"""Upload GenSegDataset (processed_unified mirror + dataset card) to the Hugging
Face Hub as a PRIVATE dataset repo. Storage is content-addressed, so fold-duplicated
images collapse to unique blobs. Resumable via upload_large_folder."""
import sys
from huggingface_hub import HfApi

REPO = "MaybeRichard/GenSegDataset"
DATA = "/home/wzhang/LSC/Dataset/Segmentation/processed_unified"
CARD = "/home/wzhang/LSC/Dataset/Segmentation/GenSegDataset_README.md"

api = HfApi()
api.create_repo(REPO, repo_type="dataset", private=True, exist_ok=True)
print("repo ready:", REPO, flush=True)

# big data upload (resumable, dedups blobs, parallel)
api.upload_large_folder(repo_id=REPO, repo_type="dataset", folder_path=DATA)
print("folder uploaded", flush=True)

# dataset card last so it is the final README at the repo root
api.upload_file(path_or_fileobj=CARD, path_in_repo="README.md",
                repo_id=REPO, repo_type="dataset",
                commit_message="add dataset card")
print("UPLOAD_DONE", flush=True)