Preprocessing IO

Source: src/single_cell_python_tools/preprocessing/_io.py

These helpers are exported through sctl.pp.

reset_cellxgene_var_names

sctl.pp.reset_cellxgene_var_names(
    adata,
    feature_name_col="feature_name",
    **parameters,
)

Reset adata.var_names for cellxgene-style datasets. The function modifies the AnnData object in place, creates a feature_name_unique column, and stores the previous index in adata.var["cellxgene_var_index"].

downcast_layers_to_float32

sctl.pp.downcast_layers_to_float32(adata, **kwargs)

Convert AnnData layers, and optionally other matrices, to float32 in place. Use this when memory pressure matters and downstream analysis accepts float32 data.