check_embeddings_table.py

"""
Проверка наличия таблицы для эмбеддингов в БД
"""
import psycopg2

DB_CONFIG = {
    'host': 'dpg-d5ht8vi4d50c739akh2g-a.virginia-postgres.render.com',
    'port': 5432,
    'database': 'lead_exchange_bk',
    'user': 'lead_exchange_bk_user',
    'password': '8m2gtTRBW0iAr7nY2Aadzz0VcZBEVKYM'
}

try:
    conn = psycopg2.connect(**DB_CONFIG)
    cursor = conn.cursor()

    print("=" * 70)
    print("CHECKING EMBEDDINGS STORAGE")
    print("=" * 70)

    # Проверяем все таблицы
    cursor.execute("""
        SELECT table_name
        FROM information_schema.tables
        WHERE table_schema = 'public'
        ORDER BY table_name
    """)

    tables = [row[0] for row in cursor.fetchall()]
    print(f"\n📋 All tables in database:")
    for table in tables:
        print(f"  - {table}")

    # Проверяем наличие pgvector extension
    cursor.execute("""
        SELECT * FROM pg_extension WHERE extname = 'vector'
    """)
    has_pgvector = cursor.fetchone() is not None
    print(f"\n🔌 pgvector extension: {'✅ Installed' if has_pgvector else '❌ Not installed'}")

    # Проверяем структуру properties
    cursor.execute("""
        SELECT column_name, data_type, character_maximum_length
        FROM information_schema.columns
        WHERE table_name = 'properties'
        ORDER BY ordinal_position
    """)

    print(f"\n🏠 Properties table structure:")
    has_embedding_column = False
    for col_name, data_type, max_length in cursor.fetchall():
        if 'embedding' in col_name.lower():
            has_embedding_column = True
            print(f"  ✅ {col_name}: {data_type}")
        else:
            print(f"  - {col_name}: {data_type}")

    if not has_embedding_column:
        print(f"\n⚠️  No embedding column found in properties table")

    # Проверяем наличие отдельной таблицы для эмбеддингов
    embedding_tables = [t for t in tables if 'embedding' in t.lower() or 'vector' in t.lower()]

    if embedding_tables:
        print(f"\n📊 Found embedding-related tables:")
        for table in embedding_tables:
            print(f"  - {table}")
            cursor.execute(f"""
                SELECT column_name, data_type
                FROM information_schema.columns
                WHERE table_name = '{table}'
                ORDER BY ordinal_position
            """)
            for col_name, data_type in cursor.fetchall():
                print(f"    - {col_name}: {data_type}")
    else:
        print(f"\n⚠️  No separate table for embeddings found")

    print("\n" + "=" * 70)
    print("RECOMMENDATION")
    print("=" * 70)

    if has_pgvector and has_embedding_column:
        print("✅ Ready to store embeddings in properties table")
        print("   Use: ALTER TABLE properties ADD COLUMN embedding vector(768)")
    elif has_pgvector and embedding_tables:
        print("✅ Can store embeddings in separate table")
    else:
        print("⚠️  Need to create storage for embeddings:")
        print("")
        print("Option 1: Add column to properties (recommended)")
        print("  ALTER TABLE properties ADD COLUMN embedding vector(768);")
        print("  CREATE INDEX ON properties USING ivfflat (embedding vector_cosine_ops);")
        print("")
        print("Option 2: Create separate table")
        print("  CREATE TABLE property_embeddings (")
        print("    property_id UUID PRIMARY KEY REFERENCES properties(property_id),")
        print("    embedding vector(768),")
        print("    created_at TIMESTAMPTZ DEFAULT NOW()")
        print("  );")
        print("")
        print("⚠️  First install pgvector: CREATE EXTENSION vector;")

    cursor.close()
    conn.close()

except Exception as e:
    print(f"❌ Error: {e}")
    import traceback
    traceback.print_exc()