Make results fully reproducible

[drbenvincent]

Problems with doctests passing revealed that we were only passing the random_seed kwarg to the pm.sample method and that this was not being passed into pm.sample_prior_predictive or pm.sample_posterior_predictive.

This will hopefully avoid any future issues in doctest fragility. But it is also a pretty useful (if not essential) fix to make.

Closes #284.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (4772b34) 76.27% compared to head (ee8515e) 76.50%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #286      +/-   ##
==========================================
+ Coverage   76.27%   76.50%   +0.22%     
==========================================
  Files          20       20              
  Lines        1332     1345      +13     
==========================================
+ Hits         1016     1029      +13     
  Misses        316      316              

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