search: Johnson all pairs shortest paths function returns incorrect path weights
kortschak opened this issue · comments
Dan Kortschak commented
The weights are not adjusted after the return from Dijkstra to account for the graph reweighting.
I will be writing a new Johnson function that will also not require copying of the input graph. This will require also rewriting BellmanFord to return a single source shortest-path tree (I have one already in the case for the single source DijkstraFrom func I have in the dijkstra branch).