yaolu-huggingface/SuperFocus

Step 1 — Install Dependencies

pip install huggingface-hub pandas

---

Step 2 — Login

from huggingface_hub import login
login(token="YOUR_HF_TOKEN")

---

Download Script

import os
from huggingface_hub import snapshot_download

def download_superfocus_dataset(ids=None, local_dir="./superfocus_prediction"):
    """
    Downloads files from the yaolu-huggingface/SuperFocus dataset.
    
    Parameters:
    - ids (list of str or str): Prefix IDs to filter files (e.g., ['INT13', 'MEND142'] or 'NCBI2').
                               If None, the entire dataset will be downloaded.
    - local_dir (str): The local directory where files should be saved.
    """
    repo_id = "yaolu-huggingface/SuperFocus"
    repo_type = "dataset"
    
    # Ensure the local directory exists
    os.makedirs(local_dir, exist_ok=True)
    
    # Handle the filtering pattern if ids are provided
    
    if len(ids)>0:
        # Build wildcard patterns (e.g., ['INT13*', 'MEND142*'])
        allow_patterns = [f"{prefix}*" for prefix in ids]
        print(f"Downloading files matching prefixes {ids} from {repo_id}...")
    else:
        allow_patterns = "*"
        print(f"Downloading the entire dataset from {repo_id} (Warning: total size is ~221 GB)...")

    # Download using snapshot_download
    path = snapshot_download(
        repo_id=repo_id,
        repo_type=repo_type,
        local_dir=local_dir,
        allow_patterns=allow_patterns
    )
    
    print(f"Download complete. Files saved to: {path}")
    return

# --- Example Usage ---

# 1. Download only specific subsets (e.g., files starting with INT13 or MEND142)
# download_superfocus_dataset(ids=["INT13", "MEND142"])

# 2. Download files matching a single ID prefix
# download_superfocus_dataset(ids="NCBI2")

# 3. Download everything (Make sure you have 221+ GB of free space!)
# download_superfocus_dataset(ids=None)

---

Usage Examples

Download One Sample

download_superfocus_dataset(
    ids=["NCBI689"]
)

---

Download Multiple Samples

download_superfocus_dataset(
    ids=["NCBI689", "MEND62"]
)

---

Download Entire Dataset

download_superfocus_dataset()

---