data/export_parquet.py

"""
export_parquet.py — Generate seed Parquet files for Kasualdad LFED.

Run once to produce enrollment, attendance, students, discipline, and grades
Parquet files. Place them in /data/ (HF Space persistent storage) or ship
with the image (Modal local dir) for fast loading.

Usage:
    python data/export_parquet.py                    # → data/
    python data/export_parquet.py --out /data        # → /data/ (HF Space)
"""

import argparse
import sys
from pathlib import Path

# Allow running from repo root
sys.path.insert(0, str(Path(__file__).parent.parent))

import duckdb
from data.generate_seed import generate_seed_data


# All 5 tables (expanded schema for Day 1)
TABLES = ["enrollment", "attendance", "students", "discipline", "grades"]


def export_parquet(out_dir: Path) -> None:
    """Generate seed data and export all tables as Parquet."""
    out_dir.mkdir(parents=True, exist_ok=True)

    conn = duckdb.connect(":memory:")
    conn.execute("SET enable_progress_bar = false;")
    generate_seed_data(conn)

    paths = {}
    for table in TABLES:
        out_path = out_dir / f"{table}.parquet"
        conn.execute(f"COPY {table} TO '{out_path}' (FORMAT PARQUET)")
        paths[table] = out_path

    total_size = 0
    print(f"✅ Exported:")
    for table, path in paths.items():
        size = path.stat().st_size
        total_size += size
        print(f"   {path}  ({size:,} bytes)")
    print(f"   Total: {total_size:,} bytes")

    # Verify round-trip
    verify = duckdb.connect(":memory:")
    for table in TABLES:
        path = paths[table]
        verify.execute(f"CREATE TABLE {table} AS SELECT * FROM read_parquet('{path}')")
        n = verify.execute(f"SELECT COUNT(*) FROM {table}").fetchone()[0]
        print(f"   ✓ {table}: {n} rows")
    verify.close()
    conn.close()


if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Export seed data to Parquet")
    parser.add_argument(
        "--out",
        type=Path,
        default=Path(__file__).parent,
        help="Output directory (default: data/)",
    )
    args = parser.parse_args()
    export_parquet(args.out)