coloring shapes by categorical variable #153

Sonja-Stockhaus · 2023-09-01T16:24:57Z

This should fix multiple problems when coloring shapes, namely:

# data reading, preprocessing
sdata = sd.read_zarr("visium_associated_xenium_io.zarr")
sdata.table.obs["cluster"] = "a"
sdata.table.obs.loc[sdata.table.obs.array_row==1.0, "cluster"] = "b"
sdata.table.obs.loc[sdata.table.obs.array_row==47.0, "cluster"] = "c"
sdata.table.obs.cluster = sdata.table.obs.cluster.astype("category")

1 Incorrect legend coloring when using the groups argument:

sdata.pl.render_shapes('CytAssist_FFPE_Human_Breast_Cancer', color="cluster", groups=["c", "b"]).pl.show("global")

2 palette parameter in render_shapes doesn't have influence on the plot (= #145)

sdata.pl.render_shapes('CytAssist_FFPE_Human_Breast_Cancer', color="cluster", groups=["c", "b"], palette=ListedColormap(['blue', 'red'])).pl.show("global")

Plus some downstream bugfixes that came up after that

codecov-commenter · 2023-09-01T16:45:27Z

Codecov Report

Merging #153 (40288fe) into main (88f6928) will increase coverage by 4.10%.
The diff coverage is 78.18%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #153      +/-   ##
==========================================
+ Coverage   74.88%   78.98%   +4.10%     
==========================================
  Files          11       11              
  Lines        1298     1342      +44     
==========================================
+ Hits          972     1060      +88     
+ Misses        326      282      -44

Files Changed	Coverage Δ
src/spatialdata_plot/pl/utils.py	`68.87% <73.68%> (+7.72%)`	⬆️
src/spatialdata_plot/pl/render.py	`89.26% <79.41%> (+1.47%)`	⬆️
src/spatialdata_plot/pl/basic.py	`85.63% <100.00%> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`

timtreis · 2023-09-04T07:53:49Z

Is this bug already solved? Or were these diagnostic plots?

Sonja-Stockhaus · 2023-09-04T08:34:05Z

Bug is solved, I still want to add some tests

for more information, see https://pre-commit.ci

timtreis · 2023-09-05T09:40:13Z

src/spatialdata_plot/pl/render.py

-        palette = (
-            ListedColormap(dict.fromkeys(color_vector)) if render_params.palette is None else render_params.palette
-        )
+        # remove the color of NaN values, else it might be assigned to a category


Better to have this comment directly next to the dict.from_keys

timtreis · 2023-09-05T09:41:26Z

src/spatialdata_plot/pl/utils.py


-        color_map = dict(zip(categories, _get_colors_for_categorical_obs(categories)))
+        color_map = dict(zip(categories, _get_colors_for_categorical_obs(categories, palette)))
        # color_map = _get_palette(
        #     adata=adata, cluster_key=value_to_plot, categories=categories, palette=palette, alpha=alpha


remove comment plz

timtreis · 2023-09-05T09:42:55Z

tests/pl/test_render_shapes.py

+        )
+
+        sdata_blobs.pl.render_shapes(
+            "blobs_polygons", color="cluster", groups=["c2", "c1"], palette=ListedColormap(["green", "yellow"])


Please take care of the ListedColormap conversion inside the code, it's unintuitive to have the user feed it in.

timtreis · 2023-09-05T09:44:04Z

CHANGELOG.md

@@ -20,6 +20,7 @@ and this project adheres to [Semantic Versioning][].
 -   Multipolygons are now handled correctly (#93)
 -   Legend order is now deterministic (#143)
 -   Images no longer normalised by default (#150)
+-   Shapes can be colored by categorical variable with the `group` and `palette` arguments (#153)


Please make clear that groups is used to filter the data and palette is used to color the (potentially remaining) categories. Also, what happens if I add a palette argument to a continous variable?

timtreis · 2023-09-05T09:58:25Z

Meeting notes @Sonja-Stockhaus

Implement also for labels using MibiTOF (this still depends on _filter_table_by_elements doesn't work for blobs dataset spatialdata#350)
Internalise ListedColormap conversion
one string is accepted for both arguments
Give tests better names -> Describe what they actually do, make distinction between groups and palette
What happens when user gives a palette arg for a continous variable? Throw error that tells user to use a cmap instead.
Docstrings
Change Changelog text

timtreis · 2023-09-05T16:49:28Z

src/spatialdata_plot/pl/basic.py

@@ -183,8 +183,11 @@ def render_shapes(
            Key in :attr:`anndata.AnnData.layers` or `None` for :attr:`anndata.AnnData.X`.
        palette
            Palette for discrete annotations, see :class:`matplotlib.colors.Colormap`.


