Fix clims when plotting shapes element annotations with matplotlib rendering #368
Conversation
melonora
changed the title
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #368 +/- ##
==========================================
+ Coverage 83.76% 83.90% +0.14%
==========================================
Files 8 8
Lines 1540 1535 -5
==========================================
- Hits 1290 1288 -2
+ Misses 250 247 -3
|
|
There were some tests that had an expected figure that was wrong in the first place so I fixed that. Think it is good to go for now. Here and there in the test code base I also saw no copies of anndata being created, spamming warnings so I silenced those. |
| elif vcenter is None: | ||
| norm = Normalize(vmin=vmin, vmax=vmax, clip=True) | ||
| else: | ||
| norm = TwoSlopeNorm(vmin=vmin, vmax=vmax, vcenter=vcenter) |
There was a problem hiding this comment.
Now that vcenter is removed, it should be removed also from the function signature. Also, kwargs is in the signature but not used, so I would remove it.
| @@ -339,7 +338,7 @@ def _get_collection_shape( | |||
| c = cmap(c) | |||
| else: | |||
| try: | |||
| norm = colors.Normalize(vmin=min(c), vmax=max(c)) | |||
| norm = colors.Normalize(vmin=min(c), vmax=max(c)) if norm is None else norm | |||
There was a problem hiding this comment.
In this function Normalize() is initialized without clip, while in _prepare_cmap_norm() the default is to set clip=True. I would choose one of the two as our default choice. The user will be able to specify clip, vcenter, etc by passing a norm object directly.
There was a problem hiding this comment.
hmm let me double check that if we don't pass norm as user, whether ultimately the norm is always created anyway, then we can get rid of normalize instance initiated here.
There was a problem hiding this comment.
ok there is code left over of when vmin and vmax were removed. Not certain whether to address this in a different PR.
There was a problem hiding this comment.
Why are the two files different?
There was a problem hiding this comment.
just different color perhaps due to made changes and then generating the labels with a different color. Other than that there are no changes.
| @@ -144,7 +144,7 @@ def test_plot_colorbar_respects_input_limits(self, sdata_blobs: SpatialData): | |||
| sdata_blobs.pl.render_shapes("blobs_polygons", color="cluster", groups=["c1"]).pl.show() | |||
|
|
|||
| def test_plot_colorbar_can_be_normalised(self, sdata_blobs: SpatialData): | |||
| sdata_blobs["table"].obs["region"] = ["blobs_polygons"] * sdata_blobs["table"].n_obs | |||
| sdata_blobs["table"].obs["region"] = "blobs_polygons" | |||
There was a problem hiding this comment.
no region is the region_key_key
| @@ -144,7 +144,7 @@ def test_plot_colorbar_respects_input_limits(self, sdata_blobs: SpatialData): | |||
| sdata_blobs.pl.render_shapes("blobs_polygons", color="cluster", groups=["c1"]).pl.show() | |||
|
|
|||
| def test_plot_colorbar_can_be_normalised(self, sdata_blobs: SpatialData): | |||
There was a problem hiding this comment.
What does groups=["c1"] do here? My understanding is that it doesn't have any effect because:
clusteris numerical and not categorical- the values of
clustersdo not includec1
But I may be missing something and be wrong.
There was a problem hiding this comment.
it does nothing. @timtreis is this purely to test that the colorbar stays limited from 1-50?
There was a problem hiding this comment.
I removed the argument now as typically it is used to just color certain categorical values within a column, which from the title of the test function definition is not the intention.
closes #324
This PR fixes the issue described in #324. When creating the patch collection a default Normalize was always created instead of using the one provided by the user. Furthermore in this PR some legacy code was removed which did not do anything.