upload_workspace.py

import os
import time
from pathlib import Path
from huggingface_hub import upload_folder

# Setup your core configurations
FOLDER_PATH = "/teamspace/studios/this_studio"
REPO_ID = "sonithoiningombam/bart"
REPO_TYPE = "model"

# Correct patterns to completely ignore all dotfiles and hidden folders (.venv, .git, etc.)
IGNORE_PATTERNS = ["**/.*", "**/.*/**"]

print("🚀 Scanning workspace directories to chunk the upload safely...")

# Step 1: Discover top-level visible items to isolate the commits
root_path = Path(FOLDER_PATH)
items_to_upload = [
    item for item in root_path.iterdir() 
    if not item.name.startswith('.')
]

print(f"📦 Found {len(items_to_upload)} root items to process sequentially.")

# Step 2: Loop through items and push them to prevent commit explosion
for index, item in enumerate(items_to_upload, start=1):
    print(f"\n🔄 [{index}/{len(items_to_upload)}] Uploading: {item.name}...")
    
    try:
        if item.is_file():
            # If it's a standalone file in the root
            upload_folder(
                folder_path=str(root_path),
                repo_id=REPO_ID,
                repo_type=REPO_TYPE,
                allow_patterns=[item.name],
                commit_message=f"Pushed file: {item.name}"
            )
        elif item.is_dir():
            # If it's a folder, upload its contents safely to its relative path
            upload_folder(
                folder_path=str(item),
                repo_id=REPO_ID,
                repo_type=REPO_TYPE,
                path_in_repo=item.name,
                ignore_patterns=IGNORE_PATTERNS,
                commit_message=f"Pushed directory segment: {item.name}"
            )
        print(f"✅ Finished uploading: {item.name}")
        
        # Step 3: Enforce a strict rest window to stay comfortably under 128 commits/hour
        print("🕒 Cooling down for 30 seconds to respect Hugging Face API rate limits...")
        time.sleep(30)

    except Exception as e:
        print(f"⚠️ Failed to upload {item.name} due to: {e}")
        print("💡 The script will continue to the next item. You can re-run later to retry failed objects.")

print("\n🎉 Workspace segment loops finished successfully!")