Brandon T. Willard

> I am struggling to setup the dev environment locally on MacOS to run the test cases. Were you able to resolve this?

> Should I go ahead and make `ChoiceRV` behave similarly to `PermutationRV` or is there a reason for this inconsistency? If it's not behaving like it does in NumPy, then...

> Also, what about my [comment](https://github.com/aesara-devs/aesara/pull/1105#issuecomment-1218334644) about `ChoiceRV` behaving differently from its NumPy equivalent? Looks like that might not have been implemented yet, so we need an issue for it.

FYI: I've been waiting to see the rendered documentation, but the job doesn't seem to finish.

> Regarding the docstring for the `size` argument. I think that the following: > > ```python > size > Sample shape. If the given size is `(m, n, k)`, then...

> The binomial is pretty clear: > > > size: int or tuple of ints, optional > > Output shape. If the given shape is, e.g., (m, n, k), then...

> Hello, @brandonwillard. Please accept my apologies for forgetting to link my analysis on this; I have now updated the description of this PR with my findings on both `solve`...

> Can we completely replace the existence `solve` with this Numpy `solve` ? I don't think so; I recall each as having distinct options/approaches.

> Although the benchmark is no longer relevant, I tested both `solve` methods out of curiosity and found that the Numpy implementation of `solve` is **2-3 times** faster than the...

Here's an example test method that can be added to `tests.test_ifelse.TestIfElse`: ```python def test_constant_folding(self): class TrackingOp(Op): n_evaled = 0 def make_node(self, inp): return Apply(self, [inp], [inp.type()]) def perform(self, node, inputs,...