add new methods and update testing for accuracy

[jc-macdonald]

This pull request expands the Operator-Partitioned Engine (OP Engine) to support new stiff ODE solvers, enhances test coverage for these methods, and improves documentation and code quality. The main changes include adding fully implicit and Rosenbrock methods, updating the test suite to validate their accuracy and requirements, and clarifying the public API in the documentation.

Solver and API enhancements:

• Added support for stiff ODE methods (implicit-euler, trapezoidal, bdf2, ros2) to CoreSolver, including optional Jacobian support, and updated the README to reflect these new capabilities. [1] [2]

Testing improvements:

• Expanded the MethodName literal and ScalarRunCase dataclass to handle new methods and Jacobians, and added comprehensive accuracy and convergence tests for the new implicit solvers. [1] [2] [3] [4] [5] [6] [7] [8] [9]
• Added new tests to ensure that implicit methods require a Jacobian and that BDF2 enforces a uniform time grid.

Code quality and linting:

• Updated pyproject.toml to ignore the PLR0914 linting rule in test files, reducing noise from parameter count warnings in pytest functions.

Minor test maintenance:

• Removed unnecessary noqa: PLR0914 comments from test functions after updating linting configuration. [1] [2]

[jc-macdonald]

supersceded by #18