Updated convert_shape and convert_size

[5hv5hvnk]

this is another PR for #5417 which I got closed because of deletion of my forked repo

convert_size wrongly assumes symbolic sizes have to be scalars

import pymc as pm
import aesara.tensor as at

s = at.scalar('s')
size = at.stack([s, s])
x = at.random.normal(size=size)
x.eval({s: 2})
# array([[-0.66285345,  0.97341598],
#        [ 0.05158132,  2.03399059]])

pm.Normal.dist(size=size)
# ValueError: The `size` parameter must be a tuple, TensorVariable, int or list. Actual: <class 'aesara.tensor.var.TensorVariable'>

fixed by changing code in

pymc/pymc/distributions/shape_utils.py

Line 475 in 9155922

if isinstance(size, int) or (isinstance(size, TensorVariable) and size.ndim == 0):

by accepting size.ndim == 1 as well as size.ndim == 0

Closes #5394

[5hv5hvnk]

@twiecki @michaeloriordan please have a look

[michaelosthege]

Good call, @5hv5hvnk to just make a new PR.

Please address thecomments & run the pre-commit before I trigger the CI

[codecov]

Codecov Report

Merging #5601 (6163ca6) into main (09bddde) will increase coverage by 0.03%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #5601      +/-   ##
==========================================
+ Coverage   87.57%   87.60%   +0.03%     
==========================================
  Files          76       76              
  Lines       13741    13745       +4     
==========================================
+ Hits        12033    12041       +8     
+ Misses       1708     1704       -4     

Impacted Files	Coverage Δ
pymc/distributions/shape_utils.py	97.07% <100.00%> (+0.07%)	⬆️
pymc/parallel_sampling.py	87.70% <0.00%> (+1.32%)	⬆️

[5hv5hvnk]

This failed Arviz compatibility checks, I manually updated Arviz but can't commit because there are no changes in any of the files, how can I fix this?

[michaelosthege]

This failed Arviz compatibility checks, I manually updated Arviz but can't commit because there are no changes in any of the files, how can I fix this?

This is a rather new problem in our CI pipeline. I think I know how to fix it.
I'll check out your branch and add the fix. Otherwise your changes look fine and I'm optimistic that we can merge it if my CI fix works.

[michaelosthege]

I force-pushed your branch after squashing your two commits.
This way we can do a rebase-merge that'll add these two logically distinct commits to our main branch.

However, it looks like GitHub Actions is currently down (https://githubstatus.com).
Let's see when it comes back and if the CI works out this time.

Your PR is most likely the next one that'll get merged 👍