pymob.sim.config package#

Submodules#

pymob.sim.config.base module#

class pymob.sim.config.base.Config(config: str | ConfigParser | None = None)#

Bases: BaseModel

Configuration manager for pymob.

This class loads a settings.cfg file, validates it against the pydantic models defined above and provides convenient accessors.

Parameters:: config (str or configparser.ConfigParser, optional) – Path to a configuration file or an already parsed ConfigParser.

case_study: Casestudy#

create_directory(directory: Literal['results', 'scenario'], force: bool = False) → None#

Create a results or scenario directory if it does not exist.

Parameters:

directory (Literal["results", "scenario"]) – Which directory to create.
force (bool, optional) – If True create without prompting.

data_structure: Datastructure#

error_model: Errormodel#

import_casestudy_modules(reset_path: bool = False) → None#

Import all modules of the current case study.

this script handles the import of a case study without the typical __init__.py file. It iterates over all .py files in the root directory of the case study (typically: sim, mod, stats, plot, data, prior) and imports them with import_module(…)

Parameters:: reset_path (bool, optional) – Reset sys.path before importing. Default is False.

import_simulation_from_case_study() → Any#

Retrieve the Simulation class defined in the case-study.

Returns:: The Simulation class object.
Return type:: Any
Raises:: ImportError – If the class cannot be found.

inference: Inference#

inference_numpyro: Numpyro#

inference_pyabc: Pyabc#

inference_pyabc_redis: Redis#

inference_pymoo: Pymoo#

property input_file_paths: List[str]#

List of all input files required for the simulation.

Return type:: List[str]

jaxsolver: Jaxsolver#

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'protected_namespaces': (), 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'case_study': FieldInfo(annotation=Casestudy, required=False, default=Casestudy(init_root='/home/docs/checkouts/readthedocs.org/user_builds/pymob/checkouts/latest/docs/source', root='.', name='unnamed_case_study', version=None, pymob_version=None, scenario='unnamed_scenario', package='case_studies', modules=['sim', 'mod', 'prob', 'data', 'plot'], simulation='Simulation', output=None, data=None, scenario_path_override=None, observations=None, logging='DEBUG', logfile=None, output_path='case_studies/unnamed_case_study/results/unnamed_scenario', data_path='case_studies/unnamed_case_study/data', default_settings_path='case_studies/unnamed_case_study/scenarios/unnamed_scenario/settings.cfg'), alias='case-study', alias_priority=2), 'data_structure': FieldInfo(annotation=Datastructure, required=False, default=Datastructure(indices=[]), alias='data-structure', alias_priority=2), 'error_model': FieldInfo(annotation=Errormodel, required=False, default=Errormodel(), alias='error-model', alias_priority=2), 'inference': FieldInfo(annotation=Inference, required=False, default=Inference(eps=1e-08, objective_function='total_average', n_objectives=1, objective_names=[], backend=None, extra_vars=[], plot=None, n_predictions=100)), 'inference_numpyro': FieldInfo(annotation=Numpyro, required=False, default=Numpyro(user_defined_probability_model=None, user_defined_error_model=None, user_defined_preprocessing=None, gaussian_base_distribution=False, kernel='nuts', init_strategy='init_to_uniform', chains=1, draws=2000, warmup=1000, thinning=1, nuts_draws=2000, nuts_step_size=0.8, nuts_max_tree_depth=10, nuts_target_accept_prob=0.8, nuts_dense_mass=True, nuts_adapt_step_size=True, nuts_adapt_mass_matrix=True, sa_adapt_state_size=None, svi_iterations=10000, svi_learning_rate=0.0001), alias='inference.numpyro', alias_priority=2), 'inference_pyabc': FieldInfo(annotation=Pyabc, required=False, default=Pyabc(sampler='SingleCoreSampler', population_size=100, minimum_epsilon=0.0, min_eps_diff=0.0, max_nr_populations=1000, database_path='/tmp/pyabc.db'), alias='inference.pyabc', alias_priority=2), 'inference_pyabc_redis': FieldInfo(annotation=Redis, required=False, default=Redis(password='nopassword', port=1111, n_predictions=50, history_id=-1, model_id=0), alias='inference.pyabc.redis', alias_priority=2), 'inference_pymoo': FieldInfo(annotation=Pymoo, required=False, default=Pymoo(algortihm='UNSGA3', population_size=100, max_nr_populations=1000, ftol=1e-05, xtol=1e-07, cvtol=1e-07, verbose=True), alias='inference.pymoo', alias_priority=2), 'jaxsolver': FieldInfo(annotation=Jaxsolver, required=False, default=Jaxsolver(diffrax_solver='Dopri5', rtol=1e-06, atol=1e-07, pcoeff=0.0, icoeff=1.0, dcoeff=0.0, max_steps=100000, throw_exception=True), alias='jax-solver', alias_priority=2), 'model_parameters': FieldInfo(annotation=Modelparameters, required=False, default=Modelparameters(), alias='model-parameters', alias_priority=2), 'multiprocessing': FieldInfo(annotation=Multiprocessing, required=False, default=Multiprocessing(cores=1)), 'report': FieldInfo(annotation=Report, required=False, default=Report(debug_report=False, pandoc_output_format='html', model=True, parameters=True, parameters_format='pandas', diagnostics=True, diagnostics_with_batch_dim_vars=False, diagnostics_exclude_vars=[], goodness_of_fit=True, goodness_of_fit_use_predictions=True, goodness_of_fit_nrmse_mode='range', table_parameter_estimates=True, table_parameter_estimates_format='csv', table_parameter_estimates_significant_figures=3, table_parameter_estimates_error_metric='sd', table_parameter_estimates_parameters_as_rows=True, table_parameter_estimates_with_batch_dim_vars=False, table_parameter_estimates_exclude_vars=[], table_parameter_estimates_override_names={}, plot_trace=True, plot_parameter_pairs=True), alias='report', alias_priority=2), 'simulation': FieldInfo(annotation=Simulation, required=False, default=Simulation(model=None, model_class=None, solver=None, y0=[], x_in=[], input_files=[], n_ode_states=-1, batch_dimension='batch_id', x_dimension='time', modeltype='deterministic', solver_post_processing=None, seed=1)), 'solverbase': FieldInfo(annotation=Solverbase, required=False, default=Solverbase(x_dim='time', exclude_kwargs_model=('t', 'time', 'x_in', 'y', 'x', 'Y', 'X'), exclude_kwargs_postprocessing=('t', 'time', 'interpolation', 'results')))}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_parameters: Modelparameters#

model_post_init(__context: Any) → None#

This function is meant to behave like a BaseModel method to initialise private attributes.

It takes context as an argument since that’s what pydantic-core passes when calling it.

Parameters:

self – The BaseModel instance.
__context – The context.

multiprocessing: Multiprocessing#

print() → None#: Print a summary of the configuration.

report: Report#

save(fp: str | None = None, force: bool = False) → None#

Save the configuration to a settings.cfg file.

Uses serializers defined at the top, which parse the options to str so they can be processed by configfile.

In case the model configuration should be stored to a json file use something like json.dumps(self.model_dump()), because the build in function, is somewhat disabled by the listparsers which are needed for configfile lists.

Parameters:

fp (Optional[str]) – File path to write the settings file to. If None the default location derived from the case study is used.
force (bool, optional) – Overwrite without prompting. Default is False.

set_option(section: str, option: str, value: str) → None#

Set a configuration option.

