scripts/upload_to_hf.py

"""Sync refresh_data.json to the PRIVATE HF dataset, then restart the Space.

Data lives in the private dataset (pramodmisra/pif-data), NOT in the public
Space repo (which is anonymously downloadable). On Space restart, start.py pulls
the dataset file via the HF_TOKEN Space secret and re-imports it.

Flow:
  1. Compare the freshly-built local file against the dataset's current copy.
  2. If unchanged, do nothing (no upload, no restart).
  3. If changed, upload to the dataset and restart the Space to re-import.

Secret: HF_TOKEN (write scope) from the environment — never hardcoded.
Dependency: huggingface_hub (CI-only; not in the Space requirements).
"""
from __future__ import annotations

import os
import sys

from huggingface_hub import HfApi, hf_hub_download

DATA_REPO_ID = "pramodmisra/pif-data"      # private dataset
SPACE_REPO_ID = "pramodmisra/pif"          # public Space (web app)
REPO_ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
REFRESH_FILE = os.path.join(REPO_ROOT, "refresh_data.json")


def _remote_bytes(api: HfApi) -> bytes | None:
    """Current dataset copy, or None if it doesn't exist yet."""
    try:
        path = hf_hub_download(
            repo_id=DATA_REPO_ID, repo_type="dataset",
            filename="refresh_data.json", token=api.token,
        )
        with open(path, "rb") as fh:
            return fh.read()
    except Exception:
        return None


def main() -> None:
    token = os.environ.get("HF_TOKEN")
    if not token:
        print("ERROR: HF_TOKEN not set in environment.", file=sys.stderr)
        sys.exit(1)
    if not os.path.exists(REFRESH_FILE):
        print(f"ERROR: {REFRESH_FILE} not found.", file=sys.stderr)
        sys.exit(1)

    api = HfApi(token=token)
    with open(REFRESH_FILE, "rb") as fh:
        local = fh.read()

    if local == _remote_bytes(api):
        print("No change vs. dataset; skipping upload + restart.")
        return

    print(f"Uploading refresh_data.json -> dataset/{DATA_REPO_ID} ...")
    api.upload_file(
        path_or_fileobj=REFRESH_FILE,
        path_in_repo="refresh_data.json",
        repo_id=DATA_REPO_ID,
        repo_type="dataset",
        commit_message="chore: daily client refresh from Google Sheet",
    )
    print(f"Restarting Space {SPACE_REPO_ID} to re-import ...")
    api.restart_space(repo_id=SPACE_REPO_ID)
    print("Done; Space will re-pull the private dataset on boot.")


if __name__ == "__main__":
    main()