Find out why DirectionalBias correction for sum of sinusoids randomly fails

[rhugonnet]

This happens even though the random_state is fixed, maybe an issue to open in SciPy's optimize.basinhopping after some digging...

Here is the documentation section I'm removing for now:

The default optimizer for directional biases fits a sum of sinusoids using 1 to 3 different frequencies and
keeps the best performing fit, which is useful for periodic along-track errors common to DEMs.

:tags: [hide-cell]
:mystnb:
:  code_prompt_show: "Show the code for adding a sum of sinusoids bias"
:  code_prompt_hide: "Hide the code for adding a sum of sinusoids bias"

# Get rotated coordinates along an angle
angle = 20
xx = gu.raster.get_xy_rotated(ref_dem, along_track_angle=angle)[0]

# One sinusoid: amplitude, phases and frequencies
params = np.array([(15, 10000, np.pi)]).flatten()

# Create a sinusoidal bias and add to the DEM
from xdem.fit import sumsin_1d
synthetic_bias = sumsin_1d(xx.flatten(), *params).reshape(np.shape(ref_dem.data))
tbc_dem_sumsin = ref_dem + synthetic_bias

# Define a directional bias correction at a certain angle (degrees), defaults to "bin_and_fit" for a sum of sinusoids
dirbias = xdem.coreg.DirectionalBias(angle=20)
# Fit and apply
corrected_dem = dirbias.fit_and_apply(ref_dem, tbc_dem_sumsin, random_state=42)

:tags: [hide-input]
:mystnb:
:  code_prompt_show: "Show plotting code"
:  code_prompt_hide: "Hide plotting code"

# Plot before and after
f, ax = plt.subplots(1, 2)
ax[0].set_title("Before directional\nde-biasing")
(tbc_dem_sumsin - ref_dem).plot(cmap='RdYlBu', vmin=-30, vmax=30, ax=ax[0])
ax[1].set_title("After directional\nde-biasing")
(corrected_dem - ref_dem).plot(cmap='RdYlBu', vmin=-30, vmax=30, ax=ax[1], cbar_title="Elevation differences (m)")
_ = ax[1].set_yticklabels([])