Update Aesara dependency

[ricardoV94]

No description provided.

[codecov]

Codecov Report

Merging #5757 (42e3745) into main (42e3745) will not change coverage.
The diff coverage is n/a.

❗ Current head 42e3745 differs from pull request most recent head ebdc1e1. Consider uploading reports for the commit ebdc1e1 to get more accurate results

@@           Coverage Diff           @@
##             main    #5757   +/-   ##
=======================================
  Coverage   88.95%   88.95%           
=======================================
  Files          74       74           
  Lines       13811    13811           
=======================================
  Hits        12286    12286           
  Misses       1525     1525           

[michaelosthege]

🛳 it

[michaelosthege]

No, wait. That failing test looks like a new bug?

[ricardoV94]

Aesara issue opened: aesara-devs/aesara#955

[twiecki]

Can we xfail or is this critical?

[michaelosthege]

Looks like some kind of tensor typing problem:
https://github.com/pymc-devs/pymc/runs/6374288465?check_suite_focus=true#step:7:1429

TypeError: The type of the replacement (TensorType(float64, (1, None))) must be compatible with the type of the original Variable (TensorType(float64, (1, 20))).

[ricardoV94]

No need to guess guys, there's a linked issue: aesara-devs/aesara#955

[ricardoV94]

This should only show up as a warning under normal aesara flags, and only in some edge cases. It will be fixed in a future release but for now I think it's fine to temporarily xfail

[ricardoV94]

The test that is failing will be removed in #5806, it was not easy to turn it off due to the way it is parametrized.

In the meantime I rebased from main to see if anything else fails.

[twiecki]

Seems still broken, or another test?

[ricardoV94]

It's the same test, it will be removed in #5806, then we can rebase and it should be good to go.