Several fixes #45
Several fixes #45
Conversation
+ Now the parameter 'MarkerSize' does also affect the scatter plot, and not only the Median. + Added a check to detect the version number of Matlab and decide to use colororder or the colormap lines as default. + Added a quick fix for datasets where not many data is available, and is quite simmilar, causing interp1 to fail. + Changed default alpha value for the shadow quartiles and the scatter, so they are more visible. modified: test_cases/testviolinplot.m
Violin.m
Outdated
| try | ||
| meanDensityWidth = interp1(value, density, meanValue)*width; | ||
| catch | ||
| % TODO: |
Violin.m
Outdated
| obj.ViolinPlot.FaceAlpha = alpha{1}; | ||
| obj.ScatterPlot.MarkerFaceAlpha = alpha{1}; | ||
| obj.ScatterPlot.MarkerFaceAlpha = 1;% reverse to alpha{1}? This increases visibility |
There was a problem hiding this comment.
I think comments such as these should be deleted from the repo before merging.
And personally, I'd be very weary about changing default values, even for good reasons. Existing users might rely on existing functionality. At the very least we should provide a clear changelog whenever a new version is released, so users can find the cause of any regression quickly.
There was a problem hiding this comment.
I was hesitant to modify this too, but markers of the scatter plot had the same alpha value as the violin, therefore they were not easily visible depending on what colours were chosen.
Violin.m
Outdated
| @@ -195,7 +201,15 @@ | |||
|
|
|||
| %% Plot the data points within the violin area | |||
| if length(density) > 1 | |||
| jitterstrength = interp1(value, density*width, data); | |||
| try | |||
There was a problem hiding this comment.
I have a feeling that there might exist a better solution than try/catch, by checking the preconditions for interp1 manually, or indeed sorting the values manually.
But maybe that solution would add too much complexity, which might not be warranted for this simple application.
|
Sy, it took me so long to reply, I have been a bit busy. So I've removed the try/catch in favour of using unique, that´s why I marked it as TODO before, I thought it left the pull request as a draft but I obviously did not. |
Now the parameter 'MarkerSize' does also affect the scatter plot, and not only the Median.
Added a check to detect the version number of Matlab and decide to use colororder or the colormap lines as default.
Added a quick fix for datasets where not many data is available, and is quite simmilar, causing interp1 to fail.
Changed default alpha value for the shadow quartiles and the scatter, so they are more visible.
modified: test_cases/testviolinplot.m