demo.py

#!/usr/bin/env python3
"""
Demonstration script for the movie recommender system.

This script runs tests to verify the system is working correctly.
Can run with or without OpenAI API key.
"""

import sys
import os
from pathlib import Path

# Add src to Python path
src_path = Path(__file__).parent / "src"
sys.path.insert(0, str(src_path))

def demo_basic_functionality():
    """Demonstrate basic functionality without requiring API keys."""
    print("🎬 Movie Recommender System Demo")
    print("=" * 50)
    
    try:
        # Test imports
        print("📦 Testing imports...")
        from movie_recommender.config import get_config
        from movie_recommender.data_processor import MovieDataProcessor
        from movie_recommender.vector_store import MovieVectorStore
        from movie_recommender.embeddings import get_embedding_provider
        print("✅ All imports successful")
        
        # Test configuration
        print("\n⚙️  Testing configuration...")
        config = get_config()
        print(f"✅ Config loaded:")
        print(f"   - Embedding provider: {config.embedding_provider}")
        print(f"   - Embedding model: {config.embedding_model_name}")
        print(f"   - LLM model: {config.llm_model_name}")
        
        # Test data processor
        print("\n📊 Testing data processor...")
        processor = MovieDataProcessor(chunk_size=100, chunk_overlap=10)
        
        # Create sample data
        import pandas as pd
        sample_data = pd.DataFrame({
            'Title': ['The Matrix', 'Inception', 'Gladiator'],
            'Year': [1999, 2010, 2000],
            'Genres': ['Action, Sci-Fi', 'Sci-Fi, Thriller', 'Action, Drama'],
            'Director': ['The Wachowskis', 'Christopher Nolan', 'Ridley Scott'],
            'Star Cast': ['Keanu Reeves, Laurence Fishburne', 'Leonardo DiCaprio, Marion Cotillard', 'Russell Crowe, Joaquin Phoenix'],
            'IMDb Rating': [8.7, 8.8, 8.5],
            'Duration (minutes)': [136, 148, 155],
            'Certificates': ['R', 'PG-13', 'R'],
            'MetaScore': [73.0, 74.0, 67.0]
        })
        
        descriptions = processor.create_movie_descriptions(sample_data)
        print(f"✅ Generated {len(descriptions)} descriptions")
        
        chunks = processor.chunk_descriptions(descriptions)
        print(f"✅ Created {len(chunks)} text chunks")
        
        # Test vector store initialization (without building index)
        print("\n🧬 Testing vector store...")
        try:
            # Try with mock API key for basic initialization test
            vector_store = MovieVectorStore(api_key="test-key")
            stats = vector_store.get_stats()
            print(f"✅ Vector store initialized - status: {stats['status']}")
        except Exception as e:
            if "api_key" in str(e).lower():
                print("✅ Vector store class working (needs real OpenAI API key for full functionality)")
            else:
                raise e
        
        print("\n🎉 Basic demo completed successfully!")
        print("\nWhat's working:")
        print("   ✅ Configuration management")
        print("   ✅ Data processing and description generation") 
        print("   ✅ Text chunking")
        print("   ✅ Vector store initialization")
        print("   ✅ Embedding provider system")
        
        return True
        
    except Exception as e:
        print(f"❌ Error during demo: {e}")
        import traceback
        traceback.print_exc()
        return False

def demo_with_sample_data():
    """Demonstrate with minimal sample data and vector search."""
    print("\n🔬 Advanced Demo with Vector Search")
    print("=" * 50)
    
    # Check for API key
    api_key = os.getenv("OPENAI_API_KEY")
    if not api_key:
        print("⚠️  OPENAI_API_KEY not set. This demo requires API access for embeddings.")
        print("   Set your API key: export OPENAI_API_KEY='your-key-here'")
        return False
    
    try:
        import pandas as pd
        from movie_recommender.data_processor import MovieDataProcessor
        from movie_recommender.vector_store import MovieVectorStore
        
        # Create sample dataset
        sample_movies = pd.DataFrame({
            'Title': ['The Matrix', 'Inception', 'Gladiator', 'Pulp Fiction', 'The Godfather'],
            'Year': [1999, 2010, 2000, 1994, 1972],
            'Genres': ['Action, Sci-Fi', 'Sci-Fi, Thriller', 'Action, Drama', 'Crime, Drama', 'Crime, Drama'],
            'Director': ['The Wachowskis', 'Christopher Nolan', 'Ridley Scott', 'Quentin Tarantino', 'Francis Ford Coppola'],
            'Star Cast': ['Keanu Reeves, Laurence Fishburne', 'Leonardo DiCaprio, Marion Cotillard', 'Russell Crowe, Joaquin Phoenix', 'John Travolta, Samuel L. Jackson', 'Marlon Brando, Al Pacino'],
            'IMDb Rating': [8.7, 8.8, 8.5, 8.9, 9.2],
            'Duration (minutes)': [136, 148, 155, 154, 175],
            'Certificates': ['R', 'PG-13', 'R', 'R', 'R'],
            'MetaScore': [73.0, 74.0, 67.0, 95.0, 100.0]
        })
        
        print(f"📊 Working with {len(sample_movies)} sample movies")
        
        # Process data
        processor = MovieDataProcessor(chunk_size=150, chunk_overlap=20)
        descriptions = processor.create_movie_descriptions(sample_movies)
        chunks = processor.chunk_descriptions(descriptions)
        
        print(f"✅ Created {len(chunks)} chunks from {len(descriptions)} descriptions")
        
        # Build vector store (this will use OpenAI embeddings)
        print("🧬 Building vector store with OpenAI embeddings...")
        vector_store = MovieVectorStore(api_key=api_key)
        vector_store.build_index(chunks)
        
        stats = vector_store.get_stats()
        print(f"✅ Vector store built: {stats['total_vectors']} vectors, {stats['unique_movies']} movies")
        
        # Test search functionality
        print("\n🔍 Testing search functionality...")
        test_queries = [
            "science fiction movie",
            "action movie with gladiators", 
            "crime drama",
            "movie with high rating"
        ]
        
        for query in test_queries:
            results = vector_store.search(query, k=2)
            print(f"Query: '{query}' → Found: {[r['Title'] for r in results[:2]]}")
        
        print("\n🎉 Advanced demo completed successfully!")
        print("The system is working correctly with OpenAI embeddings.")
        
        return True
        
    except Exception as e:
        print(f"❌ Error during advanced demo: {e}")
        import traceback
        traceback.print_exc()
        return False

if __name__ == "__main__":
    success = demo_basic_functionality()
    
    if success:
        print("\n" + "="*50)
        try:
            choice = input("Run advanced demo with OpenAI embeddings? (y/N): ").strip().lower()
            if choice in ['y', 'yes']:
                demo_with_sample_data()
            else:
                print("\n💡 To run the full system:")
                print("   1. Set OPENAI_API_KEY environment variable")
                print("   2. Run: python main.py --mode build")
                print("   3. Run: python main.py --mode ui")
        except (EOFError, KeyboardInterrupt):
            print("\n💡 To run the full system:")
            print("   1. Set OPENAI_API_KEY environment variable")
            print("   2. Run: python main.py --mode build")
            print("   3. Run: python main.py --mode ui")
    
    print("\n👋 Demo finished!")