MRG: Add support for NIRSport2 devices

[rob-luke]

Reference issue

Closes #9371
Related to mne-tools/mne-testing-data#86

What does this implement/fix?

1. Adds support for reading data from NIRSport2 devices.
2. Modifications to snirf reader such that it works with nirx created snirf files.

Additional information

This device can export simultaneously as SNIRF and vendor format. So I made a simultaneous recording and then ensure that data from the two files match if read with MNE.
The data is identical (as demonstrated in tests). However, the metadata is different across files (e.g. subject ID). I am in contact with the vendor to discuss this. So I don't think it's a reader bug.

[rob-luke]

Tests pass locally. I will need to poke CIs again when testing data is merged.

[larsoner]

https://github.com/mne-tools/mne-testing-data/releases/tag/0.121

[rob-luke]

Thanks

[rob-luke]

@larsoner could you please review this PR. I think the failing tests are due to https://github.com/mne-tools/mne-testing-data/releases/tag/0.120 not anything ive done.

[larsoner]

I think the failing tests are due to https://github.com/mne-tools/mne-testing-data/releases/tag/0.120 not anything ive done.

Yes for now just delete the line

mne\io\egi\tests\test_egi.py:372: in test_io_egi_evokeds_mff
    assert evoked_cond.info['dig'] is None

and it'll be fixed properly in #9406

I also pushed a fix that was causing failures in two of the CIs

[rob-luke]

@larsoner and @agramfort could you please review (thanks larsoner for the first pass getting things green).

[agramfort]

just a nitpick

[rob-luke]

@larsoner I think I hit all your suggestions. Please review and merge if happy.

[larsoner]

I think the last thing is latest.inc then we're good to go, thanks in advance @rob-luke !