Improving the performance of the samplings utils (see #175)

[pc494]

Description of the change

Improving the performance of the gridding utils by allowing a uniform sampling of a subspace (see #175 for details). I've also renamed a file to be more consistent with what I imagine this part of the code will look like in the future.

Progress of the PR

• Docstrings for all functions
• Unit tests with pytest for all lines
• Clean code style by running black via pre-commit

For reviewers

• The PR title is short, concise, and will make sense 1 year later.
• New functions are imported in corresponding __init__.py.
• New features, API changes, and deprecations are mentioned in the
  unreleased section in CHANGELOG.rst.

[pc494]

While I'm here I'm going to tidy up the class method "random":

orix/orix/quaternion/rotation.py

Line 530 in 692c54a

def random(cls, shape=(1,)):

which is quiet sloppy at the moment.

I'll also look into the

orix/orix/quaternion/rotation.py

Line 549 in 692c54a

def random_vonmises(cls, shape=(1,), alpha=1.0, reference=(1, 0, 0, 0)):

functionality.

[pc494]

While I'm here I'm going to tidy up the class method "random":

orix/orix/quaternion/rotation.py

Line 530 in 692c54a

def random(cls, shape=(1,)):

which is quiet sloppy at the moment.

I'll also look into the

orix/orix/quaternion/rotation.py

Line 549 in 692c54a

def random_vonmises(cls, shape=(1,), alpha=1.0, reference=(1, 0, 0, 0)):

functionality.

Going to transfer this into an issue to be dealt with when the large refactor comes in at some point in the future.

[hakonanes]

Thanks for looking into the rotation sampling, @pc494.

I haven't fully understood the maths involved here, but the sample returned with angles of about 1-7 degrees accompanied by a resolution in the order of 0.5 degrees returns an empty sampling. Could you explain this to me?

Apart from this, I like that you've renamed the utilities to SO3, so that we can easily fit SO2 in the future into this orix.sampling module.

[hakonanes]

int(num_steps * u_1_max) becomes zero when u_1_max is small (see comment to get_sample_local()). Is there some way to handle the case where max angle is small (1-7 degrees) and a resolution of below 1 degree?

[pc494]

Yup, putting a crude fix in for this

[pc494]

Thanks @hakonanes for this review, it's very good. It'll take me a little while to loop back to this but when I do I think most of the style changes will be simple to implement. The version in 0.6 is going to be a bit lacking (although we can fix the bug that gives back empty regions) and I've got a sense of what a more complete implementation will look like (probably for 0.7.0)

[hakonanes]

That sounds like a good plan. As mentioned in #175, there is no rush.

[pc494]

Working on these locally, will be up shortly.

[pc494]

Yup, putting a crude fix in for this

[hakonanes]

Good stuff! I'll let you merge yourself, @pc494.