config-ipython/profile_default/startup/init.py

import itertools
import pathlib
import typing

import numpy
import scipy.io.wavfile
from matplotlib import pyplot  # pylint: disable=unused-import; frequently used in shell


def numpy_array_from_file(
    path: typing.Union[str, pathlib.Path], dtype
) -> numpy.ndarray:
    if isinstance(path, str):
        path = pathlib.Path(path)
    return numpy.frombuffer(path.read_bytes(), dtype=dtype)


def split_sequence_by_delimiter(
    sequence: typing.Sequence, delimiter: typing.Any, delimiter_min_length: int = 1
) -> typing.Iterator[typing.Sequence]:
    slice_start_index, slice_length = 0, 0
    for is_delimiter, group in itertools.groupby(
        sequence, key=lambda item: item == delimiter
    ):
        group_length = sum(1 for _ in group)
        if is_delimiter and group_length >= delimiter_min_length:
            if slice_length > 0:
                yield sequence[slice_start_index : slice_start_index + slice_length]
            slice_start_index += slice_length + group_length
            slice_length = 0
        else:
            slice_length += group_length
    if slice_length > 0:
        yield sequence[slice_start_index : slice_start_index + slice_length]


def trim_where(
    # https://docs.python.org/3.8/library/collections.abc.html#collections-abstract-base-classes
    sequence: typing.Sequence,
    condition: typing.Sequence[bool],
) -> typing.Sequence:
    start = 0
    for item_condition in condition:
        if item_condition:
            start += 1
        else:
            break
    stop = len(sequence)
    assert stop == len(condition)
    for item_condition in condition[::-1]:
        if item_condition:
            stop -= 1
        else:
            break
    return sequence[start:stop]


def wavfile_read_mono(
    path: typing.Union[pathlib.Path, str]
) -> typing.Tuple[int, numpy.ndarray]:
    # https://docs.scipy.org/doc/scipy/reference/generated/scipy.io.wavfile.read.html
    rate, data = scipy.io.wavfile.read(path)
    if len(data.shape) == 1:
        return rate, data
    data_first_channel = data[:, 0]
    for channel_index in range(1, data.shape[1]):
        assert (data_first_channel == data[:, channel_index]).all()
    return rate, data_first_channel