Part of the framework for an alignment algorithm that's supposed to find the minimum-penalty alignment of two graphs (node-for-node) where the score function is only a function of information stored on the nodes, not the edges themselves (but it is not allowed for the edges to be rearranged in the consensus graph, only for gap nodes to be inserted). It's meant to generalize algorithms such as Smith-Waterman and Needleman-Wunsch to any graph, not just lines. I never finished it, because it was hard.
mszegedy/graphutils
Python graph utilities (mostly for finding optimal node-by-node alignments of networkx graphs)
Python