Using floats between 0 and 1 on the 2D list

[alejcas]

Hi, I need to set some fields on the 2D matrix that have a 'walk cost' of a quarter of a movement point (0.2)
But when doing so, the pathfinding algorithm starts to do some weird things.
If I instead use only integers the pathfinding works just great.

Is there any chance of using floats?

I'm using the A* algorithm.

[brean]

Interesting! In theory it should not make a difference. I tired to to recreate and fix it, but I wasn't successful (see the new test on the use_float-branch) so for now I'll mark this as a bug. I am not sure what causes the issue but I think it might have something to do with the way the path length gets calculated based on the heuristic.

The easiest workaround for now is to multiply the weight with a bigger number like 100.

[brean]

fixed in 1.0.13.
I forgot to multiply the heuristic with the weight, see:

python-pathfinding/pathfinding/finder/finder.py

Lines 77 to 81 in 89ca95c

	# in a weighted graph we also need to multiply the calculated
	# value with the weight of the node
	if self.weighted:
	nh *= node_a.weight
	return nh

Sorry it took nearly 5 years, but I am very happy to finally have figure this out!

[alejcas]

It’s never late my friend! Thanks