Summary

The Download operator currently shows up as "URIDownloader" in progress bars. This should be updated to show "Download[uri_column_name]" to make it clearer which column is being downloaded.

Current Behavior

When using the Download operator, progress bars display:

URIDownloader

Desired Behavior

Progress bars should display:

Download(uri_column_name)

Where uri_column_name is the user-specified column being downloaded.

Motivation

The current "URIDownloader" name:

• Doesn't clearly communicate what's happening in the operation
• Makes it harder to debug or understand which column is being processed
• Doesn't follow the naming pattern of other operators that show their parameters

The proposed change would:

• Make progress bars more informative and self-documenting
• Help users quickly identify which column is being downloaded
• Improve the debugging experience when multiple Download operations are running
• Align with naming conventions used by other Ray Data operators (e.g., Map(fn_name))

Example

If downloading from column "image_url", the progress bar would show:

Download(image_url)

If there are multiple URIs specified, the progress bar could show:

Download(image_url, thumbnail_url)

(Note that multiple URIs isn't currently exposed as a user-facing API yet, but it's still something you should support for internal correctness).

Implementation Sketch

Files to Modify

Primary file: /python/ray/data/_internal/planner/plan_download_op.py

Changes Required

1. Update the download operator name (Line 131 in plan_download_op.py):

   # Current code (line 131):
   name="URIDownloader",
   
   # Should be changed to:
   name=f"Download[{op.uri_column_name}]",

2. Optionally update the partition operator name (Line 89 in plan_download_op.py) for consistency:

   # Current code (line 89):
   name="URIPartitioner",
   
   # Could be changed to:
   name=f"Partition[{uri_column_name}]",

Code Context

The relevant section in plan_download_op.py (lines 127-134):

download_map_operator = MapOperator.create(
    download_map_transformer,
    partition_map_operator if partition_map_operator else input_physical_dag,
    data_context,
    name="URIDownloader",  # <- Change this line
    compute_strategy=download_compute,
    ray_remote_args=ray_remote_args,
)

Testing

Should add/update tests in /python/ray/data/tests/test_download_expression.py to verify the progress bar displays the correct operator name format.