Implement Uniform and TruncatedNormal dims distributions

[ricardoV94]

With the constraint that for free rvs the bounds must be constant. This touches on a limitation we have on transforms that depend on parameters. The previous strategy was to pass the rv inputs to the transform, but this doesn't work for derived variables like XTensorFromTensor and DimShuffled RVs, since the inputs of those are not the ones needed for the transform.

I'll be thinking about how to better handle this, so we can lift the restriction.

Also fixes a bug in eval_rv_shapes with XTensorVariables for which .shape returns tuples instead of a tensor

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.83%. Comparing base (5ff12c8) to head (97f9fda).
⚠️ Report is 4 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8065      +/-   ##
==========================================
+ Coverage   90.80%   90.83%   +0.02%     
==========================================
  Files         121      121              
  Lines       19443    19500      +57     
==========================================
+ Hits        17656    17713      +57     
  Misses       1787     1787              

Files with missing lines	Coverage Δ
pymc/dims/distributions/scalar.py	96.27% <100.00%> (+0.65%)	⬆️
pymc/dims/distributions/transforms.py	100.00% <100.00%> (ø)
pymc/model/core.py	93.30% <100.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[ricardoV94]

Also fixes a bug in eval_rv_shapes with XTensorVariables for which .shape returns tuples instead of a tensor

If you wonder why, I think that's what .shape should return so I decided to take that approach with XTensorVariables which are fresh anyway. Can reconsider.

This wouldn't have been a problem if pytensor function supported arbitrary tuple/list input/output format like any modern jit library does... We should open an issue for that

[OriolAbril]

everything looks and sounds good, I'll try to join the hackathons starting next week to go back to more active contributions on label related things

[ricardoV94]

@OriolAbril was there a reason you didn't approve the PR? Did you want someone else to review?