Upload sync_products.py

sync_products.py

@@ -0,0 +1,152 @@
+#!/usr/bin/env python3
+"""
+Product Auto-Sync Script — bep40/V.AISTUDIO ↔ bep40/vai-avatar2
+Mục đích: Tự động so sánh và cập nhật sản phẩm từ bep40/grob-products-updated dataset
+"""
+import json
+import os
+import sys
+from datetime import datetime, timezone
+
+try:
+    from huggingface_hub import HfApi, hf_hub_download
+except ImportError:
+    print("Installing huggingface_hub...")
+    os.system(f"{sys.executable} -m pip install huggingface_hub --quiet")
+    from huggingface_hub import HfApi, hf_hub_download
+
+def log(msg, level="INFO"):
+    ts = datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M:%S UTC")
+    print(f"[{level}] [{ts}] {msg}", flush=True)
+
+def slug_id(p):
+    for key in ("slug", "_source_alias", "sku", "mod", "model"):
+        val = p.get(key)
+        if val and str(val).strip():
+            return str(val).strip().lower()
+    return ""
+
+def load_dataset():
+    api = HfApi()
+    try:
+        path = hf_hub_download(repo_id="bep40/grob-products-updated", filename="products_with_slugs.json", repo_type="dataset")
+        with open(path, 'r', encoding='utf-8') as f:
+            products = json.load(f)
+        log(f"Loaded {len(products)} products from grob-products-updated dataset")
+        return products
+    except Exception as e:
+        log(f"Failed to load dataset: {e}", "ERROR")
+        return []
+
+def get_space_slugs(api, space_id):
+    slugs = set()
+    candidates = ["products_with_slugs.json", "vaix-products.json"]
+    for fname in candidates:
+        try:
+            path = hf_hub_download(repo_id=space_id, repo_type="space", filename=fname)
+            with open(path, 'r', encoding='utf-8') as f:
+                content = f.read()
+            if fname.endswith('.js'):
+                import re
+                match = re.search(r'\[.*\]', content, re.DOTALL | re.IGNORECASE)
+                if match:
+                    items = json.loads(match.group())
+                    for item in items:
+                        sid = slug_id(item)
+                        if sid: slugs.add(sid)
+            else:
+                items = json.loads(content)
+                for item in items:
+                    sid = slug_id(item)
+                    if sid: slugs.add(sid)
+            if slugs:
+                log(f"Found {len(slugs)} slugs in {space_id}/{fname}")
+                break
+        except:
+            continue
+    return slugs
+
+def sync_to_space(api, space_id, products, products_list=None):
+    """Sync a single space."""
+    if isinstance(products, dict):
+        product_list = products_list or []
+    else:
+        product_list = products
+    
+    payload = json.dumps(product_list, ensure_ascii=False, indent=2).encode('utf-8') if isinstance(product_list, list) else products.encode('utf-8') if isinstance(products, bytes) else json.dumps(products, ensure_ascii=False, indent=2).encode('utf-8')
+    
+    api.upload_file(
+        path_or_fileobj=payload,
+        path_in_repo="products_with_slugs.json",
+        repo_id=space_id,
+        repo_type="space",
+        commit_message=f"[AUTO-SYNC] Update products ({len(product_list) if isinstance(product_list, list) else len(json.loads(payload))} items)"
+    )
+    log(f"Updated {space_id}")
+    return True
+
+def sync_api_response(api, space_id, products_list=None):
+    """API Response mode for /api/vaix/json endpoint."""
+    payload = json.dumps(products_list, ensure_ascii=False, indent=2).encode('utf-8')
+    api.upload_file(
+        path_or_fileobj=payload,
+        path_in_repo="/api/vaix/json",
+        repo_id=space_id,
+        repo_type="space",
+        commit_message=f"[AUTO-SYNC] Update /api/vaix/json ({len(products_list)} items)"
+    )
+    log(f"Updated /api/vaix/json in {space_id}")
+    return True
+
+def trigger_rebuild(api, space_id):
+    ts = datetime.now(timezone.utc).isoformat()
+    content = f"# Sync rebuild at {ts}\nSpace: {space_id}\nSource: bep40/grob-products-updated\n"
+    flag = ".cache_bust" if "avatar" in space_id else "build-trigger"
+    try:
+        api.upload_file(
+            path_or_fileobj=content.encode('utf-8'),
+            path_in_repo=flag,
+            repo_id=space_id,
+            repo_type="space",
+            commit_message=f"[SYNC] Rebuild at {ts}"
+        )
+        log(f"Rebuild triggered for {space_id}")
+    except:
+        pass
+
+def main():
+    log("=" * 50)
+    log("V.AISTUDIO ↔ vai-avatar2 Auto-Sync")
+    log("=" * 50)
+    
+    api = HfApi()
+    
+    products = load_dataset()
+    if not products:
+        log("No products loaded", "ERROR")
+        sys.exit(1)
+    
+    total_prods = len(products)
+    prods_list = products if isinstance(products, list) else [products] if products else []
+    
+    # Update V.AISTUDIO with full products list
+    log("\n=== Updating V.AISTUDIO ===")
+    sync_to_space(api, "bep40/V.AISTUDIO", prods_list, prods_list)
+    trigger_rebuild(api, "bep40/V.AISTUDIO")
+    
+    # Also update /api/vaix/json for API endpoint compatibility
+    sync_api_response(api, "bep40/V.AISTUDIO", prods_list)
+    
+    # Update vai-avatar2
+    log("\n=== Updating vai-avatar2 ===")
+    sync_to_space(api, "bep40/vai-avatar2", prods_list, prods_list)
+    trigger_rebuild(api, "bep40/vai-avatar2")
+    
+    # Also update /api/vaix/json for API endpoint compatibility
+    sync_api_response(api, "bep40/vai-avatar2", prods_list)
+    
+    log(f"\n✅ Sync complete — Updated both spaces with {total_prods} products!")
+    log("=" * 50)
+
+if __name__ == "__main__":
+    main()