anndata2ri

Converter between Python’s AnnData and R’s SingleCellExperiment.

SingleCellExperiment		AnnData
assay(d, 'X')	⇄	d.X
assay(d, 'counts')	⇄	d.layers['counts']
colData(d)	⇄	d.obs
rowData(d)	⇄	d.var
metadata(d)	⇄	d.uns
reducedDim(d, 'PCA')	⇄	d.obsm['X_pca']
reducedDim(d, 'DM')	⇄	d.obsm['X_diffmap']

anndata2ri.activate()

Activate conversion for supported objects.

This includes AnnData objects as well as Array objects and pandas.DataFrames via rpy2.robjects.numpy2ri and rpy2.robjects.pandas2ri.

Does nothing if this is the active converter.

Return type:

None

anndata2ri.deactivate()

Deactivate the conversion described above if it is active.

Return type:

None

anndata2ri.py2rpy(obj)

Convert Python objects to R interface objects.

Supports: :rtype: Sexp

• AnnData → SingleCellExperiment

anndata2ri.rpy2py(obj)

Convert R interface objects to Python objects.

Supports: :rtype: AnnData | DataFrame

• SingleCellExperiment → AnnData

• S4Vectors::DataFrame → pandas.DataFrame

anndata2ri.scipy2ri

Convert scipy.sparse matrices between Python and R.

For a detailed comparison between the two languages’ sparse matrix environment, see issue #8.

Here’s an overview over the matching classes

R	Python
dgCMatrix	csc_matrix(dtype=float64)
lgCMatrix/ngCMatrix	csc_matrix(dtype=bool)
dgRMatrix	csr_matrix(dtype=float64)
lgRMatrix/ngRMatrix	csr_matrix(dtype=bool)
dgTMatrix	coo_matrix(dtype=float64)
lgTMatrix/ngTMatrix	coo_matrix(dtype=bool)
ddiMatrix	dia_matrix(dtype=float64)
ldiMatrix	dia_matrix(dtype=bool)

anndata2ri.scipy2ri.activate()

Activate conversion between sparse matrices from Scipy and R’s Matrix package.

Does nothing if this is the active conversion.

Return type:

None

anndata2ri.scipy2ri.deactivate()

Deactivate the conversion described above if it is active.

Return type:

None

anndata2ri.scipy2ri.py2rpy(obj)

Convert scipy sparse matrices objects to R sparse matrices.

Supports:

Return type:

Sexp

csc_matrix (dtype in {float32, float64, bool}) →

dgCMatrix or lgCMatrix

csr_matrix (dtype in {float32, float64, bool}) →

dgRMatrix or lgRMatrix

coo_matrix (dtype in {float32, float64, bool}) →

dgTMatrix or lgTMatrix

dia_matrix (dtype in {float32, float64, bool}) →

ddiMatrix or ldiMatrix

anndata2ri.scipy2ri.rpy2py(obj)

Convert R sparse matrices to scipy sparse matrices.

Supports:

Return type:

spmatrix

dgCMatrix, lgCMatrix, or ngCMatrix →

csc_matrix (dtype float64 or bool)

dgRMatrix, lgRMatrix, or ngRMatrix →

csr_matrix (dtype float64 or bool)

dgTMatrix, lgTMatrix, or ngTMatrix →

coo_matrix (dtype float64 or bool)

ddiMatrix or ldiMatrix →

dia_matrix (dtype float64 or bool)

anndata2ri.scipy2ri.supported_r_matrix_types = frozenset({'d', 'l', 'n'})

The Matrix data types supported by this module; Double, Logical, and patterN.

anndata2ri.scipy2ri.supported_r_matrix_storage = frozenset({'C', 'R', 'T', 'di'})

The Matrix storage types supported by this module; Column-sparse, Row-Sparse, Triplets, and DIagonal.

anndata2ri.scipy2ri.supported_r_matrix_classes(types=frozenset({'d', 'l', 'n'}), storage=frozenset({'C', 'R', 'T', 'di'}))

Get supported classes, possibly limiting data types or storage types.

Parameters:

types Union[Iterable[Literal['d', 'l', 'n']], Literal['d', 'l', 'n']] (default: frozenset({'d', 'n', 'l'}))

Data type character(s) from supported_r_matrix_types

storage Union[Iterable[Literal['C', 'R', 'T', 'di']], Literal['C', 'R', 'T', 'di']] (default: frozenset({'di', 'C', 'R', 'T'}))

Storage mode(s) from supported_r_matrix_storage

Return type:

frozenset[str]

Returns:

All supported classes with those characters