Implement as_steps in profile layer state #2292
Conversation
* actually binning logic adopted from https://github.com/astropy/specutils/blob/main/specutils/spectra/spectral_axis.py#L42
|
Nice, I didn't expect it to be as simple, LGTM, also #309 |
Codecov Report
@@ Coverage Diff @@
## main #2292 +/- ##
==========================================
+ Coverage 88.12% 88.14% +0.01%
==========================================
Files 247 247
Lines 23309 23312 +3
==========================================
+ Hits 20541 20548 +7
+ Misses 2768 2764 -4
Continue to review full report at Codecov.
|
|
As mentioned on Slack, I don't think we want to make these changes here - the Matplotlib viewers already draw things as steps by default so it is really the bqplot layer artist in glue-jupyter which should be updated to include this kind of logic. If we do want to support the option of having both draw styles, then I do support adding a drawstyle or as_steps option to the state, but it should then default to True and simply toggle the drawstyle in the Matplotlib viewer, and again the logic for glue-jupyter should live specifically in the bqplot layer artist (since the Matplotlib layer artist will be taken from glue-core and already draw as steps). |
glue/viewers/profile/state.py
Outdated
| @@ -238,6 +240,7 @@ def __init__(self, layer=None, viewer_state=None, **kwargs): | |||
|
|
|||
| self.add_callback('layer', self._on_layer_change, priority=1000) | |||
| self.add_callback('visible', self.reset_cache, priority=1000) | |||
| #self.add_callback('as_steps', self._on_as_steps_change, priority=1000) | |||
|
Can you also fix the codestyle failure (run |
kecnry
force-pushed
the
profile-viewer-as-steps
branch
from
b64e85b
to
abae5c6
Compare
Description
This implements a new boolean switch in the profile layer state to toggle displaying the profile as steps vs a line plot.
The actual binning logic is adopted from https://github.com/astropy/specutils/blob/main/specutils/spectra/spectral_axis.py#L42.(binning logic has now been moved to the bqplot artist in glue-viz/glue-jupyter#309)