+ # # Recalculate max_width and total_height for acceptable images
+ # max_width = max(img.shape[1] for img in acceptable_cropped_images)
+ # total_height = sum(img.shape[0] for img in acceptable_cropped_images)
+
+ # # Now, combine all the acceptable cropped images into a single image
+ # combined_image = np.zeros((total_height, max_width, 3), dtype=np.uint8)
+
+ # y_offset = 0
+ # for img in acceptable_cropped_images:
+ # combined_image[y_offset:y_offset+img.shape[0], :img.shape[1]] = img
+ # y_offset += img.shape[0]
+ # Start with the first image
+ # Recalculate max_width and total_height for acceptable images
+ max_width = max(img.shape[1] for img in acceptable_cropped_images)
+ total_height = sum(img.shape[0] for img in acceptable_cropped_images)
+ combined_image = np.zeros((total_height, max_width, 3), dtype=np.uint8)
+
+ y_offset = 0
+ y_offset_next_row = 0
+ x_offset = 0
+
+ # Start with the first image
+ combined_image[y_offset:y_offset+acceptable_cropped_images[0].shape[0], :acceptable_cropped_images[0].shape[1]] = acceptable_cropped_images[0]
+ y_offset_next_row += acceptable_cropped_images[0].shape[0]
+
+ # Add the second image below the first one
+ y_offset = y_offset_next_row
+ combined_image[y_offset:y_offset+acceptable_cropped_images[1].shape[0], :acceptable_cropped_images[1].shape[1]] = acceptable_cropped_images[1]
+ y_offset_next_row += acceptable_cropped_images[1].shape[0]
+
+ # Create a list to store the images that are too tall for the current row
+ too_tall_images = []
+
+ # Now try to fill in to the right with the remaining images
+ current_width = acceptable_cropped_images[1].shape[1]
+
+ for img in acceptable_cropped_images[2:]:
+ if current_width + img.shape[1] > max_width:
+ # If this image doesn't fit, start a new row
+ y_offset = y_offset_next_row
+ combined_image[y_offset:y_offset+img.shape[0], :img.shape[1]] = img
+ current_width = img.shape[1]
+ y_offset_next_row = y_offset + img.shape[0]
+ else:
+ # If this image fits, add it to the right
+ max_height = y_offset_next_row - y_offset
+ if img.shape[0] > max_height:
+ too_tall_images.append(img)
+ else:
+ combined_image[y_offset:y_offset+img.shape[0], current_width:current_width+img.shape[1]] = img
+ current_width += img.shape[1]
+
+ # Process the images that were too tall for their rows
+ for img in too_tall_images:
+ y_offset = y_offset_next_row
+ combined_image[y_offset:y_offset+img.shape[0], :img.shape[1]] = img
+ y_offset_next_row += img.shape[0]
+
+ # Trim the combined_image to remove extra black space
+ combined_image = combined_image[:y_offset_next_row]
+
+
+ # save the combined image
+ # if (detection_class in save_list) and save_per_class:
+ dir_destination = os.path.join(Dirs.save_per_annotation_class, 'label')
+ validate_dir(dir_destination)
+ # combined_image_name = '__'.join([filename, detection_class]) + '.jpg'
+ combined_image_name = '.'.join([filename,'jpg'])
+ cv2.imwrite(os.path.join(dir_destination, combined_image_name), combined_image)
+
+ original_image_name = '.'.join([filename,'jpg'])
+ cv2.imwrite(os.path.join(Dirs.save_original, original_image_name), full_image)
+
+
+
+def crop_component_from_yolo_coords(anno_type, Dirs, analysis, all_detections, full_image, filename, save_per_image, save_per_class, save_list):
+ height = analysis['height']
+ width = analysis['width']
+ if len(all_detections) < 1:
+ print(' MAKE THIS HAVE AN EMPTY PLACEHOLDER') # TODO ###################################################################################
+ else:
+ for detection in all_detections:
+ detection_class = detection[0]
+ detection_class = set_index_for_annotation(detection_class, anno_type)
+
+ if (detection_class in save_list) or ('save_all' in save_list):
+
+ location = yolo_to_position_ruler(detection, height, width)
+ ruler_polygon = [(location[1], location[2]), (location[3], location[2]), (location[3], location[4]), (location[1], location[4])]
+
+ x_coords = [x for x, y in ruler_polygon]
+ y_coords = [y for x, y in ruler_polygon]
+
+ min_x, min_y = min(x_coords), min(y_coords)
+ max_x, max_y = max(x_coords), max(y_coords)
+
+ detection_cropped = full_image[min_y:max_y, min_x:max_x]
+ loc = '-'.join([str(min_x), str(min_y), str(max_x), str(max_y)])
+ detection_cropped_name = '.'.join(['__'.join([filename, detection_class, loc]), 'jpg'])
+
+ # save_per_image
+ if (detection_class in save_list) and save_per_image:
+ dir_destination = os.path.join(Dirs.save_per_image, filename, detection_class)
+ # print(os.path.join(dir_destination,detection_cropped_name))
+ validate_dir(dir_destination)
+ cv2.imwrite(os.path.join(dir_destination,detection_cropped_name), detection_cropped)
+
+ # save_per_class
+ if (detection_class in save_list) and save_per_class:
+ dir_destination = os.path.join(Dirs.save_per_annotation_class, detection_class)
+ # print(os.path.join(dir_destination,detection_cropped_name))
+ validate_dir(dir_destination)
+ cv2.imwrite(os.path.join(dir_destination,detection_cropped_name), detection_cropped)
+ else:
+ # print(f'detection_class: {detection_class} not in save_list: {save_list}')
+ pass
+
+def yolo_to_position_ruler(annotation, height, width):
+ return ['ruler',
+ int((annotation[1] * width) - ((annotation[3] * width) / 2)),
+ int((annotation[2] * height) - ((annotation[4] * height) / 2)),
+ int(annotation[3] * width) + int((annotation[1] * width) - ((annotation[3] * width) / 2)),
+ int(annotation[4] * height) + int((annotation[2] * height) - ((annotation[4] * height) / 2))]
+
+
+class bcolors:
+ HEADER = '\033[95m'
+ OKBLUE = '\033[94m'
+ OKCYAN = '\033[96m'
+ OKGREEN = '\033[92m'
+ WARNING = '\033[93m'
+ FAIL = '\033[91m'
+ ENDC = '\033[0m'
+ BOLD = '\033[1m'
+ UNDERLINE = '\033[4m'
+ CEND = '\33[0m'
+ CBOLD = '\33[1m'
+ CITALIC = '\33[3m'
+ CURL = '\33[4m'
+ CBLINK = '\33[5m'
+ CBLINK2 = '\33[6m'
+ CSELECTED = '\33[7m'
+
+ CBLACK = '\33[30m'
+ CRED = '\33[31m'
+ CGREEN = '\33[32m'
+ CYELLOW = '\33[33m'
+ CBLUE = '\33[34m'
+ CVIOLET = '\33[35m'
+ CBEIGE = '\33[36m'
+ CWHITE = '\33[37m'
+
+ CBLACKBG = '\33[40m'
+ CREDBG = '\33[41m'
+ CGREENBG = '\33[42m'
+ CYELLOWBG = '\33[43m'
+ CBLUEBG = '\33[44m'
+ CVIOLETBG = '\33[45m'
+ CBEIGEBG = '\33[46m'
+ CWHITEBG = '\33[47m'
+
+ CGREY = '\33[90m'
+ CRED2 = '\33[91m'
+ CGREEN2 = '\33[92m'
+ CYELLOW2 = '\33[93m'
+ CBLUE2 = '\33[94m'
+ CVIOLET2 = '\33[95m'
+ CBEIGE2 = '\33[96m'
+ CWHITE2 = '\33[97m'
+
+ CGREYBG = '\33[100m'
+ CREDBG2 = '\33[101m'
+ CGREENBG2 = '\33[102m'
+ CYELLOWBG2 = '\33[103m'
+ CBLUEBG2 = '\33[104m'
+ CVIOLETBG2 = '\33[105m'
+ CBEIGEBG2 = '\33[106m'
+ CWHITEBG2 = '\33[107m'
+ CBLUEBG3 = '\33[112m'
+
+
+def set_index_for_annotation(cls,annoType):
+ if annoType == 'PLANT':
+ if cls == 0:
+ annoInd = 'Leaf_WHOLE'
+ elif cls == 1:
+ annoInd = 'Leaf_PARTIAL'
+ elif cls == 2:
+ annoInd = 'Leaflet'
+ elif cls == 3:
+ annoInd = 'Seed_Fruit_ONE'
+ elif cls == 4:
+ annoInd = 'Seed_Fruit_MANY'
+ elif cls == 5:
+ annoInd = 'Flower_ONE'
+ elif cls == 6:
+ annoInd = 'Flower_MANY'
+ elif cls == 7:
+ annoInd = 'Bud'
+ elif cls == 8:
+ annoInd = 'Specimen'
+ elif cls == 9:
+ annoInd = 'Roots'
+ elif cls == 10:
+ annoInd = 'Wood'
+ elif annoType == 'ARCHIVAL':
+ if cls == 0:
+ annoInd = 'Ruler'
+ elif cls == 1:
+ annoInd = 'Barcode'
+ elif cls == 2:
+ annoInd = 'Colorcard'
+ elif cls == 3:
+ annoInd = 'Label'
+ elif cls == 4:
+ annoInd = 'Map'
+ elif cls == 5:
+ annoInd = 'Envelope'
+ elif cls == 6:
+ annoInd = 'Photo'
+ elif cls == 7:
+ annoInd = 'Attached_item'
+ elif cls == 8:
+ annoInd = 'Weights'
+ return annoInd.lower()
+# def set_yaml(path_to_yaml, value):
+# with open('file_to_edit.yaml') as f:
+# doc = yaml.load(f)
+
+# doc['state'] = state
+
+# with open('file_to_edit.yaml', 'w') as f:
+# yaml.dump(doc, f)
\ No newline at end of file
diff --git a/vouchervision/prompt_catalog.py b/vouchervision/prompt_catalog.py
new file mode 100644
index 0000000000000000000000000000000000000000..d9462f346f33e9dd5a5dbd63196ed08afb6eff40
--- /dev/null
+++ b/vouchervision/prompt_catalog.py
@@ -0,0 +1,1313 @@
+from dataclasses import dataclass
+import yaml, json
+
+
+# catalog = PromptCatalog(OCR="Sample OCR text", domain_knowledge_example="Sample domain knowledge", similarity="0.9")
+
+@dataclass
+class PromptCatalog:
+ domain_knowledge_example: str = ""
+ similarity: str = ""
+ OCR: str = ""
+ n_fields: int = 0
+
+ # def PROMPT_UMICH_skeleton_all_asia(self, OCR=None, domain_knowledge_example=None, similarity=None):
+ def prompt_v1_verbose(self, OCR=None, domain_knowledge_example=None, similarity=None):
+ self.OCR = OCR or self.OCR
+ self.domain_knowledge_example = domain_knowledge_example or self.domain_knowledge_example
+ self.similarity = similarity or self.similarity
+ self.n_fields = 22 or self.n_fields
+
+ set_rules = """
+ Please note that your task is to generate a dictionary, following the below rules:
+ 1. Refactor the unstructured OCR text into a dictionary based on the reference dictionary structure (ref_dict).
+ 2. Each field of OCR corresponds to a column of the ref_dict. You should correctly map the values from OCR to the respective fields in ref_dict.
+ 3. If the OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None".
+ 4. If there is a field in the ref_dict that does not have a corresponding value in the OCR text, fill it based on your knowledge but don't generate new information.
+ 5. Do not use any text from the ref_dict values in the new dict, but you must use the headers from ref_dict.
+ 6. Duplicate dictionary fields are not allowed.
+ 7. Only return the new dictionary. You should not explain your answer.
+ 8. Your output should be a Python dictionary represented as a JSON string.
+ """
+
+ umich_all_asia_rules = """{
+ "Catalog Number": {
+ "format": "[Catalog Number]",
+ "null_value": "",
+ "description": "The barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits"
+ },
+ "Genus": {
+ "format": "[Genus] or '[Family] indet' if no genus",
+ "null_value": "",
+ "description": "Taxonomic determination to genus, do capitalize genus"
+ },
+ "Species": {
+ "format": "[species] or 'indet' if no species",
+ "null_value": "",
+ "description": "Taxonomic determination to species, do not capitalize species"
+ },
+ "subspecies": {
+ "format": "[subspecies]",
+ "null_value": "",
+ "description": "Taxonomic determination to subspecies (subsp.)"
+ },
+ "variety": {
+ "format": "[variety]",
+ "null_value": "",
+ "description": "Taxonomic determination to variety (var)"
+ },
+ "forma": {
+ "format": "[form]",
+ "null_value": "",
+ "description": "Taxonomic determination to form (f.)"
+ },
+ "Country": {
+ "format": "[Country]",
+ "null_value": "",
+ "description": "Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abbreviation is given"
+ },
+ "State": {
+ "format": "[Adm. Division 1]",
+ "null_value": "",
+ "description": "Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word"
+ },
+ "County": {
+ "format": "[Adm. Division 2]",
+ "null_value": "",
+ "description": "Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word"
+ },
+ "Locality Name": {
+ "format": "verbatim, if no geographic info: 'no data provided on label of catalog no: [######]', or if illegible: 'locality present but illegible/not translated for catalog no: #######', or if no named locality: 'no named locality for catalog no: #######'",
+ "description": "Description of geographic location or landscape"
+ },
+ "Min Elevation": {
+ "format": "elevation integer",
+ "null_value": "",
+ "description": "Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given"
+ },
+ "Max Elevation": {
+ "format": "elevation integer",
+ "null_value": "",
+ "description": "Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise"
+ },
+ "Elevation Units": {
+ "format": "m",
+ "null_value": "",
+ "description": "'m' only if an elevation is present"
+ },
+ "Verbatim Coordinates": {
+ "format": "[Lat, Long | UTM | TRS]",
+ "null_value": "",
+ "description": "Verbatim coordinates as they appear on the label, fix typos to match standardized GPS coordinate format"
+ },
+ "Datum": {
+ "format": "[WGS84, NAD23 etc.]",
+ "null_value": "",
+ "description": "GPS Datum of coordinates on label; empty string "" if GPS coordinates are not in OCR"
+ },
+ "Cultivated": {
+ "format": "yes",
+ "null_value": "",
+ "description": "Indicates if specimen was grown in cultivation"
+ },
+ "Habitat": {
+ "format": "verbatim",
+ "null_value": "",
+ "description": "Description of habitat or location where specimen was collected, ignore descriptions of the plant itself"
+ },
+ "Collectors": {
+ "format": "[Collector]",
+ "null_value": "not present",
+ "description": "Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas"
+ },
+ "Collector Number": {
+ "format": "[Collector No.]",
+ "null_value": "s.n.",
+ "description": "Sequential number assigned to collection, associated with the collector"
+ },
+ "Verbatim Date": {
+ "format": "verbatim",
+ "null_value": "s.d.",
+ "description": "Date of collection exactly as it appears on the label"
+ },
+ "Date": {
+ "format": "[yyyy-mm-dd]",
+ "null_value": "",
+ "description": "Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e., 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day"
+ },
+ "End Date": {
+ "format": "[yyyy-mm-dd]",
+ "null_value": "",
+ "description": "If date range is listed, later date of collection range"
+ }
+ }"""
+
+ structure = """{"Dictionary":
+ {
+ "Catalog Number": [Catalog Number],
+ "Genus": [Genus],
+ "Species": [species],
+ "subspecies": [subspecies],
+ "variety": [variety],
+ "forma": [forma],
+ "Country": [Country],
+ "State": [State],
+ "County": [County],
+ "Locality Name": [Locality Name],
+ "Min Elevation": [Min Elevation],
+ "Max Elevation": [Max Elevation],
+ "Elevation Units": [Elevation Units],
+ "Verbatim Coordinates": [Verbatim Coordinates],
+ "Datum": [Datum],
+ "Cultivated": [Cultivated],
+ "Habitat": [Habitat],
+ "Collectors": [Collectors],
+ "Collector Number": [Collector Number],
+ "Verbatim Date": [Verbatim Date],
+ "Date": [Date],
+ "End Date": [End Date]
+ },
+ "SpeciesName": {"taxonomy": [Genus_species]}}"""
+
+ prompt = f"""I'm providing you with a set of rules, an unstructured OCR text, and a reference dictionary (domain knowledge). Your task is to convert the OCR text into a structured dictionary that matches the structure of the reference dictionary. Please follow the rules strictly.
+ The rules are as follows:
+ {set_rules}
+ The unstructured OCR text is:
+ {self.OCR}
+ The reference dictionary, which provides an example of the output structure and has an embedding distance of {self.similarity} to the OCR, is:
+ {self.domain_knowledge_example}
+ Some dictionary fields have special requirements. These requirements specify the format for each field, and are given below:
+ {umich_all_asia_rules}
+ Please refactor the OCR text into a dictionary, following the rules and the reference structure:
+ {structure}
+ """
+
+ xlsx_headers = ["Catalog Number","Genus","Species","subspecies","variety","forma","Country","State","County","Locality Name","Min Elevation","Max Elevation","Elevation Units","Verbatim Coordinates","Datum","Cultivated","Habitat","Collectors","Collector Number","Verbatim Date","Date","End Date"]
+
+
+ return prompt, self.n_fields, xlsx_headers
+
+ def prompt_v1_verbose_noDomainKnowledge(self, OCR=None):
+ self.OCR = OCR or self.OCR
+ self.n_fields = 22 or self.n_fields
+
+ set_rules = """
+ Please note that your task is to generate a dictionary, following the below rules:
+ 1. Refactor the unstructured OCR text into a dictionary based on the reference dictionary structure (ref_dict).
+ 2. Each field of OCR corresponds to a column of the ref_dict. You should correctly map the values from OCR to the respective fields in ref_dict.
+ 3. If the OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None".
+ 4. If there is a field in the ref_dict that does not have a corresponding value in the OCR text, fill it based on your knowledge but don't generate new information.
+ 5. Do not use any text from the ref_dict values in the new dict, but you must use the headers from ref_dict.
+ 6. Duplicate dictionary fields are not allowed.
+ 7. Only return the new dictionary. You should not explain your answer.
+ 8. Your output should be a Python dictionary represented as a JSON string.
+ """
+
+ umich_all_asia_rules = """{
+ "Catalog Number": {
+ "format": "[Catalog Number]",
+ "null_value": "",
+ "description": "The barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits"
+ },
+ "Genus": {
+ "format": "[Genus] or '[Family] indet' if no genus",
+ "null_value": "",
+ "description": "Taxonomic determination to genus, do capitalize genus"
+ },
+ "Species": {
+ "format": "[species] or 'indet' if no species",
+ "null_value": "",
+ "description": "Taxonomic determination to species, do not capitalize species"
+ },
+ "subspecies": {
+ "format": "[subspecies]",
+ "null_value": "",
+ "description": "Taxonomic determination to subspecies (subsp.)"
+ },
+ "variety": {
+ "format": "[variety]",
+ "null_value": "",
+ "description": "Taxonomic determination to variety (var)"
+ },
+ "forma": {
+ "format": "[form]",
+ "null_value": "",
+ "description": "Taxonomic determination to form (f.)"
+ },
+ "Country": {
+ "format": "[Country]",
+ "null_value": "",
+ "description": "Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abbreviation is given"
+ },
+ "State": {
+ "format": "[Adm. Division 1]",
+ "null_value": "",
+ "description": "Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word"
+ },
+ "County": {
+ "format": "[Adm. Division 2]",
+ "null_value": "",
+ "description": "Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word"
+ },
+ "Locality Name": {
+ "format": "verbatim, if no geographic info: 'no data provided on label of catalog no: [######]', or if illegible: 'locality present but illegible/not translated for catalog no: #######', or if no named locality: 'no named locality for catalog no: #######'",
+ "description": "Description of geographic location or landscape"
+ },
+ "Min Elevation": {
+ "format": "elevation integer",
+ "null_value": "",
+ "description": "Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given"
+ },
+ "Max Elevation": {
+ "format": "elevation integer",
+ "null_value": "",
+ "description": "Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise"
+ },
+ "Elevation Units": {
+ "format": "m",
+ "null_value": "",
+ "description": "'m' only if an elevation is present"
+ },
+ "Verbatim Coordinates": {
+ "format": "[Lat, Long | UTM | TRS]",
+ "null_value": "",
+ "description": "Verbatim coordinates as they appear on the label, fix typos to match standardized GPS coordinate format"
+ },
+ "Datum": {
+ "format": "[WGS84, NAD23 etc.]",
+ "null_value": "",
+ "description": "GPS Datum of coordinates on label; empty string "" if GPS coordinates are not in OCR"
+ },
+ "Cultivated": {
+ "format": "yes",
+ "null_value": "",
+ "description": "Indicates if specimen was grown in cultivation"
+ },
+ "Habitat": {
+ "format": "verbatim",
+ "null_value": "",
+ "description": "Description of habitat or location where specimen was collected, ignore descriptions of the plant itself"
+ },
+ "Collectors": {
+ "format": "[Collector]",
+ "null_value": "not present",
+ "description": "Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas"
+ },
+ "Collector Number": {
+ "format": "[Collector No.]",
+ "null_value": "s.n.",
+ "description": "Sequential number assigned to collection, associated with the collector"
+ },
+ "Verbatim Date": {
+ "format": "verbatim",
+ "null_value": "s.d.",
+ "description": "Date of collection exactly as it appears on the label"
+ },
+ "Date": {
+ "format": "[yyyy-mm-dd]",
+ "null_value": "",
+ "description": "Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e., 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day"
+ },
+ "End Date": {
+ "format": "[yyyy-mm-dd]",
+ "null_value": "",
+ "description": "If date range is listed, later date of collection range"
+ }
+ }"""
+
+ structure = """{"Dictionary":
+ {
+ "Catalog Number": [Catalog Number],
+ "Genus": [Genus],
+ "Species": [species],
+ "subspecies": [subspecies],
+ "variety": [variety],
+ "forma": [forma],
+ "Country": [Country],
+ "State": [State],
+ "County": [County],
+ "Locality Name": [Locality Name],
+ "Min Elevation": [Min Elevation],
+ "Max Elevation": [Max Elevation],
+ "Elevation Units": [Elevation Units],
+ "Verbatim Coordinates": [Verbatim Coordinates],
+ "Datum": [Datum],
+ "Cultivated": [Cultivated],
+ "Habitat": [Habitat],
+ "Collectors": [Collectors],
+ "Collector Number": [Collector Number],
+ "Verbatim Date": [Verbatim Date],
+ "Date": [Date],
+ "End Date": [End Date]
+ },
+ "SpeciesName": {"taxonomy": [Genus_species]}}"""
+
+ prompt = f"""I'm providing you with a set of rules, an unstructured OCR text, and a reference dictionary (domain knowledge). Your task is to convert the OCR text into a structured dictionary that matches the structure of the reference dictionary. Please follow the rules strictly.
+ The rules are as follows:
+ {set_rules}
+ The unstructured OCR text is:
+ {self.OCR}
+ Some dictionary fields have special requirements. These requirements specify the format for each field, and are given below:
+ {umich_all_asia_rules}
+ Please refactor the OCR text into a dictionary, following the rules and the reference structure:
+ {structure}
+ """
+
+ xlsx_headers = ["Catalog Number","Genus","Species","subspecies","variety","forma","Country","State","County","Locality Name","Min Elevation","Max Elevation","Elevation Units","Verbatim Coordinates","Datum","Cultivated","Habitat","Collectors","Collector Number","Verbatim Date","Date","End Date"]
+
+ return prompt, self.n_fields, xlsx_headers
+
+ def prompt_v2_json_rules(self, OCR=None):
+ self.OCR = OCR or self.OCR
+ self.n_fields = 26 or self.n_fields
+
+ set_rules = """
+ 1. Refactor the unstructured OCR text into a dictionary based on the JSON structure outlined below.
