Implement is_articulation to check if vertex is articulation point

[thchr]

This factors out the DFS step from articulation into a separate function articulation_dfs! and then uses this new function to implement is_articulation(g, v) which checks whether v is an articulation point of g without necessarily computing all the articulation points.

The branching on isnothing(is_articulation_pts) in articulation_dfs! is a bit inelegant - as is the fact that articulation_dfs! doesn't mutate if is_articulation_pts === nothing, but it seemed the simplest way to avoid code duplication.

Aside from these additions, I revised the doc string of articulation(g) a bit (e.g., it 1. previously misleadingly stated that g had to be connected; it does not; and 2. it changed nomenclature from "articulation point" to "cut vertex" mid-sentence).

[thchr]

Timing-wise, is_articulation(g, v) appears to give a speed-up of about a factor of ~2 in cases I looked at, compared to v ∈ articulation(g):

julia> g = path_graph(5)
julia> articulation(g) # [2, 3, 4]
julia> @btime 2 ∈ articulation($g)   # 209.656 ns (8 allocations: 704 bytes)
julia> @btime is_articulation($g, 2) # 122.850 ns (4 allocations: 464 bytes)

julia> g = path_graph(15)
julia> articulation(g) # [2:14...]
julia> @btime 9 ∈ articulation($g)   # 453.853 ns (10 allocations: 2.16 KiB)
julia> @btime is_articulation($g, 9) # 245.286 ns (4 allocations: 624 bytes)

[codecov]

Codecov Report

Attention: Patch coverage is 98.21429% with 1 line in your changes missing coverage. Please review.

Project coverage is 97.30%. Comparing base (56e5604) to head (669093e).
Report is 4 commits behind head on master.

Files	Patch %	Lines
src/biconnectivity/articulation.jl	98.21%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #387      +/-   ##
==========================================
- Coverage   97.31%   97.30%   -0.02%     
==========================================
  Files         120      120              
  Lines        6954     6965      +11     
==========================================
+ Hits         6767     6777      +10     
- Misses        187      188       +1     

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

[thchr]

Bump.

[thchr]

Bump :)

[gdalle]

Thanks for the contribution, you were right to ping again

[gdalle]

merge default branch into this PR, the version has evolved since

[gdalle]

If the idea is to call is_articulation several times in a row for different vertices, these allocations might be reusable? Maybe we should expose articulation_dfs!?

[thchr]

Sure, that'd be alright with me also. The _dfs part is an implementation detail though, so maybe this should just be called is_articulation!, taking work-arrays for pre and low then? Let me know what you think and I'll change to that.

[gdalle]

Sounds good, but then we should at the very least explain how to initialize these arguments (and ideally what they mean)

[thchr]

See also discussion in #407 (comment): maybe we should just be calling these work_buffer1 and work_buffer2 (adding some comments to explain in the code - or reassigning them to more aptly named variables) to avoid exposing unnecessary complications and implementation details to users.

[gdalle]

Can you explain this shortcut in a comment?