Smart Document Analysis Platform

from flask import Flask, request, jsonify, render_template_string from flask_cors import CORS from google import genai from google.genai import types import os import io import httpx import uuid from datetime import datetime, timezone, timedelta from dotenv import load_dotenv import json # Load environment variables load_dotenv() # Get Google API key from environment GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY") if GOOGLE_API_KEY is None: raise ValueError("GOOGLE_API_KEY environment variable is not set. Please set it before running the script.") app = Flask(__name__) CORS(app) # Configure Flask for large file uploads (200MB for substantial documents) app.config['MAX_CONTENT_LENGTH'] = 200 * 1024 * 1024 # 200MB max file size # Initialize Gemini client with correct API key client = genai.Client(api_key=GOOGLE_API_KEY) # In-memory storage for demo (in production, use a database) document_caches = {} user_sessions = {} # HTML template for the web interface HTML_TEMPLATE = """ Smart Document Analysis Platform

📤 Upload PDF Document

📄

Drag and drop your PDF file here, or click to select

For context caching to work: Upload substantial documents (5MB+ recommended)

Maximum file size: 200MB

Or provide a URL:

💬 Ask Questions

👋 Hello! I'm ready to analyze your PDF documents. Upload a document to get started!

""" @app.route('/') def index(): return render_template_string(HTML_TEMPLATE) @app.route('/upload', methods=['POST']) def upload_file(): try: if 'file' not in request.files: return jsonify({'success': False, 'error': 'No file provided'}) file = request.files['file'] if file.filename == '': return jsonify({'success': False, 'error': 'No file selected'}) # Check file size (limit to 200MB for large documents needed for caching) file.seek(0, 2) # Seek to end file_size = file.tell() file.seek(0) # Reset to beginning # Convert to MB for display file_size_mb = file_size / (1024 * 1024) if file_size > 200 * 1024 * 1024: # 200MB limit return jsonify({'success': False, 'error': f'File too large ({file_size_mb:.1f}MB). Maximum size is 200MB.'}) # Warn about small files that might not cache if file_size < 1024 * 1024: # Less than 1MB print(f"Warning: Small file uploaded ({file_size_mb:.1f}MB). May not meet minimum token requirements for caching.") # Read file content file_content = file.read() if not file_content: return jsonify({'success': False, 'error': 'File is empty'}) file_io = io.BytesIO(file_content) # Upload to Gemini File API try: # Simple file upload - no config parameter needed document = client.files.upload(file=file_io) print(f"Document uploaded successfully: {document.name}") except Exception as upload_error: print(f"Upload error: {upload_error}") return jsonify({'success': False, 'error': f'Failed to upload file to Gemini: {str(upload_error)}'}) # Create cache with system instruction try: system_instruction = "You are an expert document analyzer. Provide detailed, accurate answers based on the uploaded document content. Always be helpful and thorough in your responses." # Use the correct model name - try 2.5 Flash first (lower token requirement) model = 'gemini-2.5-flash-001' cache = client.caches.create( model=model, config=types.CreateCachedContentConfig( display_name=f'PDF document cache - {file.filename}', system_instruction=system_instruction, contents=[document], ttl="3600s", # 1 hour TTL ) ) print(f"Cache created successfully: {cache.name}") # Store cache info cache_id = str(uuid.uuid4()) document_caches[cache_id] = { 'cache_name': cache.name, 'document_name': file.filename, 'document_file_name': document.name, 'created_at': datetime.now().isoformat() } # Get token count safely token_count = 'Unknown' if hasattr(cache, 'usage_metadata') and cache.usage_metadata: if hasattr(cache.usage_metadata, 'total_token_count'): token_count = cache.usage_metadata.total_token_count elif hasattr(cache.usage_metadata, 'cached_token_count'): token_count = cache.usage_metadata.cached_token_count return jsonify({ 'success': True, 'cache_id': cache_id, 'token_count': token_count, 'document_name': file.filename }) except Exception as cache_error: print(f"Cache error: {cache_error}") # Provide more specific error handling for token requirements error_msg = str(cache_error).lower() if "too small" in error_msg or "minimum" in