Utility functions

This module contains miscellaneous utility functions.

Array utilities

`batch_arrays_by_length`(arrays_x, y[, ...])	Batches a list of arrays of different sizes, grouping them by size.
`clip_arrays`()	Clips each array to the specified maximum sequence length.
`flat_to_inst`(x, slices)	Takes a concatenated 2D data array and converts it to either a contiguous 2D/3D array or a variable-length 3D array, with one feature vector/matrix per instance.
`frame_array`(x, frame_size, frame_shift[, ...])	Frames an array over a given axis with optional padding and copying.
`frame_arrays`(arrays, frame_size, frame_shift)	Creates sequences of frames from the given arrays.
`inst_to_flat`(x)	The inverse of flat_to_inst().
`is_mono_audio`(x)	Checks if an array represents mono audio data.
`make_array_array`(x)	Helper function to make an array of arrays.
`pad_arrays`()	Pads each array to the nearest multiple of `pad` greater than the array size.
`shuffle_multiple`(*arrays[, numpy_indexing, seed])	Shuffles multiple arrays or lists in sync.
`transpose_time`()	Transpose the time and feature axis of each array.

`batch_iterable`(it, batch_size[, return_last])	Batches an iterable into chunks of size `batch_size`.
`filter_kwargs`(kwargs, method)	Removes incompatible keyword arguments.
`itmap`(s)	Returns a new map function that additionally maps tuples to tuples and lists to lists.
`ordered_intersect`(a, b)	Returns a list of the intersection of `a` and `b`, in the order elements appear in `a`.
`ordered_subsets`(it[, max_size])	Iterate over all ordered subsets of the iterable `it`, up to a given maximum size.
`subsets`(it[, max_size])	Iterate over all subsets of the iterable `it`, up to a given maximum size.

`TqdmMultiprocessing`([total, desc, unit, ...])
`TqdmParallel`([total, desc, unit, leave])	Convenience class that acts identically to joblib.Parallel except it uses a tqdm progress bar.

`PathOrStr`
`ScoreFunction`