Datasets:

Stellaromics
/

demo

Tags:

License:

Dataset card Files Files and versions

xet

Community

Dataset Viewer

Duplicate

Split (1)

test

The dataset viewer is not available for this split.

Cannot load the dataset split (in streaming mode) to extract the first rows.

Error code:   StreamingRowsError
Exception:    TypeError
Message:      Couldn't cast array of type
struct<ome: struct<version: string, multiscales: list<item: struct<datasets: list<item: struct<path: string, coordinateTransformations: list<item: struct<type: string, scale: list<item: double>, translation: list<item: double>>>>>, name: string, axes: list<item: struct<name: string, type: string>>, coordinateTransformations: list<item: struct<type: string, transformations: list<item: struct<type: string, scale: list<item: double>, input: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>, output: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>, translation: list<item: double>>>, input: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>, output: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>>>>>, omero: struct<channels: list<item: struct<label: string>>>>, spatialdata_attrs: struct<version: string>>
to
{}
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2951, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 547, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2312, in cast_table_to_schema
                  cast_array_to_feature(
                  ~~~~~~~~~~~~~~~~~~~~~^
                      table[name] if name in table_column_names else pa.array([None] * len(table), type=schema.field(name).type),
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                      feature,
                      ^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1861, in wrapper
                  return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
                                           ~~~~^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2158, in cast_array_to_feature
                  raise TypeError(f"Couldn't cast array of type\n{_short_str(array.type)}\nto\n{_short_str(feature)}")
              TypeError: Couldn't cast array of type
              struct<ome: struct<version: string, multiscales: list<item: struct<datasets: list<item: struct<path: string, coordinateTransformations: list<item: struct<type: string, scale: list<item: double>, translation: list<item: double>>>>>, name: string, axes: list<item: struct<name: string, type: string>>, coordinateTransformations: list<item: struct<type: string, transformations: list<item: struct<type: string, scale: list<item: double>, input: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>, output: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>, translation: list<item: double>>>, input: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>, output: struct<name: string, axes: list<item: struct<name: string, type: string, unit: string>>>>>>>, omero: struct<channels: list<item: struct<label: string>>>>, spatialdata_attrs: struct<version: string>>
              to
              {}

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Stellaromics Pyxa demo dataset

Output from the Stellaromics Pyxa platform, used as a demo dataset for the spatialdata-io pyxa reader (experimental). No patient-identifiable information.

Datasets

Each folder is a self-contained dataset with the same layout.

Folder	Extent (x × y × z)	Cells	Transcripts	Size	Purpose
`colon/`	~5180 × 4230 × 128 µm	492,155	65,447,961	~30 GB	whole human colon region, 1,020-gene panel
`small/`	~1310 × 443 × 130 µm	4,372	395,667	~290 MB	full demo region
`xsmall/`	100 × 100 × 100 µm crop	187	23,495	~8 MB	CI tests and quick visual checks

colon/ is one region (Run01 / Analysis02 / A2) of the Glasgow colon H1K run, a human colon section measured for 1,020 genes. It also carries the Pyxa Studio export, pyxa_studio_v1.csv, for the 358,173 cells Pyxa Studio kept. See Colon attribution.

xsmall/ is generated from small/ by scripts/make_xsmall.py: the cube spans x ∈ [900, 1000), y ∈ [-5140, -5040), z ∈ [20, 120) µm. Cells whose centroid falls inside the cube are kept whole (all of their segmentation polygons and assigned transcripts, so the per-cell counts stay consistent with the transcripts); unassigned transcripts and the DAPI mosaic are cropped to the cube.

Files

cell_assigned_gene_v1.csv: transcript-level gene assignments (points layer). Unassigned transcripts have a cell_id ending in _-1.
cell_by_gene_v1.csv: per-cell gene expression counts (table layer).
cell_metadata_v1.csv: per-cell metadata (volume, spatial coordinates).
segmentation_geometries_v1.parquet: per-cell, per-z-plane segmentation polygons (shapes layer), in pixel units.
mosaic_3d.ome.zarr.zip: multiscale DAPI mosaic image (OME-Zarr v0.5, zipped per Hugging Face's Zarr storage guidance).
pyxa_studio_v1.csv (colon/ only): Pyxa Studio's cell filter and clusters, one row per kept cell: Cluster and a 3D UMAP (X_UMAP, Y_UMAP, Z_UMAP).

Usage

The pyxa reader (spatialdata-io, experimental) finds each file in the folder, and reads the zipped mosaic in place (no unzip needed).

from huggingface_hub import snapshot_download
from spatialdata_io.experimental import pyxa

local_dir = snapshot_download("Stellaromics/demo", repo_type="dataset", allow_patterns="xsmall/*")
sdata = pyxa(f"{local_dir}/xsmall")

Replace xsmall with small for the full region. Pass labels=True to rasterize the segmentation polygons into 3D cell_labels on the mosaic's voxel grid, annotated by the table.

For colon/, fetch only what you need: the transcripts alone are 7.7 GB and the mosaic 12 GB. pyxa_studio_v1.csv adds the Pyxa Studio clusters and UMAP.

files = ["cell_by_gene_v1.csv", "cell_metadata_v1.csv", "pyxa_studio_v1.csv",
         "segmentation_geometries_v1.parquet", "mosaic_3d.ome.zarr.zip"]
local_dir = snapshot_download(
    "Stellaromics/demo", repo_type="dataset", allow_patterns=[f"colon/{f}" for f in files]
)
sdata = pyxa(f"{local_dir}/colon", labels=True)
sdata.write("colon.sdata.zarr")  # ~8 min and ~30 GB peak on a workstation

Leave out segmentation_geometries_v1.parquet and mosaic_3d.ome.zarr.zip (and labels=True) for the table alone. The pyxa_scverse_demo colon notebook analyses this folder with scanpy and the spatial-rx Landmarks widget.

Colon attribution

School of Cancer Sciences, University of Glasgow, UK: Marta Campillo Poveda, Anthony Chalmers, Yoana Doncheva, Joanne Edwards, Andrea Gonzalez Ciscar, Nigel Jamieson, Claire Kennedy Dietrich, Ghazal Latif, Assya Legrini, Josefina Marinez Vasquez, Pamela McCall, Mari-Claire McGuigan, Luke McNickle, Tengyu Zhang

University of Edinburgh, UK: Gerry Thompson

Stellaromics Inc, Boston, MA, USA: Leah Carlson, Jeremy Lambert, Clarence Mah, Raghav Padmanabhan, Chan Park, Daphne Sze, Alexis Wong

Contact: Clarence Mah, Stellaromics (@ckmah, clarence.mah@stellaromics.com)

License

BSD 3-Clause.

Downloads last month: 203

Total file size:

30.7 GB