Fixed bug when accessing matrix variable attributes

[Joao-Dionisio]

Fix #969

There's still a bug with getCol, but that one is also a bug in the normal variables. I'm thinking it may just be about adding an assertion.

@DominikKamp do you know when SCIPvarGetCol returns NULL?

[DominikKamp]

Ideally, if there is no column to the variable for example if the variable is deleted already or no LP created yet.

[Joao-Dionisio]

@DominikKamp would it make sense to add a precondition to the method, even though it takes in SCIP_VAR* and not SCIP*? Or is returning NULL something to be expected? In the latter case, something should be said in the documentation, IMO.

[DominikKamp]

Maybe you meant to check the upper bounds of the columns after optimizing but you might need to disable presolving before.

[DominikKamp]

Returning NULL would be okay but there is also an assertion that the variable must have a column which should fail before.

[DominikKamp]

The precondition can be that the variable is in the LP.

[Copilot]

Pull Request Overview

This PR addresses a bug related to accessing matrix variable attributes. The changes include:

• Adding a new test function (test_MatrixVariable_attributes) to verify multiple matrix variable attributes.
• Adjusting the changelog to reflect the bug fix.

Reviewed Changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated no comments.

File	Description
tests/test_matrix_variable.py	Introduces tests to verify methods for matrix variable attributes.
CHANGELOG.md	Updates the changelog with the bug fix description.

Files not reviewed (1)

• src/pyscipopt/scip.pxi: Language not supported

Comments suppressed due to low confidence (1)

tests/test_matrix_variable.py:303

• The test for getCol is commented out, leaving its behavior untested; consider adding a TODO or reference to the bug tracking issue (#969) to clarify why this test is disabled and outline plans for future validation.

#assert x.getCol().tolist()# == [[5, 6], [2, 8]]

[Joao-Dionisio]

Ah no, I didn't mean to check the upper bounds of the columns, the rhs of the check was there from copy-pasting. I just wanted to access the method. I think the error is explicit enough.

[mmghannam]

Looks good, thanks @Joao-Dionisio!