Add exception handling for URI download failures

- Added try-except block in load_uri_bytes function to handle invalid URIs
- Catch specific exceptions (OSError, pa.ArrowException) for better error handling
- Failed downloads now return None instead of crashing the pipeline
- Added warning log message for failed URI downloads

Author: xyuzh

python/ray/data/_internal/planner/plan_download_op.py

@@ -189,8 +189,14 @@ def download_bytes_threaded(
         def load_uri_bytes(uri_path_iterator):
             """Function that takes an iterator of URI paths and yields downloaded bytes for each."""
             for uri_path in uri_path_iterator:
-                with fs.open_input_file(uri_path) as f:
-                    yield f.read()
+                try:
+                    with fs.open_input_file(uri_path) as f:
+                        yield f.read()
+                except (OSError, pa.ArrowException) as e:
+                    logger.warning(
+                        f"Failed to download URI '{uri_path}' from column '{uri_column_name}' with error: {e}"
+                    )
+                    yield None
 
         # Use make_async_gen to download URI bytes concurrently
         # This preserves the order of results to match the input URIs