If I have a numpy array of length n that looks like this
[2.456970453262329, 2.3738574981689453,... 1.28790283203125,]
and I want to "stretch" is so that it now has length m (m>n) but still keeps the same general distribution of values, how could I do that? Is there a numpy module for such a transformation?
e.g. for n=3 and m=5
[1, 2, 3] --> [1, 1.5, 2, 2.5, 3]
I think you are just looking for linear interpolation. np.interp will do the trick:
import numpy as np
# wrapper function for convenience
def interp1d(array: np.ndarray, new_len: int) -> np.ndarray:
la = len(array)
return np.interp(np.linspace(0, la - 1, num=new_len), np.arange(la), array)
print(interp1d(np.asarray([1, 2, 3]), new_len=5)) # [1. 1.5 2. 2.5 3. ]
print(interp1d(np.asarray([0, 2, 3]), new_len=6)) # [0. 0.8 1.6 2.2 2.6 3. ]
# can even shorten arrays as well
print(interp1d(np.asarray([0, -2, 1, 3]), new_len=3)) # [ 0. -0.5 3. ]
Provided that the spacing between each element in the array is always equal regardless of how much "stretching" you apply, it sounds like arange will work for your use case since you can just specify a step https://numpy.org/doc/stable/reference/generated/numpy.arange.html#numpy.arange