Data Compression

The library comes with some basic data compression routines. They are helpful in simple use cases like:

• compression of binary arrays

• run length encoding

• fixed length encoding of integers

These routines are primarily based on numpy arrays and bitarray based compressed bit arrays. This module doesn’t use JAX.

count_runs_values(input_arr)	Computes run lengths of an array of integers
expand_runs_values(runs, values)	Decodes run lengths to form an array of integers
encode_int_arr_sgn_mag_fl(input_arr, …)	Encodes an array of integers to a bit array using a sign bit and a fixed number of bits per sample for magnitude
decode_int_arr_sgn_mag_fl(input_bit_arr, …)	Decodes an array of integers from a bit array using a sign bit and a fixed number of bits per sample for magnitude
count_binary_runs(input_arr)	Returns the runs of 0s and 1s in a binary map
encode_binary_arr(input_arr)	Encodes a binary array into a bit array via run length encoding
decode_binary_arr(input_bit_arr)	Decodes a binary array from a bit array via run length decoding
binary_compression_ratio(input_arr, output_arr)	Returns the compression ratio of binary array compression algorithm
binary_space_saving_ratio(input_arr, output_arr)	Returns the space saving ratio of binary array compression algorithm