Integration of MatchPy's algorithms

[Upabjojr]

MatchPy is a library claiming to implement state-of-the-art associative-commutative pattern matching in Python.

• MatchPy paper on ArXiv
• MatchPy documentation

In particular, they claim:

We use the Hopcroft-Karp algorithm [HK73] to find an initial maximum
matching. However, since we are also interested in all matches
and the initial matching might have incompatible substitutions, we
use the algorithm described by Uno, Fukuda and Matsui [FM94],
[Uno97] to enumerate all maximum matchings.

MatchPy has been integrated into SymPy. MatchPy is also able to generate Python decision tree. See the generated RUBI decision tree. It would be nice to see how it would work with Julia's metaprogramming.

MatchPy's code generator has been ported into C++ as part of SymEngine (see the matchPyCpp subfolder on github).

The Hopcroft-Karp and Uno-Fukuda-Matsui algorithms have been ported into C++ (although the current lack of support for coroutines makes the code pretty slow). See the Hopcroft-Karp implementation and the Uno-Fukuda-Matsui implementation in SymEngine, which are translations of MatchPy's Python code.

Is there the possibility of integrating these algorithms into Rewrite.jl?