scripts/deploy_both.py

"""
Push to both Hugging Face Dataset and Space.

Dataset: https://huggingface.co/datasets/scholo/MMB_dataset
Space:   https://huggingface.co/spaces/scholo/Datasetviewer

Usage:
  pip install datasets pillow huggingface_hub
  huggingface-cli login

  python scripts/deploy_both.py
  python scripts/deploy_both.py --dataset-only    # Only push dataset
  python scripts/deploy_both.py --space-only      # Only push Space
"""
from __future__ import annotations

import argparse
import subprocess
import sys
from pathlib import Path

PROJECT_ROOT = Path(__file__).resolve().parent.parent
HF_DATASET_DIR = PROJECT_ROOT / "hf_dataset"
DATASET_REPO = "scholo/MMB_dataset"
SPACE_REPO = "scholo/Datasetviewer"


def find_hf_dataset_csv() -> Path | None:
    """Find CSV in hf_dataset/ (root or in subdirs)."""
    if not HF_DATASET_DIR.exists():
        return None
    csvs = list(HF_DATASET_DIR.rglob("*.csv"))
    with_q = [p for p in csvs if "question" in p.name.lower()]
    return (with_q[0] if with_q else csvs[0]) if csvs else None


def run(cmd: list[str], cwd: Path | None = None) -> bool:
    r = subprocess.run(cmd, cwd=cwd or PROJECT_ROOT)
    return r.returncode == 0


def push_dataset() -> bool:
    print("=== Pushing dataset to", DATASET_REPO, "===")
    csv_path = find_hf_dataset_csv()
    if not csv_path:
        print(f"Error: No CSV in {HF_DATASET_DIR}. Add image_mapping_with_questions.csv + images/ + scenes/", file=sys.stderr)
        return False
    print(f"Using {csv_path.relative_to(PROJECT_ROOT)}")
    return run([
        sys.executable,
        str(PROJECT_ROOT / "scripts" / "upload_to_huggingface.py"),
        str(csv_path),
        "--repo-id", DATASET_REPO,
    ])


def push_space() -> bool:
    print("=== Pushing Space to", SPACE_REPO, "===")
    try:
        from huggingface_hub import HfApi
    except ImportError:
        print("Install huggingface_hub: pip install huggingface_hub", file=sys.stderr)
        return False

    api = HfApi()

    try:
        api.upload_folder(
            folder_path=str(PROJECT_ROOT),
            path_in_repo="",
            repo_id=SPACE_REPO,
            repo_type="space",
            ignore_patterns=[
                "__pycache__", "*.pyc", ".git",
                "scripts", "requirements-upload.txt",
                "hf_dataset",  # Dataset has its own folder; Space uses data/
                "*.zip", ".cursor", "*.pyo",
            ],
            commit_message="Deploy MMB Dataset Visualizer",
        )
        print("Space pushed. View at: https://huggingface.co/spaces/" + SPACE_REPO)
        return True
    except Exception as e:
        print(f"Error pushing Space: {e}", file=sys.stderr)
        return False


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("--dataset-only", action="store_true")
    parser.add_argument("--space-only", action="store_true")
    args = parser.parse_args()

    ok = True
    if not args.space_only:
        ok = push_dataset() and ok
    if not args.dataset_only:
        ok = push_space() and ok

    sys.exit(0 if ok else 1)


if __name__ == "__main__":
    main()