Parameters:

section (str) – Name of the configuration section (e.g. simulation).
option (str) – Option name within the section.
value (str) – New value as a string; will be parsed according to the section’s type definitions.

simulation: Simulation#

solverbase: Solverbase#

pymob.sim.config.casestudy_registry module#

pymob.sim.config.casestudy_registry.get_case_study_model(name: str)#: Return the registered model class for name, or None if not registered.

pymob.sim.config.casestudy_registry.register_case_study_config(name: str, model_cls: Type[BaseModel]) → None#

Parameters:

name (str) – Identifier matching case_study.name (e.g. lotka_volterra_case_study).
model_cls (Type[BaseModel]) – The Pydantic model that defines the expected options for the section.

pymob.sim.config.parameters module#

class pymob.sim.config.parameters.Expression(expression: str | Expression)#

Bases: object

Random variables are context dependent. They may be dependent on other Variables, or datasets. In the config they represent an abstract structure, so they remain unevaluated expressions that must follow python syntax. Once, the context is available, the expressions can be evaluated by Expression.evaluate(context={…}).

evaluate(context: Dict = {}) → Any#

class pymob.sim.config.parameters.Param(*, name: str | None = None, value: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] = 0.0, dims: ~typing.Tuple[str, ...] = (), unit: str | ~typing.List[str] | None = None, prior: ~pymob.sim.config.parameters.RandomVariable | None = None, min: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None = None, max: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None = None, step: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None = None, hyper: bool = False, free: bool = True)#

Bases: BaseModel

This class serves as a Basic model for declaring parameters Including a distribution with optional depdendencies

Parameters:

value (float|NumericArray) – The parameter value. If it is not a 0-d array, float or 1-d array of length one, it should be accompanied by a dimension for each axis in the array. The array coordinates must be specified in the observation coordinates or in index coordinates
dims (Tuple[str, ...]) – Dimensions of the parameter. If the batch dimension is not specified here, it will be automatically added in the Evaluator (dispatch_constructor). If dims are specified here, they should be present in: 1) dimension in sim.config.data_structure and sim.coordinates 2) dimension in sim.config.parameters (so here) and sim.coordinates 3) dimension in sim.indices
prior (Optional[RandomVariable]) – This is a string or pymob.sim.parameters.RandomVariable that specifies the prior. Strings are automatically parsed to RandomVariables if the syntax is correct. The prior should follow the specification of scipy.stats
unit (Optional[str|List[str]]) – The unit of the parameter. The parameter can be either given explicit units or placeholders can be used, e.g. ‘{X}’ or ‘{T}’, where T is the time dimension and and X is the input dimension. These placeholders can then be replaced, if units are specified in the sim.observations.attrs section of the dataset. Here it is important, that the name of the attr matches the

dims: Tuple[str, ...]#

free: bool#

hyper: bool#

max: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None#

min: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None#

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True, 'extra': 'forbid', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'dims': FieldInfo(annotation=Tuple[str, ...], required=False, default=()), 'free': FieldInfo(annotation=bool, required=False, default=True), 'hyper': FieldInfo(annotation=bool, required=False, default=False), 'max': FieldInfo(annotation=Union[float, NDArray, NoneType], required=False), 'min': FieldInfo(annotation=Union[float, NDArray, NoneType], required=False), 'name': FieldInfo(annotation=Union[str, NoneType], required=False), 'prior': FieldInfo(annotation=Union[Annotated[RandomVariable, BeforeValidator], NoneType], required=False), 'step': FieldInfo(annotation=Union[float, NDArray, NoneType], required=False), 'unit': FieldInfo(annotation=Union[str, List[str], NoneType], required=False), 'value': FieldInfo(annotation=Union[float, NDArray], required=False, default=0.0)}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

name: str | None#

post_update()#

prior: RandomVariable | None#

step: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None#

to_xarray(coordinates)#: Converts the parameter to an xarray based on the dimensional structure and adds additional info (prior, starting values, …) as metadata.

unit: str | List[str] | None#

value: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)]#

class pymob.sim.config.parameters.RandomVariable(*, distribution: str, parameters: Dict[str, Expression], obs: Expression | None = None, obs_inv: Expression | None = None)#

Bases: BaseModel

Basic infrastructure to parse priors into their components so that they can be more easily parsed to other backends. Parsing to distributions needs more context: - other parameters (for hierarchical parameter structures) - data structure and coordinates to identify dimension sizes

classmethod check_distribution(new_value, info, **kwargs)#

classmethod check_parameters(new_value, info, **kwargs)#

distribution: str#

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True, 'extra': 'forbid', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'distribution': FieldInfo(annotation=str, required=True), 'obs': FieldInfo(annotation=Union[Expression, NoneType], required=False), 'obs_inv': FieldInfo(annotation=Union[Expression, NoneType], required=False), 'parameters': FieldInfo(annotation=Dict[str, Expression], required=True)}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_ser() → str#

obs: Expression | None#

obs_inv: Expression | None#

parameters: Dict[str, Expression]#

class pymob.sim.config.parameters.UndefinedNameFinder#

Bases: NodeVisitor

find_undefined_names(expr)#

visit_FunctionDef(node)#

visit_Name(node)#

pymob.sim.config.parameters.string_to_prior_dict(prior_str: str)#

pymob.sim.config.parameters.to_rv(option: str | RandomVariable | Dict) → RandomVariable#

pymob.sim.config.sections module#

class pymob.sim.config.sections.Casestudy(*, init_root: str = '/home/docs/checkouts/readthedocs.org/user_builds/pymob/checkouts/latest/docs/source', root: str = '.', name: str = 'unnamed_case_study', version: str | None = None, pymob_version: str | None = None, scenario: str = 'unnamed_scenario', package: str = 'case_studies', modules: List[str] = ['sim', 'mod', 'prob', 'data', 'plot'], simulation: str = 'Simulation', output: str | None = None, data: str | None = None, scenario_path_override: str | None = None, observations: str | None = None, logging: str = 'DEBUG', logfile: str | None = None)#

Bases: PymobModel

Configuration model for a case study.

init_root#

Working directory at the time of creation (excluded from serialization).

Type:: str

root#

Root directory of the repository.

Type:: str

name#

Name of the case study.

Type:: str

version#

Optional version string.

Type:: Optional[str]

pymob_version#

Version of pymob used.

Type:: Optional[str]

scenario#

Scenario name.

Type:: str

package#

Package containing the case study.

Type:: str

modules#

List of modules to import for the case study.

Type:: List[str]

simulation#

Name of the simulation class defined in sim.py.

Type:: str

output, data, scenario_path_override, observations

Paths data, scenario overrides and observation files.

Type:: Optional[str]

logging, logfile

Logging level and optional logfile.

Type:: str

data: str | None#

property data_path: str#

Path where case-study data files are stored.

Return type:: str

property default_settings_path: str#

Default location of the settings.cfg file for the case study.

Return type:: str

init_root: str#

logfile: str | None#

logging: str#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'data': FieldInfo(annotation=Union[str, NoneType], required=False), 'init_root': FieldInfo(annotation=str, required=False, default='/home/docs/checkouts/readthedocs.org/user_builds/pymob/checkouts/latest/docs/source', exclude=True), 'logfile': FieldInfo(annotation=Union[str, NoneType], required=False), 'logging': FieldInfo(annotation=str, required=False, default='DEBUG'), 'modules': FieldInfo(annotation=List[str], required=False, default=['sim', 'mod', 'prob', 'data', 'plot'], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'name': FieldInfo(annotation=str, required=False, default='unnamed_case_study'), 'observations': FieldInfo(annotation=Union[str, NoneType], required=False), 'output': FieldInfo(annotation=Union[str, NoneType], required=False), 'package': FieldInfo(annotation=str, required=False, default='case_studies'), 'pymob_version': FieldInfo(annotation=Union[str, NoneType], required=False), 'root': FieldInfo(annotation=str, required=False, default='.'), 'scenario': FieldInfo(annotation=str, required=False, default='unnamed_scenario'), 'scenario_path_override': FieldInfo(annotation=Union[str, NoneType], required=False), 'simulation': FieldInfo(annotation=str, required=False, default='Simulation', description='Simulation Class defined in sim.py module in the case study.'), 'version': FieldInfo(annotation=Union[str, NoneType], required=False)}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

modules: List[str]#

name: str#

observations: str | None#

output: str | None#

property output_path: str#

Path where results are stored.

Return type:: str

package: str#

pymob_version: str | None#

root: str#

scenario: str#

property scenario_path: str#

Path to the scenario directory, taking overrides into account.

Return type:: str

scenario_path_override: str | None#

classmethod set_root(new_value, info, **kwargs) → str#

Validate and possibly change the working directory based on root.

Returns:: The absolute path that becomes the new working directory.
Return type:: str

Note

For conditionally updating values (e.g. when data variables change) see pydantic/pydantic#7127

simulation: str#

version: str | None#

class pymob.sim.config.sections.DataVariable(*, dimensions: List[str], unit: str | List[str] | None = None, min: float = nan, max: float = nan, observed: bool = True, dimensions_evaluator: List[str] | None = None)#

Bases: BaseModel

Describe a data variable used in a case study.

Parameters:

dimensions (List[str]) – Dimensions that must be present in the observations and the dimensional order of the data variable.
min (float, optional) – Minimum value used for scaling. np.nan means the minimum is taken from the observations. Default is np.nan.
max (float, optional) – Maximum value used for scaling. np.nan means the maximum is taken from the observations. Default is np.nan.
observed (bool, optional) – Whether the variable was observed. Default is True.
dimensions_evaluator (List[str] or None, optional) – Order of dimensions returned by the evaluator. If None the order of dimensions is used.

