ehrdata.integrations.vitessce.gen_default_config

ehrdata.integrations.vitessce.gen_default_config(edata, zarr_filepath=PosixPath('adata_for_vitessce.zarr'), *, obs_columns=None, obs_embedding=None, scatter_var_cols=None, layer='tem_data', timestep=0, return_lamin_artifact=False)

Quickstart interactive Vitessce generator.

Vitessce [KGM+25] is a tool for interactive exploration of high-dimensional data, and compatible with EHRData.

While Vitessce has many features, this function provides a convenient way for an opinionated set of illustrations for an EHRData to explore interactively together(called “views”). Specifically, this function will create a Vitessce widget with multiple views:

• A view with patient groups of selected columns in edata.obs (obs_columns)

• A list of variables to display values for (obs_embedding)

• A bar plot for the number of categories of groups selected during the interactive exploration

• A scatterplot of the selected obs_embedding if provided

• A scatterplot of 2 variables if provided in scatter_var_cols

• A violin plot of selected variables across groups selected during the interactive exploration

• A histogram for selected variables during exploration

• A heatmap of the variables selected in var_cols

See the vitessce-python documentation for more details and examples.

Parameters:

edata EHRData

EHRData object to visualize

zarr_filepath Store | StorePath | FSMap | Path | str | dict[str, Buffer] | Path | str (default: PosixPath('adata_for_vitessce.zarr'))

Path to save the prepared zarr file that Vitessce can read from.

obs_columns Iterable[str] | None (default: None)

List of observation column names (without ‘obs/’ prefix)

obs_labels

Optional dict mapping column names to display labels

obs_embedding str | None (default: None)

Embedding key in edata.obsm

obs_embedding_labels

Optional dict mapping embedding keys to display labels

scatter_var_cols Iterable[str] | None (default: None)

Optional list of 2 variable columns to create ascatterplot from

layer default: 'tem_data'

Name of the layer to use for visualization. If the layer is 3D (temporal), a timestep must be selected. Default is “tem_data”

timestep default: 0

For 3D layers, the timestep index to extract. Default is 0

return_lamin_artifact bool (default: False)

If True, return a Lamin Artifact of the generated .zarr file.

Returns:

VitessceConfig object

Examples

>>> import ehrdata as ed
>>> edata = ed.dt.physionet2019(
...     layer="tem_data",
...     n_samples=4000,
... )
>>> vc = ed.integrations.vitessce.gen_default_config(
...     edata,
...     obs_columns=["Gender", "Age", "training_Set"],
...     scatter_var_cols=["HR", "MAP"],
...     layer="tem_data",
...     timestep=10,
... )
>>> vc.widget()