Spaces:
Sleeping
Sleeping
Upload 2 files
Browse files- requirements2.txt +0 -0
- segmentation.py +81 -0
requirements2.txt
ADDED
Binary file (3.93 kB). View file
|
|
segmentation.py
ADDED
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
from PIL import Image
|
3 |
+
import io
|
4 |
+
from dotenv import load_dotenv
|
5 |
+
import os
|
6 |
+
import openai
|
7 |
+
from openai import OpenAI
|
8 |
+
|
9 |
+
# Google Cloud Vision
|
10 |
+
from google.cloud import vision
|
11 |
+
|
12 |
+
# Load environment variables
|
13 |
+
load_dotenv()
|
14 |
+
|
15 |
+
# Set Google Cloud credentials in environment
|
16 |
+
service_account_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
|
17 |
+
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = r'gcv-new-project-dd6ed833cc91.json'
|
18 |
+
|
19 |
+
# Initialize Google Vision client
|
20 |
+
vision_client = vision.ImageAnnotatorClient()
|
21 |
+
|
22 |
+
def extract_text_with_google_vision_api(image):
|
23 |
+
"""Extract structured text from image using Google Cloud Vision API, with additional formatting based on bounding box analysis."""
|
24 |
+
import io
|
25 |
+
from google.cloud import vision
|
26 |
+
vision_client = vision.ImageAnnotatorClient()
|
27 |
+
|
28 |
+
img_byte_arr = io.BytesIO()
|
29 |
+
image.save(img_byte_arr, format=image.format)
|
30 |
+
image_bytes = img_byte_arr.getvalue()
|
31 |
+
|
32 |
+
image = vision.Image(content=image_bytes)
|
33 |
+
response = vision_client.document_text_detection(image=image)
|
34 |
+
|
35 |
+
structured_texts = []
|
36 |
+
for page in response.full_text_annotation.pages:
|
37 |
+
for block in page.blocks:
|
38 |
+
block_texts = []
|
39 |
+
last_paragraph_bottom = None # Store the bottom position of the last paragraph to compare spacing
|
40 |
+
for paragraph in block.paragraphs:
|
41 |
+
paragraph_text = ' '.join([''.join([symbol.text for symbol in word.symbols]) for word in paragraph.words])
|
42 |
+
paragraph_text = paragraph_text.strip()
|
43 |
+
|
44 |
+
# Example for analyzing bounding box position and size (simplified)
|
45 |
+
paragraph_bounds = paragraph.bounding_box
|
46 |
+
top_left = paragraph_bounds.vertices[0] # Example vertex
|
47 |
+
|
48 |
+
if last_paragraph_bottom is not None and (top_left.y - last_paragraph_bottom) > 10:
|
49 |
+
# Add additional break if the spacing exceeds some threshold
|
50 |
+
block_texts.append("") # This adds an extra line break to indicate a significant separation
|
51 |
+
|
52 |
+
# Update last_paragraph_bottom to the current paragraph's bottom position
|
53 |
+
last_paragraph_bottom = paragraph_bounds.vertices[2].y # Assuming 0 is top-left and going clockwise
|
54 |
+
|
55 |
+
if len(paragraph_text.split()) > 2:
|
56 |
+
block_texts.append(paragraph_text)
|
57 |
+
|
58 |
+
if block_texts:
|
59 |
+
structured_texts.append('\n'.join(block_texts))
|
60 |
+
|
61 |
+
if structured_texts:
|
62 |
+
return '\n\n'.join(structured_texts)
|
63 |
+
else:
|
64 |
+
return "No structured text found."
|
65 |
+
|
66 |
+
def main():
|
67 |
+
st.title("Article Extraction")
|
68 |
+
uploaded_file = st.file_uploader("Upload an image", type=["jpg", "jpeg", "png"])
|
69 |
+
|
70 |
+
if uploaded_file is not None:
|
71 |
+
image = Image.open(uploaded_file)
|
72 |
+
st.image(image, caption="Uploaded Image", use_column_width=True)
|
73 |
+
|
74 |
+
extracted_text = extract_text_with_google_vision_api(image)
|
75 |
+
|
76 |
+
st.success("Analysis completed successfully!")
|
77 |
+
st.header("Extracted Text:")
|
78 |
+
st.write(extracted_text if extracted_text else "No text detected.")
|
79 |
+
|
80 |
+
if __name__ == "__main__":
|
81 |
+
main()
|