Distribution of agents in areas with obstacles

[chraibi]

When setting up agent distribution in specific areas, it's crucial to ensure that spawning zones don't overlap with obstacles.

Consider a scenario where agents are evenly spread across an entire geometry, including obstacles. This is achieved effectively using the following function, which treats obstacles as part of the geometry:

distribute_agents(spawning_area=walkable_area)

In contrast, a more common situation involves distributing agents within a designated spawning_area. An example of this would be:

distribute_agents(spawning_area=spawning_area)

This method, however, may result in agents appearing inside obstacles.

The question arises: Should the jps.distribution functions automatically account for obstacles within spawning areas? Or is it more appropriate for users to meticulously select spawning areas, even though it may be less convenient?

[Ozaq]

This can be achieved by taking the intersection of your rectangle and the walkable area, see: https://shapely.readthedocs.io/en/stable/reference/shapely.intersection.html#shapely.intersection

Example:

distribute_agents(spawning_area=shapely.intersection(spawning_area, walkable_area))