Examples

>>> dv = DataVariable(dimensions=["time", "species"])
>>> dv.dimensions
['time', 'species']
>>> dv.min = 0.0
>>> dv.max = 1.0

dimensions: List[str]#

dimensions_evaluator: List[str] | None#

max: float#

min: float#

model_config: ClassVar[ConfigDict] = {'extra': 'forbid', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'dimensions': FieldInfo(annotation=List[str], required=True), 'dimensions_evaluator': FieldInfo(annotation=Union[List[str], NoneType], required=False), 'max': FieldInfo(annotation=float, required=False, default=nan), 'min': FieldInfo(annotation=float, required=False, default=nan), 'observed': FieldInfo(annotation=bool, required=False, default=True), 'unit': FieldInfo(annotation=Union[str, List[str], NoneType], required=False)}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

observed: bool#

post_update() → DataVariable#: Validate that dimensions_evaluator matches dimensions.

classmethod set_data_variable_bounds(v, info, **kwargs) → List[str] | None#

Ensure that dimensions_evaluator has the same length and names as dimensions.

Raises:: AssertionError – If the lengths differ or the sets of names differ.

unit: str | List[str] | None#

class pymob.sim.config.sections.Datastructure(*, indices: List[str] = [], **extra_data: Any)#

Bases: PymobModel

Container for data-variable definitions.

The class stores DataVariable objects in __pydantic_extra__.

property all: Dict[str, DataVariable]#: Dictionary of all data variables.

property data_variables: List[str]#: List of all data-variable names.

property data_variables_max: List[float]#: List of max values for all data variables.

property data_variables_min: List[float]#: List of min values for all data variables.

property dimdict: Dict[str, List[str]]#: Mapping of variable name to its dimensions.

property dimensions: List[str]#

Legacy method returning the union of all dimensions.

Return type:: List[str]

Note

TODO: Remove when dimensions is not accessed any longer

property evaluator_dim_order: List[str]#

Legacy method returning the order of dimensions used by the evaluator.

Return type:: List[str]

Note

TODO: Remove when dimensions is not accessed any longer

indices: List[str]#

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'indices': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')])}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

property observed_data_variables: List[str]#: List of observed data-variable names.

property observed_data_variables_max: List[float]#: List of max values for observed data variables.

property observed_data_variables_min: List[float]#: List of min values for observed data variables.

property observed_dimdict: Dict[str, List[str]]#: Mapping of observed variable name to its dimensions.

remove(key: str) → None#: Removes a data variable from the data structure.

property var_dim_mapper: Dict[str, List[str]]#

Mapping from variable name to evaluator dimension indices.

Return type:: Dict[str, List[int]]

class pymob.sim.config.sections.Errormodel(**extra_data: Any)#

Bases: PymobModel

Container for error-model specifications.

Each entry is an OptionRV describing a random variable.

property all: Dict[str, RandomVariable]#: All error-model random variables.

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

class pymob.sim.config.sections.Inference(*, eps: float = 1e-08, objective_function: str = 'total_average', n_objectives: int = 1, objective_names: List[str] = [], backend: str | None = None, extra_vars: List[str] = [], plot: str | Callable | None = None, n_predictions: int = 100)#

Bases: PymobModel

Configuration for inference settings.

eps#

Numerical epsilon for convergence.

Type:: float

objective_function#

Name of the objective function.

Type:: str

n_objectives#

Number of objectives.

Type:: int

objective_names#

Names of the objectives.

Type:: List[str]

backend#

Inference backend to use.

Type:: Optional[str]

extra_vars#

Additional variables to monitor.

Type:: List[str]

plot#

Plotting function.

Type:: Optional[Callable]

n_predictions#

Number of posterior predictive draws.

Type:: int

backend: str | None#

eps: float#

extra_vars: List[str]#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'backend': FieldInfo(annotation=Union[str, NoneType], required=False), 'eps': FieldInfo(annotation=float, required=False, default=1e-08), 'extra_vars': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'n_objectives': FieldInfo(annotation=int, required=False, default=1, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'n_predictions': FieldInfo(annotation=int, required=False, default=100, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'objective_function': FieldInfo(annotation=str, required=False, default='total_average'), 'objective_names': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'plot': FieldInfo(annotation=Union[str, Callable, NoneType], required=False)}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

n_objectives: int#

n_predictions: int#

objective_function: str#

objective_names: List[str]#

plot: str | Callable | None#

class pymob.sim.config.sections.Jaxsolver(*, diffrax_solver: str = 'Dopri5', rtol: float = 1e-06, atol: float = 1e-07, pcoeff: float = 0.0, icoeff: float = 1.0, dcoeff: float = 0.0, max_steps: int = 100000, throw_exception: bool = True)#

Bases: PymobModel

Configuration for the JAX solver.

diffrax_solver#

Name of the Diffrax solver to use.

Type:: str

rtol, atol

Relative and absolute tolerances.

Type:: float

pcoeff, icoeff, dcoeff, max_steps

Coefficients for the solver.

Type:: float or int

throw_exception#

Whether to raise an exception on solver failure.

Type:: bool

atol: float#

dcoeff: float#

diffrax_solver: str#

icoeff: float#

max_steps: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'atol': FieldInfo(annotation=float, required=False, default=1e-07), 'dcoeff': FieldInfo(annotation=float, required=False, default=0.0), 'diffrax_solver': FieldInfo(annotation=str, required=False, default='Dopri5'), 'icoeff': FieldInfo(annotation=float, required=False, default=1.0), 'max_steps': FieldInfo(annotation=int, required=False, default=100000), 'pcoeff': FieldInfo(annotation=float, required=False, default=0.0), 'rtol': FieldInfo(annotation=float, required=False, default=1e-06), 'throw_exception': FieldInfo(annotation=bool, required=False, default=True)}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

pcoeff: float#

rtol: float#

throw_exception: bool#

class pymob.sim.config.sections.ModelParameterDict#

Bases: TypedDict

Typed dictionary used for model parameters.

parameters#

Mapping of parameter names to values (float, str or int).

Type:: dict

y0#

Initial conditions for the ODE system.

Type:: xr.Dataset

x_in#

Input exposure time series.

Type:: xr.Dataset

parameters: Dict[str, float | str | int]#

x_in: Dataset#

y0: Dataset#

class pymob.sim.config.sections.Modelparameters(**extra_data: Any)#

Bases: PymobModel

Container for model parameters.

Parameters are stored in __pydantic_extra__ as Param objects.

property all: Dict[str, Param]#: All parameters.

property dimensions: List[str]#

Legacy method returning the union of all parameter dimensions.

Return type:: List[str]

Note

TODO: Remove when dimensions is not accessed any longer

property fixed: Dict[str, Param]#: Parameters marked as fixed.

property fixed_value_dict: int32'>)]]#: Dictionary of fixed parameter values.

property free: Dict[str, Param]#: Parameters marked as free.

property free_value_dict: int32'>)]]#: Dictionary of free parameter values.

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

property n_free: int#: Number of free parameters.

remove(key: str) → None#: Removes a Parameter.

reorder(keys: List[str]) → None#

Reorders model parameters. This may be necessary for hierarchical models, because priors take draws from hyperpriors to parameterize their distributions. Hence, they must be available earlier.

Parameters:: keys (List[str]) – Desired order of parameter keys. Unlisted keys are appended.

property value_dict: int32'>)]]#: Dictionary of all parameter values.

class pymob.sim.config.sections.Multiprocessing(*, cores: int = 1)#

Bases: PymobModel

Configuration for multiprocessing.

cores#

Number of cores to use. 1 means no parallelism.

Type:: int

cores: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'cores': FieldInfo(annotation=int, required=False, default=1, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')])}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_post_init(__context: Any) → None#

This function is meant to behave like a BaseModel method to initialise private attributes.

It takes context as an argument since that’s what pydantic-core passes when calling it.

Parameters:

self – The BaseModel instance.
__context – The context.

property n_cores: int#

Compute the effective number of cores.

Return type:: int

class pymob.sim.config.sections.Numpyro(*, user_defined_probability_model: str | None = None, user_defined_error_model: str | None = None, user_defined_preprocessing: str | None = None, gaussian_base_distribution: bool = False, kernel: str = 'nuts', init_strategy: str = 'init_to_uniform', chains: int = 1, draws: int = 2000, warmup: int = 1000, thinning: int = 1, nuts_draws: int = 2000, nuts_step_size: float = 0.8, nuts_max_tree_depth: int = 10, nuts_target_accept_prob: float = 0.8, nuts_dense_mass: bool = True, nuts_adapt_step_size: bool = True, nuts_adapt_mass_matrix: bool = True, sa_adapt_state_size: int | None = None, svi_iterations: int = 10000, svi_learning_rate: float = 0.0001)#

Bases: PymobModel

Configuration for the NumPyro inference backend.

user_defined_probability_model, user_defined_error_model,

user_defined_preprocessing#

Paths to user-defined modules.