+ 2. You should map the unstructured OCR text to the appropriate JSON key and then populate the field based on its rules.
+ 3. Some JSON key fields are permitted to remain empty if the corresponding information is not found in the unstructured OCR text.
+ 4. Ignore any information in the OCR text that doesn't fit into the defined JSON structure.
+ 5. Duplicate dictionary fields are not allowed.
+ 6. Ensure that all JSON keys are in lowercase.
+ 7. Ensure that new JSON field values follow sentence case capitalization.
+ 7. Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ 8. Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ 9. Only return a JSON dictionary represented as a string. You should not explain your answer.
+ """
+
+ dictionary_field_format_descriptions = """
+ The next section of instructions outlines how to format the JSON dictionary. The keys are the same as those of the final formatted JSON object.
+ For each key there is a format requirement that specifies how to transcribe the information for that key.
+ The possible formatting options are:
+ 1. "verbatim transcription" - field is populated with verbatim text from the unformatted OCR.
+ 2. "spell check transcription" - field is populated with spelling corrected text from the unformatted OCR.
+ 3. "boolean yes no" - field is populated with only yes or no.
+ 4. "integer" - field is populated with only an integer.
+ 5. "[list]" - field is populated from one of the values in the list.
+ 6. "yyyy-mm-dd" - field is populated with a date in the format year-month-day.
+ The desired null value is also given. Populate the field with the null value of the information for that key is not present in the unformatted OCR text.
+ """
+
+ json_template_rules = """
+ {"Dictionary":{
+ "catalog_number": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "The barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits."
+ },
+ "genus": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to genus. Genus must be capitalized. If genus is not present use the taxonomic family name followed by the word 'indet'."
+ },
+ "species": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to species, do not capitalize species."
+ },
+ "subspecies": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to subspecies (subsp.)."
+ },
+ "variety": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to variety (var)."
+ },
+ "forma": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to form (f.)."
+ },
+ "country": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Country that corresponds to the current geographic location of collection. Capitalize first letter of each word. If abbreviation is given populate field with the full spelling of the country's name."
+ },
+ "state": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Administrative division 1 that corresponds to the current geographic location of collection. Capitalize first letter of each word. Administrative division 1 is equivalent to a U.S. State."
+ },
+ "county": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word. Administrative division 2 is equivalent to a U.S. county, parish, borough."
+ },
+ "locality_name": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Description of geographic location, landscape, landmarks, regional features, nearby places, or any contextual information aiding in pinpointing the exact origin or site of the specimen."
+ },
+ "min_elevation": {
+ "format": "integer",
+ "null_value": "",
+ "description": "Minimum elevation or altitude in meters. Only if units are explicit then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m' or 'm.' or 'meters'). Round to integer."
+ },
+ "max_elevation": {
+ "format": "integer",
+ "null_value": "",
+ "description": "Maximum elevation or altitude in meters. If only one elevation is present, then max_elevation should be set to the null_value. Only if units are explicit then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m' or 'm.' or 'meters'). Round to integer."
+ },
+ "elevation_units": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Elevation units must be meters. If min_elevation field is populated, then elevation_units: 'm'. Otherwise elevation_units: ''."
+ },
+ "verbatim_coordinates": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Verbatim location coordinates as they appear on the label. Do not convert formats. Possible coordinate types are one of [Lat, Long, UTM, TRS]."
+ },
+ "decimal_coordinates": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Correct and convert the verbatim location coordinates to conform with the decimal degrees GPS coordinate format."
+ },
+ "datum": {
+ "format": "[WGS84, WGS72, WGS66, WGS60, NAD83, NAD27, OSGB36, ETRS89, ED50, GDA94, JGD2011, Tokyo97, KGD2002, TWD67, TWD97, BJS54, XAS80, GCJ-02, BD-09, PZ-90.11, GTRF, CGCS2000, ITRF88, ITRF89, ITRF90, ITRF91, ITRF92, ITRF93, ITRF94, ITRF96, ITRF97, ITRF2000, ITRF2005, ITRF2008, ITRF2014, Hong Kong Principal Datum, SAD69]",
+ "null_value": "",
+ "description": "Datum of location coordinates. Possible values are include in the format list. Leave field blank if unclear."
+ },
+ "cultivated": {
+ "format": "boolean yes no",
+ "null_value": "",
+ "description": "Cultivated plants are intentionally grown by humans. In text descriptions, look for planting dates, garden locations, ornamental, cultivar names, garden, or farm to indicate cultivated plant."
+ },
+ "habitat": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Description of a plant's habitat or the location where the specimen was collected. Ignore descriptions of the plant itself."
+ },
+ "plant_description": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Description of plant features such as leaf shape, size, color, stem texture, height, flower structure, scent, fruit or seed characteristics, root system type, overall growth habit and form, any notable aroma or secretions, presence of hairs or bristles, and any other distinguishing morphological or physiological characteristics."
+ },
+ "collectors": {
+ "format": "verbatim transcription",
+ "null_value": "not present",
+ "description": "Full name(s) of the individual(s) responsible for collecting the specimen. When multiple collectors are involved, their names should be separated by commas."
+ },
+ "collector_number": {
+ "format": "verbatim transcription",
+ "null_value": "s.n.",
+ "description": "Unique identifier or number that denotes the specific collecting event and associated with the collector."
+ },
+ "determined_by": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Full name of the individual responsible for determining the taxanomic name of the specimen. Sometimes the name will be near to the characters 'det' to denote determination. This name may be isolated from other names in the unformatted OCR text."
+ },
+ "multiple_names": {
+ "format": "boolean yes no",
+ "null_value": "",
+ "description": "Indicate whether multiple people or collector names are present in the unformatted OCR text. If you see more than one person's name the value is 'yes'; otherwise the value is 'no'."
+ },
+ "verbatim_date": {
+ "format": "verbatim transcription",
+ "null_value": "s.d.",
+ "description": "Date of collection exactly as it appears on the label. Do not change the format or correct typos."
+ },
+ "date": {
+ "format": "yyyy-mm-dd",
+ "null_value": "",
+ "description": "Date the specimen was collected formatted as year-month-day. If specific components of the date are unknown, they should be replaced with zeros. Examples: '0000-00-00' if the entire date is unknown, 'YYYY-00-00' if only the year is known, and 'YYYY-MM-00' if year and month are known but day is not."
+ },
+ "end_date": {
+ "format": "yyyy-mm-dd",
+ "null_value": "",
+ "description": "If a date range is provided, this represents the later or ending date of the collection period, formatted as year-month-day. If specific components of the date are unknown, they should be replaced with zeros. Examples: '0000-00-00' if the entire end date is unknown, 'YYYY-00-00' if only the year of the end date is known, and 'YYYY-MM-00' if year and month of the end date are known but the day is not."
+ },
+ },
+ "SpeciesName": {
+ "taxonomy": [Genus_species]}
+ }"""
+
+ structure = """{"Dictionary":
+ {
+ "catalog_number": "",
+ "genus": "",
+ "species": "",
+ "subspecies": "",
+ "variety": "",
+ "forma": "",
+ "country": "",
+ "state": "",
+ "county": "",
+ "locality_name": "",
+ "min_elevation": "",
+ "max_elevation": "",
+ "elevation_units": "",
+ "verbatim_coordinates": "",
+ "decimal_coordinates": "",
+ "datum": "",
+ "cultivated": "",
+ "habitat": "",
+ "plant_description": "",
+ "collectors": "",
+ "collector_number": "",
+ "determined_by": "",
+ "multiple_names": "",
+ "verbatim_date":"" ,
+ "date": "",
+ "end_date": ""
+ },
+ "SpeciesName": {"taxonomy": ""}}"""
+
+ prompt = f"""Please help me complete this text parsing task given the following rules and unstructured OCR text. Your task is to refactor the OCR text into a structured JSON dictionary that matches the structure specified in the following rules. Please follow the rules strictly.
+ The rules are:
+ {set_rules}
+ The unstructured OCR text is:
+ {self.OCR}
+ {dictionary_field_format_descriptions}
+ This is the JSON template that includes instructions for each key:
+ {json_template_rules}
+ Please populate the following JSON dictionary based on the rules and the unformatted OCR text:
+ {structure}
+ """
+
+ xlsx_headers = ["catalog_number","genus","species","subspecies","variety","forma","country","state","county","locality_name","min_elevation","max_elevation","elevation_units","verbatim_coordinates","decimal_coordinates","datum","cultivated","habitat","plant_description","collectors","collector_number","determined_by","multiple_names","verbatim_date","date","end_date"]
+
+ return prompt, self.n_fields, xlsx_headers
+
+ #############################################################################################
+ #############################################################################################
+ #############################################################################################
+ #############################################################################################
+ # These are for dynamically creating your own prompts with n-columns
+
+
+ def prompt_v2_custom(self, rules_config_path, OCR=None, is_palm=False):
+ self.OCR = OCR
+
+ self.rules_config_path = rules_config_path
+ self.rules_config = self.load_rules_config()
+
+ self.instructions = self.rules_config['instructions']
+ self.json_formatting_instructions = self.rules_config['json_formatting_instructions']
+
+ self.rules_list = self.rules_config['rules']
+ self.n_fields = len(self.rules_list['Dictionary'])
+
+ # Set the rules for processing OCR into JSON format
+ self.rules = self.create_rules(is_palm)
+
+ self.structure = self.create_structure(is_palm)
+
+ if is_palm:
+ prompt = f"""Please help me complete this text parsing task given the following rules and unstructured OCR text. Your task is to refactor the OCR text into a structured JSON dictionary that matches the structure specified in the following rules. Please follow the rules strictly.
+ The rules are:
+ {self.instructions}
+ The unstructured OCR text is:
+ {self.OCR}
+ {self.json_formatting_instructions}
+ This is the JSON template that includes instructions for each key:
+ {self.rules}
+ Please populate the following JSON dictionary based on the rules and the unformatted OCR text:
+ {self.structure}
+ {self.structure}
+ {self.structure}
+ """
+ else:
+ prompt = f"""Please help me complete this text parsing task given the following rules and unstructured OCR text. Your task is to refactor the OCR text into a structured JSON dictionary that matches the structure specified in the following rules. Please follow the rules strictly.
+ The rules are:
+ {self.instructions}
+ The unstructured OCR text is:
+ {self.OCR}
+ {self.json_formatting_instructions}
+ This is the JSON template that includes instructions for each key:
+ {self.rules}
+ Please populate the following JSON dictionary based on the rules and the unformatted OCR text:
+ {self.structure}
+ """
+ xlsx_headers = self.generate_xlsx_headers(is_palm)
+
+ return prompt, self.n_fields, xlsx_headers
+
+ def load_rules_config(self):
+ with open(self.rules_config_path, 'r') as stream:
+ try:
+ return yaml.safe_load(stream)
+ except yaml.YAMLError as exc:
+ print(exc)
+ return None
+
+ def create_rules(self, is_palm=False):
+ if is_palm:
+ # Start with a structure for the "Dictionary" section where each key contains its rules
+ dictionary_structure = {
+ key: {
+ 'description': value['description'],
+ 'format': value['format'],
+ 'null_value': value.get('null_value', '')
+ } for key, value in self.rules_list['Dictionary'].items()
+ }
+
+ # Convert the structure to a JSON string without indentation
+ structure_json_str = json.dumps(dictionary_structure, default_flow_style=False, sort_keys=False)
+ return structure_json_str
+
+ else:
+ # Start with a structure for the "Dictionary" section where each key contains its rules
+ dictionary_structure = {
+ key: {
+ 'description': value['description'],
+ 'format': value['format'],
+ 'null_value': value.get('null_value', '')
+ } for key, value in self.rules_list['Dictionary'].items()
+ }
+
+ # Combine both sections into the overall structure
+ full_structure = {
+ "Dictionary": dictionary_structure,
+ "SpeciesName": self.rules_list['SpeciesName']
+ }
+
+ # Convert the structure to a JSON string without indentation
+ structure_json_str = json.dumps(full_structure, default_flow_style=False, sort_keys=False)
+ return structure_json_str
+
+ def create_structure(self, is_palm=False):
+ if is_palm:
+ # Start with an empty structure for the "Dictionary" section
+ dictionary_structure = {key: "" for key in self.rules_list['Dictionary'].keys()}
+
+ # Convert the structure to a JSON string with indentation for readability
+ structure_json_str = json.dumps(dictionary_structure, default_flow_style=False, sort_keys=False)
+ return structure_json_str
+ else:
+ # Start with an empty structure for the "Dictionary" section
+ dictionary_structure = {key: "" for key in self.rules_list['Dictionary'].keys()}
+
+ # Manually define the "SpeciesName" section
+ species_name_structure = {"taxonomy": ""}
+
+ # Combine both sections into the overall structure
+ full_structure = {
+ "Dictionary": dictionary_structure,
+ "SpeciesName": species_name_structure
+ }
+
+ # Convert the structure to a JSON string with indentation for readability
+ structure_json_str = json.dumps(full_structure, default_flow_style=False, sort_keys=False)
+ return structure_json_str
+
+ def generate_xlsx_headers(self, is_palm):
+ # Extract headers from the 'Dictionary' keys in the JSON template rules
+ if is_palm:
+ xlsx_headers = list(self.rules_list.keys())
+ return xlsx_headers
+ else:
+ xlsx_headers = list(self.rules_list["Dictionary"].keys())
+ return xlsx_headers
+
+ def prompt_v2_custom_redo(self, incorrect_json, is_palm):
+ # Load the existing rules and structure
+ self.rules_config = self.load_rules_config()
+ # self.rules = self.create_rules(is_palm)
+ self.structure = self.create_structure(is_palm)
+
+ # Generate the prompt using the loaded rules and structure
+ if is_palm:
+ prompt = f"""This text is supposed to be JSON, but it contains an error that prevents it from loading with the Python command json.loads().
+ You need to return coorect JSON for the following dictionary. Most likely, a quotation mark inside of a field value has not been escaped properly with a backslash.
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+ Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ The incorrectly formatted JSON dictionary: {incorrect_json}
+ The output JSON structure: {self.structure}
+ The output JSON structure: {self.structure}
+ The output JSON structure: {self.structure}
+ The refactored JSON disctionary: """
+ else:
+ prompt = f"""This text is supposed to be JSON, but it contains an error that prevents it from loading with the Python command json.loads().
+ You need to return coorect JSON for the following dictionary. Most likely, a quotation mark inside of a field value has not been escaped properly with a backslash.
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+ Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ The incorrectly formatted JSON dictionary: {incorrect_json}
+ The output JSON structure: {self.structure}
+ The refactored JSON disctionary: """
+ return prompt
+
+ #############################################################################################
+ #############################################################################################
+ #############################################################################################
+ #############################################################################################
+ def prompt_gpt_redo_v1(self, incorrect_json):
+ structure = """Below is the correct JSON formatting. Modify the text to conform to the following format, fixing the incorrect JSON:
+ {"Dictionary":
+ {
+ "Catalog Number": [Catalog Number],
+ "Genus": [Genus],
+ "Species": [species],
+ "subspecies": [subspecies],
+ "variety": [variety],
+ "forma": [forma],
+ "Country": [Country],
+ "State": [State],
+ "County": [County],
+ "Locality Name": [Locality Name],
+ "Min Elevation": [Min Elevation],
+ "Max Elevation": [Max Elevation],
+ "Elevation Units": [Elevation Units],
+ "Verbatim Coordinates": [Verbatim Coordinates],
+ "Datum": [Datum],
+ "Cultivated": [Cultivated],
+ "Habitat": [Habitat],
+ "Collectors": [Collectors],
+ "Collector Number": [Collector Number],
+ "Verbatim Date": [Verbatim Date],
+ "Date": [Date],
+ "End Date": [End Date]
+ },
+ "SpeciesName": {"taxonomy": [Genus_species]}}"""
+
+ prompt = f"""This text is supposed to be JSON, but it contains an error that prevents it from loading with the Python command json.loads().
+ You need to return coorect JSON for the following dictionary. Most likely, a quotation mark inside of a field value has not been escaped properly with a backslash.
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+ Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ The incorrectly formatted JSON dictionary: {incorrect_json}
+ The output JSON structure: {structure}
+ The refactored JSON disctionary: """
+ return prompt
+
+ def prompt_gpt_redo_v2(self, incorrect_json):
+ structure = """
+ {"Dictionary":{
+ "catalog_number": "",
+ "genus": "",
+ "species": "".
+ "subspecies": "",
+ "variety": "",
+ "forma":"",
+ "country": "",
+ "state": "",
+ "county": "",
+ "locality_name": "",
+ "min_elevation": "",
+ "max_elevation": "",
+ "elevation_units": "',
+ "verbatim_coordinates": "",
+ "decimal_coordinates": "",
+ "datum": "",
+ "cultivated": "",
+ "habitat": "",
+ "plant_description": "",
+ "collectors": "",
+ "collector_number": "",
+ "determined_by": "",
+ "multiple_names": "',
+ "verbatim_date": "",
+ "date": "",
+ "end_date": "",
+ },
+ "SpeciesName": {"taxonomy": [Genus_species]}}"""
+
+ prompt = f"""This text is supposed to be JSON, but it contains an error that prevents it from loading with the Python command json.loads().
+ You need to return coorect JSON for the following dictionary. Most likely, a quotation mark inside of a field value has not been escaped properly with a backslash.
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+ Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ The incorrectly formatted JSON dictionary: {incorrect_json}
+ The output JSON structure: {structure}
+ The refactored JSON disctionary: """
+ return prompt
+ #####################################################################################################################################
+ #####################################################################################################################################
+ def prompt_v1_palm2(self, in_list, out_list, OCR=None):
+ self.OCR = OCR or self.OCR
+ set_rules = """1. Your job is to return a new dict based on the structure of the reference dict ref_dict and these are your rules.
+ 2. You must look at ref_dict and refactor the new text called OCR to match the same formatting.
+ 3. OCR contains unstructured text inside of [], use your knowledge to put the OCR text into the correct ref_dict column.
+ 4. If OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None" and skip all other steps.
+ 5. If there is a field that does not have a direct proxy in the OCR text, you can fill it in based on your knowledge, but you cannot generate new information.
+ 6. Never put text from the ref_dict values into the new dict, but you must use the headers from ref_dict.
+ 7. There cannot be duplicate dictionary fields.
+ 8. Only return the new dict, do not explain your answer.
+ 9. Do not include quotation marks in content, only use quotation marks to represent values in dictionaries.
+ 10. For GPS coordinates only use Decimal Degrees (D.D°)
+ 11. "Given the input text, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values."""
+
+ umich_all_asia_rules = """
+ "Catalog Number" - {"format": "[barcode]", "null_value": "", "description": the barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits}
+ "Genus" - {"format": "[Genus]" or "[Family] indet" if no genus", "null_value": "", "description": taxonomic determination to genus, do captalize genus}
+ "Species"- {"format": "[species]" or "indet" if no species, "null_value": "", "description": taxonomic determination to species, do not captalize species}
+ "subspecies" - {"format": "[subspecies]", "null_value": "", "description": taxonomic determination to subspecies (subsp.)}
+ "variety" - {"format": "[variety]", "null_value": "", "description": taxonomic determination to variety (var)}
+ "forma" - {"format": "[form]", "null_value": "", "description": taxonomic determination to form (f.)}
+
+ "Country" - {"format": "[Country]", "null_value": "no data", "description": Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abreviation is given}
+ "State" - {"format": "[Adm. Division 1]", "null_value": "no data", "description": Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "County" - {"format": "[Adm. Division 2]", "null_value": "no data", "description": Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "Locality Name" - {"format": "verbatim", if no geographic info: "no data provided on label of catalog no: [######]", or if illegible: "locality present but illegible/not translated for catalog no: #######", or if no named locality: "no named locality for catalog no: #######", "description": "Description of geographic location or landscape"}
+
+ "Min Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given}
+ "Max Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise}
+ "Elevation Units" - {format: "m", "null_value": "","description": "m" only if an elevation is present}
+
+ "Verbatim Coordinates" - {"format": "[Lat, Long | UTM | TRS]", "null_value": "", "description": Convert coordinates to Decimal Degrees (D.D°) format, do not use Minutes, Seconds or quotation marks}
+
+ "Datum" - {"format": "[WGS84, NAD23 etc.]", "null_value": "not present", "description": Datum of coordinates on label; "" is GPS coordinates are not in OCR}
+ "Cultivated" - {"format": "yes", "null_value": "", "description": Indicates if specimen was grown in cultivation}
+ "Habitat" - {"format": "verbatim", "null_value": "", "description": Description of habitat or location where specimen was collected, ignore descriptions of the plant itself}
+ "Collectors" - {"format": "[Collector]", "null_value": "not present", "description": Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas}
+ "Collector Number" - {"format": "[Collector No.]", "null_value": "s.n.", "description": Sequential number assigned to collection, associated with the collector}
+ "Verbatim Date" - {"format": "verbatim", "null_value": "s.d.", "description": Date of collection exactly as it appears on the label}
+ "Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e. 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day}
+ "End Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": If date range is listed, later date of collection range}
+ """
+
+ prompt = f"""Given the following set of rules:
+
+ set_rules = {set_rules}
+
+ Some dict fields have special requirements listed below. First is the column header. After the - is the format. Do not include the instructions with your response:
+
+ requirements = {umich_all_asia_rules}
+
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+
+ input: {in_list[0]}
+
+ output: {out_list[0]}
+
+ input: {in_list[1]}
+
+ output: {out_list[1]}
+
+ input: {in_list[2]}
+
+ output: {out_list[2]}
+
+ input: {self.OCR}
+
+ output:"""
+
+ return prompt
+
+ def prompt_v1_palm2_noDomainKnowledge(self, OCR=None):
+ self.OCR = OCR or self.OCR
+ set_rules = """1. Your job is to return a new dict based on the structure of the reference dict ref_dict and these are your rules.
+ 2. You must look at ref_dict and refactor the new text called OCR to match the same formatting.
+ 3. OCR contains unstructured text inside of [], use your knowledge to put the OCR text into the correct ref_dict column.
+ 4. If OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None" and skip all other steps.
+ 5. If there is a field that does not have a direct proxy in the OCR text, you can fill it in based on your knowledge, but you cannot generate new information.
+ 6. Never put text from the ref_dict values into the new dict, but you must use the headers from ref_dict.
+ 7. There cannot be duplicate dictionary fields.
+ 8. Only return the new dict, do not explain your answer.
+ 9. Do not include quotation marks in content, only use quotation marks to represent values in dictionaries.
+ 10. For GPS coordinates only use Decimal Degrees (D.D°)
+ 11. "Given the input text, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values."""
