download_dataset.py

#!/usr/bin/env python3
"""Download the SpanishBCBL dataset (MEG + EEG typing recordings) from the Hugging Face Hub.

Repository: https://huggingface.co/datasets/bcbl190626/SpanishBCBL

Examples
--------
# Download everything (~262 GB) into ./SpanishBCBL
python download_dataset.py

# Download only the MEG modality into a custom folder
python download_dataset.py --modality meg --local-dir /data/SpanishBCBL

# Pass a token explicitly (otherwise uses your cached login / HF_TOKEN env var)
python download_dataset.py --token hf_xxx

Notes
-----
- Requires `huggingface_hub`:  pip install -U "huggingface_hub[cli]"
- While the repository is private you must be authenticated: run `hf auth login`
  once, set the HF_TOKEN environment variable, or pass --token.
- Downloads resume automatically: re-running skips files already present.
"""

import argparse
import sys

REPO_ID = "bcbl190626/SpanishBCBL"
REPO_TYPE = "dataset"

# Per-modality file filters (None = no filter, i.e. download everything).
MODALITY_PATTERNS = {
    "all": None,
    "meg": ["MEG/*", "README.md"],
    "eeg": ["EEG/*", "README.md"],
}


def parse_args() -> argparse.Namespace:
    p = argparse.ArgumentParser(
        description="Download the SpanishBCBL dataset from the Hugging Face Hub.",
        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
    )
    p.add_argument(
        "--local-dir",
        default="SpanishBCBL",
        help="Destination directory for the downloaded dataset.",
    )
    p.add_argument(
        "--modality",
        choices=sorted(MODALITY_PATTERNS),
        default="all",
        help="Which modality to download.",
    )
    p.add_argument(
        "--token",
        default=None,
        help="Hugging Face access token (defaults to cached login / HF_TOKEN env var).",
    )
    p.add_argument(
        "--revision",
        default=None,
        help="Specific git revision (branch, tag, or commit) to download.",
    )
    p.add_argument(
        "--workers",
        type=int,
        default=8,
        help="Number of parallel download workers.",
    )
    return p.parse_args()


def main() -> int:
    args = parse_args()

    try:
        from huggingface_hub import snapshot_download
    except ImportError:
        sys.exit(
            'huggingface_hub is not installed. Install it with:\n'
            '    pip install -U "huggingface_hub[cli]"'
        )

    allow_patterns = MODALITY_PATTERNS[args.modality]

    print(f"Repository : {REPO_ID} ({REPO_TYPE})")
    print(f"Modality   : {args.modality}")
    print(f"Destination: {args.local_dir}")
    if allow_patterns:
        print(f"Patterns   : {allow_patterns}")
    print("Starting download (this may take a while; downloads resume if interrupted)...\n")

    try:
        path = snapshot_download(
            repo_id=REPO_ID,
            repo_type=REPO_TYPE,
            local_dir=args.local_dir,
            allow_patterns=allow_patterns,
            revision=args.revision,
            token=args.token,
            max_workers=args.workers,
        )
    except Exception as exc:  # noqa: BLE001 - surface a friendly message
        msg = str(exc)
        if "401" in msg or "403" in msg or "gated" in msg.lower() or "authenticate" in msg.lower():
            sys.exit(
                f"Authentication/authorization failed: {exc}\n\n"
                "If the dataset is private, make sure you are logged in:\n"
                "    hf auth login\n"
                "or pass a token with --token / set the HF_TOKEN environment variable."
            )
        sys.exit(f"Download failed: {exc}")

    print(f"\nDone. Dataset available at: {path}")
    return 0


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