Type:: Optional[str]

gaussian_base_distribution#

Whether to use a Gaussian base distribution.

Type:: bool

kernel, init_strategy

MCMC kernel and initialisation strategy.

Type:: str

chains, draws, warmup, thinning

MCMC sampling parameters.

Type:: int

nuts_draws, nuts_step_size, nuts_max_tree_depth,

nuts_target_accept_prob, nuts_dense_mass,

nuts_adapt_step_size, nuts_adapt_mass_matrix

NUTS sampler settings.

Type:: bool or float

sa_adapt_state_size#

Type:: Optional Simulated annealing state size.

svi_iterations, svi_learning_rate

SVI settings.

Type:: int, float

chains: int#

draws: int#

gaussian_base_distribution: bool#

init_strategy: str#

kernel: str#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'chains': FieldInfo(annotation=int, required=False, default=1, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'draws': FieldInfo(annotation=int, required=False, default=2000, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'gaussian_base_distribution': FieldInfo(annotation=bool, required=False, default=False), 'init_strategy': FieldInfo(annotation=str, required=False, default='init_to_uniform'), 'kernel': FieldInfo(annotation=str, required=False, default='nuts'), 'nuts_adapt_mass_matrix': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'nuts_adapt_step_size': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'nuts_dense_mass': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'nuts_draws': FieldInfo(annotation=int, required=False, default=2000, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'nuts_max_tree_depth': FieldInfo(annotation=int, required=False, default=10), 'nuts_step_size': FieldInfo(annotation=float, required=False, default=0.8, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'nuts_target_accept_prob': FieldInfo(annotation=float, required=False, default=0.8, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'sa_adapt_state_size': FieldInfo(annotation=Union[int, NoneType], required=False), 'svi_iterations': FieldInfo(annotation=int, required=False, default=10000, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'svi_learning_rate': FieldInfo(annotation=float, required=False, default=0.0001, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'thinning': FieldInfo(annotation=int, required=False, default=1, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'user_defined_error_model': FieldInfo(annotation=Union[str, NoneType], required=False), 'user_defined_preprocessing': FieldInfo(annotation=Union[str, NoneType], required=False), 'user_defined_probability_model': FieldInfo(annotation=Union[str, NoneType], required=False), 'warmup': FieldInfo(annotation=int, required=False, default=1000, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')])}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

nuts_adapt_mass_matrix: bool#

nuts_adapt_step_size: bool#

nuts_dense_mass: bool#

nuts_draws: int#

nuts_max_tree_depth: int#

nuts_step_size: float#

nuts_target_accept_prob: float#

sa_adapt_state_size: int | None#

svi_iterations: int#

svi_learning_rate: float#

thinning: int#

user_defined_error_model: str | None#

user_defined_preprocessing: str | None#

user_defined_probability_model: str | None#

warmup: int#

class pymob.sim.config.sections.ParameterDict(*args, **kwargs)#

Bases: dict

Dictionary that triggers a callback on mutation.

This is used to keep the configuration in sync with the underlying simulation objects.

Parameters:

callback (Callable, optional) – Function called with the dictionary after each mutation.
**kwargs (dict) – Initial key/value pairs.

update(*args, **kwargs) → None#: Update the dictionary and invoke the callback.

class pymob.sim.config.sections.Pyabc(*, sampler: str = 'SingleCoreSampler', population_size: int = 100, minimum_epsilon: float = 0.0, min_eps_diff: float = 0.0, max_nr_populations: int = 1000, database_path: str = '/tmp/pyabc.db')#

Bases: PymobModel

Configuration for the pyABC inference backend.

sampler#

Sampler class name.

Type:: str

population_size#

Number of particles per population.

Type:: int

minimum_epsilon, min_eps_diff

Epsilon settings.

Type:: float

max_nr_populations#

Maximum number of populations.

Type:: int

database_path#

Path to the SQLite database used by pyABC.

Type:: str

database_path: str#

max_nr_populations: int#

min_eps_diff: float#

minimum_epsilon: float#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'database_path': FieldInfo(annotation=str, required=False, default='/tmp/pyabc.db'), 'max_nr_populations': FieldInfo(annotation=int, required=False, default=1000, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'min_eps_diff': FieldInfo(annotation=float, required=False, default=0.0, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'minimum_epsilon': FieldInfo(annotation=float, required=False, default=0.0, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'population_size': FieldInfo(annotation=int, required=False, default=100, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'sampler': FieldInfo(annotation=str, required=False, default='SingleCoreSampler')}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

population_size: int#

sampler: str#

class pymob.sim.config.sections.PymobModel#

Bases: BaseModel

Base class for all configuration models used in pymob.

This class inherits from pydantic.BaseModel and provides dictionary-style access to its fields.

Parameters:: **kwargs (dict) – Field values passed to the underlying BaseModel.

Examples

>>> class MyModel(PymobModel):
...     a: int = 1
...     b: str = "x"
>>> m = MyModel()
>>> m["a"]
1
>>> m["b"] = "y"
>>> m.b
'y'

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

class pymob.sim.config.sections.Pymoo(*, algortihm: str = 'UNSGA3', population_size: int = 100, max_nr_populations: int = 1000, ftol: float = 1e-05, xtol: float = 1e-07, cvtol: float = 1e-07, verbose: bool = True)#

Bases: PymobModel

Configuration for the pymoo multi-objective optimisation backend.

algortihm#

Optimisation algorithm.

Type:: str

population_size, max_nr_populations

Population settings.

Type:: int

ftol, xtol, cvtol

Tolerances.

Type:: float

verbose#

Verbosity flag.

Type:: bool

algortihm: str#

cvtol: float#

ftol: float#

max_nr_populations: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'algortihm': FieldInfo(annotation=str, required=False, default='UNSGA3'), 'cvtol': FieldInfo(annotation=float, required=False, default=1e-07, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'ftol': FieldInfo(annotation=float, required=False, default=1e-05, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'max_nr_populations': FieldInfo(annotation=int, required=False, default=1000, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'population_size': FieldInfo(annotation=int, required=False, default=100, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'verbose': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'xtol': FieldInfo(annotation=float, required=False, default=1e-07, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')])}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

population_size: int#

verbose: bool#

xtol: float#

class pymob.sim.config.sections.Redis(*, password: str = 'nopassword', port: int = 1111, **extra_data: Any)#

Bases: PymobModel

Configuration for the Redis backend used by pyABC.

password#

Redis password.

Type:: str

port#

Port number.

Type:: int

n_predictions, history_id, model_id

Evaluation settings.

Type:: int

history_id: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'protected_namespaces': (), 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'history_id': FieldInfo(annotation=int, required=False, default=-1, alias='eval.history_id', alias_priority=2, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'model_id': FieldInfo(annotation=int, required=False, default=0, alias='eval.model_id', alias_priority=2, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'n_predictions': FieldInfo(annotation=int, required=False, default=50, alias='eval.n_predictions', alias_priority=2, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'password': FieldInfo(annotation=str, required=False, default='nopassword'), 'port': FieldInfo(annotation=int, required=False, default=1111, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')])}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_id: int#

model_post_init(__context: Any) → None#

This function is meant to behave like a BaseModel method to initialise private attributes.

It takes context as an argument since that’s what pydantic-core passes when calling it.

Parameters:

self – The BaseModel instance.
__context – The context.

n_predictions: int#

password: str#

port: int#

class pymob.sim.config.sections.Report(*, debug_report: bool = False, pandoc_output_format: Literal['html', 'latex-si', 'latex', 'pdf'] = 'html', model: bool = True, parameters: bool = True, parameters_format: Literal['xarray', 'pandas'] = 'pandas', diagnostics: bool = True, diagnostics_with_batch_dim_vars: bool = False, diagnostics_exclude_vars: List[str] = [], goodness_of_fit: bool = True, goodness_of_fit_use_predictions: bool = True, goodness_of_fit_nrmse_mode: Literal['mean', 'range', 'iqr'] = 'range', table_parameter_estimates: bool = True, table_parameter_estimates_format: Literal['latex', 'csv', 'tsv'] = 'csv', table_parameter_estimates_significant_figures: int = 3, table_parameter_estimates_error_metric: Literal['hdi', 'sd'] = 'sd', table_parameter_estimates_parameters_as_rows: bool = True, table_parameter_estimates_with_batch_dim_vars: bool = False, table_parameter_estimates_exclude_vars: List[str] = [], table_parameter_estimates_override_names: Mapping[str, str | float | int | List[float | int]] = {}, plot_trace: bool = True, plot_parameter_pairs: bool = True)#

Bases: PymobModel

Configuration for report generation.

debug_report#

Include debug information.

