Controllers

stochastix.controllers.Timer

Timer(
    controlled_species: str | list[str] | tuple[str, ...],
    time_triggers: ndarray,
    species_at_triggers: ndarray,
)

A controller that activates at specific time points.

This controller allows for direct manipulation of species counts at predefined time points during a simulation. When a trigger time is reached, the controller updates the counts of the specified species to predefined values.

Attributes:

• controlled_species –

  The species whose counts will be manipulated.

• time_triggers –

  The time points at which the controller activates.

• species_at_triggers –

  The new counts for the controlled species at each trigger time.

Methods:

• init –

  Initialize the controller's state for the simulation.

• step –

  Execute a step of the Timer controller.

Parameters:

• controlled_species (str | list[str] | tuple[str, ...]) –

  The name of the species or a list/tuple of species names to be controlled.

• time_triggers (ndarray) –

  A JAX array of time points at which the controller should activate.

• species_at_triggers (ndarray) –

  A JAX array where each row corresponds to a time trigger and contains the new counts for the controlled species. The order of species counts in each row must match the order of species names in controlled_species.

Raises:

• ValueError –

  If controlled_species is not a string, list, or tuple.

• ValueError –

  If the number of time triggers does not match the number of species count updates.

• ValueError –

  If the number of controlled species does not match the number of species counts in each update.

stochastix.controllers.AbstractController

AbstractController()

Base class for controllers.

A controller is a module that can be used to modify the simulation behavior during execution.

Methods:

• init –

  Initializes the controller's state.

• step –

  Performs a single control step.

init

init(
    network: ReactionNetwork,
    t: floating,
    x: ndarray,
    a: ndarray,
    *,
    key: ndarray,
) -> typing.Any

Initialize the controller's state.

This method is called once at the beginning of the simulation to set up the initial state of the controller.

Parameters:

• network (ReactionNetwork) –

  The reaction network being simulated.

• t (floating) –

  The initial time of the simulation.

• x (ndarray) –

  The initial state vector (species counts).

• a (ndarray) –

  The initial propensity vector.

• key (ndarray) –

  A JAX random key for any stochastic initialization.

Returns:

• Any –

  The initial state of the controller. Can be any PyTree.

step

step(
    t: floating,
    step_result: SimulationStep,
    controller_state: Any,
    key: ndarray,
) -> tuple[SimulationStep, typing.Any]

Perform a single control step.

This method is called at each step of the simulation, allowing the controller to inspect the simulation's progress and modify the state if necessary.

Parameters:

• t (floating) –

  The current time of the simulation.

• step_result (SimulationStep) –

  The result from the solver's step.

• controller_state (Any) –

  The current state of the controller.

• key (ndarray) –

  A JAX random key for any stochastic operations.

Returns:

• SimulationStep –

  A tuple containing the (potentially modified) simulation step result

• Any –

  and the new state of the controller.