You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Not very clear on what we want this to be. I think a good starting point would be allowing a yrelative (or some better name) in plot_dist which is then converted into an actual y value multiplying by the maximium of the kde/hist/ecdf inside plot_dist.
Idea of steps involved:
User provides yrelative values
aesthetics are generated for yrelative which can be retrived as dataset
yrelative dataset is multiplied by density.max(some_dims) -> xarray handles all aligment and broadcasting
y aesthetic is created and yrelative deleted or yrelative gets its values overwritten and the aesthetic renamed (whatever we find easier to implement)
Plotting backends only see y aesthetic
This might not 100% work with the exact relative ratios we'd want but I think it will generally be good enough and probably much easier than attempting to handle relative spacing across backends. But it is also true I have literally no idea how to use relative units in bokeh nor plotly, so it is always possible that is easier than the steps above.
The text was updated successfully, but these errors were encountered:
I agree in general, but just providing y may be enough, as we then scale them by max density. At least for plots like plot_dist, it makes sense to always think of y in relative terms.
Not very clear on what we want this to be. I think a good starting point would be allowing a
yrelative
(or some better name) inplot_dist
which is then converted into an actual y value multiplying by the maximium of the kde/hist/ecdf inside plot_dist.Idea of steps involved:
yrelative
valuesyrelative
which can be retrived as datasety
aestheticThis might not 100% work with the exact relative ratios we'd want but I think it will generally be good enough and probably much easier than attempting to handle relative spacing across backends. But it is also true I have literally no idea how to use relative units in bokeh nor plotly, so it is always possible that is easier than the steps above.
The text was updated successfully, but these errors were encountered: