system_model#

Provides the class to create an actual model of a simulation problem.

class SystemModel(name: str, specieses: list[simba_ml.simulation.species.Species], kinetic_parameters: dict[str, simba_ml.simulation.kinetic_parameters.kinetic_parameter.KineticParameter[KineticParameterType]], deriv: Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]], sparsifier: Optional[Sparsifier] = None, noiser: Optional[Noiser] = None, deriv_noiser: Optional[DerivNoiser[KineticParameterType]] = None, timestamps: Optional[Distribution[float]] = None, solver_method: str = 'LSODA')#

Bases: object

A SystemModel is the actual model of a problem.

This class solves a system of Ordinary Differential Equations(ODEs) and generates a signal for a specified number of timestamps. The signal can be modified by appliying noise or excluding a fraction of data points. The result can be exported as a csv-file.

kinetic_parameters#: The kinetic parameters of the model.

sparsifier#: A sparsifier randomly removes values from the signal. Can be None.

noiser#: A noisers applies noise to the signal. Can be None.

timestamps#: The number of timestamps the model should be capable of generating data for.

solver_method#: The solver used to solve ODEs. Corresponds to scipy.integrate.solve_ivp(). Default is LSODA.

Inits PredictionTask with the provided params.

Parameters:

name – The name of the problem.
specieses – A list of specieses.
kinetic_parameters – A dictionary of arguments, where the key is the name of the argument.
deriv – A function that takes a time, a list of species values and a dictionary of arguments and returns a tuple of derivatives.
sparsifier – A sparsifier randomly removes values from the signal. Can be None.
noiser – A noisers applies noise to the signal. Can be None.
deriv_noiser – A special noisers which applies noise to the derivative. Can be None.
timestamps – The number of timestamps the model should be capable of generating data for .
solver_method – The solver used to solve ODEs. Corresponds to scipy.integrate.solve_ivp(). Default is LSODA.

apply_noisifier(signal: DataFrame) → DataFrame#

Applies the objects noisifier to a signal.

Parameters:: signal – (pd.DataFrame) The signal.
Returns:: Signal with applied noise.
Return type:: pd.DataFrame

apply_sparsifier(signal: DataFrame) → DataFrame#

Applies the objects sparsifier to a signal.

Parameters:: signal – (pd.DataFrame) The signal.
Returns:: Signal of reduced features.
Return type:: pd.DataFrame

property deriv: Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]]#

The derivative of the model.

Returns:: A function that takes a time, a list of species values and a dictionary of arguments and returns a tuple of derivatives.

get_clean_signal(start_values: dict[str, Any], sample_id: int, deriv_noised: bool = True) → DataFrame#

Creates a clean signal.

Parameters:

start_values – Start values for the simulation.
sample_id – The id of the sample.
deriv_noised – Whether the derivative function should be noised.

Returns:

A clean signal (possibly the deriv function is noised)

property kinetic_parameters: dict[str, simba_ml.simulation.kinetic_parameters.kinetic_parameter.KineticParameter[KineticParameterType]]#

The model arguments.

Returns:: A dictionary of arguments, where the key is the name of the argument.

property name: str#

The name of the problem.

Returns:: The name of the problem.

sample_species_start_values_from_hypercube(n: int) → dict[str, list[float]]#

Samples the start values for the specieses.

Parameters:: n – the number of samples.
Returns:: The start values for the specieses, sampled from a hypercube.

sample_start_values_from_hypercube(n: int) → dict[str, Any]#

Samples the start values of the problem.

Parameters:: n – number of start_values to create

Start values are: - number of individuals per species - kinetic_parameters - number of timestamps

Returns:: The start values dict.
Return type:: dict[str, Any]

property specieses: dict[str, simba_ml.simulation.species.Species]#

The species of the model.

Returns:: A dictionary of specieses, where the key is the name of the species.