Interface SCIP logicor constraints

[mohitt31]

Adds the logicor constraint handler to the Python interface. It is one of the unchecked items in #977.

Methods added:

• addConsLogicor(vars, ...) creates the constraint
• addCoefLogicor(cons, var) adds a variable to an existing one
• getNVarsLogicor(cons) and getVarsLogicor(cons)
• getDualsolLogicor(cons) and getDualfarkasLogicor(cons)

They follow the same shape as the existing AND and OR handlers, so the keyword arguments and their defaults match addConsAnd.

Tested with SCIP 10.0.3 on macOS (arm64). The suite goes from 449 to 450 passed with the same skips, and stubs/test.sh reports no issues against the stubs I added.

One thing I ran into while writing the handler check in getVarsLogicor. The equivalent line in getVarsAnd (scip.pxi:6956) is

assert(constype == 'and', "The constraint handler %s does not have this functionality." % constype)

That asserts a two element tuple, which is always truthy, so the guard never actually fires. I wrote mine without the parentheses. I left the AND one alone to keep this PR to one thing, but happy to send a separate fix if you want it.

[Joao-Dionisio]

Hello @mohitt31 , thank you so much! I'll only be able to take a look at this over the next week, I believe, but it's looking good!

[Joao-Dionisio]

happy to send a separate fix if you want it.

I would also be happy if you did it, if you don't mind 😄
We have another issue in getDualSolKnapsack():

        if cons.isOriginal():
            transcons = cons
        else:
            transcons = <Constraint>self.getTransformedCons(cons)

This is the wrong way round, of course, and is also present in the other farkas function. Would you mind also fixing this in that separate fix?

[mohitt31]

Sure, I'll do it. I checked and getDualfarkasKnapsack has the same problem, it passes the original constraint straight to SCIP. So the separate PR will fix both knapsack functions and the getVarsAnd assert, with tests that call them on original constraints. I'll open it in the next day or two.

[Joao-Dionisio]

Thank you @mohitt31 ! I'll take a look at your other PR later, maybe this week.