datashader speedup and bugfixes by Sonja-Stockhaus · Pull Request #309 · scverse/spatialdata-plot

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

Sonja-Stockhaus · 2024-07-17T10:12:30Z

This speeds everything up for me locally, but for the benchmark (#296), I see an effect (aka datashader faster than matplotlib) for e.g. 10k points/shapes etc.

codecov-commenter · 2024-07-17T10:29:29Z

Codecov Report

Attention: Patch coverage is 83.43195% with 28 lines in your changes missing coverage. Please review.

Project coverage is 83.36%. Comparing base (80c6f77) to head (73568ec).

Files with missing lines	Patch %	Lines
src/spatialdata_plot/pl/render.py	76.59%	22 Missing ⚠️
src/spatialdata_plot/pl/utils.py	91.78%	6 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## main #309 +/- ##
==========================================
- Coverage 83.73% 83.36% -0.37% 
==========================================
Files 8 8 Lines 1543 1653 +110 ==========================================
+ Hits 1292 1378 +86 - Misses 251 275 +24

Files with missing lines	Coverage Δ
src/spatialdata_plot/pl/basic.py	`89.26% <ø> (ø)`
src/spatialdata_plot/pl/render_params.py	`100.00% <100.00%> (ø)`
src/spatialdata_plot/pl/utils.py	`77.48% <91.78%> (+1.11%)`	⬆️
src/spatialdata_plot/pl/render.py	`90.25% <76.59%> (-4.47%)`	⬇️

Sonja-Stockhaus · 2024-07-17T18:45:03Z

+ trans = mtransforms.Affine2D(matrix=affine_trans)
+ trans_data = trans + ax.transData
+
+ rgba_image = np.transpose(rgba_image.data.compute(), (1, 2, 0)) # type: ignore[attr-defined]


Here (and in render_shapes), we access the image as numpy array from the SpatialImage. mypy doesn't believe that compute() exists...

LucaMarconato · 2024-07-18T16:29:29Z

+
+ # compute canvas size in pixels close to the actual image size to speed up computation
+ plot_width = x_ext[1] - x_ext[0]
+ plot_height = y_ext[1] - y_ext[0]
+ plot_width_px = int(round(fig_params.fig.get_size_inches()[0] * fig_params.fig.dpi))
+ plot_height_px = int(round(fig_params.fig.get_size_inches()[1] * fig_params.fig.dpi))
+ factor = np.min([plot_width / plot_width_px, plot_height / plot_height_px])
+ plot_width = int(np.round(plot_width / factor))
+ plot_height = int(np.round(plot_height / factor))


I'd consider bundling this code in a private function since it's duplicate from above.

LucaMarconato · 2024-07-18T16:30:14Z


- rbga_image = np.transpose(ds_result.to_numpy().base, (0, 1, 2))
- cax = ax.imshow(rbga_image, zorder=render_params.zorder, alpha=render_params.alpha)
+ # create SpatialImage to get it back to original size


Also this code is a duplicate, I'd consider it refactoring it into a private function.

…f eq_hist for ds.shade()

…functions

…oloring

LucaMarconato · 2024-08-01T12:51:40Z

Thanks @Sonja-Stockhaus ! I have tried the PR on the Visium HD data and the speed up is significant! I switched back to using datashader instead of matplotlib in the notebook 😊

LucaMarconato · 2024-08-09T13:44:41Z

Sonja, I found a bug with this PR when the rasterized object needs to be aligned with the rest #291. The solution seems straightforward, I think one just needs to initialize the image element from datashader using the old coordinate transformations.

timtreis · 2024-08-14T23:56:54Z

Will look into it!

LucaMarconato · 2024-09-27T12:29:49Z

Thanks for the update!

outline of polygons need to be implemented in datashader

This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

Sonja-Stockhaus · 2024-09-27T12:50:47Z

outline of polygons need to be implemented in datashader
This task here seems tricky. I would consider raising a warning saying that the outline is currently not supported and workin on this on a separate PR (with low priority).

If you use cvs.line instead of cvs.polygons you get only the outline, so you should be able to overlay the two. Depending on how fast we want to merge this PR, we can of course move that to a new one

LucaMarconato · 2024-09-27T12:52:45Z

Ah ok, then it should be a fast addition, thanks for the update.

…ny reductions

Sonja-Stockhaus · 2024-10-02T16:10:08Z

General problem of datashader so far: when you render elements, coloring them by a value and a color map, you wouldn't see a single element if all of them have the same value X. This is due to the scaling of the data for the colormap that happens in datashader (scaled_data = (data - span[0])/(span[1] - span[0])) which leads to all elements being assigned alpha=0.

Fix in this PR: internally, not the whole colormap is passed to datashader, but just the color given by cmap(0.0). It is used to color all elements. Still, a colorbar of the cmap is drawn with the range [X, X+1].

