Guillaume Dalle comments

Results 1600 comments of


                                            Guillaume Dalle

Code implicitly assumes that the heuristic output is of the same type as the graph weights.

I think it is the most reasonable option for now. The heuristic has to be computed on `s` first, so we might as well use that to choose the type...

Iterator for self-avoiding walks

Hey @bramtayl, thanks for the contribution! Did you check out #106 too? Is there a way to make both approaches compatible?

Feature Request: Dominator Tree

Hi @schlichtanders :wave: I'm not an expert in dominator trees, but any addition to the library is welcome! Do you think you can revisit the old PR you mentioned and...

Roadmap for Graphs.jl 2.0

Here are a few topics that seem important to discuss, in no particular order. Feel free to add your own! **Support vertex and edge metadata** - #35 - #57 **Add...

add_vertex! should return the added vertex.

How about starting a roadmap for 2.0 with big breaking changes we would like to see happen? I have a few of my own, eg. #77

[Port] Fallback for all edge operations

Does this require duplicating all edge operations or is there a clever way to do this (eg a macro)?

Add support for induced bipartite subgraph

I agree that it would be useful functionality. @Codsilla do you want to get started on a PR so we can help?

Split out a GraphsBase package?

I for one am very much in favor of that

Standardized output for shortest path algorithms

The problem is that not all shortest path algorithms are created equal. While Dijkstra can fill a correct shortest path tree, A* only finds the shortest s-d path with no...

Standardized output for shortest path algorithms

In a future version, I think A* should return a vector of nodes instead of edges. And we could add methods for Dijkstra & Co which, when called with a...