+
+ umich_all_asia_rules = """
+ "Catalog Number" - {"format": "barcode", "null_value": "", "description": the barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits}
+ "Genus" - {"format": "Genus" or "Family indet" if no genus", "null_value": "", "description": taxonomic determination to genus, do captalize genus}
+ "Species"- {"format": "species" or "indet" if no species, "null_value": "", "description": taxonomic determination to species, do not captalize species}
+ "subspecies" - {"format": "subspecies", "null_value": "", "description": taxonomic determination to subspecies (subsp.)}
+ "variety" - {"format": "variety", "null_value": "", "description": taxonomic determination to variety (var)}
+ "forma" - {"format": "form", "null_value": "", "description": taxonomic determination to form (f.)}
+
+ "Country" - {"format": "Country", "null_value": "no data", "description": Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abreviation is given}
+ "State" - {"format": "Adm. Division 1", "null_value": "no data", "description": Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "County" - {"format": "Adm. Division 2", "null_value": "no data", "description": Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "Locality Name" - {"format": "verbatim", if no geographic info: "no data provided on label of catalog no: ######", or if illegible: "locality present but illegible/not translated for catalog no: #######", or if no named locality: "no named locality for catalog no: #######", "description": "Description of geographic location or landscape"}
+
+ "Min Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given}
+ "Max Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise}
+ "Elevation Units" - {format: "m", "null_value": "","description": "m" only if an elevation is present}
+
+ "Verbatim Coordinates" - {"format": "Lat, Long, UTM, TRS", "null_value": "", "description": Convert coordinates to Decimal Degrees (D.D°) format, do not use Minutes, Seconds or quotation marks}
+
+ "Datum" - {"format": "WGS84, NAD23 etc.", "null_value": "not present", "description": Datum of coordinates on label; "" is GPS coordinates are not in OCR}
+ "Cultivated" - {"format": "yes", "null_value": "", "description": Indicates if specimen was grown in cultivation}
+ "Habitat" - {"format": "verbatim", "null_value": "", "description": Description of habitat or location where specimen was collected, ignore descriptions of the plant itself}
+ "Collectors" - {"format": "Collector", "null_value": "not present", "description": Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas}
+ "Collector Number" - {"format": "Collector No.", "null_value": "s.n.", "description": Sequential number assigned to collection, associated with the collector}
+ "Verbatim Date" - {"format": "verbatim", "null_value": "s.d.", "description": Date of collection exactly as it appears on the label}
+ "Date" - {"format": "yyyy-mm-dd", "null_value": "", "description": Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e. 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day}
+ "End Date" - {"format": "yyyy-mm-dd", "null_value": "", "description": If date range is listed, later date of collection range}
+ """
+ structure = """{
+ "Catalog Number": "",
+ "Genus": "",
+ "Species": "",
+ "subspecies": "",
+ "variety": "",
+ "forma": "",
+ "Country": "",
+ "State": "",
+ "County": "",
+ "Locality Name": "",
+ "Min Elevation": "",
+ "Max Elevation": "",
+ "Elevation Units": "",
+ "Verbatim Coordinates": "",
+ "Datum": "",
+ "Cultivated": "",
+ "Habitat": "",
+ "Collectors": "",
+ "Collector Number": "",
+ "Verbatim Date": "",
+ "Date": "",
+ "End Date": "",
+ }"""
+ # structure = """{
+ # "Catalog Number": [Catalog Number],
+ # "Genus": [Genus],
+ # "Species": [species],
+ # "subspecies": [subspecies],
+ # "variety": [variety],
+ # "forma": [forma],
+ # "Country": [Country],
+ # "State": [State],
+ # "County": [County],
+ # "Locality Name": [Locality Name],
+ # "Min Elevation": [Min Elevation],
+ # "Max Elevation": [Max Elevation],
+ # "Elevation Units": [Elevation Units],
+ # "Verbatim Coordinates": [Verbatim Coordinates],
+ # "Datum": [Datum],
+ # "Cultivated": [Cultivated],
+ # "Habitat": [Habitat],
+ # "Collectors": [Collectors],
+ # "Collector Number": [Collector Number],
+ # "Verbatim Date": [Verbatim Date],
+ # "Date": [Date],
+ # "End Date": [End Date]
+ # }"""
+
+ prompt = f"""Given the following set of rules:
+ set_rules = {set_rules}
+ Some dict fields have special requirements listed below. First is the column header. After the - is the format. Do not include the instructions with your response:
+ requirements = {umich_all_asia_rules}
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+ The input unformatted OCR text: {self.OCR}
+ The output JSON structure: {structure}
+ The output JSON structure: {structure}
+ The output JSON structure: {structure}
+ The refactored JSON disctionary:"""
+
+ return prompt
+
+ def prompt_v2_palm2(self, OCR=None):
+ self.OCR = OCR or self.OCR
+ self.n_fields = 26 or self.n_fields
+
+ set_rules = """
+ 1. Refactor the unstructured OCR text into a dictionary based on the JSON structure outlined below.
+ 2. You should map the unstructured OCR text to the appropriate JSON key and then populate the field based on its rules.
+ 3. Some JSON key fields are permitted to remain empty if the corresponding information is not found in the unstructured OCR text.
+ 4. Ignore any information in the OCR text that doesn't fit into the defined JSON structure.
+ 5. Duplicate dictionary fields are not allowed.
+ 6. Ensure that all JSON keys are in lowercase.
+ 7. Ensure that new JSON field values follow sentence case capitalization.
+ 8. Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ 9. Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ 10. Only return a JSON dictionary represented as a string. You should not explain your answer.
+ """
+
+ dictionary_field_format_descriptions = """
+ The next section of instructions outlines how to format the JSON dictionary. The keys are the same as those of the final formatted JSON object.
+ For each key there is a format requirement that specifies how to transcribe the information for that key.
+ The possible formatting options are:
+ 1. "verbatim transcription" - field is populated with verbatim text from the unformatted OCR.
+ 2. "spell check transcription" - field is populated with spelling corrected text from the unformatted OCR.
+ 3. "boolean yes no" - field is populated with only yes or no.
+ 4. "integer" - field is populated with only an integer.
+ 5. "[list]" - field is populated from one of the values in the list.
+ 6. "yyyy-mm-dd" - field is populated with a date in the format year-month-day.
+ The desired null value is also given. Populate the field with the null value of the information for that key is not present in the unformatted OCR text.
+ """
+
+ json_template_rules = """
+ {
+ "catalog_number": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "The barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits."
+ },
+ "genus": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to genus. Genus must be capitalized. If genus is not present use the taxonomic family name followed by the word 'indet'."
+ },
+ "species": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to species, do not capitalize species."
+ },
+ "subspecies": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to subspecies (subsp.)."
+ },
+ "variety": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to variety (var)."
+ },
+ "forma": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Taxonomic determination to form (f.)."
+ },
+ "country": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Country that corresponds to the current geographic location of collection. Capitalize first letter of each word. If abbreviation is given populate field with the full spelling of the country's name. Use sentence-case to capitalize proper nouns."
+ },
+ "state": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Administrative division 1 that corresponds to the current geographic location of collection. Capitalize first letter of each word. Administrative division 1 is equivalent to a U.S. State. Use sentence-case to capitalize proper nouns."
+ },
+ "county": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word. Administrative division 2 is equivalent to a U.S. county, parish, borough. Use sentence-case to capitalize proper nouns."
+ },
+ "locality_name": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Description of geographic location, landscape, landmarks, regional features, nearby places, or any contextual information aiding in pinpointing the exact origin or site of the specimen. Use sentence-case to capitalize proper nouns."
+ },
+ "min_elevation": {
+ "format": "integer",
+ "null_value": "",
+ "description": "Minimum elevation or altitude in meters. Only if units are explicit then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m' or 'm.' or 'meters'). Round to integer."
+ },
+ "max_elevation": {
+ "format": "integer",
+ "null_value": "",
+ "description": "Maximum elevation or altitude in meters. If only one elevation is present, then max_elevation should be set to the null_value. Only if units are explicit then convert from feet ('ft' or 'ft.' or 'feet') to meters ('m' or 'm.' or 'meters'). Round to integer."
+ },
+ "elevation_units": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Elevation units must be meters. If min_elevation field is populated, then elevation_units: 'm'. Otherwise elevation_units: ''"
+ },
+ "verbatim_coordinates": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Verbatim location coordinates as they appear on the label. Do not convert formats. Possible coordinate types are one of [Lat, Long, UTM, TRS]."
+ },
+ "decimal_coordinates": {
+ "format": "spell check transcription",
+ "null_value": "",
+ "description": "Correct and convert the verbatim location coordinates to conform with the decimal degrees GPS coordinate format."
+ },
+ "datum": {
+ "format": "[WGS84, WGS72, WGS66, WGS60, NAD83, NAD27, OSGB36, ETRS89, ED50, GDA94, JGD2011, Tokyo97, KGD2002, TWD67, TWD97, BJS54, XAS80, GCJ-02, BD-09, PZ-90.11, GTRF, CGCS2000, ITRF88, ITRF89, ITRF90, ITRF91, ITRF92, ITRF93, ITRF94, ITRF96, ITRF97, ITRF2000, ITRF2005, ITRF2008, ITRF2014, Hong Kong Principal Datum, SAD69]",
+ "null_value": "",
+ "description": "Datum of location coordinates. Possible values are include in the format list. Leave field blank if unclear."
+ },
+ "cultivated": {
+ "format": "boolean yes no",
+ "null_value": "",
+ "description": "Cultivated plants are intentionally grown by humans. In text descriptions, look for planting dates, garden locations, ornamental, cultivar names, garden, or farm to indicate cultivated plant."
+ },
+ "habitat": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Description of a plant's habitat or the location where the specimen was collected. Ignore descriptions of the plant itself. Use sentence-case to capitalize proper nouns."
+ },
+ "plant_description": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Description of plant features such as leaf shape, size, color, stem texture, height, flower structure, scent, fruit or seed characteristics, root system type, overall growth habit and form, any notable aroma or secretions, presence of hairs or bristles, and any other distinguishing morphological or physiological characteristics. Use sentence-case to capitalize proper nouns."
+ },
+ "collectors": {
+ "format": "verbatim transcription",
+ "null_value": "not present",
+ "description": "Full name(s) of the individual(s) responsible for collecting the specimen. Use sentence-case to capitalize proper nouns. When multiple collectors are involved, their names should be separated by commas."
+ },
+ "collector_number": {
+ "format": "verbatim transcription",
+ "null_value": "s.n.",
+ "description": "Unique identifier or number that denotes the specific collecting event and associated with the collector."
+ },
+ "determined_by": {
+ "format": "verbatim transcription",
+ "null_value": "",
+ "description": "Full name of the individual responsible for determining the taxanomic name of the specimen. Use sentence-case to capitalize proper nouns. Sometimes the name will be near to the characters 'det' to denote determination. This name may be isolated from other names in the unformatted OCR text."
+ },
+ "multiple_names": {
+ "format": "boolean yes no",
+ "null_value": "",
+ "description": "Indicate whether multiple people or collector names are present in the unformatted OCR text. Use sentence-case to capitalize proper nouns. If you see more than one person's name the value is 'yes'; otherwise the value is 'no'."
+ },
+ "verbatim_date": {
+ "format": "verbatim transcription",
+ "null_value": "s.d.",
+ "description": "Date of collection exactly as it appears on the label. Do not change the format or correct typos."
+ },
+ "date": {
+ "format": "yyyy-mm-dd",
+ "null_value": "",
+ "description": "Date the specimen was collected formatted as year-month-day. If specific components of the date are unknown, they should be replaced with zeros. Examples: '0000-00-00' if the entire date is unknown, 'YYYY-00-00' if only the year is known, and 'YYYY-MM-00' if year and month are known but day is not."
+ },
+ "end_date": {
+ "format": "yyyy-mm-dd",
+ "null_value": "",
+ "description": "If a date range is provided, this represents the later or ending date of the collection period, formatted as year-month-day. If specific components of the date are unknown, they should be replaced with zeros. Examples: '0000-00-00' if the entire end date is unknown, 'YYYY-00-00' if only the year of the end date is known, and 'YYYY-MM-00' if year and month of the end date are known but the day is not."
+ },
+ }"""
+
+ structure = """{"catalog_number": "",
+ "genus": "",
+ "species": "".
+ "subspecies": "",
+ "variety": "",
+ "forma":"",
+ "country": "",
+ "state": "",
+ "county": "",
+ "locality_name": "",
+ "min_elevation": "",
+ "max_elevation": "",
+ "elevation_units": "',
+ "verbatim_coordinates": "",
+ "decimal_coordinates": "",
+ "datum": "",
+ "cultivated": "",
+ "habitat": "",
+ "plant_description": "",
+ "collectors": "",
+ "collector_number": "",
+ "determined_by": "",
+ "multiple_names": "',
+ "verbatim_date": "",
+ "date": "",
+ "end_date": "",
+ }"""
+ # structure = """{"catalog_number": [Catalog Number],
+ # "genus": [Genus],
+ # "species": [species],
+ # "subspecies": [subspecies],
+ # "variety": [variety],
+ # "forma": [forma],
+ # "country": [Country],
+ # "state": [State],
+ # "county": [County],
+ # "locality_name": [Locality Name],
+ # "min_elevation": [Min Elevation],
+ # "max_elevation": [Max Elevation],
+ # "elevation_units": [Elevation Units],
+ # "verbatim_coordinates": [Verbatim Coordinates],
+ # "decimal_coordinates": [Decimal Coordinates],
+ # "datum": [Datum],
+ # "cultivated": [boolean yes no],
+ # "habitat": [Habitat Description],
+ # "plant_description": [Plant Description],
+ # "collectors": [Name(s) of Collectors],
+ # "collector_number": [Collector Number],
+ # "determined_by": [Name(s) of Taxonomist],
+ # "multiple_names": [boolean yes no],
+ # "verbatim_date": [Verbatim Date],
+ # "date": [yyyy-mm-dd],
+ # "end_date": [yyyy-mm-dd],
+ # }"""
+
+ prompt = f"""Please help me complete this text parsing task given the following rules and unstructured OCR text. Your task is to refactor the OCR text into a structured JSON dictionary that matches the structure specified in the following rules. Please follow the rules strictly.
+ The rules are:
+ {set_rules}
+ The unstructured OCR text is:
+ {self.OCR}
+ {dictionary_field_format_descriptions}
+ This is the JSON template that includes instructions for each key:
+ {json_template_rules}
+ Please populate the following JSON dictionary based on the rules and the unformatted OCR text. The square brackets denote the locations that you should place the new structured text:
+ {structure}
+ {structure}
+ {structure}
+ """
+
+ return prompt
+
+ def prompt_palm_redo_v1(self, incorrect_json):
+ structure = """{
+ "Catalog Number": [Catalog Number],
+ "Genus": [Genus],
+ "Species": [species],
+ "subspecies": [subspecies],
+ "variety": [variety],
+ "forma": [forma],
+ "Country": [Country],
+ "State": [State],
+ "County": [County],
+ "Locality Name": [Locality Name],
+ "Min Elevation": [Min Elevation],
+ "Max Elevation": [Max Elevation],
+ "Elevation Units": [Elevation Units],
+ "Verbatim Coordinates": [Verbatim Coordinates],
+ "Datum": [Datum],
+ "Cultivated": [Cultivated],
+ "Habitat": [Habitat],
+ "Collectors": [Collectors],
+ "Collector Number": [Collector Number],
+ "Verbatim Date": [Verbatim Date],
+ "Date": [Date],
+ "End Date": [End Date]
+ }"""
+
+ prompt = f"""This text is supposed to be JSON, but it contains an error that prevents it from loading with the Python command json.loads().
+ You need to return coorect JSON for the following dictionary. Most likely, a quotation mark inside of a field value has not been escaped properly with a backslash.
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+ Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ The incorrectly formatted JSON dictionary: {incorrect_json}
+ The output JSON structure: {structure}
+ The output JSON structure: {structure}
+ The output JSON structure: {structure}
+ The refactored JSON disctionary: """
+ return prompt
+
+ def prompt_palm_redo_v2(self, incorrect_json):
+ structure = """{"catalog_number": "",
+ "genus": "",
+ "species": "".
+ "subspecies": "",
+ "variety": "",
+ "forma":"",
+ "country": "",
+ "state": "",
+ "county": "",
+ "locality_name": "",
+ "min_elevation": "",
+ "max_elevation": "",
+ "elevation_units": "',
+ "verbatim_coordinates": "",
+ "decimal_coordinates": "",
+ "datum": "",
+ "cultivated": "",
+ "habitat": "",
+ "plant_description": "",
+ "collectors": "",
+ "collector_number": "",
+ "determined_by": "",
+ "multiple_names": "',
+ "verbatim_date": "",
+ "date": "",
+ "end_date": "",
+ }"""
+
+ prompt = f"""This text is supposed to be JSON, but it contains an error that prevents it from loading with the Python command json.loads().
+ You need to return coorect JSON for the following dictionary. Most likely, a quotation mark inside of a field value has not been escaped properly with a backslash.
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+ Ensure all key-value pairs in the JSON dictionary strictly adhere to the format and data types specified in the template.
+ Ensure the output JSON string is valid JSON format. It should not have trailing commas or unquoted keys.
+ The incorrectly formatted JSON dictionary: {incorrect_json}
+ The output JSON structure: {structure}
+ The output JSON structure: {structure}
+ The output JSON structure: {structure}
+ The refactored JSON disctionary: """
+ return prompt
\ No newline at end of file
diff --git a/vouchervision/prompts.py b/vouchervision/prompts.py
new file mode 100644
index 0000000000000000000000000000000000000000..02b319718424fbcba3d18bd595bbc87d0ee4f582
--- /dev/null
+++ b/vouchervision/prompts.py
@@ -0,0 +1,745 @@
+'''
+###################################################################
+########################## chatGPT ##############################
+###################################################################
+
+Prompts for chatGPT
+ PROMPT_UMICH_skeleton_all_asia
+ - Designed for the "All Asia" TCN at the University of Michigan Herbarium
+ - Has 21 columns for a skeleton record
+
+ PROMPT_OCR_Organized
+ - Designed to privide human transcribers text that is grouped by category
+ so that QC of automated transcription is faster. This output is sent to
+ a custom QC GUI where human labelers can simply copy and paste raw text
+ (but organized by category) into fields that may have been transcribed
+ incorrectly by the LLM.
+
+###################################################################
+########################## chatGPT ##############################
+###################################################################
+'''
+
+def PROMPT_UMICH_skeleton_all_asia(OCR, domain_knowledge_example, similarity):
+ set_rules = """1. Your job is to return a new dict based on the structure of the reference dict ref_dict and these are your rules.
+ 2. You must look at ref_dict and refactor the new text called OCR to match the same formatting.
+ 3. OCR contains unstructured text inside of [], use your knowledge to put the OCR text into the correct ref_dict column.
+ 4. If OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None" and skip all other steps.
+ 5. If there is a field that does not have a direct proxy in the OCR text, you can fill it in based on your knowledge, but you cannot generate new information.
+ 6. Never put text from the ref_dict values into the new dict, but you must use the headers from ref_dict.
+ 7. There cannot be duplicate dictionary fields.
+ 8. Only return the new dict, do not explain your answer."""
+
+ umich_all_asia_rules = """
+ "Catalog Number" - {"format": "[Catalog Number]", "null_value": "", "description": The barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits}
+ "Genus" - {"format": "[Genus]" or "[Family] indet" if no genus", "null_value": "", "description": taxonomic determination to genus, do captalize genus}
+ "Species"- {"format": "[species]" or "indet" if no species, "null_value": "", "description": taxonomic determination to species, do not captalize species}
+ "subspecies" - {"format": "[subspecies]", "null_value": "", "description": taxonomic determination to subspecies (subsp.)}
+ "variety" - {"format": "[variety]", "null_value": "", "description": taxonomic determination to variety (var)}
+ "forma" - {"format": "[form]", "null_value": "", "description": taxonomic determination to form (f.)}
+
+ "Country" - {"format": "[Country]", "null_value": "no data", "description": Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abreviation is given}
+ "State" - {"format": "[Adm. Division 1]", "null_value": "no data", "description": Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "County" - {"format": "[Adm. Division 2]", "null_value": "no data", "description": Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "Locality Name" - {"format": "verbatim", if no geographic info: "no data provided on label of catalog no: [######]", or if illegible: "locality present but illegible/not translated for catalog no: #######", or if no named locality: "no named locality for catalog no: #######", "description": "Description of geographic location or landscape"}
+
+ "Min Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given}
+ "Max Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise}
+ "Elevation Units" - {format: "m", "null_value": "","description": "m" only if an elevation is present}
+
+ "Verbatim Coordinates" - {"format": "[Lat, Long | UTM | TRS]", "null_value": "", "description": Verbatim coordinates as they appear on the label, fix typos to match standardized GPS coordinate format}
+
+ "Datum" - {"format": "[WGS84, NAD23 etc.]", "null_value": "not present", "description": Datum of coordinates on label; "" is GPS coordinates are not in OCR}
+ "Cultivated" - {"format": "yes", "null_value": "", "description": Indicates if specimen was grown in cultivation}
+ "Habitat" - {"format": "verbatim", "null_value": "", "description": Description of habitat or location where specimen was collected, ignore descriptions of the plant itself}
+ "Collectors" - {"format": "[Collector]", "null_value": "not present", "description": Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas}
+ "Collector Number" - {"format": "[Collector No.]", "null_value": "s.n.", "description": Sequential number assigned to collection, associated with the collector}
+ "Verbatim Date" - {"format": "verbatim", "null_value": "s.d.", "description": Date of collection exactly as it appears on the label}
+ "Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e. 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day}
+ "End Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": If date range is listed, later date of collection range}
+ """
+
+ structure = """{"Dictionary":
+ {
+ "Catalog Number": [Catalog Number],
+ "Genus": [Genus],
+ "Species": [Species],
+ "subspecies": [subspecies],
+ "variety": [variety],
+ "forma": [forma],
+ "Country": [Country],
+ "State": [State],
+ "County": [County],
+ "Locality Name": [Locality Name],
+ "Min Elevation": [Min Elevation],
+ "Max Elevation": [Max Elevation],
+ "Elevation Units": [Elevation Units],
+ "Verbatim Coordinates": [Verbatim Coordinates],
+ "Datum": [Datum],
+ "Cultivated": [Cultivated],
+ "Habitat": [Habitat],
+ "Collectors": [Collectors],
+ "Collector Number": [Collector Number],
+ "Verbatim Date": [Verbatim Date],
+ "Date": [Date],
+ "End Date": [End Date]
+ },
+ "SpeciesName": {"taxonomy": "genus_species"}}"""
+
+ prompt = f"""Given the following set of rules:
+
+ set_rules = {set_rules}
+
+ The following is the raw OCR text that you must translate into a properly formatted Python dictionary based on the rules:
+
+ OCR = {OCR}
+
+ The following is an example dictionary that has an embedding distance of {similarity} compared to OCR. Use if as a guide, but never copy text directly from the domain_knowledge:
+
+ domain_knowledge = {domain_knowledge_example}
+
+ Some dict fields have special requirements listed below. First is the column header. After the - is the format. Do not include the instructions with your response:
+
+ requirements = {umich_all_asia_rules}
+
+ Please transform the OCR text into a Python dictionary following the rules to complete this dictionary, replace [] with content:
+ formatted_ocr = {structure}"""
+ # print(f'{OCR}\n\n')
+ # print(f'{domain_knowledge_example}\n\n')
+ return prompt
+
+
+def PROMPT_OCR_Organized(OCR):
+ set_rules = """1. Your job is to parse messy text and return a new dict based on these rules.
+ 2. The messy text is similar to the information contained in Darwin Core Archive files for herbarium specimens.
+ 3. You need to bin the text into 4 different information categories including TAXONOMY, GEOGRAPHY, LOCALITY, COLLECTING and MISCELLANEOUS based on symantics.
+ 4. Within each information category list separate discrete content with the comma seperator ",".
+ 5. Denote discrete content inside of each subcategory with quotation marks, like this "discrete content".
+ 6. If you can provide more detailed information for the GEOGRAPHY category, such as a more thorough location hierarchy, please include additional information along with the verbatim transcriptions.
+ 7. Transcribe verbatim unless there is a typo. You can correct typos and misspellings and you can adjust capitalization of the letters in content words to fit standard conventions given the context.
+ 8. If some content listed in the descriptions below are not present in OCR, the just skip those subfields.
+ 9. Only return the new dict, do not explain your answer."""
+
+ structure = """{"Dictionary":{
+ "TAXONOMY": ["taxonomic topics", "more taxonomic information",],
+ "GEOGRAPHY": ["geographic topics","more geographic information",],
+ "LOCALITY": ["location topics", "more location information",],
+ "COLLECTING": ["documentation and collection topics", "more documentation and collection information",],
+ "MISCELLANEOUS": ["miscellaneous topics", "more remaining miscellaneous info",]
+ },
+ "Summary": ["one sentence description of content"]}"""
+ category_rules = """
+ "TAXONOMY" - Information to include: all content relating to the name of the plant species including Order, Family, Genus, Species, Subspecies, Variety, and Forma.
