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."""
    log_debug("Downloading file from Supabase", file_path=file_path)
    response = supabase_client.storage.from_("files").download(file_path)
    log_debug("Supabase download response", status_code=response.status_code, text=response.text)
    if response.status_code != 200:
        raise Exception(f"Failed to download file from Supabase: {response.text}")
    return response.content

def analyze_pdf_layout(file_content):
    """Send PDF to Azure and get layout data."""
    log_debug("Sending PDF to Azure for analysis")
    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)
    log_debug("Azure response", status_code=response.status_code, headers=response.headers)
    if response.status_code != 202:
        raise Exception(f"Azure request failed: {response.text}")

    operation_location = response.headers.get("Operation-Location")
    log_debug("Azure operation location", operation_location=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()
        log_debug("Azure polling result", status=result.get("status"))

        if result.get("status") == "succeeded":
            log_debug("Azure analysis 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")
        log_debug("Received API request", file_id=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("filePath").eq("id", file_id).single().execute()
        log_debug("Supabase file data response", status_code=file_data.status_code, data=file_data.data)
        if file_data.status_code != 200 or not file_data.data:
            return jsonify({"error": "File not found"}), 404

        file_path = file_data.data["filePath"]
        log_debug("File path retrieved from Supabase", file_path=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)
        log_debug("Layout data retrieved", layout_data=layout_data)

        # 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
        log_debug("Extracted layout values", page_height=page_height, page_width=page_width, x1=x1, y4=y4)

        # Update the `files` table in Supabase
        update_response = supabase_client.table("files").update({
            "page_height": page_height,
            "page_width": page_width,
            "x1": x1,
            "y4": y4,
        }).eq("id", file_id).execute()
        log_debug("Supabase update response", status_code=update_response.status_code, data=update_response.data)

        if update_response.status_code != 200:
            return jsonify({"error": "Failed to update file layout data"}), 500

        return jsonify({"message": "Layout data successfully updated"}), 200

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

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