Type:: bool

pandoc_output_format#

Output format for the generated report.

Type:: Literal[“html”, “latex-si”, “latex”, “pdf”]

model, parameters, diagnostics, goodness_of_fit,

table_parameter_estimates, plot_trace, plot_parameter_pairs

Flags controlling which sections are included.

Type:: bool

debug_report: bool#

diagnostics: bool#

diagnostics_exclude_vars: List[str]#

diagnostics_with_batch_dim_vars: bool#

goodness_of_fit: bool#

goodness_of_fit_nrmse_mode: Literal['mean', 'range', 'iqr']#

goodness_of_fit_use_predictions: bool#

model: bool#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'debug_report': FieldInfo(annotation=bool, required=False, default=False, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'diagnostics': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'diagnostics_exclude_vars': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'diagnostics_with_batch_dim_vars': FieldInfo(annotation=bool, required=False, default=False, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'goodness_of_fit': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'goodness_of_fit_nrmse_mode': FieldInfo(annotation=Literal['mean', 'range', 'iqr'], required=False, default='range'), 'goodness_of_fit_use_predictions': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'model': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'pandoc_output_format': FieldInfo(annotation=Literal['html', 'latex-si', 'latex', 'pdf'], required=False, default='html'), 'parameters': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'parameters_format': FieldInfo(annotation=Literal['xarray', 'pandas'], required=False, default='pandas'), 'plot_parameter_pairs': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'plot_trace': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'table_parameter_estimates': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'table_parameter_estimates_error_metric': FieldInfo(annotation=Literal['hdi', 'sd'], required=False, default='sd'), 'table_parameter_estimates_exclude_vars': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'table_parameter_estimates_format': FieldInfo(annotation=Literal['latex', 'csv', 'tsv'], required=False, default='csv'), 'table_parameter_estimates_override_names': FieldInfo(annotation=Mapping[str, Union[str, float, int, List[Union[float, int]]]], required=False, default={}, metadata=[BeforeValidator(func=functools.partial(<function string_to_dict>, split_str=' ', sep_str='=')), PlainSerializer(func=functools.partial(<function dict_to_string>, replace_whitespace=''), return_type=<class 'str'>, when_used='json')]), 'table_parameter_estimates_parameters_as_rows': FieldInfo(annotation=bool, required=False, default=True, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'table_parameter_estimates_significant_figures': FieldInfo(annotation=int, required=False, default=3), 'table_parameter_estimates_with_batch_dim_vars': FieldInfo(annotation=bool, required=False, default=False, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')])}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

pandoc_output_format: Literal['html', 'latex-si', 'latex', 'pdf']#

parameters: bool#

parameters_format: Literal['xarray', 'pandas']#

plot_parameter_pairs: bool#

plot_trace: bool#

table_parameter_estimates: bool#

table_parameter_estimates_error_metric: Literal['hdi', 'sd']#

table_parameter_estimates_exclude_vars: List[str]#

table_parameter_estimates_format: Literal['latex', 'csv', 'tsv']#

table_parameter_estimates_override_names: Mapping[str, str | float | int | List[float | int]]#

table_parameter_estimates_parameters_as_rows: bool#

table_parameter_estimates_significant_figures: int#

table_parameter_estimates_with_batch_dim_vars: bool#

class pymob.sim.config.sections.Simulation(*, model: str | None = None, model_class: str | None = None, solver: str | None = None, y0: List[str] = [], x_in: List[str] = [], input_files: List[str] = [], n_ode_states: int = -1, batch_dimension: str = 'batch_id', x_dimension: str = 'time', modeltype: Literal['stochastic', 'deterministic'] = 'deterministic', solver_post_processing: str | None = None, seed: int = 1, **extra_data: Any)#

Bases: PymobModel

Configuration model for a simulation.

model, solver

Names of the deterministic model and solver to use.

Type:: Optional[str]

y0, x_in, input_files

Lists of initial condition files, exposure files and other input files.

Type:: List[str]

n_ode_states#

Number of ODE states; -1 means autodetect.

Type:: int

batch_dimension, x_dimension

Names of the batch and time dimensions.

Type:: str

modeltype#

Type of model.

Type:: Literal[“stochastic”, “deterministic”]

seed#

Random seed.

Type:: int

batch_dimension: str#

input_files: List[str]#

model: str | None#

model_class: str | None#

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'protected_namespaces': (), 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'batch_dimension': FieldInfo(annotation=str, required=False, default='batch_id'), 'input_files': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'model': FieldInfo(annotation=Union[str, NoneType], required=False, description='The deterministic model', validate_default=True), 'model_class': FieldInfo(annotation=Union[str, NoneType], required=False, description='A class that holds the mechanistic model. This is the path to the module class, e.g. `lotka_volterra_case_study.mod.Model`.', validate_default=True), 'modeltype': FieldInfo(annotation=Literal['stochastic', 'deterministic'], required=False, default='deterministic'), 'n_ode_states': FieldInfo(annotation=int, required=False, default=-1), 'seed': FieldInfo(annotation=int, required=False, default=1, metadata=[PlainSerializer(func=<function <lambda>>, return_type=<class 'str'>, when_used='json')]), 'solver': FieldInfo(annotation=Union[str, NoneType], required=False, validate_default=True), 'solver_post_processing': FieldInfo(annotation=Union[str, NoneType], required=False, validate_default=True), 'x_dimension': FieldInfo(annotation=str, required=False, default='time'), 'x_in': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'y0': FieldInfo(annotation=List[str], required=False, default=[], metadata=[BeforeValidator(func=<function string_to_list>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')])}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

modeltype: Literal['stochastic', 'deterministic']#

n_ode_states: int#

seed: int#

solver: str | None#

solver_post_processing: str | None#

x_dimension: str#

x_in: List[str]#

y0: List[str]#

class pymob.sim.config.sections.Solverbase(*, x_dim: str = 'time', exclude_kwargs_model: Tuple[str, ...] = ('t', 'time', 'x_in', 'y', 'x', 'Y', 'X'), exclude_kwargs_postprocessing: Tuple[str, ...] = ('t', 'time', 'interpolation', 'results'))#

Bases: PymobModel

Base configuration for solvers.

x_dim#

Name of the time dimension.

Type:: str

exclude_kwargs_model, exclude_kwargs_postprocessing

Keyword arguments to exclude when passing to the model or post-processing.

Type:: Tuple[str, …]

exclude_kwargs_model: Tuple[str, ...]#

exclude_kwargs_postprocessing: Tuple[str, ...]#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'exclude_kwargs_model': FieldInfo(annotation=Tuple[str, ...], required=False, default=('t', 'time', 'x_in', 'y', 'x', 'Y', 'X'), metadata=[BeforeValidator(func=<function string_to_tuple>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'exclude_kwargs_postprocessing': FieldInfo(annotation=Tuple[str, ...], required=False, default=('t', 'time', 'interpolation', 'results'), metadata=[BeforeValidator(func=<function string_to_tuple>), PlainSerializer(func=<function list_to_string>, return_type=<class 'str'>, when_used='json')]), 'x_dim': FieldInfo(annotation=str, required=False, default='time')}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

x_dim: str#

pymob.sim.config.sections.datavar_to_string(prm: DataVariable) → str#

Serialize a DataVariable to a key=value string.

Parameters:: prm (DataVariable) –
Return type:: str

pymob.sim.config.sections.dict_to_string(dct: Dict, replace_whitespace='', join_str=' ', sep_str='=') → str#

Serialize a dictionary to a space-separated key=value string.

Parameters:

dct (dict) – Dictionary to serialize.
replace_whitespace (str, optional) – Replacement for whitespace inside values.

Return type:

str

pymob.sim.config.sections.list_to_string(lst: List) → str#

Serialize a list to a space-separated string.

Parameters:: lst (List) –
Return type:: str

pymob.sim.config.sections.nested_dict_to_string(nested_dict: Dict[str, Dict[str, float | None]], dict_model: Type[PymobModel]) → str#

pymob.sim.config.sections.param_to_string(prm: Param) → str#

Serialize a Param to a key=value string.

Parameters:: prm (Param) –
Return type:: str

pymob.sim.config.sections.string_to_datavar(option: str | DataVariable) → DataVariable#

Convert a string or DataVariable instance to a DataVariable.

