Upgrading the package to pymc4 (or now just called pymc) will not only give us some small performance increase but also to the possibility to use the jax sampler in the future.
As far as I can see we need to finish these tasks to accomplish the port:
[x] Replace all theano calls with aesara calls as this is now the new backend
[x] Modify the plotting routines to use arviz inference data as the old pymc trace is deprecated now
[x] Rerun all example notebooks/scripts
[x] one bundesland example
[x] hierarchical bundeslaender example
[x] science paper scenarios
[x] what if
[x] Check documentation i.e. if intersphinx mapping is still working
Upgrading the package to pymc4 (or now just called pymc) will not only give us some small performance increase but also to the possibility to use the jax sampler in the future.
As far as I can see we need to finish these tasks to accomplish the port: