If you use a certain weighting for the LM preparation then the speed is only tuned for these specific settings. All options that forces to route to go off the “optimal” path will reduce performance. So if you need good performance for shortest you could do two LM preparations (one for your fastest and one for shortest).
Do you mean I can create a new vehicle profile adding distance factor, for example:
name: shortest
base: car
max_speed: 120
distance_factor: 90
then a POST including { “base”:“shortest”, “max_speed”:120 , “distance_factor”:90 } will use this specific LM preparation, correct? And same for toll (need to check yml syintax for “no_access”: { “toll”: [“all”] }).
I’ll give a try shortly! I’m curious to check how disk space and RAM consumption compares to CH.
Yes, you create a new vehicle profile and then you should be able to just POST (yaml or JSON):
base: shortest
max_speed: 120
The ideal thing would be to just specify base: shortest but at the moment I think max_speed is required.
need to check yml syintax for “no_access”: { “toll”: [“all”] }).
You can define new vehicle profiles via yaml or JSON. So I would only use yaml if you are using yaml per request. The yaml array structure is explained here.
The advantage over CH is not so much the RAM consumption but that you can modify the profile per request.
I’m getting some troubles when having two profiles, it seems jetty fails to start:
INFO [2018-10-30 19:56:26,816] com.graphhopper.reader.osm.GraphHopperOSM: graph gfast,gfastnotoll|RAM_STORE|2D|NoExt|,,,,, details:edges:0(0MB), nodes:0(0MB), name:(0MB), geo:0(0MB), bounds:1.7976931348623157E308,-1.7976931348623157E308,1.7976931348623157E308,-1.7976931348623157E308
...
INFO [2018-10-30 20:02:50,212] com.graphhopper.routing.lm.PrepareLandmarks: Calculated landmarks for 1 subnetworks, took:185.74075 =>
...
INFO [2018-10-30 20:05:51,445] org.eclipse.jetty.server.AbstractConnector: Started application@2a2843ec{HTTP/1.1,[http/1.1]}{localhost:8989}
INFO [2018-10-30 20:05:51,446] org.eclipse.jetty.server.AbstractConnector: Started admin@2042ccce{HTTP/1.1,[http/1.1]}{localhost:8990}
ERROR [2018-10-30 20:05:51,448] io.dropwizard.cli.ServerCommand: Unable to start server, shutting down
! java.lang.IllegalStateException: already initialized
! at com.graphhopper.storage.RAMDataAccess.loadExisting(RAMDataAccess.java:118)
! at com.graphhopper.routing.lm.LandmarkStorage.loadExisting(LandmarkStorage.java:705)
! at com.graphhopper.routing.lm.PrepareLandmarks.loadExisting(PrepareLandmarks.java:121)
! at com.graphhopper.routing.lm.LMAlgoFactoryDecorator$1.run(LMAlgoFactoryDecorator.java:284)
! at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
! at java.util.concurrent.FutureTask.run(FutureTask.java:266)
! at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
! at java.util.concurrent.FutureTask.run(FutureTask.java:266)
! at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
! at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
...
INFO [2018-10-30 20:05:51,452] org.eclipse.jetty.server.AbstractConnector: Stopped application@2a2843ec{HTTP/1.1,[http/1.1]}{localhost:8989}
INFO [2018-10-30 20:05:51,455] org.eclipse.jetty.server.AbstractConnector: Stopped admin@2042ccce{HTTP/1.1,[http/1.1]}{localhost:8990}
java.lang.RuntimeException: java.util.concurrent.ExecutionException: java.lang.IllegalStateException: already initialized
at com.graphhopper.routing.lm.LMAlgoFactoryDecorator.loadOrDoWork(LMAlgoFactoryDecorator.java:304)
at com.graphhopper.GraphHopper.loadOrPrepareLM(GraphHopper.java:1252)
at com.graphhopper.reader.osm.GraphHopperOSM.loadOrPrepareLM(GraphHopperOSM.java:91)
at com.graphhopper.GraphHopper.postProcessing(GraphHopper.java:912)
at com.graphhopper.GraphHopper.process(GraphHopper.java:710)
at com.graphhopper.GraphHopper.importOrLoad(GraphHopper.java:679)
at com.graphhopper.http.GraphHopperManaged.start(GraphHopperManaged.java:71)
at io.dropwizard.lifecycle.JettyManaged.doStart(JettyManaged.java:27)
...
the -gh/ folder does nto seem completed:
Oct 30 21:02 landmarks_gfast
Oct 30 21:05 landmarks_gfastnotoll
Oct 30 20:59 location_index
Oct 30 21:02 subnetwork_landmarks_gfast
Oct 30 21:05 subnetwork_landmarks_gfastnotoll
note, I believe I had the same issue when I tried leaving “graph.flag_encoders: car” in the config.yml
Now I need to find time to consolidate all the changes (and certain quicker workarounds) and also make the profile configuration more generic. This will take some time but the first work will go into 0.12 already.
Just a quick note - OSM has rather poor toll info, so in the future it would be the need to include separate source. Then such mapping can fail (different pay for certain profiles, different payment models (per distance/between points/on point/toll zone).
This sounds very promising. Still I’m not able to get it running.
I configured everything as gspeed mentioned but when running ./graphhopper.sh -a web -i europe_germany.pbf
I always get the following error: java.lang.IllegalStateException: Cannot load properties to fetch EncodingManager configuration at: ./europe_germany-gh/
while the folder “europe_germany-gh” does not exist.
BTW: I cloned the graphhopper git repo. [EDIT: and checked out the flex_vehicleprofile branch!] Is this feature there not available? When will version 0.12 be released?
Just wondering if the same performance could be achieved on master if I inject encoders I need to check max_height and a few other parameters or flex_vehicleprofile is supposed to have better performance?
RE: LM. Am I right in saying that performance deteriorates when “base profile” significantly deviates from run time parameters
I reboot the server, i still get the same question, how you do it? I just comment the routing.ch.disabling_allowed = true in the config .yml, but there shows the Disabling CH not allowed on the server-side.
