Stack slices of numpy array from given indices

Question

I'm struggling to perform the below operation on a numpy vector.

I want to take previous_n samples from vector finishing at indices.

It's like I want to perform a np.take with slicing of the previous_n samples.

Example:

import numpy as np

vector = np.array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14])

# number of previous samples
previous_n = 3

indices = np.array([ 5,  7, 12])

result

array([[ 3,  4,  5],
       [ 5,  6,  7],
       [10, 11, 12]])

@liliscent a `slice(value - previous_n, value)` for value in `indices` — James Schinner, Jan 28 '18 at 07:28
Wrapping np.array around a list comprehension is competitive. — hpaulj, Jan 28 '18 at 07:35
or broadcast [-2, -1, 0] against `indices` to create the index matrix. — hpaulj, Jan 28 '18 at 07:37

James Schinner · Accepted Answer · 2018-01-28T07:45:24.277

0

Ok, this seems to do what I want. Found here

def stack_slices(arr, previous_n, indices):
    all_idx = indices[:, None] + np.arange(previous_n) - (previous_n - 1)
    return arr[all_idx]

>>> stack_slices(vector, 3, indices)
array([[ 3,  4,  5],
       [ 5,  6,  7],
       [10, 11, 12]])

edited Jan 28 '18 at 07:45

answered Jan 28 '18 at 07:37

James Schinner

1,549
18
28

ups... I accidentally downvoted. My mistake. Is it possible to edit the answer so I can correct my mistake? – Nima Mousavi Jan 28 '18 at 17:15

Stack slices of numpy array from given indices

1 Answers1