[MRG] VolSourceEstimate: allow vector data

[christianbrodbeck]

This is the change I had to make to get vector volume source estimates. Should add a test.

CC: @agramfort @larsoner

[larsoner]

Looks reasonable to me once it's tested. The easiest way might be to apply the inverse with pick_ori=None and with pick_ori='vector' and verify that the data are allclose after you .magnitude() the vector.

[codecov]

Codecov Report

Merging #5209 into master will decrease coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #5209      +/-   ##
==========================================
- Coverage   88.16%   88.16%   -0.01%     
==========================================
  Files         358      358              
  Lines       65729    65711      -18     
  Branches    11180    11176       -4     
==========================================
- Hits        57949    57932      -17     
  Misses       4970     4970              
+ Partials     2810     2809       -1

[larsoner]

LGTM +1 for merge

[agramfort]

thx @christianbrodbeck

[larsoner]

Turns out this was probably not the right way to do it, VectorSourceEstimate is based on _BaseSourceEstimate, which is specific to surface-based source estimates. I'll work on a proper VolVectorSourceEstimate class.