bep40

/

V.AISTUDIO

like 0

Auto-Sync Infrastructure - Pull Request

Purpose

Automate nightly sync (~23,400 products) from central master dataset (bep40/grob-products-updated) to both HuggingFace storefront spaces:

• V.AISTUDIO (primary storefront)
• vai-avatar2 (secondary storefront)

Files Added

1. sync_products.py - Python script that:

   • Loads latest product data from HF dataset bep40/grob-products-updated
   • Computes diff against existing cached products
   • Uploads ONLY new/updated products to both spaces
   • Triggers Space rebuild via .gitignore touch for cache busting

2. .github/workflows/auto-sync.yml - GitHub Actions workflow that:

   • Runs daily at 00:00 UTC via cron schedule
   • Uses actions/setup-python@v5 for Python environment
   • Installs dependencies: huggingface_hub, datasets
   • Executes python3 sync_products.py
   • Includes proper error handling and logging

How It Works

malloca.com update → HF Dataset updated → Cron triggers → 
Python sync script diffs → Pushes NEW products only → 
Triggers Space rebuild on BOTH spaces

Requirements

• HF_TOKEN secret must be configured in Repo Settings → Secrets & variables → Actions
• Token needs write scope for both bep40/V.AISTUDIO and bep40/vai-avatar2 spaces

Key Features

• ✅ Incremental sync: Only uploads changed/new products (slug-based diff detection)
• ✅ Cache-busting: Touches .gitignore to force Space rebuild after sync
• ✅ Image download & upload: Downloads product images from source URLs
• ✅ Error handling: Logs warnings for failed image downloads without breaking sync
• ✅ Progress tracking: Updates every 50 products during upload
• ✅ Minimal payload: Optimized HTML templates for fast rendering
  "