Add max weight matching function

This commit adds a new python function max_weight_function() for
computing the maximum-weighted matching of a PyGraph object. The
implementation of this function is based on “Efficient Algorithms for
Finding Maximum Matching in Graphs”, Zvi Galil, ACM Computing Surveys,
1986. [1] It is basically a porting of the networkx implementation of
the algorithm [2] with some additional inspiration from the prototype for
the networkx version (it's basically the same code but some aspects were
a bit clearer to figure out from the prototype rather than networkx's
copy of it). [3][4]

Fixes Qiskit#216

[1] https://dl.acm.org/doi/10.1145/6462.6502
[2] https://github.com/networkx/networkx/blob/3351206a3ce5b3a39bb2fc451e93ef545b96c95b/networkx/algorithms/matching.py
[3] http://jorisvr.nl/article/maximum-matching
[4] http://jorisvr.nl/files/graphmatching/20130407/mwmatching.py

src/lib.rs

@@ -19,6 +19,7 @@ mod generators;
 mod graph;
 mod iterators;
 mod k_shortest_path;
+mod max_weight_matching;
 mod union;
 
 use std::cmp::{Ordering, Reverse};
@@ -2475,6 +2476,61 @@ pub fn cycle_basis(
     cycles
 }
 
+/// Compute a maximum-weighted matching for a :class:`~retworkx.PyGraph`
+///
+/// A matching is a subset of edges in which no node occurs more than once.
+/// The weight of a matching is the sum of the weights of its edges.
+/// A maximal matching cannot add more edges and still be a matching.
+/// The cardinality of a matching is the number of matched edges.
+///
+/// This function takes time :math:`O(n^3)` where ``n` is the number of nodes
+/// in the graph.
+///
+/// This method is based on the "blossom" method for finding augmenting
+/// paths and the "primal-dual" method for finding a matching of maximum
+/// weight, both methods invented by Jack Edmonds [1]_.
+///
+/// :param PyGraph graph: The undirected graph to compute the max weight
+///     matching for. Expects to have no parallel edges (multigraphs are
+///     untested currently).
+/// :param max_cardinality bool: If True, compute the maximum-cardinality
+///     matching with maximum weight among all maximum-cardinality matchings.
+///     Defaults False.
+/// :param callable weight_fn: An optional callable that will be passed a
+///     single argument the edge object for each edge in the graph. It is
+///     expected to return an ``int`` weight for that edge. For example,
+///     if the weights are all integers you can use: ``lambda x: x``. If not
+///     specified the value for ``default_weight`` will be used for all
+///     edge weights.
+/// :param int default_weight: The ``int`` value to use for all edge weights
+///     in the graph if ``weight_fn`` is not specified. Defaults to ``1``.
+///
+/// :returns: A dictionary of matching, the key is the node index and the
+///     value is it's matched node index. Note that both directions will be
+///     listed in the output, for example: ``{0: 1, 1: 0}``.
+/// :rtype: dict
+///
+/// .. [1] "Efficient Algorithms for Finding Maximum Matching in Graphs",
+///     Zvi Galil, ACM Computing Surveys, 1986.
+///
+#[pyfunction(max_cardinality = "false", default_weight = 1)]
+#[text_signature = "(graph, /, max_cardinality=False, weight_fn=None, default_weight=1)"]
+pub fn max_weight_matching(
+    py: Python,
+    graph: &graph::PyGraph,
+    max_cardinality: bool,
+    weight_fn: Option<PyObject>,
+    default_weight: i128,
+) -> PyResult<HashMap<usize, usize>> {
+    max_weight_matching::max_weight_matching(
+        py,
+        graph,
+        max_cardinality,
+        weight_fn,
+        default_weight,
+    )
+}
+
 /// Compute the strongly connected components for a directed graph
 ///
 /// This function is implemented using Kosaraju's algorithm
@@ -2644,6 +2700,7 @@ fn retworkx(py: Python<'_>, m: &PyModule) -> PyResult<()> {
     m.add_wrapped(wrap_pyfunction!(digraph_find_cycle))?;
     m.add_wrapped(wrap_pyfunction!(digraph_k_shortest_path_lengths))?;
     m.add_wrapped(wrap_pyfunction!(graph_k_shortest_path_lengths))?;
+    m.add_wrapped(wrap_pyfunction!(max_weight_matching))?;
     m.add_class::<digraph::PyDiGraph>()?;
     m.add_class::<graph::PyGraph>()?;
     m.add_class::<iterators::BFSSuccessors>()?;