Save a QueryGraph for multiple uses?

[tompeterken-os]

Relating to a previous question, I’m creating a QueryGraph with a large number (millions) of snaps. My current platform (in Databricks, distributed cloud-computing) means that I’m re-generating this QueryGraph for each instance of GraphHopper that I’m initiating (once per data partition).

Is there a way to create my QueryGraph once, write it out, and then read it in to my GraphHopper instances directly? I’m thinking a similar approach to how we generate the BaseGraph “graph-cache” once to be re-used rather than generating from scratch it every time.

[karussell]

You can surely try to (de)serialize it using the internal DataAccess classes we use to serialize the graph. See the DataAccessTest to see how it could be used.