move to v5

[Datseris]

Add changelogs and other entries in docs.

[codecov-commenter]

Codecov Report

Merging #591 (70bb028) into master (5d1182d) will decrease coverage by 0.17%.
The diff coverage is 63.63%.

@@            Coverage Diff             @@
##           master     #591      +/-   ##
==========================================
- Coverage   90.33%   90.15%   -0.18%     
==========================================
  Files          24       24              
  Lines        1562     1565       +3     
==========================================
  Hits         1411     1411              
- Misses        151      154       +3     

Impacted Files	Coverage Δ
src/Agents.jl	100.00% <ø> (ø)
src/spaces/openstreetmap.jl	71.91% <63.63%> (-0.68%)	⬇️

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[AayushSabharwal]

Is there anything else required here? What else needs to be done before we can release v5?

[Datseris]

If you don't mind, I'd like to have native osmplotting released. We are currently working on that with Fredrerik and I believe it will be done very very soon. THen, the OSM example will build its own plot. For this PR in particular I don't think there is anything more. Did you check if I forget any breakign change?

[Datseris]

The deprecations file is also empty, we can put there the names that were changed from the Pathfinding

[AayushSabharwal]

If you don't mind, I'd like to have native osmplotting released. We are currently working on that with Fredrerik and I believe it will be done very very soon. THen, the OSM example will build its own plot.

Okay, that is for the best.

[Datseris]

Actually @AayushSabharwal I'm thinking of releasing Agents.jl v5 without the plotting for open street map. The OSM plotting may take too long and is unecessarily holding back a very important release that goes into Graphs.jl. What do you think?

[AayushSabharwal]

InteractiveDynamics should still work with the current setup, bar OSMSpace right? If OSM plotting will take that long, I don't see the issue in releasing this now. Either way, that will be a new version of InteractiveDynamics and not Agents.jl, if I'm not mistaken?

[Datseris]

Yes you are correct. My rationale so far is that it would be good to state in the release message that "we also have native OSM plotting now". Let me see how this weekend goes and then I'll act accordingly.

[AayushSabharwal]

I have a couple hours now with not much to do. I've still not been able to figure out @deprecate. Is there anything else that needs cleaning up here?

[Datseris]

. I've still not been able to figure out @deprecate

Can you please remind me what this issue is?

[AayushSabharwal]

Can you please remind me what this issue is?

I think I understand what I was doing wrong now. It seems the old method used in @deprecate needs to be defined as well. I'll add dummy methods to the OSM submodule and then try add deprecations.

EDIT: Nope @deprecate still refuses to work. I tried defining intersection and road inside the OSM submodule and adding @deprecate OSM.intersection OSM.nearest_node (after merging master in, of course) and it doesn't work. I think it has problems with deprecating functions inside submodules

[Datseris]

I am hoping to merge and tag this today. Depends on how well JuliaDynamics/InteractiveDynamics.jl#87 will go. (actually, I will merge and tag this today no matter what. The plots of open street map will be updated after the 5.0 tag anyways)

[Datseris]

@AayushSabharwal please help! In the zombies.jl example, I do

step!(model, agent_step!, 1)

and I get

julia> step!(model, agent_step!, 1)
ERROR: Not implemented for space type OpenStreetMapSpace
Stacktrace:
 [1] error(s::String)
   @ Base .\error.jl:33
 [2] notimplemented(model::AgentBasedModel{Agents.OSM.OpenStreetMapSpace, Zombie, typeof(Agents.Schedulers.fastest), Dict{Symbol, Float64}, MersenneTwister})
   @ Agents C:\Users\datse\.julia\dev\Agents\src\core\space_interaction_API.jl:30
 [3] move_along_route!(agent::Zombie, model::AgentBasedModel{Agents.OSM.OpenStreetMapSpace, Zombie, typeof(Agents.Schedulers.fastest), Dict{Symbol, Float64}, MersenneTwister}, args::Nothing; kwargs::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
   @ Agents C:\Users\datse\.julia\dev\Agents\src\core\space_interaction_API.jl:132
 [4] move_along_route!(agent::Zombie, model::AgentBasedModel{Agents.OSM.OpenStreetMapSpace, Zombie, typeof(Agents.Schedulers.fastest), Dict{Symbol, Float64}, MersenneTwister}, args::Nothing)
   @ Agents C:\Users\datse\.julia\dev\Agents\src\core\space_interaction_API.jl:132
 [5] agent_step!(agent::Zombie, model::AgentBasedModel{Agents.OSM.OpenStreetMapSpace, Zombie, typeof(Agents.Schedulers.fastest), Dict{Symbol, Float64}, MersenneTwister})
   @ Main c:\Users\datse\.julia\dev\Agents\examples\zombies.jl:82
 [6] step!(model::AgentBasedModel{Agents.OSM.OpenStreetMapSpace, Zombie, typeof(Agents.Schedulers.fastest), Dict{Symbol, Float64}, MersenneTwister}, agent_step!::typeof(agent_step!), model_step!::typeof(dummystep), n::Int64, agents_first::Bool)
   @ Agents C:\Users\datse\.julia\dev\Agents\src\simulations\step.jl:55
 [7] step!(model::AgentBasedModel{Agents.OSM.OpenStreetMapSpace, Zombie, typeof(Agents.Schedulers.fastest), Dict{Symbol, Float64}, MersenneTwister}, agent_step!::Function, n::Int64, agents_first::Bool) (repeats 2 times)
   @ Agents C:\Users\datse\.julia\dev\Agents\src\simulations\step.jl:45
 [8] top-level scope
   @ REPL[6]:1

!!! Why? It is because of the distance_left = move_along_route!(agent, model, agent.speed * model.dt). Without the distance_left, i.e., using the version in src/models/zombies.jl, I don't get the error!

[Datseris]

I think is_stationary is bugged and doesn't give false even if move_along_route! returns nothiung.

[AayushSabharwal]

I'm afraid I was asleep when you pinged me. I'll try and take a look in some time, maybe even try and see what's causing #595. Good to see v5 is out though, it's been a long time coming!

[Datseris]

Notice that I've fixed it by making the function return 0 instead of missing. This means I didn't "actually" fix the issue. There were still occassions where the agent reached the destinatiion (remianing distance 0), but it was not identified as is_stationary. ANd yeah, I've noticed #595 as well when building the docs!