search

scot-dev / scot

EEG/MEG Source Connectivity Toolbox in Python
http://scot-dev.github.io/scot-doc/index.html
MIT License
58 stars 23 forks source link

Add MNE backend (Infomax and CSP) #165

Closed cbrnr closed 8 years ago

cbrnr commented 8 years ago

Fixes #163

codecov-io commented 8 years ago

Current coverage is 93.30%

Merging #165 into master will increase coverage by +0.13% as of 64aa627

@@            master   #165   diff @@
=====================================
  Files           47     48     +1
  Stmts         3471   3493    +22
  Branches       396    396       
  Methods          0      0       
=====================================
+ Hit           3234   3259    +25
  Partial         70     70       
+ Missed         167    164     -3

Review entire Coverage Diff as of 64aa627

Powered by Codecov. Updated on successful CI builds.

cbrnr commented 8 years ago

@mbillingr what do you think?

mbillingr commented 8 years ago

The implementation looks good, but I am concerned about testing.

We may accidentally introduce code that depends on MNE outside of the backend, but our tests wouldn't catch that problem because MNE is available in all testing environments. Maybe it would make sense to have an environment that checks only the minimum dependencies...

However, the same issue exists for scikit-learn. So I guess we should address this somewhere else.

:+1: for merge

cbrnr commented 8 years ago

+1 for a minimal testing env without the optional dependencies.

So only numpy and scipy, and not matplotlib, scikit-learn, and mne. BTW, it's not just scikit-learn, many of our tests already depend on matplotlib, which we do not list as a required dependency.