Add AmalgamDigraphs

[finnbuck]

Based on Issue #428. This pull request adds two functions: AmalgamDigraphs and AmalgamDigraphsIsomorphic.

Description:

The AmalgamDigraphs function takes as input two digraphs and two lists of vertices corresponding to two identical subdigraphs of the two input digraphs. It returns a new digraph which consists of the two input digraphs joined together by their common subdigraph in such a way that the edge connectivity between the vertices in the new digraph matches the edge connectivity of the two input digraphs.

AmalgamDigraphsIsomorphic functions similarly, but the two lists of vertices given as inputs need only describe subdigraphs that are isomorphic to one another rather than have them be precisely equal. AmalgamDigraphsIsomorphic then rearranges the second list of subdigraph vertices so that the induced subdigraphs become equal and then calls AmalgamDigraphs.

Both functions return a tuple of size two, with the first element being the output digraph and the second being a record which maps each vertex number in the second input digraph to its new vertex number in the output digraph. The mapping of the vertices of the first input digraph can be seen as the identity mapping.

[james-d-mitchell]

@finnbuck looks great! Lots of small things to fix, but generally this is good! Some additional things:

• please also add some tests when the arguments are mutable digraphs, I don't think this'll work quite like you expect (see comments in the PR)
• Add a version that takes 4 digraphs, D1, D2, and S and a transformation map1 (embedding S in D1), then computes map2 (an embedding of S in D2) and calls DigraphAmalgam(D1, D2, S, map1, map2);
• Add a version that takes 3 digraphs: D1, D2 and S and computes map1 and map2 (embeddings into D1 and D2 resp), and calls DigraphAmalgam(D1, D2, S, map1, map2);
• Add a version that takes 2 digraphs D1 and D2, and calls x := MaximumCommonSubdigraph(D1, D2); and then returns DigraphAmalgam(D1, D2, x[1], x[2], x[3]);

[finnbuck]

@james-d-mitchell thanks for the review, I'll get to work on that!

[wilfwilson]

I've squashed and rebased this on the latest master branch, to resolve merge conflicts and make things nicer. Hopefully we can review @finnbuck's more recent changes soon.