188 random walk refactor

[dylanhmorris]

This addresses #188 and in the process attempts to move things in the direction of the proposed RV design spec from #266

[codecov]

Codecov Report

Attention: Patch coverage is 95.55556% with 2 lines in your changes missing coverage. Please review.

Project coverage is 92.80%. Comparing base (ef87238) to head (71bd146).

Files	Patch %	Lines
model/src/pyrenew/metaclass.py	96.96%	1 Missing ⚠️
model/src/pyrenew/process/simplerandomwalk.py	88.88%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #275   +/-   ##
=======================================
  Coverage   92.79%   92.80%           
=======================================
  Files          40       39    -1     
  Lines         902      903    +1     
=======================================
+ Hits          837      838    +1     
  Misses         65       65           

Flag	Coverage Δ
unittests	92.80% <95.55%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[damonbayer]

What is the idea behind TransformedRandomVariable? Why not just create a DistributionalRV with a TransformedDistribution as the dist? Or even use TransformedDistribution as the implementation for the TransformedRandomVariable if you are trying to make a different interface.

[dylanhmorris]

What is the idea behind TransformedRandomVariable? Why not just create a DistributionalRV with a TransformedDistribution as the dist? Or even use TransformedDistribution as the implementation for the TransformedRandomVariable if you are trying to make a different interface.

We may wish to apply transformations to the output of RandomVariable.sample() calls for RVs other than DistributionalRVs. Indeed, in the examples here we use TransformedRandomVariable to transform the output of SimpleRandomWalk, which is not implemented as a numpyro.distributions.Distribution or as a DistributionalRV.

[dylanhmorris]

@damonbayer ready for re-review

[damonbayer]

Thanks @dylanhmorris! A few small comments. Nothing mandatory.