sparcs-split

Manipulate existing single cell hdf5 datasets.

usage: sparcs-split [-h] [-o OUTPUT OUTPUT] [-r] [-t THREADS] [-c] input_dataset

Positional Arguments

input_dataset: input dataset which should be split

Named Arguments

-o, --output

Output definition <name> <length>. For example -o test.h5 0.9 or or -o test.h5 1000. If the sum of all lengths is <= 1, it is interpretated as fraction. Else its used as absolute value

-r, --random

shuffle single cells randomly

Default: False

-t, --threads

number of threads

Default: 4

-c, --compression

use lzf compression

Default: False

Manipulate existing SPARCSpy single cell hdf5 datasets. sparcs-split can be used for splitting, shuffleing and compression / decompression.

Examples

Splitting with shuffle and compression:

sparcs-split single_cells.h5 -r -c -o train.h5 0.9 -o test.h5 0.05 -o validate.h5 0.05

Shuffle

sparcs-split single_cells.h5 -r -o single_cells.h5 1.0

Compression

sparcs-split single_cells.h5 -c -o single_cells.h5 1.0

Decompression

sparcs-split single_cells.h5 -o single_cells.h5 1.0