Add extra args for includeconshdlr for new SCIP version

[Opt-Mucca]

The most recent version of SCIP is incompatible with the latest PySCIPOpt version. This fixes that by adding some new arguments to the include cons handler function.

[ju-manns]

I got the following assertion error:

18:27:07 _________________________________ test_circle __________________________________
18:27:07 
18:27:07     def test_circle():
18:27:07         points =[
18:27:07                 (2.802686, 1.398947),
18:27:07                 (4.719673, 4.792101),
18:27:07                 (1.407758, 7.769566),
18:27:07                 (2.253320, 2.373641),
18:27:07                 (8.583144, 9.769102),
18:27:07                 (3.022725, 5.470335),
18:27:07                 (5.791380, 1.214782),
18:27:07                 (8.304504, 8.196392),
18:27:07                 (9.812677, 5.284600),
18:27:07                 (9.445761, 9.541600)]
18:27:07     
18:27:07         m = Model()
18:27:07         a = m.addVar('a', lb=None)
18:27:07         b = m.addVar('b', ub=None)
18:27:07         r = m.addVar('r')
18:27:07     
18:27:07         # minimize radius
18:27:07         m.setObjective(r, 'minimize')
18:27:07     
18:27:07         for i,p in enumerate(points):
18:27:07             # NOTE: SCIP will not identify this as SOC constraints!
18:27:07             m.addCons( sqrt((a - p[0])**2 + (b - p[1])**2) <= r, name = 'point_%d'%i)
18:27:07     
18:27:07         m.optimize()
18:27:07     
18:27:07         bestsol = m.getBestSol()
18:27:07         assert abs(m.getSolVal(bestsol, r) - 5.2543) < 1.0e-3
18:27:07 >       assert abs(m.getSolVal(bestsol, a) - 6.1230) < 1.0e-3
18:27:07 E       AssertionError: assert 0.0014526025465624315 < 0.001
18:27:07 E        +  where 0.0014526025465624315 = abs((6.124452602546563 - 6.123))
18:27:07 E        +    where 6.124452602546563 = <built-in method getSolVal of pyscipopt.scip.Model object at 0x7f3eb805fee8>({'a': 6.124452602546563, 'b': 5.470086589738751, 'r': 5.254360607131121}, a)
18:27:07 E        +      where <built-in method getSolVal of pyscipopt.scip.Model object at 0x7f3eb805fee8> = <pyscipopt.scip.Model object at 0x7f3eb805fee8>.getSolVal
18:27:07 
18:27:07 tests/test_nonlinear.py:140: AssertionError

[Opt-Mucca]

@ju-manns I am confident that the test failing has nothing to do with my change. These numeric errors with the non-linear tests seem to be a frequent problem. They are especially common when you test without IPOPT enabled.
I haven't spent the time exploring them to decide if it's a Python problem that is responsible for the tolerances, if it's a problem with the underlying test and the values they compare to, or if it's simply a problem with SCIP

[Joao-Dionisio]

The assertions are passing on my end. But I don't like having the fixed 1.0e-3 there, it should be dependent on the model, I believe.

[ju-manns]

I discussed this with Ksenia and we decided to raise those bounds for now.

[codecov]

Codecov Report

Attention: Patch coverage is 33.33333% with 8 lines in your changes are missing coverage. Please review.

Project coverage is 52.85%. Comparing base (06cb349) to head (fd85f23).

Files	Patch %	Lines
src/pyscipopt/conshdlr.pxi	33.33%	8 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #785      +/-   ##
==========================================
+ Coverage   52.76%   52.85%   +0.09%     
==========================================
  Files          17       17              
  Lines        3817     3829      +12     
==========================================
+ Hits         2014     2024      +10     
- Misses       1803     1805       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.