JPS compiled CPP w/python bindings- https://github.com/fgenesis/jps/blob/master/JPS.h

from Bzztploink

  • bpr, FILE , unzip into a stock win64 version of 2.79, import jps and you’re on your own from there

the info about the commands is embeeded into the .h file
located here - https://github.com/fgenesis/jps/blob/master/JPS.h

I may need help getting this working.

Attachments

jps.zip (76.3 KB)

Sorry but the wrappers are useless. I don’t know where you found them but they are for Python2.79 and BGE uses Python3.x.
The warpper is made with swig, and since you have the source code (the .h) you can rebuild them again for Python3. Check: http://www.swig.org/tutorial.html

Making this wrappers is not trivial, and I still doubt they are cross-platform (mainly because the shared libraries from C++ are not). You could however make them cross-platform if you do one wrapper for each platform and then chose which to load depending on which system the user is in.

That said, are you sure you need jps? You probably don’t.

yeah the world open dyanamic infinite procedutal terrain and a breadth first search in py was quite slow,

jps I have working in py, but it’s also quite slow when rebuilding the grid sim each pathing, right now each agent paths a 32x32 node area around them in py,

this is my current setup

with compiled code I could run the same search with more agents and a larger pathing area.

currently it’s like a very much improved seek.
but if the obstacle is too large the pathing will fail atm.

with 64x64 path or 96x96 I think the units will be quite smart witbout having to rebuild the whole world grid area each time tiles are added or removed. (202 tiles x(17x17)) vertex