from __future__ import annotations
import uuid
from itertools import groupby, chain
from collections.abc import Iterable
import numpy as np
from asr_eval.utils.types import FLOATS, INTS
__all__ = [
'new_uid',
'groupby_into_spans',
'list_join',
'rolling_window',
'locate_subarray_in_array',
]
[docs]
def new_uid() -> str:
"""A unique ID generator."""
return str(uuid.uuid4())[:16]
[docs]
def groupby_into_spans[T](
iterable: Iterable[T]
) -> Iterable[tuple[T, int, int]]:
"""Find spans of the same value in a sequence. Returns (value,
start_index, end_index).
Example:
>>> list(groupby_into_spans(['x', 'x', 'b', 'a', 'a', 'a']))
[('x', 0, 2), ('b', 2, 3), ('a', 3, 6)]
"""
for key, group in groupby(enumerate(iterable), key=lambda x: x[1]):
group = list(group)
yield key, group[0][0], group[-1][0] + 1
[docs]
def list_join[T](sep: T, iterable: Iterable[T]) -> list[T]:
"""Combines iterables via a given separator. Acts like str.join,
but for lists.
"""
it = iter(iterable)
try:
first = [next(it)]
except StopIteration:
return []
return list(chain(
first,
*([sep, x] for x in it),
))
[docs]
def rolling_window[T: (INTS, FLOATS)](arr: T, size: int) -> T:
"""Returns all subarrays of length :code:`size`, stacked together
along a new axis.
Example:
>>> rolling_window(np.array([1, 0, 2, 1, 3, 5]), 3) # doctest: +NORMALIZE_WHITESPACE
array([[1, 0, 2], [0, 2, 1], [2, 1, 3], [1, 3, 5]])
Taken from: https://stackoverflow.com/a/7100681
"""
shape = arr.shape[:-1] + (arr.shape[-1] - size + 1, size)
strides = arr.strides + (arr.strides[-1],)
return np.lib.stride_tricks.as_strided(arr, shape=shape, strides=strides) # type: ignore
[docs]
def locate_subarray_in_array[T: (INTS, FLOATS)](
arr: T, subarr: T
) -> list[int]:
"""Finds all positions X where :code:`arr[X:X+len(subarr)]` equals
:code:`subarr`, in effiecient way.
"""
assert arr.ndim == 1
assert subarr.ndim == 1
mask = np.all(rolling_window(arr, size=len(subarr)) == subarr, axis=-1)
indices, = np.where(mask)
return indices.tolist()
