import streamlit as st
import degirum as dg
from PIL import Image
import degirum_tools

# hw_location: Where you want to run inference.
#     Use "@cloud" to use DeGirum cloud.
#     Use "@local" to run on local machine.
#     Use an IP address for AI server inference.
hw_location = "@cloud"

# model_zoo_url: URL/path for the model zoo.
#     Use cloud_zoo_url for @cloud, @local, and AI server inference options.
#     Use '' for an AI server serving models from a local folder.
#     Use a path to a JSON file for a single model zoo in case of @local inference.
model_zoo_url = "https://cs.degirum.com/degirum/public"

# lp_det_model_name: Name of the model for license plate detection.
lp_det_model_name = "yolo_v5s_lp_det--512x512_quant_n2x_orca1_1"

# lp_ocr_model_name: Name of the model for license plate OCR.
lp_ocr_model_name = "yolo_v5s_lp_ocr--256x256_quant_n2x_orca1_1"

# Connect to AI inference engine
model_zoo = dg.connect(hw_location, model_zoo_url, token=st.secrets["DG_TOKEN"])

# Load models
lp_det_model = model_zoo.load_model(lp_det_model_name,
                                   image_backend='pil',
                                   overlay_color=(255,0,0),
                                   overlay_line_width=2,
                                   overlay_font_scale=2                                
                                   )
lp_ocr_model= model_zoo.load_model(lp_ocr_model_name, image_backend='pil')

# Create a compound cropping model with 5% crop extent
crop_model = degirum_tools.CroppingAndClassifyingCompoundModel(
    lp_det_model, lp_ocr_model, 5.0
    )

st.title('DeGirum Cloud Platform Demo of License Plate Detection and Recognition Models')

st.text('Upload an image. Then click on the submit button')
with st.form("model_form"):
    uploaded_file=st.file_uploader('input image')
    submitted = st.form_submit_button("Submit")
    if submitted:
        image = Image.open(uploaded_file)
        image.thumbnail((512,512), Image.Resampling.LANCZOS)
        inference_results=crop_model(image)
        st.image(inference_results.image_overlay,caption='Image with Bounding Boxes')