scripts/deploy_hf_bootstrap_catalog.py

#!/usr/bin/env python3
"""When B2 is capped: upload ``store_slim_b2.db`` to a Hub **dataset** and point the Space at it.

Uses ``[huggingface].token`` from parser ``secrets.toml``. Sets Space **variables** (not secrets):
``KINK_HF_REQUIRE_FULL_CATALOG=1``, ``KINK_HF_DATASET_REPO``, and
``KINK_HF_DATASET_FILENAME`` (default ``store_slim_bootstrap.db``) so the image downloads the
smaller full-catalog LFS object instead of copying the bundled demo seed.

After this, run ``python scripts/publish_hf_space.py --verify`` to push code and smoke the app.

Example::

  export HF_SPACE_REPO=ronheichman/kink-discovery
  python scripts/deploy_hf_bootstrap_catalog.py
"""
from __future__ import annotations

import argparse
import os
import sys
import tomllib
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
if str(ROOT) not in sys.path:
    sys.path.insert(0, str(ROOT))

from backend.repo_dotenv import load_repo_dotenv
DEFAULT_SECRETS = Path(os.environ.get("KINK_PARSER_SECRETS_TOML", Path.home() / "PycharmProjects/parser/secrets.toml"))


def main() -> int:
    load_repo_dotenv()
    ap = argparse.ArgumentParser(description=__doc__)
    ap.add_argument("--secrets-toml", type=Path, default=DEFAULT_SECRETS)
    ap.add_argument(
        "--local-path",
        type=Path,
        default=ROOT / "data" / "store_slim_b2.db",
    )
    ap.add_argument("--repo-id", default="ronheichman/kink-catalog-slim")
    ap.add_argument("--remote-filename", default="store_slim_bootstrap.db")
    ap.add_argument("--skip-hub-upload", action="store_true")
    args = ap.parse_args()

    repo_id = (os.environ.get("HF_SPACE_REPO") or "").strip()
    if not repo_id:
        print("deploy_hf_bootstrap_catalog: set HF_SPACE_REPO=owner/slug", file=sys.stderr)
        return 1

    local = args.local_path.resolve()
    if not local.is_file():
        print(f"deploy_hf_bootstrap_catalog: missing {local}", file=sys.stderr)
        return 1

    token = (os.environ.get("HF_TOKEN") or os.environ.get("HUGGING_FACE_HUB_TOKEN") or "").strip()
    if not token and args.secrets_toml.is_file():
        with args.secrets_toml.open("rb") as f:
            data = tomllib.load(f)
        token = ((data.get("huggingface") or {}).get("token") or "").strip()
    if not token:
        print(
            "deploy_hf_bootstrap_catalog: set HF_TOKEN / HUGGING_FACE_HUB_TOKEN or [huggingface].token in secrets.toml",
            file=sys.stderr,
        )
        return 1

    from huggingface_hub import HfApi

    api = HfApi(token=token)

    if not args.skip_hub_upload:
        api.create_repo(args.repo_id, repo_type="dataset", exist_ok=True)
        print(
            f"deploy_hf_bootstrap_catalog: uploading {local} → {args.repo_id}/{args.remote_filename} …",
            file=sys.stderr,
        )
        api.upload_file(
            path_or_fileobj=str(local),
            path_in_repo=args.remote_filename,
            repo_id=args.repo_id,
            repo_type="dataset",
        )

    api.add_space_variable(
        repo_id,
        "KINK_HF_DATASET_FILENAME",
        args.remote_filename,
        description="Bootstrap SQLite filename in kink-catalog-slim (slim ~280MiB)",
        token=token,
    )
    api.add_space_variable(
        repo_id,
        "KINK_HF_REQUIRE_FULL_CATALOG",
        "1",
        description="Require remote full catalog bootstrap instead of bundled demo seed",
        token=token,
    )
    api.add_space_variable(
        repo_id,
        "KINK_HF_DATASET_REPO",
        args.repo_id,
        description="Hub dataset for catalog (matches Dockerfile default)",
        token=token,
    )

    print("deploy_hf_bootstrap_catalog: Space variables updated.", file=sys.stderr)
    print(
        f"  HF_TOKEN=*** HF_SPACE_REPO={repo_id} python scripts/publish_hf_space.py --verify",
        file=sys.stderr,
    )
    return 0


if __name__ == "__main__":
    raise SystemExit(main())