anglr DEL must mesh on entire set

Possibly make this an opt-out, or maybe it's a different model?

Mar 12 '18 07:03 mdsumner

related is that DEL.SC is not normalizing in x/y, so do that first (it is for DEL.PATH causing head scratching here https://github.com/hypertidy/silicate/issues/62#issuecomment-372880366)

( I think the SC methods are wrong somewhere, it doesn't show up because the edges are fine on their own, it's only when related it comes out wrong)

Mar 14 '18 02:03 mdsumner

Here's a scgraph plot for edges and vertices

library(scgraph)
library(ggraph)
ggraph(as.igraph(x)) + 
geom_edge_link(aes(label = edge_, 
  start_cap = label_rect(node1.name)),  angle_calc = 'along', label_dodge = unit(2.5, 'mm')) +  
 geom_node_text(aes(label = name))

Mar 14 '18 03:03 mdsumner

DEL.SC now normalizes correctly, and the silicate#62 example now works.

DEL should still work set-wise, it will simplify the code a lot - but it's not clear if we should classify holes set-wise or individually (probably need it to be an option, with post-iterate normalization).

Mar 14 '18 04:03 mdsumner

We can't do this set-wise without turning SC into PATH. It has to be done

Mar 14 '18 04:03 mdsumner