chore: adding NoneConst and None check

[Hashcode-Ankit]

Description

Fixes issue: #7762

Related Issue

• Closes #
• Related to #

Checklist

• Checked that the pre-commit linting/style checks pass
• Included tests that prove the fix is effective or that the new feature works
• Added necessary documentation (docstrings and/or example notebooks)
• If you are a pro: each commit corresponds to a relevant logical change

Type of change

• New feature / enhancement
• Bug fix
• Documentation
• Maintenance
• Other (please specify):

---

📚 Documentation preview 📚: https://pymc--7764.org.readthedocs.build/en/7764/

[welcome]

]
💖 Thanks for opening this pull request! 💖 The PyMC community really appreciates your time and effort to contribute to the project. Please make sure you have read our Contributing Guidelines and filled in our pull request template to the best of your ability.

[ricardoV94]

This whole check is a bit of a mess we should think a bit more. It could also be a non-constant slice (see if not)

[Hashcode-Ankit]

@ricardoV94 by this line it could also be a non-constant slice do you mean to say that dynamic slices (non-constant slice) are also need to be skipped?

Also what is your view if we check for tensor variable and only in that case graph node get modified? (solution for mess)

[ricardoV94]

I think this check should be just something like

if any(
            (
                    isinstance(indices, TensorVariable)
                    and indices.dtype.startswith("int")
                    and any(not b for b in indices.type.broadcastable)
                )
            )
            for indices in mixing_indices
        ):

Basically rulling out potentially repeated integer indices.

[Hashcode-Ankit]

@ricardoV94 are you saying to modify graph node in above condition only,
if it is true (which i understand) then we have to use all not any:

if all( // all not any
            (
                    isinstance(indices, TensorVariable)
                    and indices.dtype.startswith("int")
                    and any(not b for b in indices.type.broadcastable)
                )
            )
            for indices in mixing_indices
        ):
        // modify the nodes 
        
 else :
     None

[ricardoV94]

No, that if any I suggested is the one leading to return None. No need for a new branch

[Hashcode-Ankit]

then in None and NoneConst cases it is going to modify node which is not recommended in issue description :

I am running below code

from pytensor.tensor.type_other import NoneConst, NoneTypeT, SliceConstant, SliceType
from pytensor.tensor import TensorVariable
mixing_indices = [None, NoneConst]

if any(
   (
      isinstance(indices, TensorVariable)
      and indices.dtype.startswith("int")
      and any(not b for b in indices.type.broadcastable)
   )
   for indices in mixing_indices
  ):
  print("return none")
else:
  print("modify node")

[ricardoV94]

The only thing we should have to worry about is repeated indices, None/NoneConst don't introduce repeated indices, just a dummy dimension

[Hashcode-Ankit]

in issue:

The [line that I quoted above](https://github.com/pymc-devs/pymc/blob/main/pymc/logprob/mixture.py#L294) needs to also check if the indices are None or [NoneConst](https://github.com/pymc-devs/pytensor/blob/main/pytensor/tensor/type_other.py#L132). That way, the rewrite will return None when it has a mixture of integer indexes, and slices or new axis.

[Hashcode-Ankit]

Hi @ricardoV94 i have done the changes, I am new to codebase (hope you understand), that is why having that much of questions.
Thanks