Path not populated until routing ends

[ludohenin]

Hi,

Thanks for your amazing work.
I'm looking at a way to define the number of steps to iterate to end routing from here rather than reaching the destination.
Is there an easy way to do that ?

The other thing (but it's maybe more a feat request) nice to have would be the possibility to have shorter steps the 1st 24h (say 20min) and 1 hour or more after.

[dakk]

Hi Ludohenin, thanks for your appreciation.
Regarding the first question, I don't really understand what are you meaning; you can stop calling Routing.step() even if you didn't reach destination.

The second thing could be a cool feature; a good start should be to add a timeDelta parameter to Router(s), and then let the library user to modify it at each step.

[ludohenin]

Stop calling the routing.step() is what I did but the path property doesn't get populated. Maybe I can do that manually.

For the new feat, I think I got how to do it, I'll drop a PR for that soon.

[dakk]

Stop calling the routing.step() is what I did but the path property doesn't get populated. Maybe I can do that manually.

Ok, now I understand. It would be great if you fix also this in your PR

For the new feat, I think I got how to do it, I'll drop a PR for that soon.

Great, thank you!
Can I ask you how are you using this library? I'm really pleased that someone is using it.

[ludohenin]

Hi @dakk
Back after a very long while.

With regard to my issue to getting the path generated, I found that in the code below, generate_path never gets called. distance_to_end_point (~3400 in my use case track = ((47.0, -3.0), (-5.0, -30.0))) is always > minIncrease param value (10.0) and nearest_solution remains falsy.

I'm lost, I don't know what I'm doing wrong.

libweatherrouting/weatherrouting/routers/linearbestisorouter.py

Lines 66 to 83 in a4fb0fa

	for p in isoc[-1]:
	distance_to_end_point = p.pointDistance(end)
	if distance_to_end_point < self.getParamValue("minIncrease"):
	# (twd,tws) = self.grib.getWindAt (time + datetime.timedelta(hours=1),
	# p.pos[0], p.pos[1])
	maxReachDistance = utils.maxReachDistance(p.pos, p.speed)
	if distance_to_end_point < abs(maxReachDistance * 1.1):
	if (
	not self.pointValidity or self.pointValidity(end[0], end[1])
	) and (
	not self.lineValidity
	or self.lineValidity(end[0], end[1], p.pos[0], p.pos[1])
	):
	if distance_to_end_point < nearest_dist:
	nearest_dist = distance_to_end_point
	nearest_solution = p
	if nearest_solution:
	generate_path(nearest_solution)

[ludohenin]

@dakk I found a ugly trick for now, adding an arg to step to set isLast and compute the path like this (from the code above):

            for p in isoc[-1]:
                distance_to_end_point = p.pointDistance(end)
                if distance_to_end_point < self.getParamValue("minIncrease"):
                    # (twd,tws) = self.grib.getWindAt (time + datetime.timedelta(hours=timedelta),
                    # p.pos[0], p.pos[1])
                    maxReachDistance = utils.maxReachDistance(p.pos, p.speed)
                    if distance_to_end_point < abs(maxReachDistance * 1.1):
                        if (
                            not self.pointValidity or self.pointValidity(end[0], end[1])
                        ) and (
                            not self.lineValidity
                            or self.lineValidity(end[0], end[1], p.pos[0], p.pos[1])
                        ):
                            if distance_to_end_point < nearest_dist:
                                nearest_dist = distance_to_end_point
                                nearest_solution = p
                
                # Added part to get path when ending steps before reaching the waypoint
                if isLast:
                    if not nearest_solution:
                        nearest_solution = p
                    else:
                        nearest_solution = p if distance_to_end_point < nearest_solution.pointDistance(end) else nearest_solution

            if nearest_solution:
                generate_path(nearest_solution)

[dakk]

Your solution works but is not very elegant; I made an alternative in #17 , please give a try and see if matches your requirement