gemseo.core.chains.initialization_chain module#

MDA input data initialization.

class MDOInitializationChain(disciplines, name='', available_data_names=())[source]#

Bases: MDOChain

An initialization process for a set of disciplines.

This MDOChain subclass computes the initialization for the computation of a set of disciplines. It is particularly useful in the case of MDAs when not all default_input_data are available, and the execution order is not obvious to compute initial values for all couplings.

From the default inputs of the disciplines, use a greedy algorithm to detect sequentially the disciplines that can be executed, and records the execution order.

The couplings are ignored, and therefore, a true MDA must be used afterward to ensure consistency.

Initialize self. See help(type(self)) for accurate signature.

Parameters:

  • disciplines (Sequence[Discipline]) -- The disciplines.

  • name (str) --

    The name of the discipline. If None, use the class name.

    By default it is set to "".

  • available_data_names (Iterable[str]) --

    The data names that are assumed to be available at runtime, in addition to the default_input_data.

    By default it is set to ().

order_disciplines_from_default_inputs(disciplines, raise_error=True, available_data_names=())[source]#

Order disciplines such that all their input values are defined.

It is particularly useful in the case of MDAs when not all default_input_data are available, and the execution order is not obvious to compute initial values for all couplings.

From the default inputs of the disciplines, use a greedy algorithm to detect sequentially the disciplines that can be executed, and records the execution order.

Raises:

ValueError -- When a discipline cannot be initialized.

Parameters:

  • disciplines (Sequence[Discipline]) -- The disciplines to compute the initialization of.

  • raise_error (bool) --

    Whether to raise an exception when the algorithm fails.

    By default it is set to True.

  • available_data_names (Iterable[str]) --

    The data names that are assumed to be available at runtime, in addition to the default_input_data.

    By default it is set to ().

Returns:

The ordered disciplines when the algorithm succeeds, or, if raise_error=False, the inputs that are missing.

Return type:

list[Discipline] | list[str]