Improve example on expanded sample_posterior_predictive

Sample posterior_predictive relies on other groups to infer volatility, so it's important to keep the whole InferenceData.

Also there's a shape error if `sample_dims` is not specified

Improve docs

Fix copying of shared variables in `fgraph_from_model` (pymc-devs#7153)

* Do not use deprecated ScalarSharedVariable

* Recreate SharedVariables with exact type in fgraph_from_model

Add test for Blockwise logp regression

Bump docker/metadata-action from 5.0.0 to 5.3.0

Bumps [docker/metadata-action](https://github.com/docker/metadata-action) from 5.0.0 to 5.3.0.
- [Release notes](https://github.com/docker/metadata-action/releases)
- [Commits](docker/metadata-action@96383f4...31cebac)

---
updated-dependencies:
- dependency-name: docker/metadata-action
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <[email protected]>

Avoid inplace mutation in replace_rvs_by_values

This would happen when transforms reference other variables

Bump Pytensor dependency to >=2.18.1,<2.19

Fix typo in distribution dimensionality (pymc-devs#7005)

The order of shape got probably mixed-up here.

Support logp derivation of `power(base, rv)` (pymc-devs#6962)

Co-authored-by: Ricardo Vieira <[email protected]>

Co-authored-by: Luke LB <[email protected]>

Fix mypy failure