Update app.py

app.py

@@ -14,6 +14,7 @@ import soundfile as sf
 from tenacity import retry, stop_after_attempt, wait_exponential
 import re
 import numpy as np
+from huggingface_hub import hf_hub_download
 
 # Log in with Hugging Face token if available
 token = os.getenv("hf_token")
@@ -36,14 +37,42 @@ audio_backup = {}  # To store original audio paths for each sample
 
 # Load saved labels from JSON file (and convert old formats if needed)
 def load_saved_labels():
+    # First, try to load from the local file system.
     if os.path.exists(SAVE_PATH):
-        with open(SAVE_PATH, "r", encoding="utf-8") as f:
-            data = json.load(f)
-            for key, value in data.items():
-                if isinstance(value, str):
-                    data[key] = {"transcript": value, "reviewer": "unreviewed"}
-            return data
-    return {}
+        try:
+            with open(SAVE_PATH, "r", encoding="utf-8") as f:
+                data = json.load(f)
+        except Exception as e:
+            print("Error loading local JSON file: " + str(e))
+            data = {}
+    else:
+        data = {}
+
+    # If no local file found or it's empty, try to download from the HF dataset repo.
+    if not data:
+        try:
+            # Download the file from HF repo (make sure repo_type="dataset" is specified)
+            hf_path = hf_hub_download(
+                repo_id=HF_DATASET_NAME,
+                filename=SAVE_PATH,
+                repo_type="dataset",
+                token=token
+            )
+            with open(hf_path, "r", encoding="utf-8") as f:
+                data = json.load(f)
+            # Write the downloaded file to the local path for caching.
+            with open(SAVE_PATH, "w", encoding="utf-8") as f:
+                json.dump(data, f, ensure_ascii=False, indent=4)
+            print("Loaded JSON file from HF dataset repository and cached it locally.")
+        except Exception as e:
+            print("Error loading JSON file from HF repo: " + str(e))
+            data = {}
+    
+    # Convert any old string-formats to the new dict format.
+    for key, value in data.items():
+        if isinstance(value, str):
+            data[key] = {"transcript": value, "reviewer": "unreviewed"}
+    return data
 
 saved_labels = load_saved_labels()