+ "GEOGRAPHY" - Information to include: The government defined names of places that would appear on a map of political boundaries including Countries, States, Prefectures, Provinces, Districts, Counties, Cities, or Adminstrative Divisions. Adjust capitalization to follow standard conventions for each.
+ "LOCALITY" - Information to include: descriptions of the landscape, habitat, surroundings or nearby places including towns, roads, buildings, geologic features, and distances.
+ "COLLECTING" - Information to include: the names of the people who collected the specimen; the collector's number; the verbatim date; all dates translated int the format [yyyy-mm-dd] with zeros replacing unknown numbers; anything relating to cultivation status or whether it was grown in a garden or captive setting; all descriptions of the habitat where the plant grows or information about the way the plant looks and behaves.
+ "MISCELLANEOUS" - Information to include: any leftover text that does not fit into the previous categories.
+ "Summary" - The second of two required keys in the output dict, fomatted_ocr. A brief one sentence summary of the content.
+ """
+
+ prompt = f"""Given the following set of rules:
+ set_rules = {set_rules}
+ You must parse the OCR content into the following formatted dictionary:
+ structure = {structure}
+ The following is the raw OCR text that you must reformat into a properly formatted Python dictionary based on the set_rules:
+ OCR = {OCR}
+ The following are descriptions of what information to bin into each TAXONOMY, GEOGRAPHY, LOCALITY, COLLECTING, and MISCELLANEOUS category, plus the Summary:
+ descriptions = {category_rules}
+ Please transform the OCR text into a Python dictionary following the rules to complete this dictionary:
+ formatted_ocr = """
+ # print(f'{OCR}\n\n')
+ return prompt
+
+### GPT4 edited PROMPT_UMICH_skeleton_all_asia to creat the following prompt:
+def PROMPT_UMICH_skeleton_all_asia_GPT4(OCR, domain_knowledge_example, similarity):
+ set_rules = """
+ Please note that your task is to generate a dictionary, following the below rules:
+ 1. Refactor the unstructured OCR text into a dictionary based on the reference dictionary structure (ref_dict).
+ 2. Each field of OCR corresponds to a column of the ref_dict. You should correctly map the values from OCR to the respective fields in ref_dict.
+ 3. If the OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None".
+ 4. If there is a field in the ref_dict that does not have a corresponding value in the OCR text, fill it based on your knowledge but don't generate new information.
+ 5. Do not use any text from the ref_dict values in the new dict, but you must use the headers from ref_dict.
+ 6. Duplicate dictionary fields are not allowed.
+ 7. Only return the new dictionary. You should not explain your answer.
+ 8. Your output should be a Python dictionary represented as a JSON string.
+ """
+
+ umich_all_asia_rules = """{
+ "Catalog Number": {
+ "format": "[Catalog Number]",
+ "null_value": "",
+ "description": "The barcode identifier, typically a number with at least 6 digits, but fewer than 30 digits"
+ },
+ "Genus": {
+ "format": "[Genus] or '[Family] indet' if no genus",
+ "null_value": "",
+ "description": "Taxonomic determination to genus, do capitalize genus"
+ },
+ "Species": {
+ "format": "[species] or 'indet' if no species",
+ "null_value": "",
+ "description": "Taxonomic determination to species, do not capitalize species"
+ },
+ "subspecies": {
+ "format": "[subspecies]",
+ "null_value": "",
+ "description": "Taxonomic determination to subspecies (subsp.)"
+ },
+ "variety": {
+ "format": "[variety]",
+ "null_value": "",
+ "description": "Taxonomic determination to variety (var)"
+ },
+ "forma": {
+ "format": "[form]",
+ "null_value": "",
+ "description": "Taxonomic determination to form (f.)"
+ },
+ "Country": {
+ "format": "[Country]",
+ "null_value": "",
+ "description": "Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abbreviation is given"
+ },
+ "State": {
+ "format": "[Adm. Division 1]",
+ "null_value": "",
+ "description": "Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word"
+ },
+ "County": {
+ "format": "[Adm. Division 2]",
+ "null_value": "",
+ "description": "Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word"
+ },
+ "Locality Name": {
+ "format": "verbatim, if no geographic info: 'no data provided on label of catalog no: [######]', or if illegible: 'locality present but illegible/not translated for catalog no: #######', or if no named locality: 'no named locality for catalog no: #######'",
+ "description": "Description of geographic location or landscape"
+ },
+ "Min Elevation": {
+ "format": "elevation integer",
+ "null_value": "",
+ "description": "Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given"
+ },
+ "Max Elevation": {
+ "format": "elevation integer",
+ "null_value": "",
+ "description": "Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise"
+ },
+ "Elevation Units": {
+ "format": "m",
+ "null_value": "",
+ "description": "'m' only if an elevation is present"
+ },
+ "Verbatim Coordinates": {
+ "format": "[Lat, Long | UTM | TRS]",
+ "null_value": "",
+ "description": "Verbatim coordinates as they appear on the label, fix typos to match standardized GPS coordinate format"
+ },
+ "Datum": {
+ "format": "[WGS84, NAD23 etc.]",
+ "null_value": "",
+ "description": "GPS Datum of coordinates on label; empty string "" if GPS coordinates are not in OCR"
+ },
+ "Cultivated": {
+ "format": "yes",
+ "null_value": "",
+ "description": "Indicates if specimen was grown in cultivation"
+ },
+ "Habitat": {
+ "format": "verbatim",
+ "null_value": "",
+ "description": "Description of habitat or location where specimen was collected, ignore descriptions of the plant itself"
+ },
+ "Collectors": {
+ "format": "[Collector]",
+ "null_value": "not present",
+ "description": "Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas"
+ },
+ "Collector Number": {
+ "format": "[Collector No.]",
+ "null_value": "s.n.",
+ "description": "Sequential number assigned to collection, associated with the collector"
+ },
+ "Verbatim Date": {
+ "format": "verbatim",
+ "null_value": "s.d.",
+ "description": "Date of collection exactly as it appears on the label"
+ },
+ "Date": {
+ "format": "[yyyy-mm-dd]",
+ "null_value": "",
+ "description": "Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e., 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day"
+ },
+ "End Date": {
+ "format": "[yyyy-mm-dd]",
+ "null_value": "",
+ "description": "If date range is listed, later date of collection range"
+ }
+ }"""
+
+ structure = """{"Dictionary":
+ {
+ "Catalog Number": [Catalog Number],
+ "Genus": [Genus],
+ "Species": [species],
+ "subspecies": [subspecies],
+ "variety": [variety],
+ "forma": [forma],
+ "Country": [Country],
+ "State": [State],
+ "County": [County],
+ "Locality Name": [Locality Name],
+ "Min Elevation": [Min Elevation],
+ "Max Elevation": [Max Elevation],
+ "Elevation Units": [Elevation Units],
+ "Verbatim Coordinates": [Verbatim Coordinates],
+ "Datum": [Datum],
+ "Cultivated": [Cultivated],
+ "Habitat": [Habitat],
+ "Collectors": [Collectors],
+ "Collector Number": [Collector Number],
+ "Verbatim Date": [Verbatim Date],
+ "Date": [Date],
+ "End Date": [End Date]
+ },
+ "SpeciesName": {"taxonomy": [Genus_species]}}"""
+
+ prompt = f"""I'm providing you with a set of rules, an unstructured OCR text, and a reference dictionary (domain knowledge). Your task is to convert the OCR text into a structured dictionary that matches the structure of the reference dictionary. Please follow the rules strictly.
+ The rules are as follows:
+ {set_rules}
+ The unstructured OCR text is:
+ {OCR}
+ The reference dictionary, which provides an example of the output structure and has an embedding distance of {similarity} to the OCR, is:
+ {domain_knowledge_example}
+ Some dictionary fields have special requirements. These requirements specify the format for each field, and are given below:
+ {umich_all_asia_rules}
+ Please refactor the OCR text into a dictionary, following the rules and the reference structure:
+ {structure}
+ """
+
+ return prompt
+
+
+### GPT4 edited PROMPT_OCR_Organized prompt:
+def PROMPT_OCR_Organized_GPT4(OCR):
+ set_rules = """
+ You need to parse a messy text and return a new dictionary, based on the following rules:
+ 1. The messy text is similar to the information contained in Darwin Core Archive files for herbarium specimens.
+ 2. You need to organize the text into 4 different information categories: TAXONOMY, GEOGRAPHY, LOCALITY, COLLECTING, and MISCELLANEOUS. Use semantic analysis to do so.
+ 3. Separate discrete content within each category with a comma separator "," and denote it with quotation marks, like this "discrete content".
+ 4. When the content falls under the GEOGRAPHY category and more detailed information is available, include the additional information.
+ 5. Transcribe the OCR text verbatim unless there is a typo. Correct any typos or misspellings and adjust the capitalization of the letters in content words to fit standard conventions.
+ 6. The output should follow the structure given in 'structure'. If the content described in the descriptions below isn't present in the OCR text, just skip those subfields.
+ 7. Your output should only be the new dictionary. You should not explain your answer.
+ 8. The output should include a 'Summary' section, providing a brief one-sentence overview of the OCR text content. This should be a general summary, touching upon the main points from all categories.
+ """
+
+ category_rules = """{
+ "TAXONOMY": {
+ "description": "Include all content that pertains to the name of the plant species, such as Order, Family, Genus, Species, Subspecies, Variety, and Forma."
+ },
+ "GEOGRAPHY": {
+ "description": "Include names of places that are government-defined and appear on a map with political boundaries, such as Countries, States, Prefectures, Provinces, Districts, Counties, Cities, or Administrative Divisions. Adjust capitalization to follow standard conventions."
+ },
+ "LOCALITY": {
+ "description": "Include descriptions of the immediate surroundings or physical landscape, including features such as roads, buildings, landmarks, natural formations, and proximities to towns. Avoid including geopolitical names that fall under the 'GEOGRAPHY' category."
+ },
+ "COLLECTING": {
+ "description": "Include names of the people who collected the specimen; the collector's number; the verbatim date; any dates translated into the format [yyyy-mm-dd] with zeros replacing unknown numbers; details relating to cultivation status or if it was grown in a garden or captive setting; all descriptions of the habitat where the plant grows or information about the plant's appearance and behavior."
+ },
+ "MISCELLANEOUS": {
+ "description": "Include any additional text that does not fit into the previous categories and does not relate directly to any other specified categories."
+ },
+ "Summary": {
+ "description": "The second of two required keys in the output dictionary, 'formatted_ocr'. This should provide a concise, one-sentence summary of the content."
+ }
+ }"""
+
+ structure = """
+ {
+ "Dictionary": {
+ "TAXONOMY": {
+ "Order": "",
+ "Family": "",
+ "Genus": "",
+ "Species": "",
+ "Subspecies": "",
+ "Variety": "",
+ "Forma": ""
+ },
+ "GEOGRAPHY": {
+ "Country": "",
+ "State": "",
+ "Prefecture": "",
+ "Province": "",
+ "District": "",
+ "County": "",
+ "City": "",
+ "Administrative Division": ""
+ },
+ "LOCALITY": {
+ "Landscape": "",
+ "Nearby Places": ""
+ },
+ "COLLECTING": {
+ "Collector": "",
+ "Collector's Number": "",
+ "Verbatim Date": "",
+ "Formatted Date": "",
+ "Cultivation Status": "",
+ "Habitat Description": ""
+ },
+ "MISCELLANEOUS": {
+ "Additional Information": ""
+ }
+ },
+ "Summary": {
+ "Content Summary": ""
+ }
+ }
+ """
+
+ prompt = f"""
+ I'm providing you with a set of rules, and an unstructured OCR text. Your task is to convert the OCR text into a structured dictionary, organized by several categories. Please follow the rules strictly.
+
+ The rules are as follows:
+
+ {set_rules}
+
+ The unstructured OCR text that needs to be restructured is:
+
+ {OCR}
+
+ The information should be organized into the following categories:
+
+ {category_rules}
+
+ The structure of the output dictionary should be as follows:
+
+ {structure}
+
+ Please transform the OCR text into a dictionary, following these rules and the provided structure.
+ """
+
+ return prompt
+
+'''
+###################################################################
+######################### PaLM ##################################
+###################################################################
+
+Prompts for PaLM
+ PROMPT_PaLM_UMICH_skeleton_all_asia
+ - Designed for the "All Asia" TCN at the University of Michigan Herbarium
+ - Has 21 columns for a skeleton record
+
+ PROMPT_PaLM_Redo
+ - PaLM version (2023/06) routinely puts quotation marks inside
+ dictionary fields without escaping the character:
+ correct: \"
+ incorrect: "
+ These appear in GPS coordinates. If json.loads() cannot parse
+ the original output, then PROMPT_PaLM_Redo is triggered, telling
+ PaLM to reformat the JSON string. Usually one redo call will suffice.
+
+ PROMPT_PaLM_OCR_Organized
+ - Similar to the chatGPT version.
+ - Designed to privide human transcribers text that is grouped by category
+ so that QC of automated transcription is faster. This output is sent to
+ a custom QC GUI where human labelers can simply copy and paste raw text
+ (but organized by category) into fields that may have been transcribed
+ incorrectly by the LLM.
+
+
+###################################################################
+######################### PaLM ##################################
+###################################################################
+'''
+def PROMPT_PaLM_UMICH_skeleton_all_asia(OCR, in_list, out_list):
+ set_rules = """1. Your job is to return a new dict based on the structure of the reference dict ref_dict and these are your rules.
+ 2. You must look at ref_dict and refactor the new text called OCR to match the same formatting.
+ 3. OCR contains unstructured text inside of [], use your knowledge to put the OCR text into the correct ref_dict column.
+ 4. If OCR is mostly empty and contains substantially less text than the ref_dict examples, then only return "None" and skip all other steps.
+ 5. If there is a field that does not have a direct proxy in the OCR text, you can fill it in based on your knowledge, but you cannot generate new information.
+ 6. Never put text from the ref_dict values into the new dict, but you must use the headers from ref_dict.
+ 7. There cannot be duplicate dictionary fields.
+ 8. Only return the new dict, do not explain your answer.
+ 9. Do not include quotation marks in content, only use quotation marks to represent values in dictionaries.
+ 10. For GPS coordinates only use Decimal Degrees (D.D°)
+ 11. "Given the input text, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values."""
+
+ umich_all_asia_rules = """
+ "Genus" - {"format": "[Genus]" or "[Family] indet" if no genus", "null_value": "", "description": taxonomic determination to genus, do captalize genus}
+ "Species"- {"format": "[species]" or "indet" if no species, "null_value": "", "description": taxonomic determination to species, do not captalize species}
+ "subspecies" - {"format": "[subspecies]", "null_value": "", "description": taxonomic determination to subspecies (subsp.)}
+ "variety" - {"format": "[variety]", "null_value": "", "description": taxonomic determination to variety (var)}
+ "forma" - {"format": "[form]", "null_value": "", "description": taxonomic determination to form (f.)}
+
+ "Country" - {"format": "[Country]", "null_value": "no data", "description": Country that corresponds to the current geographic location of collection; capitalize first letter of each word; use the entire location name even if an abreviation is given}
+ "State" - {"format": "[Adm. Division 1]", "null_value": "no data", "description": Administrative division 1 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "County" - {"format": "[Adm. Division 2]", "null_value": "no data", "description": Administrative division 2 that corresponds to the current geographic location of collection; capitalize first letter of each word}
+ "Locality Name" - {"format": "verbatim", if no geographic info: "no data provided on label of catalog no: [######]", or if illegible: "locality present but illegible/not translated for catalog no: #######", or if no named locality: "no named locality for catalog no: #######", "description": "Description of geographic location or landscape"}
+
+ "Min Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, default field for elevation if a range is not given}
+ "Max Elevation" - {format: "elevation integer", "null_value": "","description": Elevation or altitude in meters, convert from feet to meters if 'm' or 'meters' is not in the text and round to integer, maximum elevation if there are two elevations listed but '' otherwise}
+ "Elevation Units" - {format: "m", "null_value": "","description": "m" only if an elevation is present}
+
+ "Verbatim Coordinates" - {"format": "[Lat, Long | UTM | TRS]", "null_value": "", "description": Convert coordinates to Decimal Degrees (D.D°) format, do not use Minutes, Seconds or quotation marks}
+
+ "Datum" - {"format": "[WGS84, NAD23 etc.]", "null_value": "not present", "description": Datum of coordinates on label; "" is GPS coordinates are not in OCR}
+ "Cultivated" - {"format": "yes", "null_value": "", "description": Indicates if specimen was grown in cultivation}
+ "Habitat" - {"format": "verbatim", "null_value": "", "description": Description of habitat or location where specimen was collected, ignore descriptions of the plant itself}
+ "Collectors" - {"format": "[Collector]", "null_value": "not present", "description": Full name of person (i.e., agent) who collected the specimen; if more than one person then separate the names with commas}
+ "Collector Number" - {"format": "[Collector No.]", "null_value": "s.n.", "description": Sequential number assigned to collection, associated with the collector}
+ "Verbatim Date" - {"format": "verbatim", "null_value": "s.d.", "description": Date of collection exactly as it appears on the label}
+ "Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": Date of collection formatted as year, month, and day; zeros may be used for unknown values i.e. 0000-00-00 if no date, YYYY-00-00 if only year, YYYY-MM-00 if no day}
+ "End Date" - {"format": "[yyyy-mm-dd]", "null_value": "", "description": If date range is listed, later date of collection range}
+ """
+
+ prompt = f"""Given the following set of rules:
+
+ set_rules = {set_rules}
+
+ Some dict fields have special requirements listed below. First is the column header. After the - is the format. Do not include the instructions with your response:
+
+ requirements = {umich_all_asia_rules}
+
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+
+ input: {in_list[0]}
+
+ output: {out_list[0]}
+
+ input: {in_list[1]}
+
+ output: {out_list[1]}
+
+ input: {in_list[2]}
+
+ output: {out_list[2]}
+
+ input: {OCR}
+
+ output:"""
+
+ return prompt
+ # input: {in_list[3]}
+
+ # output: {out_list[3]}
+
+def PROMPT_PaLM_OCR_Organized(OCR):
+ set_rules = """1. Your job is to parse messy text and return a new dict based on these rules.
+ 2. The messy text is similar to the information contained in Darwin Core Archive files for herbarium specimens.
+ 3. You need to bin the text into 4 different information categories including TAXONOMY, GEOGRAPHY, LOCALITY, COLLECTING and MISCELLANEOUS based on symantics.
+ 4. Within each information category list separate discrete content with the comma seperator ",".
+ 5. Denote discrete content inside of each subcategory with quotation marks, like this "discrete content".
+ 6. If you can provide more detailed information for the GEOGRAPHY category, such as a more thorough location hierarchy, please include additional information along with the verbatim transcriptions.
+ 7. Transcribe verbatim unless there is a typo. You can correct typos and misspellings and you can adjust capitalization of the letters in content words to fit standard conventions given the context.
+ 8. Do not include quotation marks in content, only use quotation marks to represent values in dictionaries.
+ 9. For GPS coordinates only use Decimal Degrees (D.D°) do not use Minutes, Seconds, or quotation marks.
+ 10. If some content listed in the descriptions below are not present in OCR, the just skip those subfields.
+ 11. Only return the new dict, do not explain your answer."""
+
+
+ structure = """{"TAXONOMY": {"taxonomic topic": "relevant taxonomic info", "another taxonomic topic": "relevant taxonomic info"},
+ "GEOGRAPHY": {"geographic topic": "relevant geographic info","another geographic topic": "more relevant geographic info"},
+ "LOCALITY": {"location topic": "relevant location info","another location topic": "relevant location info"},
+ "COLLECTING": {"documentation topic": "relevant documentation info", "another documentation topic": "documentation info"},
+ "MISCELLANEOUS": {"miscellaneous topic": "remaining miscellaneous info", "another miscellaneous topic": "more remaining miscellaneous info"}}"""
+
+
+ category_rules = """"TAXONOMY" - Information to include: all content relating to the name of the plant species including Order, Family, Genus, Species, Subspecies, Variety, and Forma.
+
+ "GEOGRAPHY" - Information to include: The government defined names of places that would appear on a map of political boundaries including Countries, States, Prefectures, Provinces, Districts, Counties, Cities, or Adminstrative Divisions. Adjust capitalization to follow standard conventions for each.
+
+ "LOCALITY" - Information to include: descriptions of the landscape, habitat, surroundings or nearby places including towns, roads, buildings, geologic features, and distances.
+
+ "COLLECTING" - Information to include: the names of the people who collected the specimen; the collector's number; the verbatim date; all dates translated int the format [yyyy-mm-dd] with zeros replacing unknown numbers; anything relating to cultivation status or whether it was grown in a garden or captive setting; all descriptions of the habitat where the plant grows or information about the way the plant looks and behaves.
+
+ "MISCELLANEOUS" - Information to include: any leftover text that does not fit into the previous categories."""
+
+
+ ex_1_out = """{"TAXONOMY": {"species": "Quercus Robur C.","common name": "Ch\u00eane Robur."},"GEOGRAPHY": {"location": "Bord d'un chemin, Rymenam (Ann.), BRUXELLES"},"LOCALITY": {"distance": "300 centimeters","surroundings": "Botanic Garden, Golden Thread","miscellaneous": "I, 500 200, inches, 600 300, 700 400, is per inch (opticn), 800 500, 850 550"},"COLLECTING": {"date": "1867-07-07","collection number": "Acc. 1919","habitat": "HERBIER DU JARDIN BOTANIQUE DE L'\u00c9TAT"},"MISCELLANEOUS": {}}"""
+ ex_1_in = """Quercus Robur C. Ch\u00eane Robur. Bord d'un chemin, Rymenam (Ann.), BRUXELLES 300 centimeters surroundings Botanic Garden, Golden Thread I, 500 200, inches, 600 300, 700 400, is per inch (opticn), 800 500, 850 550 1867 July 7 HERBIER DU JARDIN BOTANIQUE DE L'\u00c9TAT 1919"""
+
+ ex_2_out = """{"TAXONOMY": {"species": "Brookea tomentosa Benth."},"GEOGRAPHY": {"country": "Malaysia","stateProvince": "Sabah","county": "Beaufort District","verbatimLocality": "Beaufort Hill. 5\u00b022'N, 115\u00b045'E. Elev. 200 m.","higherGeography": "Crocker Formation"},"LOCALITY": {"habitat": "Burned logged dipterocarp forest."},"COLLECTING": {"recordedBy": "John H. Beaman","recordNumber": "6844","verbatimEventDate": "28 August 1983","eventDate": "1983-08-28","country": "UNITED STATES","cultivationStatus": "wild","associatedTaxa": "Reed S. Beaman and Teofila E. Beaman"},"MISCELLANEOUS": {"additionalData": "Herbaria of Michigan State University (MSC) and Universiti Kebangsaan Malaysia, Sabah Campus (UKMS), centimeter, 3539788, inches, PLANTS OF BORNEO, 500 200, 600 300, US, Institution, Smithsonian, 700 400, 1, 800 500, 850 550, \u0160 8"}}"""
+ ex_2_in = """Brookea tomentosa Benth. Malaysia Sabah Beaufort District Beaufort Hill. 5\u00b022'N, 115\u00b045'E. 200 m. Crocker Formation habitat Burned logged dipterocarp forest. John H. Beaman 6844 28 August 1983 1983-08-28 UNITED STATES Reed S. Beaman and Teofila E. Beaman Herbaria of Michigan State University (MSC) and Universiti Kebangsaan Malaysia, Sabah Campus (UKMS), centimeter, 3539788, inches, PLANTS OF BORNEO, 500 200, 600 300, US, Institution, Smithsonian, 700 400, 1, 800 500, 850 550, \u0160 8"}}"""
+
+
+ prompt = f"""Given the following set of rules:
+
+ set_rules = {set_rules}
+
+ You must parse the OCR content into the following formatted dictionary:
+
+ structure = {structure}
+
+ The following are descriptions of what information to bin into each TAXONOMY, GEOGRAPHY, LOCALITY, COLLECTING, and MISCELLANEOUS category, plus the Summary:
+
+ descriptions = {category_rules}
+
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+
+ For all field values you must properly escape quotation marks using a backslash so that JSON formatting is maintained.
