#!/usr/bin/env python3
"""One-time backfill: fill missing content_id values on file-level symbol
history entries by reading the canonical source — the snapshot manifest.

Background
----------
When the indexer processes a file-level move/rename op (structured_delta op
type 'patch' with from_address), it records no new_content_id at the symbol
level.  The snapshot stored at that commit, however, has the full
{path: content_id} manifest.  This script reads those manifests and fills
the gaps.

Only file-level addresses (no '::' separator) are handled — they map directly
to a manifest path.  Symbol-level addresses are not yet backfillable this way.

Usage
-----
# Dry-run: count rows that would be changed (no writes)
docker exec musehub-blue python3 /app/deploy/backfill_content_ids_from_snapshots.py --dry-run

# Run for all repos
docker exec musehub-blue python3 /app/deploy/backfill_content_ids_from_snapshots.py

# Run for a single repo
docker exec musehub-blue python3 /app/deploy/backfill_content_ids_from_snapshots.py --repo-id <repo_id>

# Quiet (no progress output)
docker exec musehub-blue python3 /app/deploy/backfill_content_ids_from_snapshots.py -q
"""
from __future__ import annotations

import argparse
import asyncio
import sys
import time

import sqlalchemy as sa
from sqlalchemy.ext.asyncio import AsyncSession, create_async_engine
from sqlalchemy.orm import sessionmaker

sys.path.insert(0, "/app")
sys.path.insert(0, "/tmp/devpkgs")

from musehub.db.database import get_database_url
from musehub.services.musehub_symbol_indexer import backfill_content_ids_from_snapshots


async def run(dry_run: bool, quiet: bool, repo_id: str | None) -> int:
    engine = create_async_engine(get_database_url(), echo=False)
    Session = sessionmaker(engine, class_=AsyncSession, expire_on_commit=False)

    async with Session() as session:
        if not quiet:
            scope = f"repo {repo_id}" if repo_id else "all repos"
            mode = "[DRY RUN] " if dry_run else ""
            print(f"{mode}Scanning {scope} for file-level entries with missing content_id …")

        t0 = time.monotonic()
        count = await backfill_content_ids_from_snapshots(session, repo_id=repo_id, dry_run=dry_run)

        if not dry_run:
            await session.commit()

        elapsed = time.monotonic() - t0

        if dry_run:
            print(f"Would update {count} entr{'y' if count == 1 else 'ies'} ({elapsed:.1f}s)")
        else:
            print(f"Updated {count} entr{'y' if count == 1 else 'ies'} ({elapsed:.1f}s)")

    await engine.dispose()
    return count


def main() -> None:
    p = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
    p.add_argument("--dry-run", action="store_true", help="Count rows without writing")
    p.add_argument("--repo-id", metavar="REPO_ID", help="Limit to a single repo")
    p.add_argument("-q", "--quiet", action="store_true", help="Suppress progress output")
    args = p.parse_args()

    count = asyncio.run(run(dry_run=args.dry_run, quiet=args.quiet, repo_id=args.repo_id))
    sys.exit(0 if count >= 0 else 1)


if __name__ == "__main__":
    main()