L185 and L186 disagree, no?

because of the :class:matplotlib.colors.Colormap?

timtreis · 2023-09-06T12:38:05Z

tests/pl/test_render_points.py

@@ -21,3 +21,12 @@
 class TestPoints(PlotTester, metaclass=PlotTesterMeta):
    def test_plot_can_render_points(self, sdata_blobs: SpatialData):
        sdata_blobs.pl.render_points(elements="blobs_points").pl.show()
+
+    def test_plot_can_filter_with_groups(self, sdata_blobs: SpatialData):
+        sdata_blobs.pl.render_points(color="genes", groups="b", palette=["orange"]).pl.show()


Please change the palette here to be just a string so we're also testing that minor functionality here

timtreis · 2023-09-06T12:38:49Z

src/spatialdata_plot/pl/utils.py

-        cmap = plt.get_cmap(palette)
-        palette = [to_hex(x) for x in cmap(color_idx, alpha=alpha)]
+        # cmap = plt.get_cmap(palette)
+        # palette = [to_hex(x) for x in cmap(color_idx, alpha=alpha)]


remove comments

timtreis · 2023-09-06T12:39:44Z

src/spatialdata_plot/pl/basic.py

@@ -182,9 +182,13 @@ def render_shapes(
        layer
            Key in :attr:`anndata.AnnData.layers` or `None` for :attr:`anndata.AnnData.X`.
        palette
-            Palette for discrete annotations, see :class:`matplotlib.colors.Colormap`.
+            Palette for discrete annotations. List of valid color names that should be used
+            for the categories (all or as specified by `groups`). For a single category,


good descriptions 👌

coloring shapes by categorical variable

d32604c

Sonja-Stockhaus linked an issue Sep 1, 2023 that may be closed by this pull request

palette argument has no effect on plotting colors in pl.render_shapes() #145

Closed

update for case of array instead of categorical

24c1392

unittests added

d171e59

Sonja-Stockhaus requested a review from timtreis September 4, 2023 08:51

Sonja-Stockhaus marked this pull request as ready for review September 4, 2023 08:51

Sonja-Stockhaus mentioned this pull request Sep 4, 2023

Can now filter render_points using groups #154

Closed

filter points and shapes using groups

567df8e

Sonja-Stockhaus marked this pull request as draft September 4, 2023 15:33

Sonja-Stockhaus and others added 4 commits September 4, 2023 17:35

Merge branch 'main' into bugfix/145_palette_in_render_shapes

0735840

[pre-commit.ci] auto fixes from pre-commit.com hooks

9c65706

for more information, see https://pre-commit.ci

PointsModel coordinate argument

03779da

unittests for points and filtering by groups

19dd3e6

Sonja-Stockhaus linked an issue Sep 5, 2023 that may be closed by this pull request

groups still available for render_points()? #109

Closed

Sonja-Stockhaus marked this pull request as ready for review September 5, 2023 09:25

timtreis reviewed Sep 5, 2023

View reviewed changes

changing palette and cmap usage

a1cf6ae

Sonja-Stockhaus marked this pull request as draft September 5, 2023 16:39

some docstrings, changelog

14e166d

timtreis reviewed Sep 5, 2023

View reviewed changes

docstrings

b3a7805

timtreis reviewed Sep 6, 2023

View reviewed changes

Sonja Stockhaus and others added 2 commits September 6, 2023 20:40

remove comments and brackets

40288fe

Merge branch 'main' into bugfix/145_palette_in_render_shapes

f67dee5

timtreis marked this pull request as ready for review October 4, 2023 14:18

timtreis approved these changes Oct 4, 2023

View reviewed changes

timtreis merged commit 0e0ecb1 into main Oct 4, 2023

timtreis deleted the bugfix/145_palette_in_render_shapes branch October 4, 2023 14:19

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

coloring shapes by categorical variable #153

coloring shapes by categorical variable #153

Uh oh!

Sonja-Stockhaus commented Sep 1, 2023

Uh oh!

codecov-commenter commented Sep 1, 2023 •

edited

Loading

Uh oh!

timtreis commented Sep 4, 2023

Uh oh!

Sonja-Stockhaus commented Sep 4, 2023

Uh oh!

timtreis Sep 5, 2023

Uh oh!

timtreis Sep 5, 2023

Uh oh!

timtreis Sep 5, 2023

Uh oh!

timtreis Sep 5, 2023

Uh oh!

timtreis commented Sep 5, 2023 •

edited by Sonja-Stockhaus

Loading

Uh oh!

timtreis Sep 5, 2023

Uh oh!

Sonja-Stockhaus Sep 5, 2023

Uh oh!

timtreis Sep 6, 2023

Uh oh!

timtreis Sep 6, 2023

Uh oh!

timtreis Sep 6, 2023

Uh oh!

Uh oh!

coloring shapes by categorical variable #153

coloring shapes by categorical variable #153

Uh oh!

Conversation

Sonja-Stockhaus commented Sep 1, 2023

Uh oh!

codecov-commenter commented Sep 1, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

timtreis commented Sep 4, 2023

Uh oh!

Sonja-Stockhaus commented Sep 4, 2023

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

timtreis commented Sep 5, 2023 • edited by Sonja-Stockhaus Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

codecov-commenter commented Sep 1, 2023 •

edited

Loading

timtreis commented Sep 5, 2023 •

edited by Sonja-Stockhaus

Loading