fixed ellipse orientation in beam_size function

[przemororo]

Hello, I'm having trouble calculating the width of a laser beam in a specific rotated coordinate system. In function: laserbeamsize.analysis.beam_size(image, mask_diameters=3, corner_fraction=0.035, nT=3, max_iter=25, phi=None, iso_noise=True) there is phi argument which fix the ellipse azimuth. But looks like it doesn't work. Here is example based on beam generation using build in function laserbeamsize.image_tools.create_test_image(h, v, xc, yc, dx, dy, phi, noise=0, ntype='poisson', max_value=255)

import numpy as np
import laserbeamsize as lbs

beam = lbs.create_test_image(h = 600, v = 600, xc = 300, yc = 300, dx = 150, dy = 100, phi = np.pi/4)

xc, yc, dx, dy, phi = lbs.beam_size(beam, phi = 0)
print("Phi = 0")
print(xc, yc, dx, dy, phi/np.pi)
xc, yc, dx, dy, phi = lbs.beam_size(beam, phi = np.pi/2)
print("Phi = pi/4")
print(xc, yc, dx, dy, phi/np.pi)
xc, yc, dx, dy, phi = lbs.beam_size(beam, phi = np.pi/4)
print("Phi = pi/4")
print(xc, yc, dx, dy, phi/np.pi)

The beam is titled by -45 degrees from x axis and looks:

The output is:

Phi = 0
300.7037284975563 300.0015631385715 146.96479605836058 97.9867640173793 0.0
Phi = pi/4
300.7037284975563 300.0015631385715 146.96479605836058 97.9867640173793 0.5
Phi = pi/4
300.7037284975563 300.0015631385715 146.96479605836058 97.9867640173793 0.25

It seems that regardless of the declared value of the ellipse orientation, it always looks for the optimal one even if you want to calculate in specified axis. For example if I want calculate beam width in x and y axis (phi = 0 or pi), the width should be the same:

xc = 300
yc = 300
x = np.arange(0, 600)
X, Y = np.meshgrid(x, x)
dx = 4 * np.sqrt(np.sum((X - xc)**2 * beam) / np.sum(beam))
dy = 4 * np.sqrt(np.sum((Y - yc)**2 * beam) / np.sum(beam))
print(dx, dy)

The output is:
124.93714721914387 124.89471657232691