JuliaGraphs/Graphs.jl v1.5.2 on GitHub

Graphs v1.5.2

Summary of changes:

Potential breaking changes:

Parallel.gdistances had some concurrency issues that lead to wrong results and was therefore moved to Experimental.Parallel.gdistances.

Deprecations:

squash(g) for a general AbstractGraph g is now deprecated. Packages that implement some subtypes of AbstractGraph should implement their own method of squash. squash(g::SimpleGraph) and squash(g::SimpleDiGraph) can still be used.

New features:

The random functions label_propagation, diffusion and diffusion_rate now allow one to optionally pass a random generator argument.
A new type IncrementalCycleTracker and a function add_edge_checked! were added to solve the incremental (online) cycle detection problem.
squash(g::SimpleGraph) and squash(g::SimpleDiGraph) now have an optional keyword argument alwayscopy (true by default) to specify if the resulting graph should by a copy of the original graph, even if the eltype of that graph does not change.

Bug fixes:

Fix a bug where the function period(g) would fail if g is not a SimpleDiGraph but some other graph. (#89)
Some tests for graph matrices are now deterministic in order to get rid of flaky tests.
Fix a bug in the documentation that lead to missing documentation (#99, #101)
Fix a bug where the function biconnected_components would not return all edges (#13)

Auto generated summary:

Merged pull requests:

fix period (#90) (@etiennedeg)
Move broken Parallel.gdistances function to Experimental submodule (@simonschoelly)
Make squash specialized on simple graphs (#93) (@simonschoelly)
Allow passing random number generator to label-propagation (#95) (@simonschoelly)
Add a random number keyword argument for diffusion and diffusion_rate (#96) (@simonschoelly)
Make random numbers deterministic for graph matrices tests (#97) (@simonschoelly)
fix community docs (#102) (@etiennedeg)
Add incremental cycle detection algorithm (#36) (@Keno, @oscardssmith)
Fix biconnected_components loses some edges (#100) (@etiennedeg)
Bump patch version to v1.5.2 (@simonschoelly)