Parameters:: option (str | DataVariable) –
Return type:: DataVariable

pymob.sim.config.sections.string_to_dict(option: Dict[str, float | str | int] | str, split_str=' ', sep_str='=') → Dict[str, float | str | int]#

Parse a configuration string into a dictionary.

The string must consist of space-separated key=value pairs. Values are interpreted as float, NumericArray or Python literals when possible.

Parameters:: option (Union[dict, str]) – Either an existing dictionary or a string to be parsed.
Returns:: Mapping of keys to parsed values.
Return type:: dict

Examples

>>> string_to_dict("a=1 b=2")
{'a': 1.0, 'b': 2.0}

pymob.sim.config.sections.string_to_list(option: List | str) → List#

Convert a string or list/tuple to a list of strings.

Parameters:: option (Union[List, str]) – Input that may already be a list/tuple or a space-separated string.
Returns:: List of strings.
Return type:: List[str]

Examples

>>> string_to_list("a b c")
['a', 'b', 'c']
>>> string_to_list(["x", "y"])
['x', 'y']

pymob.sim.config.sections.string_to_param(option: str | Param) → Param#

Convert a string or Param instance to a Param.

Parameters:: option (str | Param) –
Return type:: Param

pymob.sim.config.sections.string_to_tuple(option: List | str) → Tuple#

Convert a string or list/tuple to a tuple of strings.

Parameters:: option (Union[List, str]) –
Return type:: Tuple[str, …]

pymob.sim.config.sections.to_nested_dict(option: str | Dict[str, str | Dict[str, float | None]], dict_model: Type[PymobModel]) → Dict[str, Dict[str, float | None]]#

Module contents#

class pymob.sim.config.Casestudy(*, init_root: str = '/home/docs/checkouts/readthedocs.org/user_builds/pymob/checkouts/latest/docs/source', root: str = '.', name: str = 'unnamed_case_study', version: str | None = None, pymob_version: str | None = None, scenario: str = 'unnamed_scenario', package: str = 'case_studies', modules: List[str] = ['sim', 'mod', 'prob', 'data', 'plot'], simulation: str = 'Simulation', output: str | None = None, data: str | None = None, scenario_path_override: str | None = None, observations: str | None = None, logging: str = 'DEBUG', logfile: str | None = None)#

Bases: PymobModel

Configuration model for a case study.

init_root#

Working directory at the time of creation (excluded from serialization).

Type:: str

root#

Root directory of the repository.

Type:: str

name#

Name of the case study.

Type:: str

version#

Optional version string.

Type:: Optional[str]

pymob_version#

Version of pymob used.

Type:: Optional[str]

scenario#

Scenario name.

Type:: str

package#

Package containing the case study.

Type:: str

modules#

List of modules to import for the case study.

Type:: List[str]

simulation#

Name of the simulation class defined in sim.py.

Type:: str

output, data, scenario_path_override, observations

Paths data, scenario overrides and observation files.

Type:: Optional[str]

logging, logfile

Logging level and optional logfile.

Type:: str

data: str | None#

property data_path: str#

Path where case-study data files are stored.

Return type:: str

property default_settings_path: str#

Default location of the settings.cfg file for the case study.

Return type:: str

init_root: str#

logfile: str | None#

logging: str#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

modules: List[str]#

name: str#

observations: str | None#

output: str | None#

property output_path: str#

Path where results are stored.

Return type:: str

package: str#

pymob_version: str | None#

root: str#

scenario: str#

property scenario_path: str#

Path to the scenario directory, taking overrides into account.

Return type:: str

scenario_path_override: str | None#

classmethod set_root(new_value, info, **kwargs) → str#

Validate and possibly change the working directory based on root.

Returns:: The absolute path that becomes the new working directory.
Return type:: str

Note

For conditionally updating values (e.g. when data variables change) see pydantic/pydantic#7127

simulation: str#

version: str | None#

class pymob.sim.config.Config(config: str | ConfigParser | None = None)#

Bases: BaseModel

Configuration manager for pymob.

This class loads a settings.cfg file, validates it against the pydantic models defined above and provides convenient accessors.

Parameters:: config (str or configparser.ConfigParser, optional) – Path to a configuration file or an already parsed ConfigParser.

case_study: Casestudy#

create_directory(directory: Literal['results', 'scenario'], force: bool = False) → None#

Create a results or scenario directory if it does not exist.

Parameters:

directory (Literal["results", "scenario"]) – Which directory to create.
force (bool, optional) – If True create without prompting.

data_structure: Datastructure#

error_model: Errormodel#

import_casestudy_modules(reset_path: bool = False) → None#

Import all modules of the current case study.

Parameters:: reset_path (bool, optional) – Reset sys.path before importing. Default is False.

import_simulation_from_case_study() → Any#

Retrieve the Simulation class defined in the case-study.

Returns:: The Simulation class object.
Return type:: Any
Raises:: ImportError – If the class cannot be found.

inference: Inference#

inference_numpyro: Numpyro#

inference_pyabc: Pyabc#

inference_pyabc_redis: Redis#

inference_pymoo: Pymoo#

property input_file_paths: List[str]#

List of all input files required for the simulation.

Return type:: List[str]

jaxsolver: Jaxsolver#

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'protected_namespaces': (), 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_parameters: Modelparameters#

model_post_init(__context: Any) → None#

This function is meant to behave like a BaseModel method to initialise private attributes.

It takes context as an argument since that’s what pydantic-core passes when calling it.

Parameters:

self – The BaseModel instance.
__context – The context.

multiprocessing: Multiprocessing#

print() → None#: Print a summary of the configuration.

report: Report#

save(fp: str | None = None, force: bool = False) → None#

Save the configuration to a settings.cfg file.

Uses serializers defined at the top, which parse the options to str so they can be processed by configfile.

Parameters:

fp (Optional[str]) – File path to write the settings file to. If None the default location derived from the case study is used.
force (bool, optional) – Overwrite without prompting. Default is False.

set_option(section: str, option: str, value: str) → None#

Set a configuration option.

Parameters:

section (str) – Name of the configuration section (e.g. simulation).
option (str) – Option name within the section.
value (str) – New value as a string; will be parsed according to the section’s type definitions.

simulation: Simulation#

solverbase: Solverbase#

class pymob.sim.config.DataVariable(*, dimensions: List[str], unit: str | List[str] | None = None, min: float = nan, max: float = nan, observed: bool = True, dimensions_evaluator: List[str] | None = None)#

Bases: BaseModel

Describe a data variable used in a case study.

Parameters:

dimensions (List[str]) – Dimensions that must be present in the observations and the dimensional order of the data variable.
min (float, optional) – Minimum value used for scaling. np.nan means the minimum is taken from the observations. Default is np.nan.
max (float, optional) – Maximum value used for scaling. np.nan means the maximum is taken from the observations. Default is np.nan.
observed (bool, optional) – Whether the variable was observed. Default is True.
dimensions_evaluator (List[str] or None, optional) – Order of dimensions returned by the evaluator. If None the order of dimensions is used.

Examples

>>> dv = DataVariable(dimensions=["time", "species"])
>>> dv.dimensions
['time', 'species']
>>> dv.min = 0.0
>>> dv.max = 1.0

dimensions: List[str]#

dimensions_evaluator: List[str] | None#

max: float#

min: float#

model_config: ClassVar[ConfigDict] = {'extra': 'forbid', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

observed: bool#

post_update() → DataVariable#: Validate that dimensions_evaluator matches dimensions.

classmethod set_data_variable_bounds(v, info, **kwargs) → List[str] | None#

Ensure that dimensions_evaluator has the same length and names as dimensions.

Raises:: AssertionError – If the lengths differ or the sets of names differ.

unit: str | List[str] | None#

class pymob.sim.config.Datastructure(*, indices: List[str] = [], **extra_data: Any)#

Bases: PymobModel

Container for data-variable definitions.

The class stores DataVariable objects in __pydantic_extra__.

property all: Dict[str, DataVariable]#: Dictionary of all data variables.

property data_variables: List[str]#: List of all data-variable names.

property data_variables_max: List[float]#: List of max values for all data variables.

property data_variables_min: List[float]#: List of min values for all data variables.

property dimdict: Dict[str, List[str]]#: Mapping of variable name to its dimensions.

property dimensions: List[str]#

Legacy method returning the union of all dimensions.

Return type:: List[str]

Note

TODO: Remove when dimensions is not accessed any longer

property evaluator_dim_order: List[str]#

Legacy method returning the order of dimensions used by the evaluator.

