Skip to content

project_checks

The checks of a project as functions: the registry, the experiments, the fits.

A model package, or a repository of model packages, calls them from its own tests, so a package which is exported into a repository of its own keeps them (check_model checks that the project belongs to its package):

# not executed
@pytest.mark.parametrize("project", PROJECTS)
def test_experiments_registered(project: Project) -> None:
    project_checks.check_registry(project)

The known exceptions of the checks (experiments which are not registered on purpose, modules which do not import, the reason to run the experiments one at a time, the checks of a battery which fail) are the checks of the project, sbmlsim.project.CheckSettings, each with its reason.

The functions raise AssertionError with a message which names every problem, also for experiments which fail to initialize or to run.

import sbmlsim does not import this module; it imports the experiments and the fits when a check runs.

BatteryResult

Bases: Protocol

The result of a check of a battery, see check_battery.

Attributes:

Name Type Description
check object

the check, its name as its value (an Enum) or as str.
passed bool | None

whether the check passed, None if it does not apply.
status object

the status of the check, as its value (an Enum) or as str, e.g. not_applicable.
problems Sequence[str]

the problems of a failed check, one line each.

check property 

check

The check.

passed property 

passed

Whether the check passed, None if it does not apply.

status property 

status

The status of the check.

problems property 

problems

The problems of a failed check.

Battery

Bases: Protocol

A battery of checks of the models of a project, see check_battery.

results property 

results

The results of the checks.

experiment_modules 

experiment_modules(project)

Import the modules of the experiments of a project.

The module is CheckSettings.experiment_modules below the package of the project, a package whose modules are imported with it, or a module. The modules in CheckSettings.not_importable are skipped.

Parameters:

Name Type Description Default
project Project

the project.

 required

Returns:

Type Description
list[ModuleType]

The imported modules and the traceback of every module which does not
dict[str, str]

import, by its name relative to the package of the project.

Raises:

Type Description
AssertionError

for a project without a package.

defined_experiments 

defined_experiments(modules)

Get the experiments defined in modules, without the base experiments.

A base experiment is a class from which another of the found classes derives.

registered_experiments 

registered_experiments(project)

Get the experiments of the group all of a project.

check_registry 

check_registry(project)

Check that every experiment a project defines is one of its experiments.

Every module of the experiments imports (experiment_modules), every experiment is in the group all once, and the known exceptions of project.checks are still exceptions (a listed class is defined and not registered, a listed module exists).

Parameters:

Name Type Description Default
project Project

the project.

 required

Raises:

Type Description
AssertionError

naming every problem.

check_model 

check_model(project)

Check that a project belongs to its package and its references hold.

A slip in package or root of a project resolves the references of another project, or names the paths wrongly, without an error otherwise. With a package, the package imports, has the project as an attribute, its directory is root, and the name of the project is the last component of the package (or the component ends with _<name>). The variants, the directory of the factory, the directory of the fit records and the file of the adopted fits lie below root, the variants and the data paths exist, the references (the factory, the experiments, the fits, the credibility declaration, the release hook, the actions) resolve and the release hook is one, every action takes the arguments of its call, the experiments load (Project.load_experiments), the fits of a reference are not empty, and every parameter of an adopted fit has its value and a distribution in every variant (sbmlsim.experiment.draws.check_fit_parameters).

Parameters:

Name Type Description Default
project Project

the project.

 required

Raises:

Type Description
AssertionError

naming every problem.

initialize_experiments 

initialize_experiments(project)

Initialize the experiments with their data, figures, mappings.

Every experiment is initialized, also after one which cannot be.

Returns:

Type Description
dict[str, SimulationExperiment]

The experiments by name.

Raises:

Type Description
AssertionError

naming every experiment which cannot be initialized with its error; the tracebacks are in the log of sbmlsim.

run_all_experiments 

run_all_experiments(project, output_dir, *, n_workers=1)

Run the experiments of a project, as the kind simulate does.

Every experiment runs, also after one which failed, with the settings of the simulator and the figures of the project (Project.run_experiments, the implementation of Project).

Parameters:

Name Type Description Default
project Project

the project.

 required
output_dir Path

the directory of the results. With CheckSettings.one_at_a_time every experiment runs on its own, into a directory named after it: the memory of a run of all experiments of a large model exceeds that of a runner of a continuous integration; else they run together into all, with their figures collected.

 required
n_workers int | None

the processes of the run, serial by default, since the tests of a package run in processes of their own.

 1

Raises:

Type Description
AssertionError

naming every failed experiment and figure, and every run which wrote no report.

check_battery 

check_battery(project, run)

Run a battery of the checks of the models of a project.

A failed check is a problem unless the project knows it (CheckSettings.battery); a known failure whose check passes or does not apply no longer applies and is a problem as well.

Parameters:

Name Type Description Default
project Project

the project.

 required
run Callable[[Project], B]

runs the battery of the project, e.g. the battery of a framework with its settings.

 required

Returns:

Type Description
B

The battery.

Raises:

Type Description
AssertionError

naming every problem.

start_fits 

start_fits(project, output_dir, max_nfev=2)

Start a fit of every fit problem: one start with a few evaluations.

Parameters:

Name Type Description Default
project Project

the project.

 required
output_dir Path

the directory of the results of the fits.

 required
max_nfev int

evaluations of the cost per start.

 2

Returns:

Type Description
dict[str, dict[str, FitRun]]

The runs of every fit problem by its subset.

Raises:

Type Description
AssertionError

if a fit did not run.