Release 0.16.0 - #1956
Conversation
More informative `__repr__` for `CvodeSolver`, `IDASolver`, and `SolverPtr`. Closes #1926 Output of `print(amici.CVodeSolver())`: ``` <Swig Object of type 'amici::CVodeSolver *' at 0x7f988f70f090 reporting_mode: <RDataReporting.full: 0> sens_meth: <SensitivityMethod.forward: 1> sens_order: <SensitivityOrder.none: 0> sens_meth_preeq: <SensitivityMethod.forward: 1> maxsteps: 10000 maxtime: 1.7976931348623157e+308s abs_tol: 1e-16 rel_tol: 1e-08 abs_tol_b: 1e-16 rel_tol_b: 1e-08 abs_tol_fsa: 1e-16 rel_tol_fsa: 1e-08 abs_tol_quad: 1e-12 rel_tol_quad: 1e-08 abs_tol_ss: 1e-14 rel_tol_ss: 1e-06 abs_tol_sss: 1e-14 rel_tol_sss: 1e-06 int_sens_meth: <InternalSensitivityMethod.simultaneous: 1> int_type: <InterpolationType.hermite: 1> linsol: <LinearSolver.KLU: 9> lmm: <LinearMultistepMethod.BDF: 2> newton_damp_mode: <NewtonDampingFactorMode.on: 1> newton_damp_lb: 1e-08 newton_maxsteps: 0 newton_ss_check: False sens_ss_check: True interpolation_type: <InterpolationType.hermite: 1> ism: <InternalSensitivityMethod.simultaneous: 1> nlsol_iter: <NonlinearSolverIteration.newton: 2> stability_limit: True state_ordering: 0 > ```
Quite helpful for debugging.
This makes it more convenient to create a new (unmodified) model instance, or to find out where the model files are located. Also adds a snake_case version of `getModel`.
AMICI_FIRST_RHSFUNC_ERR
* Skip irrelevant frames in stacktraces * Add optional startlevel for stacktrace * Fix maxLevel, which should be interpreted relative to the startlevel
Use apt, which wasn't possible back then...
* move .binder/ to binder/, as jupyter lab won't allow us accessing hidden directories by default * add landing page for easier access of different notebooks [](https://mybinder.org/v2/gh/AMICI-dev/AMICI/binder_landing?labpath=binder%2Foverview.ipynb) * binder link to readme / doc
* log all nested execution times as debug * set logging for performance test to debug
after #1939, we can simply install it via rtd_requirements.txt
Fixes: ``` [~] /home/runner/work/AMICI/AMICI/include/amici/edata.h:118: warning: Member operator==(const ExpData &lhs, const ExpData &rhs) (friend) of class amici::ExpData is not documented. [~] /home/runner/work/AMICI/AMICI/include/amici/forwardproblem.h:295: warning: amici::ForwardProblem::applyEventBolus has @PARAM documentation sections but no arguments [~] /home/runner/work/AMICI/AMICI/include/amici/steadystateproblem.h:188: warning: argument 'newtonSolver' of command @PARAM is not found in the argument list of amici::SteadystateProblem::computeSteadyStateQuadrature(const Solver &solver, Model &model) [~] /home/runner/work/AMICI/AMICI/include/amici/amici.h:47: warning: argument 'return_code' of command @PARAM is not found in the argument list of amici::simulation_status_to_str(int status) [~] /home/runner/work/AMICI/AMICI/include/amici/amici.h:47: warning: The following parameter of amici::simulation_status_to_str(int status) is not documented: [~] parameter 'status' ``` - https://github.com/AMICI-dev/AMICI/actions/runs/3978146452/jobs/6819757866 +clang-format
Fixes `warning: comparison of integer expressions of different signedness`
* add edata __repr__ fixes #1941 * add test * improve tests * fix * fix test * added __repr__ to SwigPtrView * fix counts * add info from SimulationParameters * fixup * Update python/sdist/amici/numpy.py Co-authored-by: Daniel Weindl <dweindl@users.noreply.github.com> Co-authored-by: Daniel Weindl <dweindl@users.noreply.github.com>
* GHA: Run tests on Python3.11 * Fixes related to Python3.11 changes to enums (https://docs.python.org/3/whatsnew/3.11.html#enum) * Fix unreliable string `is` comparison
cd43b8d to
0e4bb13
Compare
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## master #1956 +/- ##
==========================================
+ Coverage 75.96% 75.98% +0.01%
==========================================
Files 76 76
Lines 12967 12982 +15
==========================================
+ Hits 9851 9864 +13
- Misses 3116 3118 +2
Flags with carried forward coverage won't be shown. Click here to find out more.
|
Initial states should always be checked for finite values, not only with `Model.setAlwaysCheckFinite(True)`. As this only happens once per simulation, the overhead is negligible. Closes #1952
Notebook demonstrating common simulation failures and show how to analyze / fix them. Co-authored-by: Fabian Fröhlich <fabian@schaluck.com>
0ecf53e to
3aa4579
Compare
|
SonarCloud Quality Gate failed. |
FFroehlich
left a comment
There was a problem hiding this comment.
👍 is this really a minor release? To me this primarily looks like fixes and documentation updates, no?
Define "fix", define "new functionality". 🤷♂️ |








No description provided.