Skip to content

[BUG] waveorder writes ome-zarr data as XY images rather than XYZ chunks #550

Closed

Description

@ieivanov

In apply_inverse_transger_function.py we have

def get_reconstruction_output_metadata(position_path: Path, config_path: Path):
    ...
    
    return {
        "shape": (T, len(channel_names), output_z_shape, Y, X),
        "chunks": (1, 1, 1, Y, X),  # <--- chunks set to XY planes
        "scale": input_dataset.scale,
        "channel_names": channel_names,
        "dtype": np.float32,
        "plate_metadata": plate_metadata,
        "version": input_version,
    }

Is there a reason we don't use iohub create_empty_plate here?

Activity

  1. ieivanov commented on Apr 15, 2026

    @ieivanov
    ContributorAuthor

    Looks like this has been the case for a while. This creates a fare bit of file pollution on bruno.

  2. talonchandler commented on Apr 15, 2026

    @talonchandler
    Collaborator

    Thanks for flagging. Related to #540.

  3. ieivanov commented on Apr 15, 2026

    @ieivanov
    ContributorAuthor

    Very likely I'll take care of this before Taylla get back, otherwise she can resolve this next week

  4. linked a pull request that will close this issueDefer chunk/shard choice to iohub #555on Apr 23, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions