New Penalty Tests Logic

[ArthurCRodrigues]

We want to split the template library function tests into two: normal ones and penalty ones. The why of that is because we can have two approaches: if the test pass it is okay or if the test fails it is okay. So we decided to implement if the test pass it is okay, but we will split into two types of test to help the user to understand how it needs to be configured, as it will highlight how the process works.

[ArthurCRodrigues]

@matheusmra this was an old issue for fixing the way you currently handle penalty tests.
I noticed that recently you made some fixes to the existing penalty logic, could you give us a brief overview of that fix and explain how penalty tests should be like now?

[matheusmra]

@matheusmra this was an old issue for fixing the way you currently handle penalty tests. I noticed that recently you made some fixes to the existing penalty logic, could you give us a brief overview of that fix and explain how penalty tests should be like now?

Hey! Yeah, the penalty logic had a flaw in the way penalty_score was being interpreted.
The old formula was: (penalty_score * penalty_weight / 100)
The problem is that penalty_score represents how well the student passed the penalty tests (0 = failed all, 100 = passed all). So the old formula was doing the opposite of what we wanted it was deducting the most points when the student passed all penalty tests, and deducting nothing when they failed all of them.
The fix:

The ((100 - penalty_score) / 100) * penalty_weight) formula inverts the score to represent how much the student failed, which is what should drive the deduction.
Now:
Student passed all penalty tests → [penalty_score = 100] → [penalty_points = 0] (no deduction)
Student failed all penalty tests → [penalty_score = 0] → [penalty_points = penalty_weight] (full deduction)