Improve implementation of has_ghost_vertices and is_boundary_node

[DanielVandH]

No description provided.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.90%. Comparing base (29cf338) to head (7ca46e6).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #240      +/-   ##
==========================================
+ Coverage   94.87%   94.90%   +0.02%     
==========================================
  Files         102      102              
  Lines       10370    10422      +52     
==========================================
+ Hits         9839     9891      +52     
  Misses        531      531              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[rkube]

This is great, thanks for looking into this.
I benchmarked this branch on some simulation data:
v1.6.6:

[ Info: build_grid: 29841 points, 58432 triangles, 39.202s triangulation, 39.971s total

This PR:

[ Info: build_grid: 29841 points, 58432 triangles, 32.907s triangulation, 33.433s total

so roughly 20% faster.

[DanielVandH]

Hm. That's something, I suppose. I wonder what else could be done to cut the time down. Will have to give it a think, although I think technically (?) this addresses the possible fault mentioned in your original issue...

I do know where a lot of performance issues currently are - it's in the representation of triangles. I had a v2 branch some time ago that aimed to solve a big performance issue (you can see what I had in mind here https://github.com/JuliaGeometry/DelaunayTriangulation.jl/blob/v2/NEWS.md) but I ran into some issue that I can't exactly recall now, and I ran out of time to really delve into it halfway through. Would love to return to that sometime.