Closed rht closed 2 weeks ago
I think the run.py was originally created for 2 reasons...
I am not against your suggestion especially now that the runserver command exists.
I simplified the Boltzmann wealth model layout in https://github.com/projectmesa/mesa-examples/pull/44 into examples/boltzmann_wealth_model/:
__init__.py
(so that the model is importable)To run the GUI, simply do solara run app.py
.
How does this work with mesa runserver?
mesa runserver
could later be a proxy to solara run
, but probably later on in 3.0.
this is all deprecated with MESA 3.
run.py always contains a few lines of code. Should we merge it into server.py to simplify the coding convention? If people need to call
server.launch
from a different files (e.g. running Mesa simulation as a part of a larger web application), they can define their separate run.py. I'm citing the 80-20 UX rule (see https://lawsofux.com/pareto-principle/) that we should optimize for the 80% of the use case, and hide the remaining 20% until when it is actually needed.