error_msg: return jsonify({ 'success': False, 'error': f'Document content is insufficient for caching. Gemini 2.5 Flash requires minimum 1,024 tokens (~2-3 pages of text). Your document: {file.filename} ({file_size_mb:.1f}MB)', 'suggestion': 'Upload a longer document with more text content (recommended: 5MB+ with substantial text).', 'fallback': 'You can still use the document without caching by implementing direct file processing.' }) elif "invalid" in error_msg or "model" in error_msg: # Try fallback to 2.0 Flash try: cache_fallback = client.caches.create( model='gemini-2.0-flash-001', config=types.CreateCachedContentConfig( display_name=f'PDF document cache - {file.filename}', system_instruction=system_instruction, contents=[document], ttl="3600s", ) ) print(f"Fallback cache created with 2.0 Flash: {cache_fallback.name}") # Store with fallback model info cache_id = str(uuid.uuid4()) document_caches[cache_id] = { 'cache_name': cache_fallback.name, 'document_name': file.filename, 'document_file_name': document.name, 'model': 'gemini-2.0-flash-001', 'created_at': datetime.now().isoformat() } token_count = 'Unknown' if hasattr(cache_fallback, 'usage_metadata') and cache_fallback.usage_metadata: if hasattr(cache_fallback.usage_metadata, 'total_token_count'): token_count = cache_fallback.usage_metadata.total_token_count return jsonify({ 'success': True, 'cache_id': cache_id, 'token_count': token_count, 'document_name': file.filename, 'model_used': 'gemini-2.0-flash-001' }) except Exception as fallback_error: print(f"Fallback cache error: {fallback_error}") return jsonify({'success': False, 'error': f'Failed to create cache with both models: {str(fallback_error)}'}) else: return jsonify({'success': False, 'error': f'Failed to create cache: {str(cache_error)}'}) except Exception as e: print(f"General error: {e}") return jsonify({'success': False, 'error': f'Server error: {str(e)}'}) @app.route('/upload-url', methods=['POST']) def upload_from_url(): try: data = request.get_json() url = data.get('url') if not url: return jsonify({'success': False, 'error': 'No URL provided'}) # Download file from URL with timeout and size limits try: with httpx.Client(timeout=30.0) as client_http: response = client_http.get(url) response.raise_for_status() # Check content type content_type = response.headers.get('content-type', '').lower() if 'pdf' not in content_type and not url.lower().endswith('.pdf'): return jsonify({'success': False, 'error': 'URL does not point to a PDF file'}) # Check file size content_length = len(response.content) content_length_mb = content_length / (1024 * 1024) if content_length > 200 * 1024 * 1024: # 200MB limit return jsonify({'success': False, 'error': f'File too large ({content_length_mb:.1f}MB). Maximum size is 200MB.'}) # Warn about small files if content_length < 1024 * 1024: # Less than 1MB print(f"Warning: Small file from URL ({content_length_mb:.1f}MB). May not meet minimum token requirements for caching.") file_io = io.BytesIO(response.content) except httpx.TimeoutException: return jsonify({'success': False, 'error': 'Request timeout. Please try a different URL.'}) except httpx.HTTPError as e: return jsonify({'success': False, 'error': f'Failed to download file: {str(e)}'}) # Extract filename from URL filename = url.split('/')[-1] if not filename.endswith('.pdf'): filename += '.pdf' # Upload to Gemini File API try: # Simple file upload - no config parameter needed document = client.files.upload(file=file_io) print(f"Document uploaded successfully: {document.name}") except Exception as upload_error: print(f"Upload error: {upload_error}") return jsonify({'success': False, 'error': f'Failed to upload file to Gemini: {str(upload_error)}'}) # Create cache with system instruction try: system_instruction = "You are an expert document analyzer. Provide detailed, accurate answers based on the uploaded document content. Always be helpful and thorough in your responses." # Use the correct model name - try 2.5 Flash first (lower token requirement) model = 'gemini-2.5-flash-001' cache = client.caches.create( model=model, config=types.CreateCachedContentConfig( display_name=f'PDF document cache - {filename}', system_instruction=system_instruction, contents=[document], ttl="3600s", # 1 hour TTL ) ) print(f"Cache created successfully: {cache.name}") # Store cache