tfmindi.save_h5ad

Contents

tfmindi.save_h5ad#

tfmindi.save_h5ad(adata, filename, compression=None, compression_opts=None, as_dense=None, **kwargs)#

Save AnnData object to H5AD format with proper handling of numpy arrays in .obs and .var.

This function wraps AnnData.write_h5ad() with additional preprocessing to handle numpy arrays stored in .obs and .var columns, which would otherwise cause HDF5 serialization errors. The numpy arrays are temporarily converted to string representations for serialization, with metadata stored to restore them during loading.

Parameters:

  • adata (AnnData) – AnnData object to save

  • filename (str | Path) – Path to the output H5AD file

  • compression (str | None (default: None)) – Compression algorithm to use (e.g., ‘gzip’, ‘lzf’)

  • compression_opts (int | None (default: None)) – Compression options

  • as_dense (str | None (default: None)) – Write sparse data as dense arrays

  • **kwargs – Additional arguments passed to AnnData.write_h5ad()

Return type:

None

Examples

>>> import tfmindi as tm
>>> tm.save_h5ad(adata, "my_data.h5ad")
>>> tm.save_h5ad(adata, "my_data.h5ad", compression="gzip")
Contents