OWAnnotator: add heuristics for choosing the right x and y #1046
Conversation
|
The test is currently not working. It continues to throw "PC1" as |
| widget = self.create_widget(OWAnnotator) | ||
| c = preprocess(Corpus.from_file("slo-opinion-corpus.tab")) | ||
| self.send_signal(widget.Inputs.corpus, c[:20]) | ||
| self.wait_until_finished() |
There was a problem hiding this comment.
When you create a new widget within a test you need to pass it to the wait_until_finished:
self.wait_until_finished(widget=widget)
Beside that, the wait_until_finished has nothing to do with the attributes. It waits for run to finish, i.e. on_done to happen.
However, it should be called (wait_until_finished) anyway not to mess up the other tests.
| self.send_signal(widget.Inputs.corpus, c[:20]) | ||
| self.wait_until_finished() | ||
| controls = widget.controls | ||
| self.assertEqual(controls.attr_x, "t-SNE-x") |
There was a problem hiding this comment.
controls.attr_x is a QCombo and not a string.
string is a corresponing property widget.attr_x.name.
So you can either do:
self.assertEqual(controls.attr_x.currentText(), "t-SNE-x")
or
self.assertEqual(widget.attr_x.name, "t-SNE-x").
Another thing, your data has PCA instead of tSNE.
ajdapretnar
force-pushed
the
annot-heuristics
branch
from
e1e97e1 to
edf91c8
Compare
ajdapretnar
force-pushed
the
annot-heuristics
branch
from
edf91c8 to
a399811
Compare
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## master #1046 +/- ##
==========================================
- Coverage 82.51% 82.25% -0.27%
==========================================
Files 92 93 +1
Lines 12400 12448 +48
Branches 1695 1699 +4
==========================================
+ Hits 10232 10239 +7
- Misses 1855 1896 +41
Partials 313 313 |
Issue
When a corpus contains continuous metas, they are used for plotting instead of t-SNE variables.
Description of changes
Use heuristics to set t-SNE variables as default variables for plotting.
Includes