test_corrected_english.py

#!/usr/bin/env python3
"""
Test the corrected English articles collection.
"""

import os
import sys

# Add the parent directory to Python path
current_dir = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, current_dir)

def test_english_articles():
    """Test the corrected English articles collection."""
    print("🎯 Testing Corrected English Articles Collection")
    print("=" * 60)
    
    try:
        from cve_factchecker.firebase_loader import FirebaseNewsLoader
        
        loader = FirebaseNewsLoader()
        
        print(f"📡 Project ID: {loader.project_id}")
        print(f"📚 English Collection: {loader.config.ENGLISH_ARTICLES_COLLECTION}")
        
        # Test fetching English articles
        print(f"\n📊 Fetching English articles...")
        english_articles = loader.fetch_english_articles(limit=10)
        
        print(f"✅ Successfully fetched {len(english_articles)} English articles")
        
        if english_articles:
            print(f"\n📖 Sample articles:")
            for i, article in enumerate(english_articles[:3], 1):
                print(f"   {i}. {article.title[:80]}...")
                print(f"      📝 Content: {len(article.content)} chars")
                if hasattr(article, 'language') and article.language:
                    print(f"      🌐 Language: {article.language}")
                if hasattr(article, 'url') and article.url:
                    print(f"      🔗 URL: {article.url}")
                print()
            
            # Test larger fetch
            print(f"📊 Testing larger fetch (100 articles)...")
            large_batch = loader.fetch_english_articles(limit=100)
            print(f"✅ Large batch: {len(large_batch)} articles")
            
            # Test unlimited fetch
            print(f"📊 Testing unlimited fetch...")
            all_articles = loader.fetch_english_articles(limit=None)
            print(f"✅ All articles: {len(all_articles)} articles")
            
            return len(all_articles)
        else:
            print("❌ No English articles found")
            return 0
            
    except Exception as e:
        print(f"❌ Test failed: {e}")
        import traceback
        traceback.print_exc()
        return 0

def test_integration_with_system():
    """Test integration with the rest of the system."""
    print(f"\n🔧 Testing System Integration")
    print("=" * 60)
    
    try:
        from cve_factchecker.orchestrator import FactCheckSystem
        
        system = FactCheckSystem()
        
        print(f"📊 Testing English articles ingestion...")
        result = system.ingest_firebase(
            collection="english_articles",  # Use English collection
            limit=20,  # Small test batch
            language="English"
        )
        
        print(f"✅ Ingestion result: {result}")
        
        if result.get("success"):
            synced_count = result.get("synced", 0)
            
            print(f"📚 Articles synced: {synced_count}")
            
            # Check if vector database has chunks now
            try:
                # Try a test search to see if chunks were stored
                test_results = system.retriever.semantic_search("test", k=1)
                chunks_count = len(test_results)
                print(f"🧩 Searchable chunks: {chunks_count}")
                
                if synced_count > 0:
                    print(f"✅ System integration successful!")
                    return True
                else:
                    print(f"⚠️ No articles were synced")
                    return False
            except Exception as e:
                print(f"⚠️ Could not test vector search: {e}")
                # Still consider success if articles were synced
                return synced_count > 0
        else:
            error = result.get("error", "Unknown error")
            print(f"❌ Ingestion failed: {error}")
            return False
            
    except Exception as e:
        print(f"❌ Integration test failed: {e}")
        import traceback
        traceback.print_exc()
        return False

def main():
    """Main test function."""
    print("🧪 CVE Fact Checker - English Articles Collection Test")
    print("=" * 80)
    
    # Test English articles collection
    article_count = test_english_articles()
    
    # Test system integration if we have articles
    if article_count > 0:
        integration_success = test_integration_with_system()
    else:
        integration_success = False
    
    print(f"\n📋 Test Summary:")
    print(f"   English articles found: {article_count}")
    print(f"   System integration: {'✅ OK' if integration_success else '❌ Failed'}")
    
    if article_count > 0 and integration_success:
        print(f"\n🎉 Success! English articles collection is working correctly.")
        print(f"💡 You can now use the system to fetch English articles from Firebase.")
    elif article_count > 0:
        print(f"\n⚠️ Articles found but integration needs work.")
    else:
        print(f"\n❌ No English articles found. Check Firebase collection.")
    
    return article_count > 0

if __name__ == "__main__":
    success = main()
    sys.exit(0 if success else 1)