import streamlit as st
from PIL import Image
import io
from dotenv import load_dotenv
import os
import openai
from openai import OpenAI

# Google Cloud Vision
from google.cloud import vision

# Load environment variables
load_dotenv()

# Set Google Cloud credentials in environment
service_account_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = r'gcv-new-project-dd6ed833cc91.json'

# Initialize Google Vision client
vision_client = vision.ImageAnnotatorClient()

def extract_text_with_google_vision_api(image):
    """Extract structured text from image using Google Cloud Vision API, with additional formatting based on bounding box analysis."""
    import io
    from google.cloud import vision
    vision_client = vision.ImageAnnotatorClient()

    img_byte_arr = io.BytesIO()
    image.save(img_byte_arr, format=image.format)
    image_bytes = img_byte_arr.getvalue()

    image = vision.Image(content=image_bytes)
    response = vision_client.document_text_detection(image=image)
    
    structured_texts = []
    for page in response.full_text_annotation.pages:
        for block in page.blocks:
            block_texts = []
            last_paragraph_bottom = None  # Store the bottom position of the last paragraph to compare spacing
            for paragraph in block.paragraphs:
                paragraph_text = ' '.join([''.join([symbol.text for symbol in word.symbols]) for word in paragraph.words])
                paragraph_text = paragraph_text.strip()
                
                # Example for analyzing bounding box position and size (simplified)
                paragraph_bounds = paragraph.bounding_box
                top_left = paragraph_bounds.vertices[0]  # Example vertex
                
                if last_paragraph_bottom is not None and (top_left.y - last_paragraph_bottom) > 10:
                    # Add additional break if the spacing exceeds some threshold
                    block_texts.append("")  # This adds an extra line break to indicate a significant separation
                
                # Update last_paragraph_bottom to the current paragraph's bottom position
                last_paragraph_bottom = paragraph_bounds.vertices[2].y  # Assuming 0 is top-left and going clockwise
                
                if len(paragraph_text.split()) > 2:
                    block_texts.append(paragraph_text)
                    
            if block_texts:
                structured_texts.append('\n'.join(block_texts))
    
    if structured_texts:
        return '\n\n'.join(structured_texts)
    else:
        return "No structured text found."

def main():
    st.title("Article Extraction")
    uploaded_file = st.file_uploader("Upload an image", type=["jpg", "jpeg", "png"])

    if uploaded_file is not None:
        image = Image.open(uploaded_file)
        st.image(image, caption="Uploaded Image", use_column_width=True)

        extracted_text = extract_text_with_google_vision_api(image)

        st.success("Analysis completed successfully!")
        st.header("Extracted Text:")
        st.write(extracted_text if extracted_text else "No text detected.")

if __name__ == "__main__":
    main()