Sascha

@Alucard IMHO Macbooks are better then your run of the mill windows notebook even when you plan on running windows on it.

I mainly use StringTemplate because I like the syntax and never noticed this issue ...until I called it a few hundred thousand times in one go

@Kuba Thanks for confirming this

takes 0.0005ish seconds which does not sound too much but is reeeeeaaaallly slow compared to basic arithmetic

    StringTemplate["first `a` then `b`"][<|"a" -> 1234,
   "b" -> 5678|>] // AbsoluteTiming

Is it only me or is StringTemplate horribly slow? On my machine

@Szabolcs I also noticed that the syntax PropertyValue[{g, someEdge}, {edgeProp1, edgeProp2, ...}] does not evaluate. It is mentioned in the documentation as the third syntax variant (but no examples is given). Can you (or anyone else confirm this)?

@Szabolcs I reported my issue with Subgraph (together with the obligatory rage rant at the state of the graph functionality in Mathematica)

@Szabolcs No I have not reported my issues with Subgraph yet

I haven't tried all kinds of quirky edge cases but for my use case it works well and performance is fine

g // IGEdgeSelect[# === Green &, EdgeStyle ]

g = CompleteGraph[5,
  EdgeStyle -> {1 <-> 2 -> Green, 3 <-> 4 -> Green}]

Usage example:

    IGEdgeSelect[function_, property_][g_] :=
 With[{selection =
    EdgeList[
     g, _?(Function[edge,
        Not@function@PropertyValue[{g, edge}, property]])]},
  EdgeDelete[g, selection] ]

@Szabolcs I tinkered a bit with my idea for a utility function IGEdgeSelect and came up with the following straight forward implementation:

Similar to the build-in Subgraph

For generality it might be nice if the result of such a function would not be a list of the edges but a graph containing only the selected edges. One would then have to call EdgeList on the Graph to get just the list of course

or IGEdgeSelect[function, edgeProperty]

Syntax could be something like IGEdgeMap e.g. graph // IGEdgeCases[pattern, edgeProperty]

Or in this case rather a Cases-like function

@Szabolcs Concerning my issue from before, a Select-like utility function would be a nice replacement for using the second argument of EdgeList (and VertexList).

I general I'd like to see more work done on anything optimization related be it flow in graphs or better/faster optimizers for NMinimize

Exactly

(for large graphs)

When comparing the result of FindIndependentEdgeSet and equivalent formulation for FindMinimumCostFlow for weighted bipartide graphs at some point FindMinimumCostFlow does give a different result than FindIndependentEdgeSet

Not that you mention OptimimFlowData, i also recently encountered an issue with FindMinimumCostFlow

This could all be so easy if WRI would just rewrite Properties to use associations

I can imagine that would be usefull

you mean like Merge does for associations?

Despite your philosophy to not duplicate build-in functionality in IGraphM, have you though about a IGPropertyValue function that remedies at least the syntactic issues with PropertyValue?

I personally find PropertyValue a pain to use in general

This is again something that should be easy but isn't. I don't get why the Patterntest for EdgeList does not get passed a "self reference" i.e. {graph, edge} instead of just edge

I mostly use IGraphM for its utility functions e.g IGEdgeMap and friends

@Szabolcs As far as I could test mathematica's build-in FindIndependentEdgeSet works reasonably well. FindMinimumCostFlow on the other hand is not very reliable.

In the above picture every blue line is a possible match between to samples, the line thickness indicates the goodness of the match and the color i.e. red which matches were selected

I use FindIndependentEdgeSet to find matches of produced samples that best fit together to reduce tolerances in a final product

@Szabolcs BTW if you are interested in what I am doing:

IGEdgeApply might not be the right name though since what I want is not to apply a function Apply[f]@ someProperty but rather f@someProperty

another example instead of Rescale would be (max-#)& where max is the largest weight.

that is exactly what I mean

Yes

@Szabolcs I have played around with IGraphM in the last week and I am very impressed! I hope you won't mind comments/feature requests from me in the future as I already have one: IGEdgeApply would be a useful function to have e.g. to use Rescale on the EdgeWeights as a whole.

Have you fiddeled with finding maximum maximal matchings?

I currently use FindIndependendEdgeSet and this seems to solve the problem but since the documentation is so thin here (and I am no expert on graphs) I doubt myself a bit.

I work with weighted, undirected adjacency graphs for solving a combinatorical problem (matching from a pool of individual samples by some distance metric).