idx = (minimum_yaw_angle_subset > 0.0) | (maximum_yaw_angle_subset < 0.0)
if np.any(idx):
# Find bounds closest to 0.0 deg
combined_bounds = np.concatenate(
(
np.expand_dims(minimum_yaw_angle_subset, axis=3),
np.expand_dims(maximum_yaw_angle_subset, axis=3)
),
axis=3
)
# Overwrite all values that are not allowed to be 0.0 with bound value closest to zero
ids_closest = np.expand_dims(np.argmin(np.abs(combined_bounds), axis=3), axis=3)
yaw_mb = np.squeeze(np.take_along_axis(combined_bounds, ids_closest, axis=3))
yaw_angles_template_subset[idx] = yaw_mb[idx]
were missed in the v4 dimension reduction process #764 , and can be fixed by replacing with
idx = (minimum_yaw_angle_subset > 0.0) | (maximum_yaw_angle_subset < 0.0)
if np.any(idx):
# Find bounds closest to 0.0 deg
combined_bounds = np.concatenate(
(
np.expand_dims(minimum_yaw_angle_subset, axis=2),
np.expand_dims(maximum_yaw_angle_subset, axis=2)
),
axis=2
)
# Overwrite all values that are not allowed to be 0.0 with bound value closest to zero
ids_closest = np.expand_dims(np.argmin(np.abs(combined_bounds), axis=2), axis=2)
yaw_mb = np.squeeze(np.take_along_axis(combined_bounds, ids_closest, axis=2), axis=2)
yaw_angles_template_subset[idx] = yaw_mb[idx]
(that is, switching from axis=3 to axis=2 throughout, and also specifying the axis for the np.squeeze operation).
The reason this hasn't come up before is that this code only runs if minimum_yaw_angle is strictly greater 0 or maximum_yaw_angle is strictly less than 0, which is an unusual circumstance.
I will create a bugfix PR in due course that will include tests for these "unusual" minimum and maximum angles.
These lines
were missed in the v4 dimension reduction process #764 , and can be fixed by replacing with
(that is, switching from
axis=3toaxis=2throughout, and also specifying the axis for thenp.squeezeoperation).The reason this hasn't come up before is that this code only runs if
minimum_yaw_angleis strictly greater 0 ormaximum_yaw_angleis strictly less than 0, which is an unusual circumstance.I will create a bugfix PR in due course that will include tests for these "unusual" minimum and maximum angles.