Related to:#321

I have my own algorithm to calculate the border of points group, but would like to use your label location procedure. Do you think it'd be possible to extract something likegeom_mark_polygon from the current code?

This is the plot I'm interested to add labels to:

library(ggplot2)library(ggforce)library(data.table)load(gzfile("border_example.rda.gz"))ggplot(data, aes(x=UMAP_1,y=UMAP_2))+     geom_point(color="grey")+     geom_shape(data=maskTable, aes(group=group,color=cluster),linewidth=1,fill=NA,expand=unit(-1,"pt"))+    coord_fixed()+     theme_classic()

border_example.rda.gz

Currently, the best I can do is something like this, relying ongeom_mark_hull:

ggplot(data, aes(x=UMAP_1,y=UMAP_2))+    geom_point(color="grey")+    geom_shape(data=maskTable, aes(group=group,color=cluster),linewidth=1,fill=NA,expand=unit(-1,"pt"))+    geom_mark_hull(data=maskTable, aes(fill=NA,group=cluster,label=cluster),radius=0,expand=0,con.cap=0)+    coord_fixed()+    theme_classic()

However, I don't need to actually build a hull, since I have already have a proper polygon, and I don't want to have an additional dependency.

There are also couple of other questions:

1. Several label connectors (for example, NK, CD8 T, and others) lead to a split point between two clusters, resulting in a bit of ambiguity. Is there a way to fix it?
2. Is there a way to remove fill?fill=NA doesn't seem to work as I would expect.
3. Is there a way to decrease the length of the connectors? They seem to be longer that I'd like, and I couldn't find an option to control it.