Calling sdata.pl.render_points("transcripts").pl.show() on a SpatialData object containing Xenium data raises:
ValueError: cannot reindex on an axis with duplicate labels
Observation:
I noticed that the points element index of the SpatialData object is not unique. More generally, when reading the transcripts.parquet file into a Dask DataFrame (Dask 2026.1.1), the resulting index is not globally unique. Instead, indices appear to be enumerated independently within partitions, leading to duplicate labels across the full dataframe.
This may be related to the error above, although I am not certain that this is the index involved in the failing reindex operation.