Ignore flake8 E741 (ambiguous variable name)

[hoechenberger]

Running flake8 on master gives me:

./tutorials/source-modeling/plot_background_freesurfer_mne.py:133:22: E741 ambiguous variable name 'I'
./tutorials/source-modeling/plot_background_freesurfer_mne.py:136:22: E741 ambiguous variable name 'I'

mne-python/tutorials/source-modeling/plot_background_freesurfer_mne.py

Lines 132 to 137 in d839a6c

	ori_slice = dict(P='Coronal', A='Coronal',
	I='Axial', S='Axial',
	L='Sagittal', R='Saggital')
	ori_names = dict(P='posterior', A='anterior',
	I='inferior', S='superior',
	L='left', R='right')

I don't think this rule is very useful (as we'll spot badly-named variables during the code review anyway), so this PR disables it.

[hoechenberger]

@larsoner Travis fails with

[gw0] linux -- Python 3.6.7 /home/travis/miniconda/envs/test/bin/python3.6
4043mne/tests/test_docstring_parameters.py:161: in test_docstring_parameters
4044    raise AssertionError(msg)
4045E   AssertionError: 
4046E   mne.source_space.SourceSpaces.clear : SS03 : Summary does not end with a period
4047E   mne.source_space.SourceSpaces.reverse : SS03 : Summary does not end with a period
4048E   2 errors
4049
4050

I checked and the docstrings look alright on Py 3.7 and 3.8, so this seems to be a Py 3.6 thing…
(SourceSpaces inherits from list and we don't have our own clear and reverse implementations)

[larsoner]

I would rather rename the variable and keep the check

[larsoner]

Or noqa it if it's a valuable exception

[hoechenberger]

@larsoner
But I think that the naming that causes the messages here makes total sense in the given context (dict key), and we don't want #noqa strings in a tutorial, do we?

That said, maybe there's an easier solution:
instead of creating the dict via

 ori_slice = dict(P='Coronal', A='Coronal', 
                  I='Axial', S='Axial', 
                  L='Sagittal', R='Saggital') 

we could use the {k: v} syntax, so the names flake8 complains about would become strings. And now thinking about this some more it actually seems like a flake8 bug to me in the first place!!

[hoechenberger]

Seeing that the docstyle tests pass for #7991 after it has been rebased on master, I have to conclude that the flake complaint is probably something that only happens on my system -- at least it doesn't seem to happen on CI (anymore). Therefore, I'm closing this PR.

Still-- what do we do about the Travis poblem?

[hoechenberger]

Just for reference, I've updated to:

flake8-3.8.3
pyflakes-2.2.0

and I don't see this error anymore.