Allow reading of EDAX TSL .ang files with ten columns

[hakonanes]

Description of the change

EDAX TSL .ang files with ten data columns should be read correctly with this patch.

Fixes #413, where it was identified that two openly available .ang files, shown in #411 (comment), were read incorrectly. With this patch, there is a correct correspondence between phase ID and phase list, and the scan unit is "um". Big thanks to @argerlt who pointed me to these datasets.

I've made another small change to IO scan units. Previously, ASTAR/orix/unknown .ang files returned maps with "nm" as scan unit. Now, all but ASTAR .ang files return maps with "um" as scan unit. As far as I know there is no scan unit information .ang files. Hence, we have to guess, and I think the new guess is better.

I'd like to release this as part of a 0.10.3 patch soon (see #415).

Progress of the PR

• [n/a] Docstrings for all functions
• Unit tests with pytest for all lines
• Clean code style by running black via pre-commit
• "Revert" 0.10.3 commits (hope to make a 0.11 release instead)

Minimal example of the bug fix or new feature

Previously

>>> from orix import io

>>> xmap = io.load("IN100_Raw/ANG_Series/Slice_001.ang")
/home/hakon/kode/orix/orix/io/plugins/ang.py:268: UserWarning: Number of columns, 10, in the file is not equal to the expected number of columns, 14, for the 
assumed vendor 'tsl'. Will therefore assume the following columns: euler1, euler2, euler3, x, y, unknown1, unknown2, phase_id, unknown3, unknown4, etc.
  warnings.warn(
>>> xmap
Phase    Orientations    Name  Space group  Point group  Proper point group     Color
    0  37989 (100.0%)  Nickel         None          432                 432  tab:blue
Properties: unknown1, unknown2, unknown3, unknown4
Scan unit: nm

>>> xmap2 = io.load("Field of view 1_EBSD data_Raw.ang")
/home/hakon/kode/orix/orix/io/plugins/ang.py:268: UserWarning: Number of columns, 10, in the file is not equal to the expected number of columns, 14, for the 
assumed vendor 'tsl'. Will therefore assume the following columns: euler1, euler2, euler3, x, y, unknown1, unknown2, phase_id, unknown3, unknown4, etc.
  warnings.warn(
>>> xmap2
Phase      Orientations       Name  Space group  Point group  Proper point group       Color
    0         29 (0.0%)  Austenite         None          432                 432  tab:orange
    1      14900 (0.7%)    Ferrite         None          432                 432    tab:blue
    2   2202639 (99.3%)       None         None         None                None   tab:green
Properties: unknown1, unknown2, unknown3, unknown4
Scan unit: nm

Now

>>> from orix import io

>>> xmap = io.load("IN100_Raw/ANG_Series/Slice_001.ang")
>>> xmap
Phase    Orientations         Name  Space group  Point group  Proper point group     Color
   -1    7912 (20.8%)  not_indexed         None         None                None         w
    0   30077 (79.2%)       Nickel         None          432                 432  tab:blue
Properties: iq, ci, unknown1, fit
Scan unit: um

>>> xmap2 = io.load("Field of view 1_EBSD data_Raw.ang")
>>> xmap2
Phase      Orientations         Name  Space group  Point group  Proper point group       Color
   -1         29 (0.0%)  not_indexed         None         None                None           w
    1      14900 (0.7%)    Austenite         None          432                 432  tab:orange
    2   2202639 (99.3%)      Ferrite         None          432                 432    tab:blue
Properties: iq, ci, unknown1, fit
Scan unit: um

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.
• Contributor(s) are listed correctly in __credits__ in orix/__init__.py and in
  .zenodo.json.

[hakonanes]

The two failing tests on Windows Py 3.9/10 are fixed on develop in #410:

I've also bumped the minimum version of diffpy.structure to 3.0.2 in setup.py. This ensures that Phase.from_cif() works on Windows with Python 3.10. I've added a changelog entry for this change.

I think this fix should be included in a potential patch release.

[argerlt]

I believe changing unknown1 to sem is the correct way to read TSL files. example below:

            0: [
                "euler1",
                "euler2",
                "euler3",
                "x",
                "y",
                "iq",  # Image quality from Hough transform
                "ci",  # Confidence index
                "phase_id",
                "sem",
                "fit",  # Pattern fit
                "unknown1",
                "unknown2",
                "unknown3",
                "unknown4",
            ],
            1: [
                "euler1",
                "euler2",
                "euler3",
                "x",
                "y",
                "iq",
                "ci",
                "phase_id",
                "sem",
                "fit",
            ],

[hakonanes]

You're right, thanks! I've renamed it to "detector_signal", as I find this more descriptive.

[hakonanes]

The failing tests are test_symmetry_plot() in which the number of Matplotlib collections are checked in the axis returned from Symmetry.plot() (via the figure). The expected number does not match the actual in these tests.

I can reproduce these errors locally after updating to Matplotlib 3.6.3. When running the code in the tests outside of tests, I get the expected number of collections, though. So I do not know what causes the extra collections when running with pytest.

The errors happen when running on both one or more cores (via pytest-xdist).

I think easing the test criteria is the way to go. Will try to find a fix for this in this PR.

[pc494]

Looks good to me, glad the workflows have got an update as well :)

[hakonanes]

Thank you, @pc494!