+
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+
+ input: {ex_1_in}
+
+ output: {ex_1_out}
+
+ input: {ex_2_in}
+
+ output: {ex_2_out}
+
+ input: {OCR}
+
+ output:"""
+
+ return prompt
+
+
+def PROMPT_PaLM_Redo(bad_response):
+ # GPS coordinates are the problem, so skip them
+ ex_1_in = """{"TAXONOMY": {"species": "Quercus Robur C.", "common name": "Ch\u00eane Robur."} "GEOGRAPHY": "location": "Bord d'un chemin, Rymenam (Ann.), BRUXELLES"}, "LOCALITY": {"distance": "300 centimeters", "surroundings": "Botanic Garden, Golden Thread", "miscellaneous": "I, 500 200, inches, 600 300, 700 400, is per inch (opticn), 800 500, 850 550"}, "COLLECTING": {"date": "1867-07-07", "collection number": "Acc. 1919", "habitat": "HERBIER DU JARDIN BOTANIQUE DE L'\u00c9TAT"}, "MISCELLANEOUS": {}}"""
+ ex_1_out = """{"TAXONOMY": {"species": "Quercus Robur C.", "common name": "Ch\u00eane Robur."}, "GEOGRAPHY": {"location": "Bord d'un chemin, Rymenam (Ann.), BRUXELLES"}, "LOCALITY": {"distance": "300 centimeters", "surroundings": "Botanic Garden, Golden Thread", "miscellaneous": "I, 500 200, inches, 600 300, 700 400, is per inch (opticn), 800 500, 850 550"}, "COLLECTING": {"date": "1867-07-07", "collection number": "Acc. 1919", "habitat": "HERBIER DU JARDIN BOTANIQUE DE L'\u00c9TAT"}, "MISCELLANEOUS": {}}"""
+
+ '''for just skipping the verbatim coordinates'''
+ ex_2_in = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "vari"ety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+ ex_2_out = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+
+ ex_3_in = """{"Genus": "Forchammeria" "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+ ex_3_out = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+
+ ex_4_in = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "" "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+ ex_4_out = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+
+ '''for trying to fix the escape chars'''
+ # ex_2_in = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "vari"ety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+ # ex_2_out = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+
+ # ex_3_in = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "-34°6'15"N, 119°45'0"W",, "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+ # ex_3_out = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "-34°6\'15\"N, 119°45\'0\"W",, "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+
+ # ex_4_in = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "-25°3'24"N, 109°50'0"W", "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+ # ex_4_out = """{"Genus": "Forchammeria", "Species": "Watsonii", "subspecies": "", "variety": "", "forma": "", "Country": "Mexico", "State": "Baja California Sur", "County": "Cerralvo Island", "Locality Name": "South end of Cerralvo Island", "Min Elevation": "", "Max Elevation": "", "Elevation Units": "", "Verbatim Coordinates": "-25°3\'24\"N, 109°50\'0\"W",, "Datum": "", "Cultivated": "", "Habitat": "", "Collectors": "Reid Moran", "Collector Number": "3592", "Verbatim Date": "3. April... 1952", "End Date": ""}"""
+
+ prompt = f"""This text is supposed to be JSON, but it contains an error that prevents it from loading with the Python command json.loads().
+
+ You need to return coorect JSON for the following dictionary. Most likely, a quotation mark inside of a field value has not been escaped properly with a backslash.
+
+ Given the input, please generate a JSON response. Please note that the response should not contain any special characters, including quotation marks (single ' or double \"), within the JSON values.
+
+ Escape all JSON control characters that appear in input including ampersand (&) and other control characters.
+
+ input: {ex_1_in}
+
+ output: {ex_1_out}
+
+ input: {ex_2_in}
+
+ output: {ex_2_out}
+
+ input: {ex_3_in}
+
+ output: {ex_3_out}
+
+ input: {ex_4_in}
+
+ output: {ex_4_out}
+
+ input: {bad_response}
+
+ output:"""
+
+ return prompt
+
+
+def PROMPT_JSON(opt, bad_response=''):
+ if opt == 'dict':
+
+ guide = f"""This is the JSON text that contains an error, typically there is an errant quotation mark inside of value, so escape with a backslash any quotation marks that occur in the middle of a value field:
+ {bad_response}"""
+
+ structure = """Below is the correct JSON formatting. Modify the text to conform to the following format, fixing the incorrect JSON:
+ {"Dictionary":
+ {
+ "Catalog Number": [Catalog Number],
+ "Genus": [Genus],
+ "Species": [species],
+ "subspecies": [subspecies],
+ "variety": [variety],
+ "forma": [forma],
+ "Country": [Country],
+ "State": [State],
+ "County": [County],
+ "Locality Name": [Locality Name],
+ "Min Elevation": [Min Elevation],
+ "Max Elevation": [Max Elevation],
+ "Elevation Units": [Elevation Units],
+ "Verbatim Coordinates": [Verbatim Coordinates],
+ "Datum": [Datum],
+ "Cultivated": [Cultivated],
+ "Habitat": [Habitat],
+ "Collectors": [Collectors],
+ "Collector Number": [Collector Number],
+ "Verbatim Date": [Verbatim Date],
+ "Date": [Date],
+ "End Date": [End Date]
+ },
+ "SpeciesName": {"taxonomy": [Genus_species]}}"""
+ prompt = '\n'.join([guide, structure])
+ return prompt
+
+ elif opt == 'helper':
+ guide = f"""This is the JSON text that contains an 1error, typically there is an errant quotation mark inside of value, so escape with a backslash any quotation marks that occur in the middle of a value field:
+ {bad_response}"""
+
+ structure = """Below is the correct JSON formatting. Modify the text to conform to the following format, fixing the incorrect JSON:
+ {
+ "Dictionary": {
+ "TAXONOMY": {
+ "Order": "",
+ "Family": "",
+ "Genus": "",
+ "Species": "",
+ "Subspecies": "",
+ "Variety": "",
+ "Forma": ""
+ },
+ "GEOGRAPHY": {
+ "Country": "",
+ "State": "",
+ "Prefecture": "",
+ "Province": "",
+ "District": "",
+ "County": "",
+ "City": "",
+ "Administrative Division": ""
+ },
+ "LOCALITY": {
+ "Landscape": "",
+ "Nearby Places": ""
+ },
+ "COLLECTING": {
+ "Collector": "",
+ "Collector's Number": "",
+ "Verbatim Date": "",
+ "Formatted Date": "",
+ "Cultivation Status": "",
+ "Habitat Description": ""
+ },
+ "MISCELLANEOUS": {
+ "Additional Information": ""
+ }
+ },
+ "Summary": {
+ "Content Summary": ""
+ }
+ }
+ """
+ prompt = '\n'.join([guide, structure])
+ return prompt
+
+
+
+
diff --git a/vouchervision/utils_GBIF.py b/vouchervision/utils_GBIF.py
new file mode 100644
index 0000000000000000000000000000000000000000..2ddb101b177c8a82b81b9457bb131545db3a9f9c
--- /dev/null
+++ b/vouchervision/utils_GBIF.py
@@ -0,0 +1,944 @@
+import os, time, requests, yaml, re, csv, sys, inspect
+from dataclasses import dataclass, field
+# from difflib import diff_bytes
+import pandas as pd
+import numpy as np
+from PIL import Image
+import matplotlib.pyplot as plt
+from urllib.parse import urlparse
+from requests.adapters import HTTPAdapter
+from urllib3.util import Retry
+from torch import ge
+from re import S
+from threading import Lock
+from random import shuffle
+from collections import defaultdict
+
+currentdir = os.path.dirname(os.path.dirname(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+sys.path.append(currentdir)
+from concurrent.futures import ThreadPoolExecutor as th
+
+
+from vouchervision.general_utils import bcolors, validate_dir
+
+'''
+For download parallelization, I followed this guide https://rednafi.github.io/digressions/python/2020/04/21/python-concurrent-futures.html
+'''
+
+'''
+####################################################################################################
+Read config files
+####################################################################################################
+'''
+def get_cfg_from_full_path(path_cfg):
+ with open(path_cfg, "r") as ymlfile:
+ cfg = yaml.full_load(ymlfile)
+ return cfg
+
+'''
+Classes
+'''
+@dataclass
+class ImageCandidate:
+ cfg: str = ''
+ herb_code: str = ''
+ specimen_id: str = ''
+ family: str = ''
+ genus: str = ''
+ species: str = ''
+ fullname: str = ''
+
+ filename_image: str = ''
+ filename_image_jpg: str = ''
+
+ url: str = ''
+ headers_occ: str = ''
+ headers_img: str = ''
+
+ occ_row: list = field(init=False,default_factory=None)
+ image_row: list = field(init=False,default_factory=None)
+
+
+ def __init__(self, cfg, image_row, occ_row, url, lock):
+ # self.headers_occ = list(occ_row.columns.values)
+ # self.headers_img = list(image_row.columns.values)
+ self.headers_occ = occ_row
+ self.headers_img = image_row
+ self.occ_row = occ_row # pd.DataFrame(data=occ_row,columns=self.headers_occ)
+ self.image_row = image_row # pd.DataFrame(data=image_row,columns=self.headers_img)
+ self.url = url
+ self.cfg = cfg
+
+ self.filename_image, self.filename_image_jpg, self.herb_code, self.specimen_id, self.family, self.genus, self.species, self.fullname = generate_image_filename(occ_row)
+ self.download_image(lock)
+
+ def download_image(self, lock) -> None:
+ dir_destination = self.cfg['dir_destination_images']
+ MP_low = self.cfg['MP_low']
+ MP_high = self.cfg['MP_high']
+ # Define URL get parameters
+ sep = '_'
+ session = requests.Session()
+ retry = Retry(connect=1) #2, backoff_factor=0.5)
+ adapter = HTTPAdapter(max_retries=retry)
+ session.mount('http://', adapter)
+ session.mount('https://', adapter)
+
+ print(f"{bcolors.BOLD} {self.fullname}{bcolors.ENDC}")
+ print(f"{bcolors.BOLD} URL: {self.url}{bcolors.ENDC}")
+ try:
+ response = session.get(self.url, stream=True, timeout=1.0)
+ img = Image.open(response.raw)
+ self._save_matching_image(img, MP_low, MP_high, dir_destination, lock)
+ print(f"{bcolors.OKGREEN} SUCCESS{bcolors.ENDC}")
+ except Exception as e:
+ print(f"{bcolors.FAIL} SKIP No Connection or ERROR --> {e}{bcolors.ENDC}")
+ print(f"{bcolors.WARNING} Status Code --> {response.status_code}{bcolors.ENDC}")
+ print(f"{bcolors.WARNING} Reasone --> {response.reason}{bcolors.ENDC}")
+
+ def _save_matching_image(self, img, MP_low, MP_high, dir_destination, lock) -> None:
+ img_mp, img_w, img_h = check_image_size(img)
+ if img_mp < MP_low:
+ print(f"{bcolors.WARNING} SKIP < {MP_low}MP: {img_mp}{bcolors.ENDC}")
+
+ elif MP_low <= img_mp <= MP_high:
+ image_path = os.path.join(dir_destination,self.filename_image_jpg)
+ img.save(image_path)
+
+ #imgSaveName = pd.DataFrame({"image_path": [image_path]})
+ self._add_occ_and_img_data(lock)
+
+ print(f"{bcolors.OKGREEN} Regular MP: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKGREEN} Image Saved: {image_path}{bcolors.ENDC}")
+
+ elif img_mp > MP_high:
+ if self.cfg['do_resize']:
+ [img_w, img_h] = calc_resize(img_w, img_h)
+ newsize = (img_w, img_h)
+ img = img.resize(newsize)
+ image_path = os.path.join(dir_destination,self.filename_image_jpg)
+ img.save(image_path)
+
+ #imgSaveName = pd.DataFrame({"imgSaveName": [imgSaveName]})
+ self._add_occ_and_img_data(lock)
+
+ print(f"{bcolors.OKGREEN} {MP_high}MP+ Resize: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKGREEN} Image Saved: {image_path}{bcolors.ENDC}")
+ else:
+ print(f"{bcolors.OKCYAN} {MP_high}MP+ Resize: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKCYAN} SKIP: {image_path}{bcolors.ENDC}")
+
+ def _add_occ_and_img_data(self, lock) -> None:
+ self.image_row = self.image_row.to_frame().transpose().rename(columns={"identifier": "url"})
+ self.image_row = self.image_row.rename(columns={"gbifID": "gbifID_images"})
+
+ new_data = {'fullname': [self.fullname], 'filename_image': [self.filename_image], 'filename_image_jpg': [self.filename_image_jpg]}
+ new_data = pd.DataFrame(data=new_data)
+
+ all_data = [new_data.reset_index(), self.image_row.reset_index(), self.occ_row.reset_index()]
+ combined = pd.concat(all_data,ignore_index=False, axis=1)
+
+ w_1 = new_data.shape[1] + 1
+ w_2 = self.image_row.shape[1] + 1
+ w_3 = self.occ_row.shape[1]
+
+ combined.drop([combined.columns[0], combined.columns[w_1], combined.columns[w_1 + w_2]], axis=1, inplace=True)
+ headers = np.hstack((new_data.columns.values, self.image_row.columns.values, self.occ_row.columns.values))
+ combined.columns = headers
+ self._append_combined_occ_image(self.cfg, combined, lock)
+
+ def _append_combined_occ_image(self, cfg, combined, lock) -> None:
+ path_csv_combined = os.path.join(cfg['dir_destination_csv'], cfg['filename_combined'])
+ with lock:
+ try:
+ # Add row once the file exists
+ csv_combined = pd.read_csv(path_csv_combined,dtype=str)
+ combined.to_csv(path_csv_combined, mode='a', header=False, index=False)
+ print(f'{bcolors.OKGREEN} Added 1 row to combined CSV: {path_csv_combined}{bcolors.ENDC}')
+
+ except Exception as e:
+ print(f"{bcolors.WARNING} Initializing new combined .csv file: [occ,images]: {path_csv_combined}{bcolors.ENDC}")
+ combined.to_csv(path_csv_combined, mode='w', header=True, index=False)
+
+
+
+@dataclass
+class ImageCandidateMulti:
+ cfg: str = ''
+ herb_code: str = ''
+ specimen_id: str = ''
+ family: str = ''
+ genus: str = ''
+ species: str = ''
+ fullname: str = ''
+
+ filename_image: str = ''
+ filename_image_jpg: str = ''
+
+ url: str = ''
+ headers_occ: str = ''
+ headers_img: str = ''
+
+ occ_row: list = field(init=False,default_factory=None)
+ image_row: list = field(init=False,default_factory=None)
+
+ download_success: bool = False
+
+
+ def __init__(self, cfg, image_row, occ_row, url, dir_destination, lock):
+ # Convert the Series to a DataFrame with one row
+ try:
+ # Now, you can access columns and data as you would in a DataFrame
+ self.headers_occ = occ_row
+ self.headers_img = image_row
+ except Exception as e:
+ print(f"Exception occurred: {e}")
+
+
+ self.occ_row = occ_row # pd.DataFrame(data=occ_row,columns=self.headers_occ)
+ self.image_row = image_row # pd.DataFrame(data=image_row,columns=self.headers_img)
+ self.url = url
+ self.cfg = cfg
+
+ self.filename_image, self.filename_image_jpg, self.herb_code, self.specimen_id, self.family, self.genus, self.species, self.fullname = generate_image_filename(occ_row)
+
+ self.download_success = self.download_image(dir_destination, lock)
+
+
+
+ def download_image(self, dir_destination, lock) -> None:
+ # dir_destination = self.cfg['dir_destination_images']
+ MP_low = self.cfg['MP_low']
+ MP_high = self.cfg['MP_high']
+ # Define URL get parameters
+ sep = '_'
+ session = requests.Session()
+ retry = Retry(connect=1) #2, backoff_factor=0.5)
+ adapter = HTTPAdapter(max_retries=retry)
+ session.mount('http://', adapter)
+ session.mount('https://', adapter)
+
+ print(f"{bcolors.BOLD} {self.fullname}{bcolors.ENDC}")
+ print(f"{bcolors.BOLD} URL: {self.url}{bcolors.ENDC}")
+ try:
+ response = session.get(self.url, stream=True, timeout=1.0)
+ img = Image.open(response.raw)
+ self._save_matching_image(img, MP_low, MP_high, dir_destination, lock)
+ print(f"{bcolors.OKGREEN} SUCCESS{bcolors.ENDC}")
+ return True
+ except Exception as e:
+ print(f"{bcolors.FAIL} SKIP No Connection or ERROR --> {e}{bcolors.ENDC}")
+ print(f"{bcolors.WARNING} Status Code --> {response.status_code}{bcolors.ENDC}")
+ print(f"{bcolors.WARNING} Reasone --> {response.reason}{bcolors.ENDC}")
+ return False
+
+ def _save_matching_image(self, img, MP_low, MP_high, dir_destination, lock) -> None:
+ img_mp, img_w, img_h = check_image_size(img)
+ if img_mp < MP_low:
+ print(f"{bcolors.WARNING} SKIP < {MP_low}MP: {img_mp}{bcolors.ENDC}")
+
+ elif MP_low <= img_mp <= MP_high:
+ image_path = os.path.join(dir_destination,self.filename_image_jpg)
+ img.save(image_path)
+
+ #imgSaveName = pd.DataFrame({"image_path": [image_path]})
+ self._add_occ_and_img_data(lock)
+
+ print(f"{bcolors.OKGREEN} Regular MP: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKGREEN} Image Saved: {image_path}{bcolors.ENDC}")
+
+ elif img_mp > MP_high:
+ if self.cfg['do_resize']:
+ [img_w, img_h] = calc_resize(img_w, img_h)
+ newsize = (img_w, img_h)
+ img = img.resize(newsize)
+ image_path = os.path.join(dir_destination,self.filename_image_jpg)
+ img.save(image_path)
+
+ #imgSaveName = pd.DataFrame({"imgSaveName": [imgSaveName]})
+ self._add_occ_and_img_data(lock)
+
+ print(f"{bcolors.OKGREEN} {MP_high}MP+ Resize: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKGREEN} Image Saved: {image_path}{bcolors.ENDC}")
+ else:
+ print(f"{bcolors.OKCYAN} {MP_high}MP+ Resize: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKCYAN} SKIP: {image_path}{bcolors.ENDC}")
+
+ def _add_occ_and_img_data(self, lock) -> None:
+ self.image_row = self.image_row.to_frame().transpose().rename(columns={"identifier": "url"})
+ self.image_row = self.image_row.rename(columns={"gbifID": "gbifID_images"})
+
+ new_data = {'fullname': [self.fullname], 'filename_image': [self.filename_image], 'filename_image_jpg': [self.filename_image_jpg]}
+ new_data = pd.DataFrame(data=new_data)
+
+ all_data = [new_data.reset_index(), self.image_row.reset_index(), self.occ_row.reset_index()]
+ combined = pd.concat(all_data,ignore_index=False, axis=1)
+
+ w_1 = new_data.shape[1] + 1
+ w_2 = self.image_row.shape[1] + 1
+ w_3 = self.occ_row.shape[1]
+
+ combined.drop([combined.columns[0], combined.columns[w_1], combined.columns[w_1 + w_2]], axis=1, inplace=True)
+ headers = np.hstack((new_data.columns.values, self.image_row.columns.values, self.occ_row.columns.values))
+ combined.columns = headers
+ self._append_combined_occ_image(self.cfg, combined, lock)
+
+ def _append_combined_occ_image(self, cfg, combined, lock) -> None:
+ path_csv_combined = os.path.join(cfg['dir_destination_csv'], cfg['filename_combined'])
+ with lock:
+ try:
+ # Add row once the file exists
+ csv_combined = pd.read_csv(path_csv_combined,dtype=str)
+ combined.to_csv(path_csv_combined, mode='a', header=False, index=False)
+ print(f'{bcolors.OKGREEN} Added 1 row to combined CSV: {path_csv_combined}{bcolors.ENDC}')
+
+ except Exception as e:
+ print(f"{bcolors.WARNING} Initializing new combined .csv file: [occ,images]: {path_csv_combined}{bcolors.ENDC}")
+ combined.to_csv(path_csv_combined, mode='w', header=True, index=False)
+
+class SharedCounter:
+ def __init__(self):
+ self.img_count_dict = {}
+ self.lock = Lock()
+
+ def increment(self, key, value=1):
+ with self.lock:
+ self.img_count_dict[key] = self.img_count_dict.get(key, 0) + value
+
+ def get_count(self, key):
+ with self.lock:
+ return self.img_count_dict.get(key, 0)
+
+
+
+@dataclass
+class ImageCandidateCustom:
+ cfg: str = ''
+ # herb_code: str = ''
+ # specimen_id: str = ''
+ # family: str = ''
+ # genus: str = ''
+ # species: str = ''
+ fullname: str = ''
+
+ filename_image: str = ''
+ filename_image_jpg: str = ''
+
+ url: str = ''
+ # headers_occ: str = ''
+ headers_img: str = ''
+
+ # occ_row: list = field(init=False,default_factory=None)
+ image_row: list = field(init=False,default_factory=None)
+
+
+ def __init__(self, cfg, image_row, url, col_name, lock):
+ # self.headers_occ = list(occ_row.columns.values)
+ # self.headers_img = list(image_row.columns.values)
+ self.image_row = image_row # pd.DataFrame(data=image_row,columns=self.headers_img)
+
+ self.url = url
+ self.cfg = cfg
+ self.col_name = col_name
+
+ self.fullname = image_row[col_name]
+ self.filename_image = image_row[col_name]
+ self.filename_image_jpg = ''.join([image_row[col_name], '.jpg'])
+
+ self.download_image(lock)
+
+ def download_image(self, lock) -> None:
+ dir_destination = self.cfg['dir_destination_images']
+ MP_low = self.cfg['MP_low']
+ MP_high = self.cfg['MP_high']
+ # Define URL get parameters
+ sep = '_'
+ session = requests.Session()
+ retry = Retry(connect=1) #2, backoff_factor=0.5)
+ adapter = HTTPAdapter(max_retries=retry)
+ session.mount('http://', adapter)
+ session.mount('https://', adapter)
+
+ print(f"{bcolors.BOLD} {self.fullname}{bcolors.ENDC}")
+ print(f"{bcolors.BOLD} URL: {self.url}{bcolors.ENDC}")
+ try:
+ response = session.get(self.url, stream=True, timeout=1.0)
+ img = Image.open(response.raw)
+ self._save_matching_image(img, MP_low, MP_high, dir_destination, lock)
+ print(f"{bcolors.OKGREEN} SUCCESS{bcolors.ENDC}")
+ except Exception as e:
+ print(f"{bcolors.FAIL} SKIP No Connection or ERROR --> {e}{bcolors.ENDC}")
+ print(f"{bcolors.WARNING} Status Code --> {response.status_code}{bcolors.ENDC}")
+ print(f"{bcolors.WARNING} Reasone --> {response.reason}{bcolors.ENDC}")
+
+ def _save_matching_image(self, img, MP_low, MP_high, dir_destination, lock) -> None:
+ img_mp, img_w, img_h = check_image_size(img)
+ if img_mp < MP_low:
+ print(f"{bcolors.WARNING} SKIP < {MP_low}MP: {img_mp}{bcolors.ENDC}")
+
+ elif MP_low <= img_mp <= MP_high:
+ image_path = os.path.join(dir_destination,self.filename_image_jpg)
+ img.save(image_path)
+
+ print(f"{bcolors.OKGREEN} Regular MP: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKGREEN} Image Saved: {image_path}{bcolors.ENDC}")
+
+ elif img_mp > MP_high:
+ if self.cfg['do_resize']:
