update

app.py

@@ -5,11 +5,50 @@ import os
 import shutil
 import uuid
 import glob
-
-from huggingface_hub import CommitScheduler, HfApi
+from huggingface_hub import CommitScheduler, HfApi, snapshot_download
+from pathlib import Path
 
 api = HfApi(token=os.environ["HF_TOKEN"])
 
+
+# Download existing data from hub
+def sync_with_hub():
+    """
+    Synchronize local data with the hub by downloading latest dataset
+    """
+    print("Starting sync with hub...")
+    data_dir = Path("./data")
+    if data_dir.exists():
+        # Backup existing data
+        backup_dir = Path("./data_backup")
+        if backup_dir.exists():
+            shutil.rmtree(backup_dir)
+        shutil.copytree(data_dir, backup_dir)
+
+    # Download latest data from hub
+    repo_path = snapshot_download(
+        repo_id="taesiri/zb_dataset_storage", repo_type="dataset", local_dir="hub_data"
+    )
+
+    # Merge hub data with local data
+    hub_data_dir = Path(repo_path) / "data"
+    if hub_data_dir.exists():
+        # Create data dir if it doesn't exist
+        data_dir.mkdir(exist_ok=True)
+
+        # Copy files from hub
+        for item in hub_data_dir.glob("*"):
+            if item.is_dir():
+                dest = data_dir / item.name
+                if not dest.exists():  # Only copy if doesn't exist locally
+                    shutil.copytree(item, dest)
+
+    # Clean up downloaded repo
+    if Path("hub_data").exists():
+        shutil.rmtree("hub_data")
+    print("Finished syncing with hub!")
+
+
 scheduler = CommitScheduler(
     repo_id="taesiri/zb_dataset_storage",
     repo_type="dataset",
@@ -884,4 +923,8 @@ with gr.Blocks() as demo:
         ],
     )
 
-demo.launch()
+if __name__ == "__main__":
+    print("Initializing app...")
+    sync_with_hub()  # Sync before launching the app
+    print("Starting Gradio interface...")
+    demo.launch()