#!/usr/bin/env python3
"""One-off migration: `edgar_filings` primary key `accn` → `(cik, accn)` (co-registrants share accessions).

    # production (M3U96b) — BEFORE deploying the code that expects the composite key; stop the writers first:
    pm2 stop hfmarketdata-edgar-incremental hfmarketdata-edgar-backfill
    cd ~/hfmarketdata && HFMD_DATA_ROOT=~/firstratedata venv/bin/python scripts/migrate_edgar_filings.py
    pm2 start hfmarketdata-edgar-incremental

The API may keep serving during the copy (WAL: readers are not blocked; a write to edgar_filings would wait on
the busy timeout, hence stopping the ingestion jobs). The table is rebuilt in ONE transaction
(`edgar_filings_new` ← INSERT OR IGNORE, drop, rename, 4 indexes) — on failure nothing changes. Idempotent:
running it on a migrated database prints the state and exits 0. `--check` only reports.

Expected duration: 1.37 M rows ≈ 212 MB → copy ~5–10 s + indexes ~10–20 s on the M3 Ultra SSD (≈ 30–60 s
worst case); the WAL grows by ~300 MB and is checkpointed at the end.

Author: Simon-Pierre Boucher <contact@spboucher.ai>
"""
from __future__ import annotations

import argparse
import logging
import sys
import time
from pathlib import Path

HERE = Path(__file__).resolve().parent
sys.path.insert(0, str(HERE.parent / "hfmarketdata" / "api"))

log = logging.getLogger("migrate_edgar_filings")


def main() -> int:
    ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
    ap.add_argument("--check", action="store_true", help="report the current key shape and exit")
    ap.add_argument("-v", "--verbose", action="store_true")
    args = ap.parse_args()
    logging.basicConfig(level=logging.DEBUG if args.verbose else logging.INFO, format="%(asctime)s %(levelname)s %(name)s: %(message)s")

    from sqlalchemy import text

    from core.config import settings
    from core.db import engine
    from fundamentals import migrations

    log.info("database: %s", settings.state_db)
    with engine.connect() as con:
        state = migrations.filings_pk_is_composite(con)
        n = migrations.filings_row_count(con) if state is not None else 0
    if state is None:
        log.info("edgar_filings does not exist yet — nothing to migrate (init_db will create it with the composite key)")
        return 0
    if state:
        log.info("edgar_filings already keyed on (cik, accn) — %d rows, nothing to do", n)
        return 0
    log.info("edgar_filings keyed on accn alone: %d rows to copy", n)
    if args.check:
        return 2
    t0 = time.time()
    with engine.begin() as con:
        con.execute(text("PRAGMA busy_timeout=600000"))
        copied = migrations.migrate_filings_pk(con)
    with engine.connect().execution_options(isolation_level="AUTOCOMMIT") as con:
        con.execute(text("PRAGMA wal_checkpoint(TRUNCATE)"))
        ok = migrations.filings_pk_is_composite(con)
    log.info("done: %d rows copied (%d dropped as exact duplicates), composite key: %s, %.1fs", copied, n - copied, ok, time.time() - t0)
    return 0 if ok else 1


if __name__ == "__main__":
    sys.exit(main())