+ [img_w, img_h] = calc_resize(img_w, img_h)
+ newsize = (img_w, img_h)
+ img = img.resize(newsize)
+ image_path = os.path.join(dir_destination,self.filename_image_jpg)
+ img.save(image_path)
+
+ print(f"{bcolors.OKGREEN} {MP_high}MP+ Resize: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKGREEN} Image Saved: {image_path}{bcolors.ENDC}")
+ else:
+ print(f"{bcolors.OKCYAN} {MP_high}MP+ Resize: {img_mp}{bcolors.ENDC}")
+ print(f"{bcolors.OKCYAN} SKIP: {image_path}{bcolors.ENDC}")
+
+
+'''
+####################################################################################################
+General Functions
+####################################################################################################
+'''
+# If image is larger than MP max, downsample to have long side = 5000
+def calc_resize(w,h):
+ if h > w:
+ ratio = h/w
+ new_h = 5000
+ new_w = round(5000/ratio)
+ elif w >= h:
+ ratio = w/h
+ new_w = 5000
+ new_h = round(5000/ratio)
+ return new_w, new_h
+
+def check_image_size(img):
+ [img_w, img_h] = img.size
+ img_mp = round(img_w * img_h / 1000000,1)
+ return img_mp, img_w, img_h
+
+def check_n_images_in_group(detailedOcc,N):
+ fam = detailedOcc['fullname'].unique()
+ for f in fam:
+ ct = len(detailedOcc[detailedOcc['fullname'].str.match(f)])
+ if ct == N:
+ print(f"{bcolors.OKGREEN}{f}: {ct}{bcolors.ENDC}")
+ else:
+ print(f"{bcolors.FAIL}{f}: {ct}{bcolors.ENDC}")
+
+
+
+'''
+####################################################################################################
+Functions for --> download_GBIF_from_user_file.py
+####################################################################################################
+'''
+
+# def download_subset_images_user_file(dir_home,dir_destination,n_already_downloaded,MP_low,MP_high,wishlist,filename_occ,filename_img):
+# # (dirWishlists,dirNewImg,alreadyDownloaded,MP_Low,MP_High,wishlist,aggOcc_filename,aggImg_filename):
+# sep = '_'
+# aggOcc = pd.DataFrame()
+# aggImg = pd.DataFrame()
+
+# # Define URL get parameters
+# session = requests.Session()
+# retry = Retry(connect=1) #2, backoff_factor=0.5)
+# adapter = HTTPAdapter(max_retries=retry)
+# session.mount('http://', adapter)
+# session.mount('https://', adapter)
+
+# listMax = wishlist.shape[0]
+# for index, spp in wishlist.iterrows():
+# imageFound = False
+# currentFamily = spp['family']
+# # currentSpecies = spp['genus'] + ' ' + spp['species']
+# currentFullname = spp['fullname']
+# currentURL = spp['url']
+# currentBarcode = spp['barcode']
+# currentHerb = spp['herbCode']
+# print(f"{bcolors.BOLD}Family: {currentFamily}{bcolors.ENDC}")
+# print(f"{bcolors.BOLD} {currentFullname}{bcolors.ENDC}")
+# print(f"{bcolors.BOLD} In Download List: {index} / {listMax}{bcolors.ENDC}")
+
+# imgFilename = [currentHerb, currentBarcode, currentFullname]
+# imgFilename = sep.join(imgFilename)
+# imgFilenameJPG = imgFilename + ".jpg"
+# print(f"{bcolors.BOLD} URL: {currentURL}{bcolors.ENDC}")
+# try:
+# img = Image.open(session.get(currentURL, stream=True, timeout=1.0).raw)
+# imageFound, alreadyDownloaded, aggOcc, aggImg = save_matching_image_user_file(alreadyDownloaded,img,MP_Low,MP_High,dirNewImg,imgFilenameJPG)
+# print(f"{bcolors.OKGREEN} SUCCESS{bcolors.ENDC}")
+# except Exception as e:
+# print(f"{bcolors.WARNING} SKIP No Connection or ERROR{bcolors.ENDC}")
+
+
+# aggOcc.to_csv(os.path.join(dir_home,aggOcc_filename),index=False)
+# aggImg.to_csv(os.path.join(dir_home,aggImg_filename),index=False)
+
+# return alreadyDownloaded, aggOcc, aggImg
+
+
+# Return entire row of file_to_search that matches the gbif_id, else return []
+def find_gbifID(gbif_id,file_to_search):
+ row_found = file_to_search.loc[file_to_search['gbifID'].astype(str).str.match(str(gbif_id)),:]
+ if row_found.empty:
+ print(f"{bcolors.WARNING} gbif_id: {gbif_id} not found in occurrences file{bcolors.ENDC}")
+ row_found = None
+ else:
+ print(f"{bcolors.OKGREEN} gbif_id: {gbif_id} successfully found in occurrences file{bcolors.ENDC}")
+ return row_found
+
+def validate_herb_code(occ_row):
+ # print(occ_row)
+ # Herbarium codes are not always in the correct column, we need to find the right one
+ try:
+ opts = [occ_row['institutionCode'],
+ occ_row['institutionID'],
+ occ_row['ownerInstitutionCode'],
+ occ_row['collectionCode'],
+ occ_row['publisher'],
+ occ_row['occurrenceID']]
+ opts = [item for item in opts if not(pd.isnull(item.values)) == True]
+ except:
+ opts = [str(occ_row['institutionCode']),
+ str(occ_row['institutionID']),
+ str(occ_row['ownerInstitutionCode']),
+ str(occ_row['collectionCode']),
+ str(occ_row['publisher']),
+ str(occ_row['occurrenceID'])]
+ opts = pd.DataFrame(opts)
+ opts = opts.dropna()
+ opts = opts.apply(lambda x: x[0]).tolist()
+
+ opts_short = []
+
+ for word in opts:
+ #print(word)
+ if len(word) <= 8:
+ if word is not None:
+ opts_short = opts_short + [word]
+
+ if len(opts_short) == 0:
+ try:
+ herb_code = occ_row['publisher'].values[0].replace(" ","-")
+ except:
+ try:
+ herb_code = occ_row['publisher'].replace(" ","-")
+ except:
+ herb_code = "ERROR"
+ try:
+ inst_ID = occ_row['institutionID'].values[0]
+ occ_ID = occ_row['occurrenceID'].values[0]
+ except:
+ inst_ID = occ_row['institutionID']
+ occ_ID = occ_row['occurrenceID']
+ if inst_ID == "UBC Herbarium":
+ herb_code = "UBC"
+ elif inst_ID == "Naturalis Biodiversity Center":
+ herb_code = "L"
+ elif inst_ID == "Forest Herbarium Ibadan (FHI)":
+ herb_code = "FHI"
+ elif 'id.luomus.fi' in occ_ID:
+ herb_code = "FinBIF"
+ else:
+ if len(opts_short) > 0:
+ herb_code = opts_short[0]
+
+ try:
+ herb_code = herb_code.values[0]
+ except:
+ herb_code = herb_code
+
+ # Specific cases that require manual overrides
+ # If you see an herbarium DWC file with a similar error, add them here
+ if herb_code == "Qarshi-Botanical-Garden,-Qarshi-Industries-Pvt.-Ltd,-Pakistan":
+ herb_code = "Qarshi-Botanical-Garden"
+ elif herb_code == "12650":
+ herb_code = "SDSU"
+ elif herb_code == "322":
+ herb_code = "SDSU"
+ elif herb_code == "GC-University,-Lahore":
+ herb_code = "GC-University-Lahore"
+ elif herb_code == "Institute-of-Biology-of-Komi-Scientific-Centre-of-the-Ural-Branch-of-the-Russian-Academy-of-Sciences":
+ herb_code = "Komi-Scientific-Centre"
+
+ return herb_code
+
+def remove_illegal_chars(text):
+ cleaned = re.sub(r"[^a-zA-Z0-9_-]","",text)
+ return cleaned
+
+def keep_first_word(text):
+ if (' ' in text) == True:
+ cleaned = text.split(' ')[0]
+ else:
+ cleaned = text
+ return cleaned
+
+# Create a filename for the downloaded image
+# In the case sensitive format:
+# HERBARIUM_barcode_Family_Genus_species.jpg
+def generate_image_filename(occ_row):
+ herb_code = remove_illegal_chars(validate_herb_code(occ_row))
+ try:
+ specimen_id = str(occ_row['gbifID'].values[0])
+ family = remove_illegal_chars(occ_row['family'].values[0])
+ genus = remove_illegal_chars(occ_row['genus'].values[0])
+ species = remove_illegal_chars(keep_first_word(occ_row['specificEpithet'].values[0]))
+ except:
+ specimen_id = str(occ_row['gbifID'])
+ family = remove_illegal_chars(occ_row['family'])
+ genus = remove_illegal_chars(occ_row['genus'])
+ species = remove_illegal_chars(keep_first_word(occ_row['specificEpithet']))
+ fullname = '_'.join([family, genus, species])
+
+ filename_image = '_'.join([herb_code, specimen_id, fullname])
+ filename_image_jpg = '.'.join([filename_image, 'jpg'])
+
+ return filename_image, filename_image_jpg, herb_code, specimen_id, family, genus, species, fullname
+
+def read_DWC_file(cfg):
+ dir_home = cfg['dir_home']
+ filename_occ = cfg['filename_occ']
+ filename_img = cfg['filename_img']
+ # read the images.csv or occurences.csv file. can be txt ro csv
+ occ_df = ingest_DWC(filename_occ,dir_home)
+ images_df = ingest_DWC(filename_img,dir_home)
+ return occ_df, images_df
+
+def read_DWC_file_multiDirs(cfg, dir_sub):
+ filename_occ = cfg['filename_occ']
+ filename_img = cfg['filename_img']
+ # read the images.csv or occurences.csv file. can be txt ro csv
+ occ_df = ingest_DWC(filename_occ,dir_sub)
+ images_df = ingest_DWC(filename_img,dir_sub)
+ return occ_df, images_df
+
+def ingest_DWC(DWC_csv_or_txt_file,dir_home):
+ if DWC_csv_or_txt_file.split('.')[1] == 'txt':
+ df = pd.read_csv(os.path.join(dir_home,DWC_csv_or_txt_file), sep="\t",header=0, low_memory=False, dtype=str)
+ elif DWC_csv_or_txt_file.split('.')[1] == 'csv':
+ df = pd.read_csv(os.path.join(dir_home,DWC_csv_or_txt_file), sep=",",header=0, low_memory=False, dtype=str)
+ else:
+ print(f"{bcolors.FAIL}DWC file {DWC_csv_or_txt_file} is not '.txt' or '.csv' and was not opened{bcolors.ENDC}")
+ return df
+
+'''
+#######################################################################
+Main function for the config_download_from_GBIF_all_images_in_file.yml
+see yml for details
+#######################################################################
+'''
+def download_all_images_in_images_csv_multiDirs(cfg):
+ dir_destination_parent = cfg['dir_destination_images']
+ dir_destination_csv = cfg['dir_destination_csv']
+ n_already_downloaded = cfg['n_already_downloaded']
+ n_max_to_download = cfg['n_max_to_download']
+ n_imgs_per_species = cfg['n_imgs_per_species']
+ MP_low = cfg['MP_low']
+ MP_high = cfg['MP_high']
+ do_shuffle_occurrences = cfg['do_shuffle_occurrences']
+
+ shared_counter = SharedCounter()
+
+ # (dirWishlists,dirNewImg,alreadyDownloaded,MP_Low,MP_High,aggOcc_filename,aggImg_filename):
+
+
+ # Get DWC files
+ for dir_DWC, dirs_sub, __ in os.walk(cfg['dir_home']):
+ for dir_sub in dirs_sub:
+ dir_home = os.path.join(dir_DWC, dir_sub)
+ dir_destination = os.path.join(dir_destination_parent, dir_sub)
+
+ validate_dir(dir_destination)
+ validate_dir(dir_destination_csv)
+
+ occ_df, images_df = read_DWC_file_multiDirs(cfg, dir_home)
+
+ # Shuffle the order of the occurrences DataFrame if the flag is set
+ if do_shuffle_occurrences:
+ occ_df = occ_df.sample(frac=1).reset_index(drop=True)
+
+ # Report summary
+ print(f"{bcolors.BOLD}Beginning of images file:{bcolors.ENDC}")
+ print(images_df.head())
+ print(f"{bcolors.BOLD}Beginning of occurrence file:{bcolors.ENDC}")
+ print(occ_df.head())
+
+ # Ignore problematic Herbaria
+ if cfg['ignore_banned_herb']:
+ for banned_url in cfg['banned_url_stems']:
+ images_df = images_df[~images_df['identifier'].str.contains(banned_url, na=False)]
+
+ # Report summary
+ n_imgs = images_df.shape[0]
+ n_occ = occ_df.shape[0]
+ print(f"{bcolors.BOLD}Number of images in images file: {n_imgs}{bcolors.ENDC}")
+ print(f"{bcolors.BOLD}Number of occurrence to search through: {n_occ}{bcolors.ENDC}")
+
+ results = process_image_batch_multiDirs(cfg, images_df, occ_df, dir_destination, shared_counter, n_imgs_per_species, do_shuffle_occurrences)
+
+
+def download_all_images_in_images_csv(cfg):
+ dir_destination = cfg['dir_destination_images']
+ dir_destination_csv = cfg['dir_destination_csv']
+
+ # (dirWishlists,dirNewImg,alreadyDownloaded,MP_Low,MP_High,aggOcc_filename,aggImg_filename):
+ validate_dir(dir_destination)
+ validate_dir(dir_destination_csv)
+
+ if cfg['is_custom_file']:
+ download_from_custom_file(cfg)
+ else:
+ # Get DWC files
+ occ_df, images_df = read_DWC_file(cfg)
+
+ # Report summary
+ print(f"{bcolors.BOLD}Beginning of images file:{bcolors.ENDC}")
+ print(images_df.head())
+ print(f"{bcolors.BOLD}Beginning of occurrence file:{bcolors.ENDC}")
+ print(occ_df.head())
+
+ # Ignore problematic Herbaria
+ if cfg['ignore_banned_herb']:
+ for banned_url in cfg['banned_url_stems']:
+ images_df = images_df[~images_df['identifier'].str.contains(banned_url, na=False)]
+
+ # Report summary
+ n_imgs = images_df.shape[0]
+ n_occ = occ_df.shape[0]
+ print(f"{bcolors.BOLD}Number of images in images file: {n_imgs}{bcolors.ENDC}")
+ print(f"{bcolors.BOLD}Number of occurrence to search through: {n_occ}{bcolors.ENDC}")
+
+ results = process_image_batch(cfg, images_df, occ_df)
+
+def process_image_batch(cfg, images_df, occ_df):
+ futures_list = []
+ results = []
+
+ # single threaded, useful for debugging
+ # for index, image_row in images_df.iterrows():
+ # futures = process_each_image_row( cfg, image_row, occ_df)
+ # futures_list.append(futures)
+ # for future in futures_list:
+ # try:
+ # result = future.result(timeout=60)
+ # results.append(result)
+ # except Exception:
+ # results.append(None)
+ lock = Lock()
+
+ with th(max_workers=13) as executor:
+ for index, image_row in images_df.iterrows():
+ futures = executor.submit(process_each_image_row, cfg, image_row, occ_df, lock)
+ futures_list.append(futures)
+
+ for future in futures_list:
+ try:
+ result = future.result(timeout=60)
+ results.append(result)
+ except Exception:
+ results.append(None)
+ return results
+
+
+def process_image_batch_multiDirs(cfg, images_df, occ_df, dir_destination, shared_counter, n_imgs_per_species, do_shuffle_occurrences):
+ futures_list = []
+ results = []
+
+ lock = Lock()
+
+ if do_shuffle_occurrences:
+ images_df = images_df.sample(frac=1).reset_index(drop=True)
+
+ # Partition occ_df based on the first word of the 'specificEpithet' column
+ partition_dict = defaultdict(list)
+ for index, row in occ_df.iterrows():
+ first_word = row['specificEpithet'] # Assuming keep_first_word is defined
+ partition_dict[first_word].append(row)
+
+ # Convert lists to DataFrames
+ for key in partition_dict.keys():
+ partition_dict[key] = pd.DataFrame(partition_dict[key])
+
+ num_workers = 13
+
+ with th(max_workers=num_workers) as executor:
+ for specific_epithet, partition in partition_dict.items():
+ future = executor.submit(process_occ_chunk_multiDirs, cfg, images_df, partition, dir_destination, shared_counter, n_imgs_per_species, do_shuffle_occurrences, lock)
+ futures_list.append(future)
+
+ for future in futures_list:
+ try:
+ result = future.result(timeout=60)
+ results.append(result)
+ except Exception:
+ results.append(None)
+ return results
+
+def process_occ_chunk_multiDirs(cfg, images_df, occ_chunk, dir_destination, shared_counter, n_imgs_per_species, do_shuffle_occurrences, lock):
+ results = []
+ for index, occ_row in occ_chunk.iterrows():
+ result = process_each_occ_row_multiDirs(cfg, images_df, occ_row, dir_destination, shared_counter, n_imgs_per_species, do_shuffle_occurrences, lock)
+ results.append(result)
+ return results
+
+def process_each_occ_row_multiDirs(cfg, images_df, occ_row, dir_destination, shared_counter, n_imgs_per_species, do_shuffle_occurrences, lock):
+ print(f"{bcolors.BOLD}Working on occurrence: {occ_row['gbifID']}{bcolors.ENDC}")
+ gbif_id = occ_row['gbifID']
+
+ image_row = find_gbifID_in_images(gbif_id, images_df) # New function to find the image_row
+
+ if image_row is not None:
+ filename_image, filename_image_jpg, herb_code, specimen_id, family, genus, species, fullname = generate_image_filename(occ_row)
+
+ current_count = shared_counter.get_count(fullname)
+
+ # If the fullname is not in the counter yet, increment it
+ if current_count == 0:
+ shared_counter.increment(fullname)
+
+ print(shared_counter.get_count(fullname))
+ if shared_counter.get_count(fullname) > n_imgs_per_species:
+ print(f"Reached image limit for {fullname}. Skipping.")
+ return
+ else:
+
+ gbif_url = image_row['identifier']
+
+ image_candidate = ImageCandidateMulti(cfg, image_row, occ_row, gbif_url, dir_destination, lock)
+ if image_candidate.download_success:
+ shared_counter.increment(fullname)
+ else:
+ pass
+
+def find_gbifID_in_images(gbif_id, images_df):
+ image_row = images_df[images_df['gbifID'] == gbif_id]
+ if image_row.empty:
+ return None
+ return image_row.iloc[0]
+
+
+def process_each_image_row_multiDirs(cfg, image_row, occ_df, dir_destination, shared_counter, n_imgs_per_species, do_shuffle_occurrences, lock):
+ print(f"{bcolors.BOLD}Working on image: {image_row['gbifID']}{bcolors.ENDC}")
+ gbif_id = image_row['gbifID']
+ gbif_url = image_row['identifier']
+
+ occ_row = find_gbifID(gbif_id,occ_df)
+
+ if occ_row is not None:
+ filename_image, filename_image_jpg, herb_code, specimen_id, family, genus, species, fullname = generate_image_filename(occ_row)
+
+ current_count = shared_counter.get_count(fullname)
+
+ # If the fullname is not in the counter yet, increment it
+ if current_count == 0:
+ shared_counter.increment(fullname)
+
+ print(shared_counter.get_count(fullname))
+ if shared_counter.get_count(fullname) > n_imgs_per_species:
+ print(f"Reached image limit for {fullname}. Skipping.")
+ return
+
+ image_candidate = ImageCandidateMulti(cfg, image_row, occ_row, gbif_url, dir_destination, lock)
+ if image_candidate.download_success:
+ shared_counter.increment(fullname)
+ else:
+ pass
+
+
+def process_each_image_row(cfg, image_row, occ_df, lock):
+ print(f"{bcolors.BOLD}Working on image: {image_row['gbifID']}{bcolors.ENDC}")
+ gbif_id = image_row['gbifID']
+ gbif_url = image_row['identifier']
+
+ occ_row = find_gbifID(gbif_id,occ_df)
+
+ if occ_row is not None:
+ ImageInfo = ImageCandidate(cfg, image_row, occ_row, gbif_url, lock)
+ # ImageInfo.download_image(cfg, occ_row, image_row)
+ else:
+ pass
+
+def download_from_custom_file(cfg):
+ # Get DWC files
+ images_df = read_custom_file(cfg)
+
+ col_url = cfg['col_url']
+ col_name = cfg['col_name']
+ if col_url == None:
+ col_url = 'identifier'
+ else:
+ col_url = col_url
+
+ # Report summary
+ print(f"{bcolors.BOLD}Beginning of images file:{bcolors.ENDC}")
+ print(images_df.head())
+
+ # Ignore problematic Herbaria
+ if cfg['ignore_banned_herb']:
+ for banned_url in cfg['banned_url_stems']:
+ images_df = images_df[~images_df[col_url].str.contains(banned_url, na=False)]
+
+ # Report summary
+ n_imgs = images_df.shape[0]
+ print(f"{bcolors.BOLD}Number of images in images file: {n_imgs}{bcolors.ENDC}")
+
+ results = process_custom_image_batch(cfg, images_df)
+
+def read_custom_file(cfg):
+ dir_home = cfg['dir_home']
+ filename_img = cfg['filename_img']
+ # read the images.csv or occurences.csv file. can be txt ro csv
+ images_df = ingest_DWC(filename_img,dir_home)
+ return images_df
+
+# def ingest_DWC(DWC_csv_or_txt_file,dir_home):
+# if DWC_csv_or_txt_file.split('.')[1] == 'txt':
+# df = pd.read_csv(os.path.join(dir_home,DWC_csv_or_txt_file), sep="\t",header=0, low_memory=False, dtype=str)
+# elif DWC_csv_or_txt_file.split('.')[1] == 'csv':
+# df = pd.read_csv(os.path.join(dir_home,DWC_csv_or_txt_file), sep=",",header=0, low_memory=False, dtype=str)
+# else:
+# print(f"{bcolors.FAIL}DWC file {DWC_csv_or_txt_file} is not '.txt' or '.csv' and was not opened{bcolors.ENDC}")
+# return df
+
+def process_custom_image_batch(cfg, images_df):
+ futures_list = []
+ results = []
+
+ lock = Lock()
+
+ with th(max_workers=13) as executor:
+ for index, image_row in images_df.iterrows():
+ futures = executor.submit(process_each_custom_image_row, cfg, image_row, lock)
+ futures_list.append(futures)
+
+ for future in futures_list:
+ try:
+ result = future.result(timeout=60)
+ results.append(result)
+ except Exception:
+ results.append(None)
+ return results
+
+def process_each_custom_image_row(cfg, image_row, lock):
+ col_url = cfg['col_url']
+ col_name = cfg['col_name']
+
+ if col_url == None:
+ col_url = 'identifier'
+ else:
+ col_url = col_url
+
+ gbif_url = image_row[col_url]
+
+ print(f"{bcolors.BOLD}Working on image: {image_row[col_name]}{bcolors.ENDC}")
+ if image_row is not None:
+ ImageInfo = ImageCandidateCustom(cfg, image_row, gbif_url, col_name, lock)
+ else:
+ pass
\ No newline at end of file
diff --git a/vouchervision/utils_VoucherVision.py b/vouchervision/utils_VoucherVision.py
new file mode 100644
index 0000000000000000000000000000000000000000..a50396e0c789504aecaecd871043e94c6390a1d6
--- /dev/null
+++ b/vouchervision/utils_VoucherVision.py
@@ -0,0 +1,793 @@
+import openai
+import os, sys, json, inspect, glob, tiktoken, shutil, yaml
+import openpyxl
+from openpyxl import Workbook, load_workbook
+import google.generativeai as palm
+from langchain.chat_models import AzureChatOpenAI
+
+currentdir = os.path.dirname(os.path.abspath(
+ inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+parentdir = os.path.dirname(parentdir)
+sys.path.append(parentdir)
+
+from general_utils import get_cfg_from_full_path, num_tokens_from_string
+from embeddings_db import VoucherVisionEmbedding
+from OCR_google_cloud_vision import detect_text, overlay_boxes_on_image
+from LLM_chatGPT_3_5 import OCR_to_dict, OCR_to_dict_16k
+from LLM_PaLM import OCR_to_dict_PaLM
+# from LLM_Falcon import OCR_to_dict_Falcon
+from prompts import PROMPT_UMICH_skeleton_all_asia, PROMPT_OCR_Organized, PROMPT_UMICH_skeleton_all_asia_GPT4, PROMPT_OCR_Organized_GPT4, PROMPT_JSON
+from prompt_catalog import PromptCatalog
+'''
+* For the prefix_removal, the image names have 'MICH-V-' prior to the barcode, so that is used for matching
+ but removed for output.
