make_intervals

dtaianomaly.utils.make_intervals(x: ~numpy.array) -> (<built-in function array>, <built-in function array>)

Numba implementation to convert an array into intervals.

Given a binary array, i.e., an array consisting of only 0’s and 1’s, this method will extract all intervals of consecutive o1’s from the array. The start-index corresponds to the index (0-based) of the first 1 in the interval, the end-index is the index of the last 1 in the interval. The method outputs two independent arrays, one for the start indices and one for the end incides.

Parameters:

xnp.array

The array from which the intervals must be computed. The array should consist of solely 0’s and 1’s.

Returns:

starts: np.array of shape (I,)

The start index of each interval. In total, there are I intervals.

ends: np.array of shape (I,)

The end index of each interval. In total, there are I intervals.

Examples

>>> import numpy as np
>>> from dtaianomaly.utils import make_intervals
>>> starts, ends = make_intervals(np.array([0, 1, 1, 1, 0, 0, 1, 0]))
>>> starts
array([1, 6])
>>> ends
array([3, 6])