publish_cleaned_to_hf.py

#!/usr/bin/env python3
"""
publish_cleaned_to_hf.py — Upload a single cleaned.csv to Hugging Face dataset repo.

- Ensures the dataset repo exists (public by default).
- Uploads the specified file (default: cleaned.csv) to the repo root or given path.
- Optionally uploads a minimal README.md dataset card referencing cleaned.csv.

Authentication:
- Set env var `HF_TOKEN` or `HUGGINGFACE_TOKEN`, or pass `--token`.

Usage:
  python publish_cleaned_to_hf.py --file cleaned.csv --repo-id savedata101/repair_data
  python publish_cleaned_to_hf.py --file report/cleaned.csv --path-in-repo cleaned.csv
  python publish_cleaned_to_hf.py --add-readme --title "Repair Data (Cleaned)"
"""

from __future__ import annotations

import argparse
import os
import sys
from typing import Optional

try:
    from huggingface_hub import HfApi
except Exception:
    print("Please install huggingface_hub: pip install huggingface_hub", file=sys.stderr)
    raise


def get_token(cli_token: Optional[str]) -> Optional[str]:
    if cli_token:
        return cli_token
    for var in ("HF_TOKEN", "HUGGINGFACE_TOKEN"):
        t = os.environ.get(var)
        if t:
            return t
    return None


def main(argv=None) -> int:
    ap = argparse.ArgumentParser(description="Upload cleaned.csv to Hugging Face dataset repo")
    ap.add_argument("--repo-id", default="savedata101/repair_data", help="Target repo id")
    ap.add_argument("--file", default="cleaned.csv", help="Local CSV file to upload")
    ap.add_argument("--path-in-repo", default="", help="Destination path in repo (default root)")
    ap.add_argument("--token", default=None, help="Hugging Face access token")
    ap.add_argument("--private", action="store_true", help="Create as private repo (default public)")
    ap.add_argument("--add-readme", action="store_true", help="Upload a minimal README.md dataset card")
    ap.add_argument("--title", default="Repair Data — cleaned.csv", help="Dataset card title")
    args = ap.parse_args(argv)

    token = get_token(args.token)
    if not token:
        print("Missing token. Set HF_TOKEN or pass --token.", file=sys.stderr)
        return 2

    if not os.path.isfile(args.file):
        print(f"File not found: {args.file}", file=sys.stderr)
        return 2

    api = HfApi(token=token)

    # Ensure repo exists (dataset type)
    api.create_repo(repo_id=args.repo_id, repo_type="dataset", private=args.private, exist_ok=True)

    # Upload cleaned.csv
    dest = args.path_in_repo if args.path_in_repo else os.path.basename(args.file)
    api.upload_file(
        path_or_fileobj=args.file,
        repo_id=args.repo_id,
        repo_type="dataset",
        path_in_repo=dest,
        commit_message=f"Upload {dest}",
    )

    # Optionally add a minimal dataset card pointing to cleaned.csv
    if args.add_readme:
        card = f"""---
license: other
tags:
  - csv
pretty_name: {args.title}
---

# {args.title}

This dataset hosts a cleaned CSV file for UI preview and use with `datasets`.

## Files

- `{dest}` — main data file.

## Load with datasets

```python
from datasets import load_dataset
ds = load_dataset(
    "{args.repo_id}",
    data_files={{{"train": "{dest}"}}},
)
print(ds["train"])  # Dataset
```
"""
        api.upload_file(
            path_or_fileobj=card.encode("utf-8"),
            repo_id=args.repo_id,
            repo_type="dataset",
            path_in_repo="README.md",
            commit_message="Add dataset card",
        )

    url = f"https://huggingface.co/datasets/{args.repo_id}"
    print(f"Uploaded {dest}. View dataset: {url}")
    return 0


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