Update handler.py

handler.py

@@ -213,14 +213,29 @@ def TemporaryDirectory():
         shutil.rmtree(str(tmpdir))
 
 
-def download_file(s3_public_url: str, output_fn) -> Path:
-    """Download file from a public S3 bucket URL."""
+def download_file(s3_public_url: str, output_fn: str) -> Path:
+    """Download file from a public S3 bucket URL and ensure UTF-8 encoding."""
+    temp_fn = output_fn + ".temp"
+
+    # Download the file as binary data
     with requests.get(s3_public_url, stream=True) as r:
         r.raise_for_status()
-        with open(output_fn, "wb") as f:
+        with open(temp_fn, "wb") as f:
             for chunk in r.iter_content(chunk_size=8192):
                 f.write(chunk)
-    return output_fn
+
+    # Read the downloaded file and ensure UTF-8 encoding
+    with open(temp_fn, 'r', encoding='utf-8', errors='replace') as f:
+        content = f.read()
+
+    # Write the content back with correct encoding
+    with open(output_fn, 'w', encoding='utf-8') as f:
+        f.write(content)
+
+    # Clean up: remove the temporary file
+    Path(temp_fn).unlink()
+
+    return Path(output_fn)
 
 def _run_align_script(bo_fn, en_fn, output_dir):
     start = time.time()