Current situation

• Running the full test suite on master takes about 2 minutes.
• Most of this time is spent on the solver-related tests.
• However, for these cases we’re only running static tests, not testing solver convergence.

Proposal

• Introduce lightweight Foo problem classes and Foo models (for solver tests).
• This would significantly reduce runtime, while still verifying the core logic.
• Place these helpers in a central tests/helpers.py module to keep test code clean and DRY.

Future idea

• Add dynamic tests to verify solver convergence.
• We already do this in-house, but having a structured approach in the repo would be valuable.

Questions

1. Should we add lightweight helper classes/models for faster static tests?
2. Should we define a structured way to run convergence tests separately from the main suite?