Return type:: List[str]

Note

TODO: Remove when dimensions is not accessed any longer

indices: List[str]#

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

property observed_data_variables: List[str]#: List of observed data-variable names.

property observed_data_variables_max: List[float]#: List of max values for observed data variables.

property observed_data_variables_min: List[float]#: List of min values for observed data variables.

property observed_dimdict: Dict[str, List[str]]#: Mapping of observed variable name to its dimensions.

remove(key: str) → None#: Removes a data variable from the data structure.

property var_dim_mapper: Dict[str, List[str]]#

Mapping from variable name to evaluator dimension indices.

Return type:: Dict[str, List[int]]

class pymob.sim.config.Errormodel(**extra_data: Any)#

Bases: PymobModel

Container for error-model specifications.

Each entry is an OptionRV describing a random variable.

property all: Dict[str, RandomVariable]#: All error-model random variables.

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

class pymob.sim.config.Expression(expression: str | Expression)#

Bases: object

evaluate(context: Dict = {}) → Any#

class pymob.sim.config.Inference(*, eps: float = 1e-08, objective_function: str = 'total_average', n_objectives: int = 1, objective_names: List[str] = [], backend: str | None = None, extra_vars: List[str] = [], plot: str | Callable | None = None, n_predictions: int = 100)#

Bases: PymobModel

Configuration for inference settings.

eps#

Numerical epsilon for convergence.

Type:: float

objective_function#

Name of the objective function.

Type:: str

n_objectives#

Number of objectives.

Type:: int

objective_names#

Names of the objectives.

Type:: List[str]

backend#

Inference backend to use.

Type:: Optional[str]

extra_vars#

Additional variables to monitor.

Type:: List[str]

plot#

Plotting function.

Type:: Optional[Callable]

n_predictions#

Number of posterior predictive draws.

Type:: int

backend: str | None#

eps: float#

extra_vars: List[str]#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

n_objectives: int#

n_predictions: int#

objective_function: str#

objective_names: List[str]#

plot: str | Callable | None#

class pymob.sim.config.Jaxsolver(*, diffrax_solver: str = 'Dopri5', rtol: float = 1e-06, atol: float = 1e-07, pcoeff: float = 0.0, icoeff: float = 1.0, dcoeff: float = 0.0, max_steps: int = 100000, throw_exception: bool = True)#

Bases: PymobModel

Configuration for the JAX solver.

diffrax_solver#

Name of the Diffrax solver to use.

Type:: str

rtol, atol

Relative and absolute tolerances.

Type:: float

pcoeff, icoeff, dcoeff, max_steps

Coefficients for the solver.

Type:: float or int

throw_exception#

Whether to raise an exception on solver failure.

Type:: bool

atol: float#

dcoeff: float#

diffrax_solver: str#

icoeff: float#

max_steps: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

pcoeff: float#

rtol: float#

throw_exception: bool#

class pymob.sim.config.Modelparameters(**extra_data: Any)#

Bases: PymobModel

Container for model parameters.

Parameters are stored in __pydantic_extra__ as Param objects.

property all: Dict[str, Param]#: All parameters.

property dimensions: List[str]#

Legacy method returning the union of all parameter dimensions.

Return type:: List[str]

Note

TODO: Remove when dimensions is not accessed any longer

property fixed: Dict[str, Param]#: Parameters marked as fixed.

property fixed_value_dict: int32'>)]]#: Dictionary of fixed parameter values.

property free: Dict[str, Param]#: Parameters marked as free.

property free_value_dict: int32'>)]]#: Dictionary of free parameter values.

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

property n_free: int#: Number of free parameters.

remove(key: str) → None#: Removes a Parameter.

reorder(keys: List[str]) → None#

Reorders model parameters. This may be necessary for hierarchical models, because priors take draws from hyperpriors to parameterize their distributions. Hence, they must be available earlier.

Parameters:: keys (List[str]) – Desired order of parameter keys. Unlisted keys are appended.

property value_dict: int32'>)]]#: Dictionary of all parameter values.

class pymob.sim.config.Multiprocessing(*, cores: int = 1)#

Bases: PymobModel

Configuration for multiprocessing.

cores#

Number of cores to use. 1 means no parallelism.

Type:: int

cores: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_post_init(__context: Any) → None#

This function is meant to behave like a BaseModel method to initialise private attributes.

It takes context as an argument since that’s what pydantic-core passes when calling it.

Parameters:

self – The BaseModel instance.
__context – The context.

property n_cores: int#

Compute the effective number of cores.

Return type:: int

class pymob.sim.config.Numpyro(*, user_defined_probability_model: str | None = None, user_defined_error_model: str | None = None, user_defined_preprocessing: str | None = None, gaussian_base_distribution: bool = False, kernel: str = 'nuts', init_strategy: str = 'init_to_uniform', chains: int = 1, draws: int = 2000, warmup: int = 1000, thinning: int = 1, nuts_draws: int = 2000, nuts_step_size: float = 0.8, nuts_max_tree_depth: int = 10, nuts_target_accept_prob: float = 0.8, nuts_dense_mass: bool = True, nuts_adapt_step_size: bool = True, nuts_adapt_mass_matrix: bool = True, sa_adapt_state_size: int | None = None, svi_iterations: int = 10000, svi_learning_rate: float = 0.0001)#

Bases: PymobModel

Configuration for the NumPyro inference backend.

user_defined_probability_model, user_defined_error_model,

user_defined_preprocessing#

Paths to user-defined modules.

Type:: Optional[str]

gaussian_base_distribution#

Whether to use a Gaussian base distribution.

Type:: bool

kernel, init_strategy

MCMC kernel and initialisation strategy.

Type:: str

chains, draws, warmup, thinning

MCMC sampling parameters.

Type:: int

nuts_draws, nuts_step_size, nuts_max_tree_depth,

nuts_target_accept_prob, nuts_dense_mass,

nuts_adapt_step_size, nuts_adapt_mass_matrix

NUTS sampler settings.

Type:: bool or float

sa_adapt_state_size#

Type:: Optional Simulated annealing state size.

svi_iterations, svi_learning_rate

SVI settings.

Type:: int, float

chains: int#

draws: int#

gaussian_base_distribution: bool#

init_strategy: str#

kernel: str#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

nuts_adapt_mass_matrix: bool#

nuts_adapt_step_size: bool#

nuts_dense_mass: bool#

nuts_draws: int#

nuts_max_tree_depth: int#

nuts_step_size: float#

nuts_target_accept_prob: float#

sa_adapt_state_size: int | None#

svi_iterations: int#

svi_learning_rate: float#

thinning: int#

user_defined_error_model: str | None#

user_defined_preprocessing: str | None#

user_defined_probability_model: str | None#

warmup: int#

class pymob.sim.config.Param(*, name: str | None = None, value: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] = 0.0, dims: ~typing.Tuple[str, ...] = (), unit: str | ~typing.List[str] | None = None, prior: ~pymob.sim.config.parameters.RandomVariable | None = None, min: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None = None, max: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None = None, step: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None = None, hyper: bool = False, free: bool = True)#

Bases: BaseModel

This class serves as a Basic model for declaring parameters Including a distribution with optional depdendencies

Parameters:

value (float|NumericArray) – The parameter value. If it is not a 0-d array, float or 1-d array of length one, it should be accompanied by a dimension for each axis in the array. The array coordinates must be specified in the observation coordinates or in index coordinates
dims (Tuple[str, ...]) – Dimensions of the parameter. If the batch dimension is not specified here, it will be automatically added in the Evaluator (dispatch_constructor). If dims are specified here, they should be present in: 1) dimension in sim.config.data_structure and sim.coordinates 2) dimension in sim.config.parameters (so here) and sim.coordinates 3) dimension in sim.indices
prior (Optional[RandomVariable]) – This is a string or pymob.sim.parameters.RandomVariable that specifies the prior. Strings are automatically parsed to RandomVariables if the syntax is correct. The prior should follow the specification of scipy.stats
unit (Optional[str|List[str]]) – The unit of the parameter. The parameter can be either given explicit units or placeholders can be used, e.g. ‘{X}’ or ‘{T}’, where T is the time dimension and and X is the input dimension. These placeholders can then be replaced, if units are specified in the sim.observations.attrs section of the dataset. Here it is important, that the name of the attr matches the

dims: Tuple[str, ...]#

free: bool#

hyper: bool#

max: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None#

min: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None#

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True, 'extra': 'forbid', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

name: str | None#

post_update()#

prior: RandomVariable | None#

step: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)] | None#

to_xarray(coordinates)#: Converts the parameter to an xarray based on the dimensional structure and adds additional info (prior, starting values, …) as metadata.

unit: str | List[str] | None#

value: float | ~numpydantic.vendor.nptyping.base_meta_classes.NDArray[~typing.Any, (<class 'numpy.float64'>, <class 'numpy.int64'>, <class 'numpy.float32'>, <class 'numpy.int32'>)]#

class pymob.sim.config.Pyabc(*, sampler: str = 'SingleCoreSampler', population_size: int = 100, minimum_epsilon: float = 0.0, min_eps_diff: float = 0.0, max_nr_populations: int = 1000, database_path: str = '/tmp/pyabc.db')#

Bases: PymobModel

Configuration for the pyABC inference backend.

sampler#

Sampler class name.

