qxmt.experiment.evaluation_factory module

qxmt.experiment.evaluation_factory module#

class qxmt.experiment.evaluation_factory.EvaluationFactory

Bases: object

Factory class that instantiates and executes appropriate evaluation.

This class centralizes the mapping between (model_type, task_type) and concrete Evaluation implementation, decoupling Experiment/Executor classes from evaluation details and making it easy to add new metrics or model types.

Class Attributes:

QKERNEL_MODEL_TYPE_NAME (str): Constant representing the qkernel model type. VQE_MODEL_TYPE_NAME (str): Constant representing the VQE model type.

static evaluate(*, model_type, task_type, params, default_metrics_name=None, custom_metrics=None)

Perform evaluation and return result as plain dictionary.

Parameters:

  • model_type (str) – Type of the model to evaluate. Must be either “qkernel” or “vqe”.

  • task_type (Optional[str]) – Type of the task. For qkernel, must be “classification” or “regression”. For vqe, must be None.

  • params (dict[str, Any]) – Arguments forwarded to the evaluation class. For supervised tasks, should include “actual” and “predicted”. For VQE, should include “cost_history”.

  • default_metrics_name (Optional[list[str]]) – List of default metric names to use.

  • custom_metrics (Optional[list[dict[str, Any]]]) – List of custom metrics to use.

Returns:

Dictionary containing the evaluation results.

Return type:

dict[str, Any]

Raises:

ValueError – If the combination of model_type and task_type is invalid.

Examples

>>> result = EvaluationFactory.evaluate(
...     model_type="qkernel",
...     task_type="classification",
...     params={"actual": [0, 1], "predicted": [0, 1]},
...     default_metrics_name=["accuracy"]
... )