API

The central class:

AnnData([X, obs, var, uns, obsm, varm, ...])

An annotated data matrix.

Combining

Combining AnnData objects. See also the section on concatenation.

concat(adatas, *[, axis, join, merge, ...])

Concatenates AnnData objects along an axis.

Reading anndata’s native file format .h5ad.

read_h5ad(filename[, backed, as_sparse, ...])

Read .h5ad-formatted hdf5 file.

Reading other file formats.

`read_csv`(filename[, delimiter, ...])	Read .csv file.
`read_excel`(filename, sheet[, dtype])	Read .xlsx (Excel) file.
`read_hdf`(filename, key)	Read .h5 (hdf5) file.
`read_loom`(filename, *[, sparse, cleanup, ...])	Read .loom-formatted hdf5 file.
`read_mtx`(filename[, dtype])	Read .mtx file.
`read_text`(filename[, delimiter, ...])	Read .txt, .tab, .data (text) file.
`read_umi_tools`(filename[, dtype])	Read a gzipped condensed count matrix from umi_tools.
`read_zarr`(store)	Read from a hierarchical Zarr array store.

Writing to anndata’s native file format .h5ad.

AnnData.write([filename, compression, ...])

Write .h5ad-formatted hdf5 file.

Writing to other formats.

`AnnData.write_csvs`(dirname[, skip_data, sep])	Write annotation to `.csv` files.
`AnnData.write_loom`(filename[, write_obsm_varm])	Write `.loom`-formatted hdf5 file.
`AnnData.write_zarr`(store[, chunks])	Write a hierarchical Zarr array store.

Warning

API’s in the experimenal module are currently in development and subject to change at any time.

Two classes for working with batched access to collections of many AnnData objects or h5ad files. In paritcular, for pytorch-based models.

`experimental.AnnCollection`(adatas[, ...])	Lazily concatenate AnnData objects along the `obs` axis.
`experimental.AnnLoader`(adatas[, batch_size, ...])	PyTorch DataLoader for AnnData objects.

Low level methods for reading and writing elements of an AnnData` object to a store:

`experimental.read_elem`(elem)	Read an element from a store.
`experimental.write_elem`(store, k, elem, *[, ...])	Write an element to a storage group using anndata encoding.

Utilities for customizing the IO process:

`experimental.read_dispatched`(elem, callback)	Read elem, calling the callback at each sub-element.
`experimental.write_dispatched`(store, key, ...)	Write elem to store, recusively calling callback at each sub-element.
`experimental.IOSpec`(encoding_type, ...)

Raised whenever initializing an object or assigning a property changes the type of a part of a parameter or the value being assigned.