Daniel Weindl

Would be great to set up benchmarks/workflows with https://github.com/benchmark-action/github-action-benchmark / pytest-benchmark.

Perform gradient checks for SBML test suite

5

Motivated by https://github.com/AMICI-dev/AMICI/issues/1230 Should be done for both forward and adjoint sensitivities (#18).

Would be great to clean up error handling in AMICI. Currently, the different exception classes are used rather inconsistently. In particular, it would be quite helpful to distinguish more clearly...

E.g. most of ModelState is missed in Model serialization. Probably others as well.

Run full model code generation for test models during CI from Python

1

Currently model code is pregenerated from Matlab, due to Matlab not being available on CI servers. We should implement those models so that they are importable from Python. (Some features...

e.g. ``` TestCase 01256 was skipped: Encountered unsupported expression "rateOf" of type "" as part of a Rule: "rateOf(S1)"! ```

Currently all observables defined in the SBML model are used during model import. There should be a option to filter based on the available measurements. This would be quite helpful...

Improper handling of reserved symbols in SBML import

3

Using a parameter with ID `k` in an SBML model results in compiler errors. Did not yet investigate in detail, but the problem seems to be that `_clean_reserved_symbols()` does not...

Add test case comparing -O3 model simulation with -O2

3

Would be good to add a test case comparing simulation results from -O3 compiled model to those from -O2 for a non-trivial model. For repeated simulations during parameter estimation, -O3...

Use SUNLinSolWrapper in NewtonSolver

1

We have C++ wrappers for the Sundials linear solvers, but the `NewtonSolver`s still use the raw objects. This makes those classes unnecessarily complex. Should use those wrappers there.