bpcells.experimental.DirMatrix#

class bpcells.experimental.DirMatrix(dir: str)[source]#

Disk-backed BPCells integer matrix

This reads BPCells-format matrices, returning scipy.sparse.csc_matrix objects when sliced.

Parameters:: dir (str) – Path of the matrix directory

Examples

>>> from bpcells import DirMatrix
>>> mat = DirMatrix("/path/to/matrix")
>>> mat[:,[1,3,2,4]]
<3x4 sparse matrix of type '<class 'numpy.uint32'>'
    with 6 stored elements in Compressed Sparse Column format>

Attributes

`DirMatrix.T`	Return a transposed view of the matrix
`DirMatrix.shape`	Dimensions of matrix
`DirMatrix.threads`	Number of threads to use for reading (default=1)

Methods

`DirMatrix.from_h5ad`(h5ad_path, out_dir[, group])	Create a DirMatrix from an h5ad file.
`DirMatrix.from_hstack`(mats, out_dir)	Create a DirMatrix by concatenating a list of DirMatrix objects horizontally (column wise)
`DirMatrix.from_scipy_sparse`(scipy_mat, dir)	Create a DirMatrix from a scipy sparse matrix.
`DirMatrix.from_vstack`(mats, out_dir)	Create a DirMatrix by concatenating a list of DirMatrix objects vertically (row wise)
`DirMatrix.transpose`()	Return at transposed view of the matrix