Type:: str

population_size#

Number of particles per population.

Type:: int

minimum_epsilon, min_eps_diff

Epsilon settings.

Type:: float

max_nr_populations#

Maximum number of populations.

Type:: int

database_path#

Path to the SQLite database used by pyABC.

Type:: str

database_path: str#

max_nr_populations: int#

min_eps_diff: float#

minimum_epsilon: float#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

population_size: int#

sampler: str#

class pymob.sim.config.PymobModel#

Bases: BaseModel

Base class for all configuration models used in pymob.

This class inherits from pydantic.BaseModel and provides dictionary-style access to its fields.

Parameters:: **kwargs (dict) – Field values passed to the underlying BaseModel.

Examples

>>> class MyModel(PymobModel):
...     a: int = 1
...     b: str = "x"
>>> m = MyModel()
>>> m["a"]
1
>>> m["b"] = "y"
>>> m.b
'y'

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {}#

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

class pymob.sim.config.Pymoo(*, algortihm: str = 'UNSGA3', population_size: int = 100, max_nr_populations: int = 1000, ftol: float = 1e-05, xtol: float = 1e-07, cvtol: float = 1e-07, verbose: bool = True)#

Bases: PymobModel

Configuration for the pymoo multi-objective optimisation backend.

algortihm#

Optimisation algorithm.

Type:: str

population_size, max_nr_populations

Population settings.

Type:: int

ftol, xtol, cvtol

Tolerances.

Type:: float

verbose#

Verbosity flag.

Type:: bool

algortihm: str#

cvtol: float#

ftol: float#

max_nr_populations: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

population_size: int#

verbose: bool#

xtol: float#

class pymob.sim.config.RandomVariable(*, distribution: str, parameters: Dict[str, Expression], obs: Expression | None = None, obs_inv: Expression | None = None)#

Bases: BaseModel

classmethod check_distribution(new_value, info, **kwargs)#

classmethod check_parameters(new_value, info, **kwargs)#

distribution: str#

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True, 'extra': 'forbid', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_ser() → str#

obs: Expression | None#

obs_inv: Expression | None#

parameters: Dict[str, Expression]#

class pymob.sim.config.Redis(*, password: str = 'nopassword', port: int = 1111, **extra_data: Any)#

Bases: PymobModel

Configuration for the Redis backend used by pyABC.

password#

Redis password.

Type:: str

port#

Port number.

Type:: int

n_predictions, history_id, model_id

Evaluation settings.

Type:: int

history_id: int#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'protected_namespaces': (), 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

model_id: int#

model_post_init(__context: Any) → None#

This function is meant to behave like a BaseModel method to initialise private attributes.

It takes context as an argument since that’s what pydantic-core passes when calling it.

Parameters:

self – The BaseModel instance.
__context – The context.

n_predictions: int#

password: str#

port: int#

class pymob.sim.config.Report(*, debug_report: bool = False, pandoc_output_format: Literal['html', 'latex-si', 'latex', 'pdf'] = 'html', model: bool = True, parameters: bool = True, parameters_format: Literal['xarray', 'pandas'] = 'pandas', diagnostics: bool = True, diagnostics_with_batch_dim_vars: bool = False, diagnostics_exclude_vars: List[str] = [], goodness_of_fit: bool = True, goodness_of_fit_use_predictions: bool = True, goodness_of_fit_nrmse_mode: Literal['mean', 'range', 'iqr'] = 'range', table_parameter_estimates: bool = True, table_parameter_estimates_format: Literal['latex', 'csv', 'tsv'] = 'csv', table_parameter_estimates_significant_figures: int = 3, table_parameter_estimates_error_metric: Literal['hdi', 'sd'] = 'sd', table_parameter_estimates_parameters_as_rows: bool = True, table_parameter_estimates_with_batch_dim_vars: bool = False, table_parameter_estimates_exclude_vars: List[str] = [], table_parameter_estimates_override_names: Mapping[str, str | float | int | List[float | int]] = {}, plot_trace: bool = True, plot_parameter_pairs: bool = True)#

Bases: PymobModel

Configuration for report generation.

debug_report#

Include debug information.

Type:: bool

pandoc_output_format#

Output format for the generated report.

Type:: Literal[“html”, “latex-si”, “latex”, “pdf”]

model, parameters, diagnostics, goodness_of_fit,

table_parameter_estimates, plot_trace, plot_parameter_pairs

Flags controlling which sections are included.

Type:: bool

debug_report: bool#

diagnostics: bool#

diagnostics_exclude_vars: List[str]#

diagnostics_with_batch_dim_vars: bool#

goodness_of_fit: bool#

goodness_of_fit_nrmse_mode: Literal['mean', 'range', 'iqr']#

goodness_of_fit_use_predictions: bool#

model: bool#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

pandoc_output_format: Literal['html', 'latex-si', 'latex', 'pdf']#

parameters: bool#

parameters_format: Literal['xarray', 'pandas']#

plot_parameter_pairs: bool#

plot_trace: bool#

table_parameter_estimates: bool#

table_parameter_estimates_error_metric: Literal['hdi', 'sd']#

table_parameter_estimates_exclude_vars: List[str]#

table_parameter_estimates_format: Literal['latex', 'csv', 'tsv']#

table_parameter_estimates_override_names: Mapping[str, str | float | int | List[float | int]]#

table_parameter_estimates_parameters_as_rows: bool#

table_parameter_estimates_significant_figures: int#

table_parameter_estimates_with_batch_dim_vars: bool#

class pymob.sim.config.Simulation(*, model: str | None = None, model_class: str | None = None, solver: str | None = None, y0: List[str] = [], x_in: List[str] = [], input_files: List[str] = [], n_ode_states: int = -1, batch_dimension: str = 'batch_id', x_dimension: str = 'time', modeltype: Literal['stochastic', 'deterministic'] = 'deterministic', solver_post_processing: str | None = None, seed: int = 1, **extra_data: Any)#

Bases: PymobModel

Configuration model for a simulation.

model, solver

Names of the deterministic model and solver to use.

Type:: Optional[str]

y0, x_in, input_files

Lists of initial condition files, exposure files and other input files.

Type:: List[str]

n_ode_states#

Number of ODE states; -1 means autodetect.

Type:: int

batch_dimension, x_dimension

Names of the batch and time dimensions.

Type:: str

modeltype#

Type of model.

Type:: Literal[“stochastic”, “deterministic”]

seed#

Random seed.

Type:: int

batch_dimension: str#

input_files: List[str]#

model: str | None#

model_class: str | None#

model_config: ClassVar[ConfigDict] = {'extra': 'allow', 'protected_namespaces': (), 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

modeltype: Literal['stochastic', 'deterministic']#

n_ode_states: int#

seed: int#

solver: str | None#

solver_post_processing: str | None#

x_dimension: str#

x_in: List[str]#

y0: List[str]#

class pymob.sim.config.Solverbase(*, x_dim: str = 'time', exclude_kwargs_model: Tuple[str, ...] = ('t', 'time', 'x_in', 'y', 'x', 'Y', 'X'), exclude_kwargs_postprocessing: Tuple[str, ...] = ('t', 'time', 'interpolation', 'results'))#

Bases: PymobModel

Base configuration for solvers.

x_dim#

Name of the time dimension.

Type:: str

exclude_kwargs_model, exclude_kwargs_postprocessing

Keyword arguments to exclude when passing to the model or post-processing.

Type:: Tuple[str, …]

exclude_kwargs_model: Tuple[str, ...]#

exclude_kwargs_postprocessing: Tuple[str, ...]#

model_config: ClassVar[ConfigDict] = {'extra': 'ignore', 'validate_assignment': True}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

x_dim: str#

pymob.sim.config.register_case_study_config(name: str, model_cls: Type[BaseModel]) → None#

Parameters:

name (str) – Identifier matching case_study.name (e.g. lotka_volterra_case_study).
model_cls (Type[BaseModel]) – The Pydantic model that defines the expected options for the section.