(I am trying to make a feature request, hope this is the right place. Happy to repost on github/somewhere else if needed.)
legend_loc="on data" and use it very often. However, when I have lots of (sub)clusters or long cluster names, they often overlap, making the plot unreadable. Are there plans to address this? For the sake of comparison/an example, Seurat has a “repel” argument to deal with this.