utilities/extract_text.py

# import os
# import requests

# EXTRACT_TEXT_API = "https://point9-extract-text-and-table.hf.space/api/text"  # Replace with your space URL

# def extract_text_remote(state):
#     filename = state["filename"]
#     path = state["temp_files"][filename]

#     with open(path, "rb") as f:
#         files = {"file": (filename, f, "application/pdf")}
#         data = {
#             "filename": filename,
#             "start_page": state.get("start_page", 1),
#             "end_page": state.get("end_page", 1)
#         }
#         headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_TOKEN')}"}
#         resp = requests.post(EXTRACT_TEXT_API, files=files, data=data, headers=headers)

#     if resp.status_code != 200:
#         raise RuntimeError(f"Extract text API failed: {resp.text}")

#     state["text"] = resp.json().get("text", "")
#     return state

import os
import requests

# Hardcoded API URL - DO NOT CHANGE
EXTRACT_TEXT_API = "https://point9-extract-text-and-table.hf.space/api/text"

def extract_text_remote(state):
    """
    Extract text from documents via API.
    All file types are sent to the API (no local processing).
    
    Args:
        state: Dict with 'filename', 'temp_files', 'start_page', 'end_page'
    
    Returns:
        state: Dict with 'text' key containing extracted text
    """
    filename = state["filename"]
    path = state["temp_files"][filename]

    if not os.path.exists(path):
        raise RuntimeError(f"File not found: {path}")

    # Get file size for debugging
    file_size = os.path.getsize(path)
    
    # Extract just the filename (not full path) to match curl format
    file_basename = os.path.basename(path)
    
    print(f"\n{'='*60}")
    print(f"📄 EXTRACT TEXT API CALL")
    print(f"{'='*60}")
    print(f"File: {filename}")
    print(f"Basename: {file_basename}")
    print(f"Path: {path}")
    print(f"Size: {file_size} bytes")
    print(f"Start Page: {state.get('start_page', 1)}")
    print(f"End Page: {state.get('end_page', 1)}")
    print(f"API URL: {EXTRACT_TEXT_API}")
    print(f"Auth Token: {'✓ Set' if os.getenv('HUGGINGFACE_API_TOKEN') else '✗ Not Set'}")

    with open(path, "rb") as f:
        # IMPORTANT: Use basename for the file tuple (matches curl format)
        files = {"file": (file_basename, f, "application/pdf")}
        
        # IMPORTANT: Convert page numbers to strings (matches curl -F format)
        data = {
            "filename": file_basename,  # Just filename, not full path
            "start_page": str(state.get("start_page", 1)),  # String, not int
            "end_page": str(state.get("end_page", 1))  # String, not int
        }
        headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_TOKEN')}"}
        # DON'T set Content-Type - let requests handle it for multipart/form-data
        
        print(f"\n🚀 Sending request to API...")
        print(f"File tuple: ('file', ('{file_basename}', <binary>, 'application/pdf'))")
        print(f"Data params: {data}")
        print(f"Data types: start_page={type(data['start_page'])}, end_page={type(data['end_page'])}")
        print(f"Headers: {headers}")
        print(f"File size in bytes: {file_size}")
        
        # Call API and wait for response
        # NOTE: Increased timeout to 300 seconds (5 minutes) for large files
        try:
            resp = requests.post(EXTRACT_TEXT_API, files=files, data=data, headers=headers, timeout=300)
        except requests.exceptions.Timeout:
            print(f"❌ Request timed out after 300 seconds")
            raise RuntimeError("API request timed out after 5 minutes")
        except requests.exceptions.RequestException as e:
            print(f"❌ Request exception: {str(e)}")
            raise RuntimeError(f"API request failed: {str(e)}")

    print(f"\n📥 API Response:")
    print(f"Status Code: {resp.status_code}")
    print(f"Response Headers: {dict(resp.headers)}")
    
    if resp.status_code != 200:
        print(f"❌ Error Response: {resp.text[:500]}")
        raise RuntimeError(f"Extract text API failed with status {resp.status_code}: {resp.text}")

    try:
        response_json = resp.json()
        print(f"Response JSON keys: {list(response_json.keys())}")
        print(f"Response JSON: {str(response_json)[:500]}")
        
        # IMPORTANT: The API returns {"status": "completed", "result": "text here", "process_id": "..."}
        # NOT {"text": "..."}
        extracted_text = response_json.get("result", "") or response_json.get("text", "")
        text_length = len(extracted_text) if extracted_text else 0
        
        print(f"\n📊 Extraction Result:")
        print(f"API Status: {response_json.get('status', 'unknown')}")
        print(f"Process ID: {response_json.get('process_id', 'none')}")
        print(f"Text Length: {text_length} characters")
        if text_length > 0:
            print(f"First 200 chars: {extracted_text[:200]}")
        else:
            print(f"⚠️  WARNING: API returned EMPTY text!")
            print(f"Full response: {response_json}")
        
        state["text"] = extracted_text
        print(f"{'='*60}\n")
        return state
        
    except Exception as e:
        print(f"❌ Error parsing response: {str(e)}")
        print(f"Raw response: {resp.text[:500]}")
        raise RuntimeError(f"Failed to parse API response: {str(e)}")