+* There is also code active to replace the LLM-predicted "Catalog Number" with the correct number since it is known.
+ The LLMs to usually assign the barcode to the correct field, but it's not needed since it is already known.
+ - Look for ####################### Catalog Number pre-defined
+'''
+
+'''
+Prior to StructuredOutputParser:
+ response = openai.ChatCompletion.create(
+ model=MODEL,
+ temperature = 0,
+ messages=[
+ {"role": "system", "content": "You are a helpful assistant acting as a transcription expert and your job is to transcribe herbarium specimen labels based on OCR data and reformat it to meet Darwin Core Archive Standards into a Python dictionary based on certain rules."},
+ {"role": "user", "content": prompt},
+ ],
+ max_tokens=2048,
+ )
+ # print the model's response
+ return response.choices[0].message['content']
+'''
+
+class VoucherVision():
+
+ def __init__(self, cfg, logger, dir_home, path_custom_prompts, Project, Dirs):
+ self.cfg = cfg
+ self.logger = logger
+ self.dir_home = dir_home
+ self.path_custom_prompts = path_custom_prompts
+ self.Project = Project
+ self.Dirs = Dirs
+ self.headers = None
+ self.prompt_version = None
+
+ self.set_API_keys()
+ self.setup()
+
+
+ def setup(self):
+ self.logger.name = f'[Transcription]'
+ self.logger.info(f'Setting up OCR and LLM')
+
+ self.db_name = self.cfg['leafmachine']['project']['embeddings_database_name']
+ self.path_domain_knowledge = self.cfg['leafmachine']['project']['path_to_domain_knowledge_xlsx']
+ self.build_new_db = self.cfg['leafmachine']['project']['build_new_embeddings_database']
+
+ self.continue_run_from_partial_xlsx = self.cfg['leafmachine']['project']['continue_run_from_partial_xlsx']
+
+ self.prefix_removal = self.cfg['leafmachine']['project']['prefix_removal']
+ self.suffix_removal = self.cfg['leafmachine']['project']['suffix_removal']
+ self.catalog_numerical_only = self.cfg['leafmachine']['project']['catalog_numerical_only']
+
+ self.prompt_version0 = self.cfg['leafmachine']['project']['prompt_version']
+ self.use_domain_knowledge = self.cfg['leafmachine']['project']['use_domain_knowledge']
+
+ self.catalog_name_options = ["Catalog Number", "catalog_number"]
+
+ self.utility_headers = ["tokens_in", "tokens_out", "path_to_crop","path_to_original","path_to_content","path_to_helper",]
+
+ self.map_prompt_versions()
+ self.map_dir_labels()
+ self.map_API_options()
+ self.init_embeddings()
+ self.init_transcription_xlsx()
+
+ '''Logging'''
+ self.logger.info(f'Transcribing dataset --- {self.dir_labels}')
+ self.logger.info(f'Saving transcription batch to --- {self.path_transcription}')
+ self.logger.info(f'Saving individual transcription files to --- {self.Dirs.transcription_ind}')
+ self.logger.info(f'Starting transcription...')
+ self.logger.info(f' LLM MODEL --> {self.version_name}')
+ self.logger.info(f' Using Azure API --> {self.is_azure}')
+ self.logger.info(f' Model name passed to API --> {self.model_name}')
+ self.logger.info(f' API access token is found in PRIVATE_DATA.yaml --> {self.has_key}')
+
+ def map_API_options(self):
+ self.chat_version = self.cfg['leafmachine']['LLM_version']
+ version_mapping = {
+ 'GPT 4': ('OpenAI GPT 4', False, 'GPT_4', self.has_key_openai),
+ 'GPT 3.5': ('OpenAI GPT 3.5', False, 'GPT_3_5', self.has_key_openai),
+ 'Azure GPT 3.5': ('(Azure) OpenAI GPT 3.5', True, 'Azure_GPT_3_5', self.has_key_azure_openai),
+ 'Azure GPT 4': ('(Azure) OpenAI GPT 4', True, 'Azure_GPT_4', self.has_key_azure_openai),
+ 'PaLM 2': ('Google PaLM 2', None, None, self.has_key_palm2)
+ }
+ if self.chat_version not in version_mapping:
+ supported_LLMs = ", ".join(version_mapping.keys())
+ raise Exception(f"Unsupported LLM: {self.chat_version}. Requires one of: {supported_LLMs}")
+
+ self.version_name, self.is_azure, self.model_name, self.has_key = version_mapping[self.chat_version]
+
+ def map_prompt_versions(self):
+ self.prompt_version_map = {
+ "Version 1": "prompt_v1_verbose",
+ "Version 1 No Domain Knowledge": "prompt_v1_verbose_noDomainKnowledge",
+ "Version 2": "prompt_v2_json_rules",
+ "Version 1 PaLM 2": 'prompt_v1_palm2',
+ "Version 1 PaLM 2 No Domain Knowledge": 'prompt_v1_palm2_noDomainKnowledge',
+ "Version 2 PaLM 2": 'prompt_v2_palm2',
+ }
+ self.prompt_version = self.prompt_version_map.get(self.prompt_version0, self.path_custom_prompts)
+ self.is_predefined_prompt = self.is_in_prompt_version_map(self.prompt_version)
+
+ def is_in_prompt_version_map(self, value):
+ return value in self.prompt_version_map.values()
+
+ def init_embeddings(self):
+ if self.use_domain_knowledge:
+ self.logger.info(f'*** USING DOMAIN KNOWLEDGE ***')
+ self.logger.info(f'*** Initializing vector embeddings database ***')
+ self.initialize_embeddings()
+ else:
+ self.Voucher_Vision_Embedding = None
+
+ def map_dir_labels(self):
+ if self.cfg['leafmachine']['use_RGB_label_images']:
+ self.dir_labels = os.path.join(self.Dirs.save_per_annotation_class,'label')
+ else:
+ self.dir_labels = self.Dirs.save_original
+
+ # Use glob to get all image paths in the directory
+ self.img_paths = glob.glob(os.path.join(self.dir_labels, "*"))
+
+ def load_rules_config(self):
+ with open(self.path_custom_prompts, 'r') as stream:
+ try:
+ return yaml.safe_load(stream)
+ except yaml.YAMLError as exc:
+ print(exc)
+ return None
+
+ def generate_xlsx_headers(self):
+ # Extract headers from the 'Dictionary' keys in the JSON template rules
+ xlsx_headers = list(self.rules_config_json['rules']["Dictionary"].keys())
+ xlsx_headers = xlsx_headers + self.utility_headers
+ return xlsx_headers
+
+ def init_transcription_xlsx(self):
+ self.HEADERS_v1_n22 = ["Catalog Number","Genus","Species","subspecies","variety","forma","Country","State","County","Locality Name","Min Elevation","Max Elevation","Elevation Units","Verbatim Coordinates","Datum","Cultivated","Habitat","Collectors","Collector Number","Verbatim Date","Date","End Date"]
+ self.HEADERS_v2_n26 = ["catalog_number","genus","species","subspecies","variety","forma","country","state","county","locality_name","min_elevation","max_elevation","elevation_units","verbatim_coordinates","decimal_coordinates","datum","cultivated","habitat","plant_description","collectors","collector_number","determined_by","multiple_names","verbatim_date","date","end_date"]
+ self.HEADERS_v1_n22 = self.HEADERS_v1_n22 + self.utility_headers
+ self.HEADERS_v2_n26 = self.HEADERS_v2_n26 + self.utility_headers
+ # Initialize output file
+ self.path_transcription = os.path.join(self.Dirs.transcription,"transcribed.xlsx")
+
+ if self.prompt_version in ['prompt_v2_json_rules','prompt_v2_palm2']:
+ self.headers = self.HEADERS_v2_n26
+ self.headers_used = 'HEADERS_v2_n26'
+
+ elif self.prompt_version in ['prompt_v1_verbose', 'prompt_v1_verbose_noDomainKnowledge','prompt_v1_palm2', 'prompt_v1_palm2_noDomainKnowledge']:
+ self.headers = self.HEADERS_v1_n22
+ self.headers_used = 'HEADERS_v1_n22'
+
+ else:
+ if not self.is_predefined_prompt:
+ # Load the rules configuration
+ self.rules_config_json = self.load_rules_config()
+ # Generate the headers from the configuration
+ self.headers = self.generate_xlsx_headers()
+ # Set the headers used to the dynamically generated headers
+ self.headers_used = 'CUSTOM'
+ else:
+ # If it's a predefined prompt, raise an exception as we don't have further instructions
+ raise ValueError("Predefined prompt is not handled in this context.")
+
+ self.create_or_load_excel_with_headers(os.path.join(self.Dirs.transcription,"transcribed.xlsx"), self.headers)
+
+
+ def pick_model(self, vendor, nt):
+ if vendor == 'GPT_3_5':
+ if nt > 6000:
+ return "gpt-3.5-turbo-16k-0613", True
+ else:
+ return "gpt-3.5-turbo", False
+ if vendor == 'GPT_4':
+ return "gpt-4", False
+ if vendor == 'Azure_GPT_3_5':
+ return "gpt-35-turbo", False
+ if vendor == 'Azure_GPT_4':
+ return "gpt-4", False
+
+ def create_or_load_excel_with_headers(self, file_path, headers, show_head=False):
+ output_dir_names = ['Archival_Components', 'Config_File', 'Cropped_Images', 'Logs', 'Original_Images', 'Transcription']
+ self.completed_specimens = []
+
+ # Check if the file exists and it's not None
+ if self.continue_run_from_partial_xlsx is not None and os.path.isfile(self.continue_run_from_partial_xlsx):
+ workbook = load_workbook(filename=self.continue_run_from_partial_xlsx)
+ sheet = workbook.active
+ show_head=True
+ # Identify the 'path_to_crop' column
+ try:
+ path_to_crop_col = headers.index('path_to_crop') + 1
+ path_to_original_col = headers.index('path_to_original') + 1
+ path_to_content_col = headers.index('path_to_content') + 1
+ path_to_helper_col = headers.index('path_to_helper') + 1
+ # self.completed_specimens = list(sheet.iter_cols(min_col=path_to_crop_col, max_col=path_to_crop_col, values_only=True, min_row=2))
+ except ValueError:
+ print("'path_to_crop' not found in the header row.")
+
+
+ path_to_crop = list(sheet.iter_cols(min_col=path_to_crop_col, max_col=path_to_crop_col, values_only=True, min_row=2))
+ path_to_original = list(sheet.iter_cols(min_col=path_to_original_col, max_col=path_to_original_col, values_only=True, min_row=2))
+ path_to_content = list(sheet.iter_cols(min_col=path_to_content_col, max_col=path_to_content_col, values_only=True, min_row=2))
+ path_to_helper = list(sheet.iter_cols(min_col=path_to_helper_col, max_col=path_to_helper_col, values_only=True, min_row=2))
+ others = [path_to_crop_col, path_to_original_col, path_to_content_col, path_to_helper_col]
+ jsons = [path_to_content_col, path_to_helper_col]
+
+ for cell in path_to_crop[0]:
+ old_path = cell
+ new_path = file_path
+ for dir_name in output_dir_names:
+ if dir_name in old_path:
+ old_path_parts = old_path.split(dir_name)
+ new_path_parts = new_path.split('Transcription')
+ updated_path = new_path_parts[0] + dir_name + old_path_parts[1]
+ self.completed_specimens.append(os.path.basename(updated_path))
+ print(f"{len(self.completed_specimens)} images are already completed")
+
+ ### Copy the JSON files over
+ for colu in jsons:
+ cell = next(sheet.iter_rows(min_row=2, min_col=colu, max_col=colu))[0]
+ old_path = cell.value
+ new_path = file_path
+
+ old_path_parts = old_path.split('Transcription')
+ new_path_parts = new_path.split('Transcription')
+ updated_path = new_path_parts[0] + 'Transcription' + old_path_parts[1]
+
+ # Copy files
+ old_dir = os.path.dirname(old_path)
+ new_dir = os.path.dirname(updated_path)
+
+ # Check if old_dir exists and it's a directory
+ if os.path.exists(old_dir) and os.path.isdir(old_dir):
+ # Check if new_dir exists. If not, create it.
+ if not os.path.exists(new_dir):
+ os.makedirs(new_dir)
+
+ # Iterate through all files in old_dir and copy each to new_dir
+ for filename in os.listdir(old_dir):
+ shutil.copy2(os.path.join(old_dir, filename), new_dir) # copy2 preserves metadata
+
+ ### Update the file names
+ for colu in others:
+ for row in sheet.iter_rows(min_row=2, min_col=colu, max_col=colu):
+ for cell in row:
+ old_path = cell.value
+ new_path = file_path
+ for dir_name in output_dir_names:
+ if dir_name in old_path:
+ old_path_parts = old_path.split(dir_name)
+ new_path_parts = new_path.split('Transcription')
+ updated_path = new_path_parts[0] + dir_name + old_path_parts[1]
+ cell.value = updated_path
+ show_head=True
+
+
+ else:
+ # Create a new workbook and select the active worksheet
+ workbook = Workbook()
+ sheet = workbook.active
+
+ # Write headers in the first row
+ for i, header in enumerate(headers, start=1):
+ sheet.cell(row=1, column=i, value=header)
+ self.completed_specimens = []
+
+ # Save the workbook
+ workbook.save(file_path)
+
+ if show_head:
+ print("continue_run_from_partial_xlsx:")
+ for i, row in enumerate(sheet.iter_rows(values_only=True)):
+ print(row)
+ if i == 3: # print the first 5 rows (0-indexed)
+ print("\n")
+ break
+
+
+
+ def add_data_to_excel_from_response(self, path_transcription, response, filename_without_extension, path_to_crop, path_to_content, path_to_helper, nt_in, nt_out):
+ wb = openpyxl.load_workbook(path_transcription)
+ sheet = wb.active
+
+ # find the next empty row
+ next_row = sheet.max_row + 1
+
+ if isinstance(response, str):
+ try:
+ response = json.loads(response)
+ except json.JSONDecodeError:
+ print(f"Failed to parse response: {response}")
+ return
+
+ # iterate over headers in the first row
+ for i, header in enumerate(sheet[1], start=1):
+ # check if header value is in response keys
+ if (header.value in response) and (header.value not in self.catalog_name_options): ####################### Catalog Number pre-defined
+ # check if the response value is a dictionary
+ if isinstance(response[header.value], dict):
+ # if it is a dictionary, extract the 'value' field
+ cell_value = response[header.value].get('value', '')
+ else:
+ # if it's not a dictionary, use it directly
+ cell_value = response[header.value]
+
+ try:
+ # write the value to the cell
+ sheet.cell(row=next_row, column=i, value=cell_value)
+ except:
+ sheet.cell(row=next_row, column=i, value=cell_value[0])
+
+ elif header.value in self.catalog_name_options:
+ # if self.prefix_removal:
+ # filename_without_extension = filename_without_extension.replace(self.prefix_removal, "")
+ # if self.suffix_removal:
+ # filename_without_extension = filename_without_extension.replace(self.suffix_removal, "")
+ # if self.catalog_numerical_only:
+ # filename_without_extension = self.remove_non_numbers(filename_without_extension)
+ sheet.cell(row=next_row, column=i, value=filename_without_extension)
+ elif header.value == "path_to_crop":
+ sheet.cell(row=next_row, column=i, value=path_to_crop)
+ elif header.value == "path_to_original":
+ if self.cfg['leafmachine']['use_RGB_label_images']:
+ fname = os.path.basename(path_to_crop)
+ base = os.path.dirname(os.path.dirname(os.path.dirname(os.path.dirname(path_to_crop))))
+ path_to_original = os.path.join(base, 'Original_Images', fname)
+ sheet.cell(row=next_row, column=i, value=path_to_original)
+ else:
+ fname = os.path.basename(path_to_crop)
+ base = os.path.dirname(os.path.dirname(path_to_crop))
+ path_to_original = os.path.join(base, 'Original_Images', fname)
+ sheet.cell(row=next_row, column=i, value=path_to_original)
+ elif header.value == "path_to_content":
+ sheet.cell(row=next_row, column=i, value=path_to_content)
+ elif header.value == "path_to_helper":
+ sheet.cell(row=next_row, column=i, value=path_to_helper)
+ elif header.value == "tokens_in":
+ sheet.cell(row=next_row, column=i, value=nt_in)
+ elif header.value == "tokens_out":
+ sheet.cell(row=next_row, column=i, value=nt_out)
+ # save the workbook
+ wb.save(path_transcription)
+
+
+
+
+ def has_API_key(self, val):
+ if val != '':
+ return True
+ else:
+ return False
+
+ def set_API_keys(self):
+ self.dir_home = os.path.dirname(os.path.dirname(__file__))
+ self.path_cfg_private = os.path.join(self.dir_home, 'PRIVATE_DATA.yaml')
+ self.cfg_private = get_cfg_from_full_path(self.path_cfg_private)
+
+ self.has_key_openai = self.has_API_key(self.cfg_private['openai']['OPENAI_API_KEY'])
+
+ self.has_key_azure_openai = self.has_API_key(self.cfg_private['openai_azure']['api_version'])
+
+ self.has_key_palm2 = self.has_API_key(self.cfg_private['google_palm']['google_palm_api'])
+
+ self.has_key_google_OCR = self.has_API_key(self.cfg_private['google_cloud']['path_json_file'])
+
+ if self.has_key_openai:
+ openai.api_key = self.cfg_private['openai']['OPENAI_API_KEY']
+ os.environ["OPENAI_API_KEY"] = self.cfg_private['openai']['OPENAI_API_KEY']
+
+
+ if self.has_key_azure_openai:
+ # os.environ["OPENAI_API_KEY"] = self.cfg_private['openai_azure']['openai_api_key']
+ self.llm = AzureChatOpenAI(
+ deployment_name='gpt-35-turbo',
+ openai_api_version=self.cfg_private['openai_azure']['api_version'],
+ openai_api_key=self.cfg_private['openai_azure']['openai_api_key'],
+ openai_api_base=self.cfg_private['openai_azure']['openai_api_base'],
+ openai_organization=self.cfg_private['openai_azure']['openai_organization'],
+ openai_api_type=self.cfg_private['openai_azure']['openai_api_type']
+ )
+
+ if self.has_key_google_OCR:
+ os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = self.cfg_private['google_cloud']['path_json_file']
+
+ if self.has_key_palm2:
+ os.environ['PALM'] = self.cfg_private['google_palm']['google_palm_api']
+ palm.configure(api_key=os.environ['PALM'])
+
+
+ def initialize_embeddings(self):
+ '''Loading embedding search __init__(self, db_name, path_domain_knowledge, logger, build_new_db=False, model_name="hkunlp/instructor-xl", device="cuda")'''
+ self.Voucher_Vision_Embedding = VoucherVisionEmbedding(self.db_name, self.path_domain_knowledge, logger=self.logger, build_new_db=self.build_new_db)
+
+ def clean_catalog_number(self, data, filename_without_extension):
+ #Cleans up the catalog number in data if it's a dict
+
+ def modify_catalog_key(catalog_key, filename_without_extension, data):
+ # Helper function to apply modifications on catalog number
+ if catalog_key not in data:
+ new_data = {catalog_key: None}
+ data = {**new_data, **data}
+
+ if self.prefix_removal:
+ filename_without_extension = filename_without_extension.replace(self.prefix_removal, "")
+ if self.suffix_removal:
+ filename_without_extension = filename_without_extension.replace(self.suffix_removal, "")
+ if self.catalog_numerical_only:
+ filename_without_extension = self.remove_non_numbers(data[catalog_key])
+ data[catalog_key] = filename_without_extension
+ return data
+
+ if isinstance(data, dict):
+ if self.headers_used == 'HEADERS_v1_n22':
+ return modify_catalog_key("Catalog Number", filename_without_extension, data)
+ elif self.headers_used in ['HEADERS_v2_n26', 'CUSTOM']:
+ return modify_catalog_key("catalog_number", filename_without_extension, data)
+ else:
+ raise ValueError("Invalid headers used.")
+ else:
+ raise TypeError("Data is not of type dict.")
+
+
+ def write_json_to_file(self, filepath, data):
+ '''Writes dictionary data to a JSON file.'''
