app.py

from flask import Flask, request, jsonify
import requests
import time
import json
import supabase
import logging

# Configure logging
logging.basicConfig(level=logging.INFO)

# Azure Document Intelligence setup
AZURE_ENDPOINT = "https://gosignpdf.cognitiveservices.azure.com/"
AZURE_KEY = "2nUifMPmbS35qkiFr5OjgzDw7ooE5Piw5892GQgyWZHe0oNRIBJHJQQJ99AKACfhMk5XJ3w3AAALACOGkANC"

# Supabase setup
SUPABASE_URL = "https://dtzuqtvroalrjhgdcowq.supabase.co/"
SUPABASE_KEY = "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJzdXBhYmFzZSIsInJlZiI6ImR0enVxdHZyb2FscmpoZ2Rjb3dxIiwicm9sZSI6ImFub24iLCJpYXQiOjE3MjU0NDk3MzIsImV4cCI6MjA0MTAyNTczMn0.WrIvwEOq4CqCb8IkU8G4jiWkf9DM1JxGd2_aTN4vlV4"
supabase_client = supabase.create_client(SUPABASE_URL, SUPABASE_KEY)

app = Flask(__name__)

def log_debug(message, **kwargs):
    """Log debug messages for tracking."""
    print(f"[DEBUG] {message}")
    if kwargs:
        for key, value in kwargs.items():
            print(f"  - {key}: {value}")

def download_file_from_supabase(file_path):
    # """Download file from Supabase storage."""
    response = supabase_client.storage.from_("files").download(file_path)
    # No need to check status_code here, as response.content is the file content (bytes).
    if isinstance(response, bytes):  # Direct check if response is file content.
        return response
    else:
        raise Exception(f"Failed to download file from Supabase: {response.text}")


def analyze_pdf_layout(file_content):
    # """Send PDF to Azure and get layout data."""
    url = f"{AZURE_ENDPOINT}/formrecognizer/documentModels/prebuilt-layout:analyze?api-version=2023-07-31"
    headers = {
        "Ocp-Apim-Subscription-Key": AZURE_KEY,
        "Content-Type": "application/pdf",
    }

    response = requests.post(url, headers=headers, data=file_content)
    if response.status_code != 202:
        raise Exception(f"Azure request failed: {response.text}")

    operation_location = response.headers.get("Operation-Location")
    if not operation_location:
        raise Exception("Operation-Location header not found in response.")

    while True:
        result_response = requests.get(operation_location, headers={"Ocp-Apim-Subscription-Key": AZURE_KEY})
        result = result_response.json()

        if result.get("status") == "succeeded":
            return result["analyzeResult"]
        elif result.get("status") == "failed":
            raise Exception("Analysis failed.")
        time.sleep(8)

@app.route("/analyze", methods=["POST"])
def analyze():
    try:
        # Get file ID from request
        file_id = request.json.get("file_id")
        if not file_id:
            return jsonify({"error": "File ID is required"}), 400

        # Fetch file path from Supabase
        file_data = supabase_client.table("files").select("file_path").eq("id", file_id).single().execute()

        # Check if file_data is not None and contains valid data
        if not file_data or not file_data.data:
            return jsonify({"error": "File not found or Supabase query failed"}), 404

        file_path = file_data.data["file_path"]
        # Download the file from Supabase
        file_content = download_file_from_supabase(file_path)

        # Analyze the PDF layout with Azure
        layout_data = analyze_pdf_layout(file_content)

        # Extract required layout values
        page_data = layout_data.get("pages", [])[0]  # Assuming single-page PDF for simplicity
        first_word = page_data.get("words", [])[0]
        last_word = page_data.get("words", [])[-1]

        page_height = page_data["height"]
        page_width = page_data["width"]
        x1 = first_word["polygon"][0]  # X1 of first word
        y4 = last_word["polygon"][-1]  # Y4 of last word

        # Commenting out the Supabase file update logic
        update_response = supabase_client.table("files").update({
            "page_height": page_height,
            "page_width": page_width,
            "x1": x1,
            "y4": y4,
        }).eq("id", file_id).execute()
        
        # Check if update was successful by checking if data is present and valid
        if not update_response.data:
            return jsonify({"error": "Failed to update file layout data"}), 500
        
        # Check if there is any error message in the response
        if hasattr(update_response, 'error') and update_response.error:
            return jsonify({"error": "Failed to update file layout data", "details": update_response.error}), 500
        
        return jsonify({
            "message": "Layout data analyzed successfully",
            "page_height": page_height,
            "page_width": page_width,
            "x1": x1,
            "y4": y4
        }), 200


    except Exception as e:
        return jsonify({"error": str(e)}), 500




if __name__ == "__main__":
    app.run(host="0.0.0.0", port=8000)