`random_nearby_position()` might return out-of-bounds positions

[sosiristseng]

Describe the bug

I found that the random_nearby_position() function returned out-of-bounds positions when I failed to run the event ABM example.

Minimal Working Example

https://github.com/sosiristseng/agents-out-of-bounds/blob/main/main.jl

The results: https://github.com/sosiristseng/agents-out-of-bounds/actions/runs/8627726973/job/23648255748

Agents.jl version

Agents v6.0.7

• Project.toml: https://github.com/sosiristseng/agents-out-of-bounds/blob/main/Project.toml
• Manifest.toml: https://github.com/sosiristseng/agents-out-of-bounds/blob/main/Manifest.toml

[sosiristseng]

Digging deeper, I found that IteratorSampling.jl might return an out-of-bound output when given an empty iterator. Maybe this issue should be transferred there.

iter = nearby_positions(model[1], model, 1)
iter_filtered = Iterators.filter(pos -> isempty(pos, model), iter)

# If I run collect(iter_filtered), it will return an empty array

# Returns a random, maybe out-of-bound output
IteratorSampling.itsample(abmrng(model), iter_filtered; method=:alg_R) 

[Tortar]

my mistake :q

I made it during a major overall of the IteratorSampling library, should be fixed by JuliaDynamics/StreamSampling.jl#48

[sosiristseng]

@Tortar thank you for your reply. It's resolved after IteratorSampling v0.2.10.

https://github.com/sosiristseng/agents-out-of-bounds/actions/runs/8628924786/job/23652013609