+ with open(filepath, 'w') as txt_file:
+ if isinstance(data, dict):
+ data = json.dumps(data, indent=4)
+ txt_file.write(data)
+
+ def create_null_json(self):
+ return {}
+
+ def remove_non_numbers(self, s):
+ return ''.join([char for char in s if char.isdigit()])
+
+ def create_null_row(self, filename_without_extension, path_to_crop, path_to_content, path_to_helper):
+ json_dict = {header: '' for header in self.headers}
+ for header, value in json_dict.items():
+ if header in self.catalog_name_options:
+ if self.prefix_removal:
+ json_dict[header] = filename_without_extension.replace(self.prefix_removal, "")
+ if self.suffix_removal:
+ json_dict[header] = filename_without_extension.replace(self.suffix_removal, "")
+ if self.catalog_numerical_only:
+ json_dict[header] = self.remove_non_numbers(json_dict[header])
+ elif header == "path_to_crop":
+ json_dict[header] = path_to_crop
+ elif header == "path_to_original":
+ fname = os.path.basename(path_to_crop)
+ base = os.path.dirname(os.path.dirname(os.path.dirname(os.path.dirname(path_to_crop))))
+ path_to_original = os.path.join(base, 'Original_Images', fname)
+ json_dict[header] = path_to_original
+ elif header == "path_to_content":
+ json_dict[header] = path_to_content
+ elif header == "path_to_helper":
+ json_dict[header] = path_to_helper
+ return json_dict
+
+
+ def setup_GPT(self, prompt_version, gpt):
+ Catalog = PromptCatalog()
+ self.logger.info(f'Length of OCR raw -- {len(self.OCR)}')
+
+ # if prompt_version == 'prompt_v1_verbose':
+ if self.is_predefined_prompt:
+ if self.use_domain_knowledge:
+ # Find a similar example from the domain knowledge
+ domain_knowledge_example = self.Voucher_Vision_Embedding.query_db(self.OCR, 1)
+ similarity= self.Voucher_Vision_Embedding.get_similarity()
+
+ if prompt_version == 'prompt_v1_verbose':
+ prompt, n_fields, xlsx_headers = Catalog.prompt_v1_verbose(OCR=self.OCR,domain_knowledge_example=domain_knowledge_example,similarity=similarity)
+
+ else:
+ if prompt_version == 'prompt_v1_verbose_noDomainKnowledge':
+ prompt, n_fields, xlsx_headers = Catalog.prompt_v1_verbose_noDomainKnowledge(OCR=self.OCR)
+
+ elif prompt_version == 'prompt_v2_json_rules':
+ prompt, n_fields, xlsx_headers = Catalog.prompt_v2_json_rules(OCR=self.OCR)
+ else:
+ prompt, n_fields, xlsx_headers = Catalog.prompt_v2_custom(self.path_custom_prompts, OCR=self.OCR)
+
+
+
+ nt = num_tokens_from_string(prompt, "cl100k_base")
+ self.logger.info(f'Prompt token length --- {nt}')
+
+ MODEL, use_long_form = self.pick_model(gpt, nt)
+ self.logger.info(f'Waiting for {gpt} API call --- Using {MODEL}')
+
+ return MODEL, prompt, use_long_form, n_fields, xlsx_headers, nt
+
+
+ # def setup_GPT(self, opt, gpt):
+ # if opt == 'dict':
+ # # Find a similar example from the domain knowledge
+ # domain_knowledge_example = self.Voucher_Vision_Embedding.query_db(self.OCR, 1)
+ # similarity= self.Voucher_Vision_Embedding.get_similarity()
+
+ # self.logger.info(f'Length of OCR raw -- {len(self.OCR)}')
+
+ # # prompt = PROMPT_UMICH_skeleton_all_asia_GPT4(self.OCR, domain_knowledge_example, similarity)
+ # prompt, n_fields, xlsx_headers =
+
+ # nt = num_tokens_from_string(prompt, "cl100k_base")
+ # self.logger.info(f'Prompt token length --- {nt}')
+
+ # MODEL, use_long_form = self.pick_model(gpt, nt)
+
+ # ### Direct GPT ###
+ # self.logger.info(f'Waiting for {MODEL} API call --- Using chatGPT --- Content')
+
+ # return MODEL, prompt, use_long_form
+
+ # elif opt == 'helper':
+ # prompt = PROMPT_OCR_Organized_GPT4(self.OCR)
+ # nt = num_tokens_from_string(prompt, "cl100k_base")
+
+ # MODEL, use_long_form = self.pick_model(gpt, nt)
+
+ # self.logger.info(f'Length of OCR raw -- {len(self.OCR)}')
+ # self.logger.info(f'Prompt token length --- {nt}')
+ # self.logger.info(f'Waiting for {MODEL} API call --- Using chatGPT --- Helper')
+
+ # return MODEL, prompt, use_long_form
+
+
+ def use_chatGPT(self, is_azure, progress_report, gpt):
+ total_tokens_in = 0
+ total_tokens_out = 0
+ final_JSON_response = None
+ if progress_report is not None:
+ progress_report.set_n_batches(len(self.img_paths))
+ for i, path_to_crop in enumerate(self.img_paths):
+ if progress_report is not None:
+ progress_report.update_batch(f"Working on image {i+1} of {len(self.img_paths)}")
+
+ if os.path.basename(path_to_crop) in self.completed_specimens:
+ self.logger.info(f'[Skipping] specimen {os.path.basename(path_to_crop)} already processed')
+ else:
+ filename_without_extension, txt_file_path, txt_file_path_OCR, txt_file_path_OCR_bounds, jpg_file_path_OCR_helper = self.generate_paths(path_to_crop, i)
+
+ # Use Google Vision API to get OCR
+ # self.OCR = detect_text(path_to_crop)
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Starting OCR')
+ self.OCR, self.bounds, self.text_to_box_mapping = detect_text(path_to_crop)
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Finished OCR')
+ if len(self.OCR) > 0:
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Creating OCR Overlay Image')
+ self.overlay_image = overlay_boxes_on_image(path_to_crop, self.bounds)
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Saved OCR Overlay Image')
+
+ self.write_json_to_file(txt_file_path_OCR, {"OCR":self.OCR})
+ self.write_json_to_file(txt_file_path_OCR_bounds, {"OCR_Bounds":self.bounds})
+ self.overlay_image.save(jpg_file_path_OCR_helper)
+
+ # Setup Dict
+ MODEL, prompt, use_long_form, n_fields, xlsx_headers, nt_in = self.setup_GPT(self.prompt_version, gpt)
+
+ if is_azure:
+ self.llm.deployment_name = MODEL
+ else:
+ self.llm = None
+
+ # Send OCR to chatGPT and return formatted dictonary
+ if use_long_form:
+ response_candidate = OCR_to_dict_16k(is_azure, self.logger, MODEL, prompt, self.llm, self.prompt_version)
+ nt_out = num_tokens_from_string(response_candidate, "cl100k_base")
+ else:
+ response_candidate = OCR_to_dict(is_azure, self.logger, MODEL, prompt, self.llm, self.prompt_version)
+ nt_out = num_tokens_from_string(response_candidate, "cl100k_base")
+ else:
+ response_candidate = None
+ nt_out = 0
+
+ total_tokens_in += nt_in
+ total_tokens_out += nt_out
+
+ final_JSON_response0 = self.save_json_and_xlsx(response_candidate, filename_without_extension, path_to_crop, txt_file_path, jpg_file_path_OCR_helper, nt_in, nt_out)
+ if response_candidate is not None:
+ final_JSON_response = final_JSON_response0
+
+ self.logger.info(f'Formatted JSON\n{final_JSON_response}')
+ self.logger.info(f'Finished {MODEL} API calls\n')
+
+ if progress_report is not None:
+ progress_report.reset_batch(f"Batch Complete")
+ try:
+ final_JSON_response = json.loads(final_JSON_response.strip('```').replace('json\n', '', 1).replace('json', '', 1))
+ except:
+ pass
+ return final_JSON_response, total_tokens_in, total_tokens_out
+
+
+
+ def use_PaLM(self, progress_report):
+ total_tokens_in = 0
+ total_tokens_out = 0
+ final_JSON_response = None
+ if progress_report is not None:
+ progress_report.set_n_batches(len(self.img_paths))
+ for i, path_to_crop in enumerate(self.img_paths):
+ if progress_report is not None:
+ progress_report.update_batch(f"Working on image {i+1} of {len(self.img_paths)}")
+ if os.path.basename(path_to_crop) in self.completed_specimens:
+ self.logger.info(f'[Skipping] specimen {os.path.basename(path_to_crop)} already processed')
+ else:
+ filename_without_extension, txt_file_path, txt_file_path_OCR, txt_file_path_OCR_bounds, jpg_file_path_OCR_helper = self.generate_paths(path_to_crop, i)
+
+ # Use Google Vision API to get OCR
+ self.OCR, self.bounds, self.text_to_box_mapping = detect_text(path_to_crop)
+ if len(self.OCR) > 0:
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Starting OCR')
+ self.OCR = self.OCR.replace("\'", "Minutes").replace('\"', "Seconds")
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Finished OCR')
+
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Creating OCR Overlay Image')
+ self.overlay_image = overlay_boxes_on_image(path_to_crop, self.bounds)
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- Saved OCR Overlay Image')
+
+ self.write_json_to_file(txt_file_path_OCR, {"OCR":self.OCR})
+ self.write_json_to_file(txt_file_path_OCR_bounds, {"OCR_Bounds":self.bounds})
+ self.overlay_image.save(jpg_file_path_OCR_helper)
+
+ # Send OCR to chatGPT and return formatted dictonary
+ response_candidate, nt_in = OCR_to_dict_PaLM(self.logger, self.OCR, self.prompt_version, self.Voucher_Vision_Embedding)
+ nt_out = num_tokens_from_string(response_candidate, "cl100k_base")
+
+ else:
+ response_candidate = None
+ nt_out = 0
+
+ total_tokens_in += nt_in
+ total_tokens_out += nt_out
+
+ final_JSON_response0 = self.save_json_and_xlsx(response_candidate, filename_without_extension, path_to_crop, txt_file_path, jpg_file_path_OCR_helper, nt_in, nt_out)
+ if response_candidate is not None:
+ final_JSON_response = final_JSON_response0
+ self.logger.info(f'Formatted JSON\n{final_JSON_response}')
+ self.logger.info(f'Finished PaLM 2 API calls\n')
+
+ if progress_report is not None:
+ progress_report.reset_batch(f"Batch Complete")
+ return final_JSON_response, total_tokens_in, total_tokens_out
+
+
+ '''
+ def use_falcon(self, progress_report):
+ for i, path_to_crop in enumerate(self.img_paths):
+ progress_report.update_batch(f"Working on image {i+1} of {len(self.img_paths)}")
+ if os.path.basename(path_to_crop) in self.completed_specimens:
+ self.logger.info(f'[Skipping] specimen {os.path.basename(path_to_crop)} already processed')
+ else:
+ filename_without_extension = os.path.splitext(os.path.basename(path_to_crop))[0]
+ txt_file_path = os.path.join(self.Dirs.transcription_ind, filename_without_extension + '.json')
+ txt_file_path_helper = os.path.join(self.Dirs.transcription_ind_helper, filename_without_extension + '.json')
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- {filename_without_extension}')
+
+ # Use Google Vision API to get OCR
+ self.OCR, self.bounds, self.text_to_box_mapping = detect_text(path_to_crop)
+ if len(self.OCR) > 0:
+ self.OCR = self.OCR.replace("\'", "Minutes").replace('\"', "Seconds")
+
+ # Send OCR to Falcon and return formatted dictionary
+ response = OCR_to_dict_Falcon(self.logger, self.OCR, self.Voucher_Vision_Embedding)
+ # response_helper = OCR_to_helper_Falcon(self.logger, OCR) # Assuming you have a similar helper function for Falcon
+ response_helper = None
+
+ self.logger.info(f'Finished Falcon API calls\n')
+ else:
+ response = None
+
+ if (response is not None) and (response_helper is not None):
+ # Save transcriptions to json files
+ self.write_json_to_file(txt_file_path, response)
+ # self.write_json_to_file(txt_file_path_helper, response_helper)
+
+ # add to the xlsx file
+ self.add_data_to_excel_from_response(self.path_transcription, response, filename_without_extension, path_to_crop, txt_file_path, txt_file_path_helper)
+ progress_report.reset_batch()
+ '''
+
+ def generate_paths(self, path_to_crop, i):
+ filename_without_extension = os.path.splitext(os.path.basename(path_to_crop))[0]
+ txt_file_path = os.path.join(self.Dirs.transcription_ind, filename_without_extension + '.json')
+ txt_file_path_OCR = os.path.join(self.Dirs.transcription_ind_OCR, filename_without_extension + '.json')
+ txt_file_path_OCR_bounds = os.path.join(self.Dirs.transcription_ind_OCR_bounds, filename_without_extension + '.json')
+ jpg_file_path_OCR_helper = os.path.join(self.Dirs.transcription_ind_OCR_helper, filename_without_extension + '.jpg')
+
+ self.logger.info(f'Working on {i+1}/{len(self.img_paths)} --- {filename_without_extension}')
+
+ return filename_without_extension, txt_file_path, txt_file_path_OCR, txt_file_path_OCR_bounds, jpg_file_path_OCR_helper
+
+ def save_json_and_xlsx(self, response, filename_without_extension, path_to_crop, txt_file_path, jpg_file_path_OCR_helper, nt_in, nt_out):
+ if response is None:
+ response = self.create_null_json()
+ self.write_json_to_file(txt_file_path, response)
+
+ # Then add the null info to the spreadsheet
+ response_null = self.create_null_row(filename_without_extension, path_to_crop, txt_file_path, jpg_file_path_OCR_helper)
+ self.add_data_to_excel_from_response(self.path_transcription, response_null, filename_without_extension, path_to_crop, txt_file_path, jpg_file_path_OCR_helper, nt_in=0, nt_out=0)
+
+ ### Set completed JSON
+ else:
+ response = self.clean_catalog_number(response, filename_without_extension)
+ self.write_json_to_file(txt_file_path, response)
+ # add to the xlsx file
+ self.add_data_to_excel_from_response(self.path_transcription, response, filename_without_extension, path_to_crop, txt_file_path, jpg_file_path_OCR_helper, nt_in, nt_out)
+ return response
+
+ def process_specimen_batch(self, progress_report):
+ try:
+ if self.has_key:
+ if self.model_name:
+ final_json_response, total_tokens_in, total_tokens_out = self.use_chatGPT(self.is_azure, progress_report, self.model_name)
+ else:
+ final_json_response, total_tokens_in, total_tokens_out = self.use_PaLM(progress_report)
+ return final_json_response, total_tokens_in, total_tokens_out
+ else:
+ self.logger.info(f'No API key found for {self.version_name}')
+ raise Exception(f"No API key found for {self.version_name}")
+ except:
+ if progress_report is not None:
+ progress_report.reset_batch(f"Batch Failed")
+ self.logger.error("LLM call failed. Ending batch. process_specimen_batch()")
+ for handler in self.logger.handlers[:]:
+ handler.close()
+ self.logger.removeHandler(handler)
+ raise
+
+ def process_specimen_batch_OCR_test(self, path_to_crop):
+ for img_filename in os.listdir(path_to_crop):
+ img_path = os.path.join(path_to_crop, img_filename)
+ self.OCR, self.bounds, self.text_to_box_mapping = detect_text(img_path)
+
+
+
+def space_saver(cfg, Dirs, logger):
+ dir_out = cfg['leafmachine']['project']['dir_output']
+ run_name = Dirs.run_name
+
+ path_project = os.path.join(dir_out, run_name)
+
+ if cfg['leafmachine']['project']['delete_temps_keep_VVE']:
+ logger.name = '[DELETE TEMP FILES]'
+ logger.info("Deleting temporary files. Keeping files required for VoucherVisionEditor.")
+ delete_dirs = ['Archival_Components', 'Config_File']
+ for d in delete_dirs:
+ path_delete = os.path.join(path_project, d)
+ if os.path.exists(path_delete):
+ shutil.rmtree(path_delete)
+
+ elif cfg['leafmachine']['project']['delete_all_temps']:
+ logger.name = '[DELETE TEMP FILES]'
+ logger.info("Deleting ALL temporary files!")
+ delete_dirs = ['Archival_Components', 'Config_File', 'Original_Images', 'Cropped_Images']
+ for d in delete_dirs:
+ path_delete = os.path.join(path_project, d)
+ if os.path.exists(path_delete):
+ shutil.rmtree(path_delete)
+
+ # Delete the transctiption folder, but keep the xlsx
+ transcription_path = os.path.join(path_project, 'Transcription')
+ if os.path.exists(transcription_path):
+ for item in os.listdir(transcription_path):
+ item_path = os.path.join(transcription_path, item)
+ if os.path.isdir(item_path): # if the item is a directory
+ if os.path.exists(item_path):
+ shutil.rmtree(item_path) # delete the directory
diff --git a/vouchervision/utils_embeddings.py b/vouchervision/utils_embeddings.py
new file mode 100644
index 0000000000000000000000000000000000000000..3251a48b6aa0472220121727040d58de1bffd186
--- /dev/null
+++ b/vouchervision/utils_embeddings.py
@@ -0,0 +1,6 @@
+from InstructorEmbedding import INSTRUCTOR
+model = INSTRUCTOR('hkunlp/instructor-xl')
+instruction = "Represent a row from a table describing and herbarium specimen."
+sentence = "3D ActionSLAM: wearable person tracking in multi-floor environments"
+embeddings = model.encode([[instruction,sentence]])
+print(embeddings)
diff --git a/vouchervision/vouchervision_main.py b/vouchervision/vouchervision_main.py
new file mode 100644
index 0000000000000000000000000000000000000000..62b164c85bb40447993b00f8f50cbf57ba96b4af
--- /dev/null
+++ b/vouchervision/vouchervision_main.py
@@ -0,0 +1,164 @@
+'''
+VoucherVision - based on LeafMachine2 Processes
+'''
+import os, inspect, sys, logging, subprocess
+from time import perf_counter
+currentdir = os.path.dirname(os.path.dirname(inspect.getfile(inspect.currentframe())))
+parentdir = os.path.dirname(currentdir)
+sys.path.append(parentdir)
+sys.path.append(currentdir)
+from vouchervision.component_detector.component_detector import detect_plant_components, detect_archival_components
+from general_utils import add_to_expense_report, save_token_info_as_csv, print_main_start, check_for_subdirs_VV, load_config_file, load_config_file_testing, report_config, save_config_file, subset_dir_images, crop_detections_from_images_VV
+from directory_structure_VV import Dir_Structure
+from data_project import Project_Info
+from LM2_logger import start_logging
+from fetch_data import fetch_data
+from utils_VoucherVision import VoucherVision, space_saver
+
+
+def voucher_vision(cfg_file_path, dir_home, path_custom_prompts, cfg_test, progress_report, path_api_cost=None, test_ind = None):
+ # get_n_overall = progress_report.get_n_overall()
+ # progress_report.update_overall(f"Working on {test_ind+1} of {get_n_overall}")
+
+ t_overall = perf_counter()
+
+ # Load config file
+ report_config(dir_home, cfg_file_path, system='VoucherVision')
+
+ if cfg_test is None:
+ cfg = load_config_file(dir_home, cfg_file_path, system='VoucherVision') # For VoucherVision
+ else:
+ cfg = cfg_test
+ # user_cfg = load_config_file(dir_home, cfg_file_path)
+ # cfg = Config(user_cfg)
+
+ # Check to see if there are subdirs
+ # Yes --> use the names of the subsirs as run_name
+ run_name, dirs_list, has_subdirs = check_for_subdirs_VV(cfg)
+ print(f"run_name {run_name} dirs_list{dirs_list} has_subdirs{has_subdirs}")
+
+ # for dir_ind, dir_in in enumerate(dirs_list):
+ # if has_subdirs:
+ # cfg['leafmachine']['project']['dir_images_local'] = dir_in
+ # cfg['leafmachine']['project']['run_name'] = run_name[dir_ind]
+
+ # Dir structure
+ print_main_start("Creating Directory Structure")
+ Dirs = Dir_Structure(cfg)
+
+ # logging.info("Hi")
+ logger = start_logging(Dirs, cfg)
+
+ # Check to see if required ML files are ready to use
+ ready_to_use = fetch_data(logger, dir_home, cfg_file_path)
+ assert ready_to_use, "Required ML files are not ready to use!\nThe download may have failed,\nor\nthe directory structure of LM2 has been altered"
+
+ # Wrangle images and preprocess
+ print_main_start("Gathering Images and Image Metadata")
+ Project = Project_Info(cfg, logger, dir_home, Dirs) # Where file names are modified
+
+ # Save config file
+ save_config_file(cfg, logger, Dirs)
+
+ # Detect Archival Components
+ print_main_start("Locating Archival Components")
+ Project = detect_archival_components(cfg, logger, dir_home, Project, Dirs)
+
+ # Save cropped detections
+ crop_detections_from_images_VV(cfg, logger, dir_home, Project, Dirs)
+
+ # Process labels
+ Voucher_Vision = VoucherVision(cfg, logger, dir_home, path_custom_prompts, Project, Dirs)
+ n_images = len(Voucher_Vision.img_paths)
+ last_JSON_response, total_tokens_in, total_tokens_out = Voucher_Vision.process_specimen_batch(progress_report)
+
+ if path_api_cost:
+ cost_summary, data, total_cost = save_token_info_as_csv(Dirs, cfg['leafmachine']['LLM_version'], path_api_cost, total_tokens_in, total_tokens_out, n_images)
+ add_to_expense_report(dir_home, data)
+ logger.info(cost_summary)
+ else:
+ total_cost = None #TODO add config tests to expense_report
+
+ t_overall_s = perf_counter()
+ logger.name = 'Run Complete! :)'
+ logger.info(f"[Total elapsed time] {round((t_overall_s - t_overall)/60)} minutes")
+ space_saver(cfg, Dirs, logger)
+
+ for handler in logger.handlers[:]:
+ handler.close()
+ logger.removeHandler(handler)
+
+ return last_JSON_response, total_cost
+
+def voucher_vision_OCR_test(cfg_file_path, dir_home, cfg_test, path_to_crop):
+ # get_n_overall = progress_report.get_n_overall()
+ # progress_report.update_overall(f"Working on {test_ind+1} of {get_n_overall}")
+
+ # Load config file
+ report_config(dir_home, cfg_file_path, system='VoucherVision')
+
+ if cfg_test is None:
+ cfg = load_config_file(dir_home, cfg_file_path, system='VoucherVision') # For VoucherVision
+ else:
+ cfg = cfg_test
+ # user_cfg = load_config_file(dir_home, cfg_file_path)
+ # cfg = Config(user_cfg)
+
+ # Check to see if there are subdirs
+ # Yes --> use the names of the subsirs as run_name
+ run_name, dirs_list, has_subdirs = check_for_subdirs_VV(cfg)
+ print(f"run_name {run_name} dirs_list{dirs_list} has_subdirs{has_subdirs}")
+
+ # for dir_ind, dir_in in enumerate(dirs_list):
+ # if has_subdirs:
+ # cfg['leafmachine']['project']['dir_images_local'] = dir_in
+ # cfg['leafmachine']['project']['run_name'] = run_name[dir_ind]
+
+ # Dir structure
+ print_main_start("Creating Directory Structure")
+ Dirs = Dir_Structure(cfg)
+
+ # logging.info("Hi")
+ logger = start_logging(Dirs, cfg)
+
+ # Check to see if required ML files are ready to use
+ ready_to_use = fetch_data(logger, dir_home, cfg_file_path)
+ assert ready_to_use, "Required ML files are not ready to use!\nThe download may have failed,\nor\nthe directory structure of LM2 has been altered"
+
+ # Wrangle images and preprocess
+ print_main_start("Gathering Images and Image Metadata")
+ Project = Project_Info(cfg, logger, dir_home, Dirs) # Where file names are modified
+
+ # Save config file
+ save_config_file(cfg, logger, Dirs)
+
+ # Detect Archival Components
+ print_main_start("Locating Archival Components")
+ Project = detect_archival_components(cfg, logger, dir_home, Project, Dirs)
+
+ # Save cropped detections
+ crop_detections_from_images_VV(cfg, logger, dir_home, Project, Dirs)
+
+ # Process labels
+ Voucher_Vision = VoucherVision(cfg, logger, dir_home, None, Project, Dirs)
+ last_JSON_response = Voucher_Vision.process_specimen_batch_OCR_test(path_to_crop)
+
+if __name__ == '__main__':
+ is_test = False
+
+ # Set LeafMachine2 dir
+ dir_home = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
+
+ if is_test:
+ cfg_file_path = os.path.join(dir_home, 'demo','demo.yaml') #'D:\Dropbox\LeafMachine2\LeafMachine2.yaml'
+ # cfg_file_path = 'test_installation'
+
+ cfg_testing = load_config_file_testing(dir_home, cfg_file_path)
+ cfg_testing['leafmachine']['project']['dir_images_local'] = os.path.join(dir_home, cfg_testing['leafmachine']['project']['dir_images_local'][0], cfg_testing['leafmachine']['project']['dir_images_local'][1])
+ cfg_testing['leafmachine']['project']['dir_output'] = os.path.join(dir_home, cfg_testing['leafmachine']['project']['dir_output'][0], cfg_testing['leafmachine']['project']['dir_output'][1])
+
+ last_JSON_response = voucher_vision(cfg_file_path, dir_home, cfg_testing, None)
+ else:
+ cfg_file_path = None
+ cfg_testing = None
+ last_JSON_response = voucher_vision(cfg_file_path, dir_home, cfg_testing, None)
\ No newline at end of file