[MAINT] Refactor and move computation functions for epochs and time in spectral submodule to a sep file

[adam2392]

Discussed in call w/ Thomas Binns that I will move refactor and clean up the epochs.py and time.py files inside spectral submodule to make the code cleaner to read

[adam2392]

@tsbinns is this accurate?

[tsbinns]

Yes, thank you @adam2392!

[tsbinns]

@adam2392 Would this be the right place to discuss the other refactoring points raised by @drammock in #142, or would a new issue be better?

[adam2392]

Sure we can discuss the refactoring of the files here.

[larsoner]

+1 for a new PR that just copy-pastes code and adds a few import statements to glue everything together

[tsbinns]

+1 for a new PR that just copy-pastes code and adds a few import statements to glue everything together

So something along the lines of:

• epochs.py handling the code required for both bivariate and multivariate methods
• epochs_bivariate.py handling the code specific for the bivariate methods
• epochs_multivariate.py handling the code specific for the multivariate methods

Then a similar thing for spectral_connectivity_time.

Would this be okay?

[adam2392]

That seems reasonable to me

[adam2392]

@tsbinns thanks for all the hard work in working w/ us on getting the multivariate methods into proper shape and also assisting w/ the refactoring.

FYI I sent an invite to collaborate on the mne-connectivity repo, so you can have better control over GH issues

[tsbinns]

Thanks very much @adam2392, happy to help!

epochs.py handling the code required for both bivariate and multivariate methods
epochs_bivariate.py handling the code specific for the bivariate methods
epochs_multivariate.py handling the code specific for the multivariate methods

From the looks of things, people are happy with this approach. The next days are pretty busy for me, but I will try to have something for later next week.

[tsbinns]

The changes discussed here were merged in #156.

@adam2392 Do you think there is anything more that needs to be done for this issue?

[adam2392]

Yes we can close this now! Thanks!