Add BDF export

[cbrnr]

Now that edfio 0.4.10 added support for BDF, we should support that in MNE as well. The current implementation duplicates a lot of EDF code, which I think is OK for now (I'd prefer to refactor in a follow-up PR).

[cbrnr]

Any ideas why the ultraslow_pg job is failing?

The failing Windows pip pre job is due to pyvistaqt DeprecationWarnings, so this is unrelated I guess. The Ubuntu pip pre job segfaults, also due to some VTK issue.

[larsoner]

The current implementation duplicates a lot of EDF code, which I think is OK for now (I'd prefer to refactor in a follow-up PR).

It looks like _bdf.py and _edf.py only differ by < 10 lines. Can I push a commit just to take care of this now? Seems simpler and cleaner than adding a bunch of lines just to almost immediately remove it

[cbrnr]

Sure, but if you want I can also refactor it.

[larsoner]

Sure feel free! On Azure this failure does look related:

_____________________ ERROR collecting mne/export/_bdf.py ______________________
mne/export/_bdf.py:13: in <module>
    _check_edfio_installed()
mne/utils/check.py:451: in _check_edfio_installed
    return _soft_import("edfio", "exporting to EDF", strict=strict)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
mne/utils/check.py:429: in _soft_import
    raise RuntimeError(
E   RuntimeError: For exporting to EDF to work, the module edfio is needed, but it could not be imported. Use the following installation method appropriate for your environment:
E   
E       pip install edfio
E       conda install -c conda-forge edfio
---------- generated xml file: /home/vsts/work/1/s/junit-results.xml -----------
=========================== short test summary info ============================
ERROR mne/export/_bdf.py - RuntimeError: For exporting to EDF to work, the module edfio is needed, but it could not be imported. Use the following installation method appropriate for your environment:

    pip install edfio
    conda install -c conda-forge edfio

but maybe the refactoring already took care of it. I'll tackle the unrelated errors in #13434

[cbrnr]

I don't understand this error. The new test_export_raw_bdf function has the exact same markers as test_export_raw_edf (i.e., @edfio_mark() and pytest.mark.slowtest in particular). So why is this test collected in the first place?

[cbrnr]

Found it and hopefully fixed it.

[cbrnr]

Looks good, feel free to merge @larsoner!

[drammock]

LGTM, just a couple minor things

[drammock]

IIUC, doing it this way will change the order of the items in the annotations list (previously, all the BAD_ACQ_SKIP` would be at the beginning, now they will be at the end). Does that matter? Or does EDF/BDF not care (or does the writer sort the annots in any case)?

[cbrnr]

I thought it didn't really matter (the docs are not clear or I couldn't find that info). @hofaflo?