LucaMarconato · 2024-10-02T16:50:53Z

Thanks for the explanation, I think the approach that you implemented is a good one!

melonora

Hi @Sonja-Stockhaus, thanks for the amazing work. I think it is very far already, but would like some minor changes here and there, mainly to shorten the code a bit and some refactorings to keep it more maintainable.

timtreis

as discussed

Sonja-Stockhaus · 2024-10-16T17:56:10Z

As discussed with @timtreis, we now use sum as default datashader aggregation for points and mean for shapes

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

speed up datashader by using canvas size equal to image size
50808b3

Sonja-Stockhaus linked an issue Jul 17, 2024 that may be closed by this pull request
Worse performance using datashader? #296
Closed

_cax definition bugfix
759ecd8

Sonja-Stockhaus mentioned this pull request Jul 17, 2024
Error in Squidpy notebook due to rasterization not handling transformations #291
Closed

attempt mypy error fixes
ea718bf

Sonja-Stockhaus commented Jul 17, 2024
View reviewed changes

delete comments
eef7b8b

Sonja-Stockhaus marked this pull request as ready for review July 17, 2024 18:57

Sonja-Stockhaus requested a review from timtreis July 17, 2024 18:57

LucaMarconato reviewed Jul 18, 2024
View reviewed changes

continuous agg with mean instead of sum and use linear cmap instead o…
702a226
…f eq_hist for ds.shade()

Sonja-Stockhaus linked an issue Jul 19, 2024 that may be closed by this pull request
Wrong colors when using method='datashader'#311
Closed

Sonja-Stockhaus marked this pull request as draft July 19, 2024 15:48

LucaMarconatoand others added 2 commits July 22, 2024 14:50

Merge branch 'main' into feature/296-datashader-canvas-size
420772a

switch back to sum() and move duplicate datashader code into private …
48c1c52
…functions

Sonja-Stockhaus linked an issue Jul 30, 2024 that may be closed by this pull request
Datashader aggregation function can't be specified by user #314
Closed
3 tasks

make ds reduction a kwarg and fix ds colorbar limits for continuous c…
fc89462
…oloring

Sonja-Stockhaus linked an issue Jul 31, 2024 that may be closed by this pull request
Point sizes disagree between matplotlib and datashader #315
Closed

adapt ds px using dpi/100
a0dac08

LucaMarconato mentioned this pull request Aug 1, 2024
Visium hd notebook adjust for squares instead of circles scverse/spatialdata-tutorials#112
Merged

spread how kw refactoring
c8b0b34

This was referenced Aug 12, 2024
figures sharpness #317
Closed
Alignment image with xenium scverse/spatialdata-io#168
Open

timtreis self-assigned this Aug 14, 2024

timtreis marked this pull request as ready for review August 14, 2024 23:59

timtreis assigned Sonja-Stockhaus and unassigned timtreis Sep 30, 2024

Sonja Stockhaus added 3 commits October 1, 2024 15:36

add ds shapes outlines, fix shapes fill_alpha behavior, fix std/var/a…
59a19da
…ny reductions

add test images
22cdabc

update test images
1d07544

Sonja-Stockhaus marked this pull request as ready for review October 2, 2024 07:20

changelog update and documentation
febd424

Sonja-Stockhaus mentioned this pull request Oct 2, 2024
Datashader shapes outlines and fill_alpha documentation #367
Closed

LucaMarconato mentioned this pull request Oct 3, 2024
Using groups in pl.render_shapes() fails with large data #357
Closed

Sonja-Stockhausand others added 3 commits October 11, 2024 16:39

Merge branch 'main' into feature/296-datashader-canvas-size
2a95236

fix resolved merge conflict
1d871ed

fix resolved merge conflict
0bf8c35

melonora requested changes Oct 14, 2024
View reviewed changes

Sonja Stockhaus added 2 commits October 14, 2024 22:20

Wouter's feedback
4391b81

fix coordinate system
a06400b

timtreis requested changes Oct 16, 2024
View reviewed changes

Sonja-Stockhaus mentioned this pull request Oct 16, 2024
Missing tests for norm=Normalize(...) #372
Closed

Sonja Stockhaus added 2 commits October 16, 2024 18:02

passing Normalize with datashader has an effect now
4b52fe7

fix norm behavior
cd0f68b

Sonja Stockhaus added 3 commits October 16, 2024 19:59

Tim's feedback
214cb9d

fix test
ddc7927

add test image
73568ec

timtreis merged commit bc2db2c into mainOct 23, 2024

timtreis deleted the feature/296-datashader-canvas-size branch October 23, 2024 22:16

Uh oh!

Conversation

Sonja-Stockhaus commented Jul 17, 2024

Uh oh!

codecov-commenter commented Jul 17, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Sonja-StockhausJul 17, 2024

Choose a reason for hiding this comment

Uh oh!

LucaMarconatoJul 18, 2024

Choose a reason for hiding this comment

Uh oh!

LucaMarconatoJul 18, 2024

Choose a reason for hiding this comment

Uh oh!

LucaMarconato commented Aug 1, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

LucaMarconato commented Aug 9, 2024• edited by timtreis Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

timtreis commented Aug 14, 2024

Uh oh!

LucaMarconato commented Sep 27, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Sonja-Stockhaus commented Sep 27, 2024

Uh oh!

LucaMarconato commented Sep 27, 2024

Uh oh!

Sonja-Stockhaus commented Oct 2, 2024

Uh oh!

LucaMarconato commented Oct 2, 2024

Uh oh!

melonora left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

timtreis left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Sonja-Stockhaus commented Oct 16, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Uh oh!

Conversation

Sonja-Stockhaus commented Jul 17, 2024

Uh oh!

codecov-commenter commented Jul 17, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Sonja-StockhausJul 17, 2024

Choose a reason for hiding this comment

Uh oh!

LucaMarconatoJul 18, 2024

Choose a reason for hiding this comment

codecov-commenter commented Jul 17, 2024•
edited
Loading

LucaMarconato commented Aug 1, 2024•
edited
Loading

LucaMarconato commented Aug 9, 2024•
edited by timtreis
Loading

LucaMarconato commented Sep 27, 2024•
edited
Loading

codecov-commenter commented Jul 17, 2024•
edited
Loading

LucaMarconato commented Aug 1, 2024•
edited
Loading

LucaMarconato commented Aug 9, 2024•
edited by timtreis
Loading

LucaMarconato commented Sep 27, 2024•
edited
Loading