info cache_id = str(uuid.uuid4()) document_caches[cache_id] = { 'cache_name': cache.name, 'document_name': filename, 'document_file_name': document.name, 'source_url': url, 'created_at': datetime.now().isoformat() } # Get token count safely token_count = 'Unknown' if hasattr(cache, 'usage_metadata') and cache.usage_metadata: if hasattr(cache.usage_metadata, 'total_token_count'): token_count = cache.usage_metadata.total_token_count elif hasattr(cache.usage_metadata, 'cached_token_count'): token_count = cache.usage_metadata.cached_token_count return jsonify({ 'success': True, 'cache_id': cache_id, 'token_count': token_count, 'document_name': filename }) except Exception as cache_error: print(f"Cache error: {cache_error}") # If caching fails due to small content, provide alternative approach if "too small" in str(cache_error).lower(): return jsonify({ 'success': False, 'error': 'PDF content is too small for caching. Please upload a larger document with more text content.', 'suggestion': 'Try uploading a longer document or combine multiple documents.' }) else: return jsonify({'success': False, 'error': f'Failed to create cache: {str(cache_error)}'}) except Exception as e: print(f"General error: {e}") return jsonify({'success': False, 'error': f'Server error: {str(e)}'}) @app.route('/ask', methods=['POST']) def ask_question(): try: data = request.get_json() question = data.get('question') cache_id = data.get('cache_id') if not question or not cache_id: return jsonify({'success': False, 'error': 'Missing question or cache_id'}) if cache_id not in document_caches: return jsonify({'success': False, 'error': 'Cache not found. Please upload a document first.'}) cache_info = document_caches[cache_id] # Generate response using cached content with correct model format try: response = client.models.generate_content( model='gemini-2.5-flash-001', # Use 2.5 Flash for consistency contents=question, config=types.GenerateContentConfig( cached_content=cache_info['cache_name'] ) ) if response and response.text: return jsonify({ 'success': True, 'answer': response.text }) else: return jsonify({ 'success': False, 'error': 'No response generated from the model' }) except Exception as gen_error: print(f"Generation error: {gen_error}") return jsonify({'success': False, 'error': f'Failed to generate response: {str(gen_error)}'}) except Exception as e: print(f"General error in ask_question: {e}") return jsonify({'success': False, 'error': f'Server error: {str(e)}'}) @app.route('/caches', methods=['GET']) def list_caches(): try: caches = [] for cache_id, cache_info in document_caches.items(): caches.append({ 'cache_id': cache_id, 'document_name': cache_info['document_name'], 'created_at': cache_info['created_at'] }) return jsonify({'success': True, 'caches': caches}) except Exception as e: return jsonify({'success': False, 'error': str(e)}) @app.route('/cache/', methods=['DELETE']) def delete_cache(cache_id): try: if cache_id not in document_caches: return jsonify({'success': False, 'error': 'Cache not found'}) cache_info = document_caches[cache_id] # Delete from Gemini API try: client.caches.delete(cache_info['cache_name']) except Exception as delete_error: print(f"Error deleting cache from Gemini API: {delete_error}") # Continue to remove from local storage even if API deletion fails # Remove from local storage del document_caches[cache_id] return jsonify({'success': True, 'message': 'Cache deleted successfully'}) except Exception as e: return jsonify({'success': False, 'error': str(e)}) # Health check endpoint @app.route('/health', methods=['GET']) def health_check(): return jsonify({'status': 'healthy', 'service': 'Smart Document Analysis Platform'}) # Error handlers @app.errorhandler(413) def too_large(e): return jsonify({'success': False, 'error': 'File too large. Maximum size is 200MB for substantial documents needed for context caching.'}), 413 @app.errorhandler(500) def internal_error(e): return jsonify({'success': False, 'error': 'Internal server error'}), 500 if __name__ == '__main__': import os port = int(os.environ.get("PORT", 7860)) print(f"Starting server on port {port}") print(f"Google API Key configured: {'Yes' if GOOGLE_API_KEY else 'No'}") app.run(debug=False, host='0.0.0.0', port=port)

📚 Smart Document Analysis Platform

Or provide a URL:

✅ Document Cached Successfully!