scripts/eval_tables.py

from data_access import get_pg_sync_connection

conn = get_pg_sync_connection()



def create_eval_database():
    """Create SQLite database with a proper relational structure."""
    # Connect to the database (creates it if it doesn't exist)
    cursor = conn.cursor()
    cursor.execute('''
    CREATE TABLE IF NOT EXISTS question_groups (
        id SERIAL PRIMARY KEY,
        group_name TEXT NOT NULL,
        description TEXT
    );
    ''')


    # Create questions table
    cursor.execute('''
    CREATE TABLE IF NOT EXISTS questions (
        id SERIAL PRIMARY KEY,
        question_text TEXT NOT NULL,
        question_group_id INTEGER,
    CONSTRAINT unique_question_text UNIQUE (question_text),
    FOREIGN KEY (question_group_id) REFERENCES question_groups (id);
    ''')


    cursor.execute('''
        CREATE TABLE IF NOT EXISTS source_finders (
            id SERIAL PRIMARY KEY,
            source_finder_type TEXT NOT NULL,
            description TEXT,
            source_finder_version TEXT NOT NULL
        );
        ''')

    cursor.execute('''
        CREATE TABLE IF NOT EXISTS source_finder_runs (
            id SERIAL PRIMARY KEY,
            run_id INTEGER NOT NULL,
            source_finder_id INTEGER NOT NULL,
            description TEXT,
            FOREIGN KEY (source_finder_id) REFERENCES source_finders(id),
            CONSTRAINT unique_source_per_run_id UNIQUE(run_id, source_finder_id)
        );
        ''')

    cursor.execute('''
        CREATE TABLE IF NOT EXISTS source_finder_run_question_metadata (
            id SERIAL PRIMARY KEY,
            question_id INTEGER NOT NULL,
            source_finder_run_id INTEGER NOT NULL,            
            metadata JSON,
            FOREIGN KEY (source_finder_run_id) REFERENCES source_finder_runs(id),d4pnrdvjqlt9gg
            FOREIGN KEY (question_id) REFERENCES questions(id),
            CONSTRAINT unique_question_per_run_id UNIQUE(question_id, source_finder_run_id)
        );
        ''')


    # Create table for logging all sources from each run
    cursor.execute('''
    CREATE TABLE IF NOT EXISTS source_run_results (
        id SERIAL PRIMARY KEY,
        source_finder_run_id INTEGER NOT NULL,
        run_timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
        question_id INTEGER NOT NULL,
        tractate TEXT NOT NULL,
        folio TEXT NOT NULL,
        sugya_id TEXT NOT NULL,
        rank INTEGER NOT NULL,
        reason TEXT,
        FOREIGN KEY (source_finder_run_id) REFERENCES source_finder_runs(id),
        FOREIGN KEY (question_id) REFERENCES questions(id),
        CONSTRAINT constraint source_run_results_pk UNIQUE(source_finder_run_id, question_id, sugya_id)
    );
    ''')

    conn.commit()
    conn.close()

def load_source_finders():
    cursor = conn.cursor()
    for item in ["claude_sources", "keywords", "lenses"]:
        cursor.execute("INSERT INTO source_finders (source_finder_type, source_finder_version) VALUES (%s, 1)", (item,))
    conn.commit()




if __name__ == '__main__':
    # Create the database
    create_eval_database()