There is a lot of transit-specific logic for computing the estimated remaining distance heuristic in AStar right now. I propose refactoring that functionality into a pluggable strategy class. This functionality would better separate the algorithm vs the transit-specific tweaks and would also allow others plug in their own strategies (something I'm looking to do with OBA), as a param to TraverseOptions.
See the attached patch for the proposed functionality.
There is a lot of transit-specific logic for computing the estimated remaining distance heuristic in AStar right now. I propose refactoring that functionality into a pluggable strategy class. This functionality would better separate the algorithm vs the transit-specific tweaks and would also allow others plug in their own strategies (something I'm looking to do with OBA), as a param to TraverseOptions.
See the attached patch for the proposed functionality.