WIP: added colors option to grow_labels #8519
Conversation
|
can you update a test? |
mne/label.py
Outdated
| else: | ||
| # use specified colors | ||
| label_colors = colors |
There was a problem hiding this comment.
There is no guarantee or check that label_colors is the right size. I think you had the right approach before, namely to do:
label_colors = np.empty((len(labels), 3))
label_colors[:] = colors
This guarantees that whatever colors is, it can broadcast and you will end up with label_colors of the right shape.
There was a problem hiding this comment.
Fine by me, but _n_colors() (label.py) returns array, shape (n, 4).
There was a problem hiding this comment.
Then you can do label_colors = np.ones((len(labels), 4)); label_colors[:, :3] = label_colors if you want. I doubt anyone wants an alpha < 1
There was a problem hiding this comment.
Then you can do
label_colors = np.ones((len(labels), 4)); label_colors[:, :3] = label_colorsif you want. I doubt anyone wants an alpha < 1
I'm now allowing four columns and added info about RGB/alpha values to doc-string.
|
LGTM -- can you update |
| else: | ||
| # use specified colors | ||
| label_colors = np.empty((len(labels), 4)) | ||
| label_colors[:] = colors |
There was a problem hiding this comment.
label_colors = np.array(colors)
seems easier and equivalent.
also do you test somewhere that len(colors) == len(labels) ?
There was a problem hiding this comment.
No it's not equivalent. The point of doing the np.empty is that
- It allows broadcasting like
colors=0.if you don't care about the color they can all beblacktransparent, orcolor=(1, 0, 0, 1)to make them all red, and - It automatically makes sure
len(labels) == len(colors)
EDIT: in other words, we let NumPy do the sanity checking for us by using np.empty
There was a problem hiding this comment.
ok then. +1 for MRG if CIs are green.
thx @olafhauk !
There was a problem hiding this comment.
colors can also be of shape (, 4) (one color for all labels). I amended doc-string and added checks/tests.
olafhauk
force-pushed
the
grow_labels_colors
branch
from
46c3704
to
daed8ce
Compare
mne/source_space.py
Outdated
| @@ -3197,3 +3197,110 @@ def _get_src_nn(s, use_cps=True, vertices=None): | |||
| else: | |||
| nn = s['nn'][vertices, :] | |||
| return nn | |||
|
|
|||
|
|
|||
| def vertex_depth(inst, info=None, picks=None, trans=None, mode='dist', | |||
There was a problem hiding this comment.
I never know what to do about the codecov errors. Otherwise I think it's fine now.
|
Thx @olafhauk ! |
Closes #8517
What does this implement/fix?
Added
colorsoption to mne.grow_label() to avoid error message for large number of seeds.Additional information