""" download_from_hf.py Run this on your new (i9/3080ti) laptop to pull down the trimmed UCSF-PDGM dataset you uploaded to HuggingFace (unknownBrain/ucsf). Install first: pip install -U huggingface_hub If the dataset repo is private, also run: huggingface-cli login Run in two steps: python download_from_hf.py list -> just prints file list/structure python download_from_hf.py download -> actually downloads everything """ import sys from huggingface_hub import HfApi, snapshot_download HF_REPO_ID = "unknownBrain/ucsf" LOCAL_DIR = r"D:\jainam\UCSF-PDGM-v5" # where you want it locally api = HfApi() def list_files(): files = api.list_repo_files(repo_id=HF_REPO_ID, repo_type="dataset") print(f"total files: {len(files)}") for f in files[:30]: print(" ", f) if len(files) > 30: print(f" ... and {len(files) - 30} more") # This tells you whether folder structure survived the upload # (e.g. 'UCSF-PDGM-004/UCSF-PDGM-004_T1c.nii.gz') or got flattened # (e.g. 'UCSF-PDGM-004_T1c.nii.gz' with no subfolder) -- update # dataset.py's build_records() path template to match whichever it is. def download(): path = snapshot_download( repo_id=HF_REPO_ID, repo_type="dataset", local_dir=LOCAL_DIR, local_dir_use_symlinks=False, # actual files, not symlinks -- safer on Windows # allow_patterns=["*_T1c.nii.gz", "*_T2.nii.gz", "*_tumor_segmentation.nii.gz"], # ^ uncomment + edit if you only want specific modalities right now ) print("Downloaded to:", path) if __name__ == "__main__": mode = sys.argv[1] if len(sys.argv) > 1 else "list" if mode == "list": list_files() elif mode == "download": download() else: print("usage: python download_from_hf.py [list|download]")