`threshold_firm` returns NaN at the threshold when `value_low == value_high`

[whyvineet]

pywt.threshold_firm returns NaN for input values whose magnitude is exactly equal to the threshold in the case where value_low is equal to value_high. This is unexpected because the function's docstring clearly states that when value_high equals value_low, threshold_firm is equivalent to hard thresholding.

pywt/pywt/_thresholding.py

Line 212 in cc45b0d

If ``value_high == value_low`` this function becomes hard-thresholding.

Reproducer

import numpy as np
import pywt

data = np.array([1.0, 2.0, -2.0, 3.0])

print(pywt.threshold_firm(data, 2.0, 2.0))

Current result:

[ 0. nan nan  3.]

Expected result:

[ 0.  2. -2.  3.]

which is the same result as:

pywt.threshold(data, 2.0, mode="hard")