Thanks @mlewe, and thanks for explaining that the license makes the Artifact impossible. Maybe I (or someone) will get around to an open-licensed implementation in pure Julia.

Not sure if an issue is the right place, since this is more of a question/suggestion. But I was trying to figure out the final cost of the solution. Currently, we can get the edge matches for the solution, and it's pretty straightforward to simply write a script that will calculate the total cost after calling get_all_matches, but I figured this information should be available in the structure PerfectMatchingCtx.ptr and should be able to be retrieved with a ccall, I just have no idea how to do that.

Is there a way to see the contents of what the object PerfectMatchingCtx.ptr is pointing to contains?

Ok yea, iterating over the solution is easier. Thanks!

Registration pull request created: JuliaRegistries/General/25384

After the above pull request is merged, it is recommended that a tag is created on this repository for the registered package version.

This will be done automatically if the Julia TagBot GitHub Action is installed, or can be done manually through the github interface, or via:

git tag -a v0.4.3 -m "<description of version>" fca1d332d98259c3928f4eecebb574ee9554d340
git push origin v0.4.3

Not sure if an issue is the right place, since this is more of a question/suggestion. But I was trying to figure out the final cost of the solution. Currently, we can get the edge matches for the solution, and it's pretty straightforward to simply write a script that will calculate the total cost after calling get_all_matches, but I figured this information should be available in the structure PerfectMatchingCtx.ptr and should be able to be retrieved with a ccall, I just have no idea how to do that.

Is there a way to see the contents of what the object PerfectMatchingCtx.ptr is pointing to contains?

I'm new to this and am trying to understand how to use this wrapper. I'm trying to run the examples from the c++ blossom5 algorithm, one of which is the graph in GRAPH1.TXT

6 9
0 1 3
0 3 10
0 4 7
1 2 -1
1 4 5
1 5 4
2 5 -7
3 4 0
4 5 4

Running this with the c++ version as ./blossim5 -e GRAPH1.TXT returns

perfect matching with 6 nodes and 9 edges
    starting init...done [0.000 secs]. 0 trees
    .
done [0.000 secs]. 0 grows, 0 expands, 0 shrinks
    expands: [0.000 secs], shrinks: [0.000 secs], dual updates: [0.000 secs]
cost = -4.0

So what I have in Julia is

using BlossomV

differences = [
	0 1 3
	0 3 10
	0 4 7
	1 2 -1
	1 4 5
	1 5 4
	2 5 -7
	3 4 0
	4 5 4
]

m = Matching(6)
for row in 1:size(differences,1)
	vertex1, vertex2, weight = differences[row,:]
	add_edge(m, vertex1, vertex2, weight)
end

solve(m)
for i in 1:6
	get_match(m,i)
end

But I get the error

ERROR: LoadError: Cost must be positive. Edge between 1 and 2 has cost -1.

Is this by design? Seems that a wrapper should perform exactly what the original one performs. Let me know if I'm missing something. Thanks!

Sounds good, I'll go ahead and close this with the pending pull request https://github.com/mlewe/BlossomV.jl/pull/18.

Simple edits to fix the function get_all_matches and allowing a graph with negative cost.

I'm new to this and am trying to understand how to use this wrapper. I'm trying to run the examples from the c++ blossom5 algorithm, one of which is the graph in GRAPH1.TXT

6 9
0 1 3
0 3 10
0 4 7
1 2 -1
1 4 5
1 5 4
2 5 -7
3 4 0
4 5 4

Running this with the c++ version as ./blossim5 -e GRAPH1.TXT returns

perfect matching with 6 nodes and 9 edges
    starting init...done [0.000 secs]. 0 trees
    .
done [0.000 secs]. 0 grows, 0 expands, 0 shrinks
    expands: [0.000 secs], shrinks: [0.000 secs], dual updates: [0.000 secs]
cost = -4.0

So what I have in Julia is

using BlossomV

differences = [
	0 1 3
	0 3 10
	0 4 7
	1 2 -1
	1 4 5
	1 5 4
	2 5 -7
	3 4 0
	4 5 4
]

m = Matching(6)
for row in 1:size(differences,1)
	vertex1, vertex2, weight = differences[row,:]
	add_edge(m, vertex1, vertex2, weight)
end

solve(m)
for i in 1:6
	get_match(m,i)
end

But I get the error

ERROR: LoadError: Cost must be positive. Edge between 1 and 2 has cost -1.

Is this by design? Seems that a wrapper should perform exactly what the original one performs. Let me know if I'm missing something. Thanks!

It would be nice to update this package to more modern Julia. Specifically, it would be nice to generate a Project.toml file, and especially nice to use the artifact system (Julia 1.3+). There is a wizard for generating the build: https://github.com/JuliaPackaging/BinaryBuilder.jl. I played with it briefly and it seemed fairly straightforward, but the first hurdle is figuring out how to handle your customizations (sharedlib.patch and interface/).