Update translator_utils.py

translator_utils.py

@@ -6,10 +6,16 @@ from huggingface_hub import HfApi
 import spacy
 from spacy.lang.en import English
 
-# URL of the dataset
-DATASET_URL = "https://huggingface.co/datasets/Brian045/data_berita.csv/resolve/main/data_berita.csv"
-REPO_ID = "Brian045/data_berita"
-LOCAL_FILENAME = "data_berita.csv"
+# --- Configuration & Globals ---
+# Using persistent storage path if available in typical HF spaces, or fallback
+DATA_FILE = "data_berita.csv" 
+# Initial Env Vars (Fallback)
+ENV_HF_TOKEN = os.getenv("HF_TOKEN", "")
+# Default to Brian045/data_berita if env var is not set, but prefer env var
+ENV_REPO_ID = os.getenv("DATASET_REPO_ID", "Brian045/data_berita")
+
+DATASET_URL = f"https://huggingface.co/datasets/{ENV_REPO_ID}/resolve/main/{DATA_FILE}"
+LOCAL_FILENAME = DATA_FILE
 
 # Global model cache
 _translator = None
@@ -27,7 +33,8 @@ def download_dataset():
             print("Download complete.")
         except Exception as e:
             print(f"Error downloading dataset: {e}")
-            raise e
+            # Do not raise, just print error and create empty if needed or handle gracefully
+            # raise e 
     else:
         # print(f"File {LOCAL_FILENAME} already exists.")
         pass
@@ -36,19 +43,22 @@ def upload_dataset(token):
     """
     Uploads the local CSV file back to the Hugging Face Hub.
     """
-    if not token:
+    # Use passed token, or fallback to env var
+    token_to_use = token if token else ENV_HF_TOKEN
+    
+    if not token_to_use:
         print("No HF Token provided. Skipping upload.")
         return "Skipped Upload (No Token)"
     
-    print(f"Uploading {LOCAL_FILENAME} to {REPO_ID}...")
+    print(f"Uploading {LOCAL_FILENAME} to {ENV_REPO_ID}...")
     try:
         api = HfApi()
         api.upload_file(
             path_or_fileobj=LOCAL_FILENAME,
-            path_in_repo="data_berita.csv",
-            repo_id=REPO_ID,
+            path_in_repo=DATA_FILE,
+            repo_id=ENV_REPO_ID,
             repo_type="dataset",
-            token=token,
+            token=token_to_use,
             commit_message="Update translations via Auto AI Translator"
         )
         print("Upload successful!")
@@ -191,8 +201,14 @@ def process_rows(indices, token=None, progress=None):
     df.to_csv(LOCAL_FILENAME, index=False)
     
     upload_status = "Skipped Upload"
-    if token:
-        upload_status = upload_dataset(token)
+    
+    # Use explicit token if provided, else rely on upload_dataset to pick up env var if None passed
+    # But UI usually passes "" if empty. 
+    token_arg = token if token and token.strip() else None
+    
+    # Always attempt upload if we have a token from somewhere (arg or env)
+    # logic is handled in upload_dataset
+    upload_status = upload_dataset(token_arg)
         
     return df, f"Processed {count} rows. {upload_status}"
 
@@ -217,8 +233,7 @@ def revert_rows(indices, token=None, progress=None):
     print("Saving to CSV...")
     df.to_csv(LOCAL_FILENAME, index=False)
     
-    upload_status = "Skipped Upload"
-    if token:
-        upload_status = upload_dataset(token)
+    token_arg = token if token and token.strip() else None
+    upload_status = upload_dataset(token_arg)
         
     return df, f"Reverted {len(indices)} rows. {upload_status}"