Fix regression in scipy 1.5.0 in dendogram when labels is a numpy array. #12421
Conversation
lesteve
changed the title
AtsushiSakai
added
scipy.cluster
defect
tylerjereddy
added
the
backport-candidate
|
I added the backport label and the |
| @@ -812,6 +812,13 @@ def test_valid_orientation(self): | |||
| Z = linkage(hierarchy_test_data.ytdist, 'single') | |||
| assert_raises(ValueError, dendrogram, Z, orientation="foo") | |||
|
|
|||
| def test_can_take_label_as_array_or_list(self): | |||
| Z = linkage(hierarchy_test_data.ytdist, 'single') | |||
| labels = np.array([1, 3, 2, 6, 4, 5]) | |||
There was a problem hiding this comment.
testing with an empty NumPy array might also be useful, though maybe not required
adding a comment to the test like # test for gh-12418 can be nice too, though again perhaps not required
There was a problem hiding this comment.
I added a test for empty labels. I am not sure which behaviour we want:
- an error "Dimensions of Z and labels must be consistent." (this is the current behaviour)
labels=[]is the same aslabels=None(labels=Noneis the default)
There was a problem hiding this comment.
I also added # test for gh-12418.
In general I feel git blame works reasonably well to find the associated PR with the discussion. At the same time it may be annoying if the history is complicated so I can see why the comment is useful.
There was a problem hiding this comment.
I'd prefer an error, as it is now.
There was a problem hiding this comment.
OK great, let me know if you need anything else, maybe a changelog entry somewhere?
There was a problem hiding this comment.
I think this is it, our changelog for the release notes will be auto-generated.
Not a big deal at all. IMO the strength of an ecosystem is not that its members never make any mistake but instead that people are motivated to work together and fix others mistakes when they bump into it and overall make the ecosystem a better place to live in. |
tylerjereddy
removed
the
backport-candidate
Reference issue
Fix #12418.
What does this implement/fix?
This fixes a regression when
labelsis a numpy array.I added a test but I am not too familiar with dendrogram so let me know if the labels I chose do not make sense.