app.py

import base64
import streamlit as st
from openai import OpenAI
import os
from dotenv import load_dotenv
import fitz
from PIL import Image
import io
import tempfile

# Load environment variables
load_dotenv()
# Initialize OpenAI client
client = OpenAI(api_key=os.getenv('OPENAI_API_KEY'))

def convert_pdf_to_images(pdf_file):
    """Convert PDF to list of images using PyMuPDF"""
    images = []
    with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as tmp_file:
        tmp_file.write(pdf_file.getvalue())
        pdf_path = tmp_file.name
        
        pdf_document = fitz.open(pdf_path)
        for page_number in range(pdf_document.page_count):
            page = pdf_document[page_number]
            pix = page.get_pixmap()
            img_data = pix.tobytes("png")
            image = Image.open(io.BytesIO(img_data))
            images.append(image)
            
        pdf_document.close()
    os.unlink(pdf_path)
    return images


def format_response(text):
    """Format the analysis response with clean styling"""
    formatted_text = ""
    
    # Split into pages
    pages = text.split("Page")
    
    for page_num, page_content in enumerate(pages[1:], 1):  # Skip first empty split
        formatted_text += f'\n### Page {page_num}\n'
        
        # Process each line
        lines = page_content.split('\n')
        for line in lines:
            # Skip empty lines and lines with asterisks
            if line.strip() and not line.strip().startswith('*') and not line.strip().startswith('Here'):
                # Remove asterisks and dashes
                line = line.replace('**', '').replace('- ', '')
                
                if ':' in line:
                    label, value = line.split(':', 1)
                    formatted_text += f'- *{label.strip()}*: {value.strip()}\n'
        
    return formatted_text

def analyze_image(image):
    """Analyze image using OpenAI API"""
    try:
        img_byte_arr = io.BytesIO()
        image.save(img_byte_arr, format='PNG')
        img_byte_arr = img_byte_arr.getvalue()
        
        base64_image = base64.b64encode(img_byte_arr).decode("utf-8")
        
        response = client.chat.completions.create(
            model="gpt-4o-mini",  # Update to the correct model name
            messages=[
                {
                    "role": "user",
                    "content": [
                        {
                            "type": "text",
                            "text": """Please analyze the image and extract the following information:
                            - Sender information
                            - Recipient information
                            - Container details
                            - Weights and measurements
                            - Dates and reference numbers
                            - Cargo details
                            
                            Format the response as 'Label: Value' pairs."""
                        },
                        {
                            "type": "image_url",
                            "image_url": {
                                "url": f"data:image/jpeg;base64,{base64_image}"
                            },
                        },
                    ],
                }
            ],
            max_tokens=1000
        )
        
        return response.choices[0].message.content
    except Exception as e:
        return f"An error occurred: {str(e)}"

def main():
    st.set_page_config(page_title="Document Analysis App", layout="wide")

    st.title("Document Analysis App")
    uploaded_file = st.file_uploader("Upload document (PDF/Image)", type=['pdf', 'png', 'jpg', 'jpeg'])

    if uploaded_file is not None:
        if uploaded_file.type == "application/pdf":
            # Handle PDF
            with st.spinner("Processing PDF..."):
                images = convert_pdf_to_images(uploaded_file)

                if st.button("Extract Information"):
                    with st.spinner("Analyzing document..."):
                        all_results = []
                        for i, image in enumerate(images, 1):
                            result = analyze_image(image)
                            all_results.append(f"Page {i} Information:\n{result}")

                        combined_results = "\n\n".join(all_results)
                        st.markdown(format_response(combined_results))

        else:
            # Handle single image
            image = Image.open(uploaded_file)

            if st.button("Extract Information"):
                with st.spinner("Analyzing document..."):
                    result = analyze_image(image)
                    st.markdown(format_response(result))

# Call the main function directly (no need for __name__ == "__main__" in Hugging Face Spaces)
main()