pydruglogics.execution.Executor

Provides training and prediction functions for Boolean models.

pydruglogics.execution.Executor.train(*args, **kwargs)

Train a Boolean Model using genetic algorithm and evolution strategy. Finds the models with the best fitness score.

Parameters

  • boolean_model (BooleanModel): The model to be trained.

  • model_outputs (ModelOutputs): The outputs to be optimized.

  • ga_args (dict): Genetic algorithm parameters.

  • ev_args (dict): Evolution strategy parameters.

  • training_data (TrainingData, optional): Training data for the model.

  • save_best_models (bool): Whether to save the best models.

  • save_path (str): Path to save models.

Returns

  • List[BooleanModel]: List of models with the best fitness.

pydruglogics.execution.Executor.predict(*args, **kwargs)

Predict model outcomes and plot the results.

Parameters

  • best_boolean_models (list of BooleanModel, optional): Models to use for prediction.

  • model_outputs (ModelOutputs): The outputs to be predicted.

  • perturbations (Perturbation): Perturbation scenarios.

  • observed_synergy_scores (list of str): Observed synergy scores.

  • synergy_method (str): Method for calculating synergy (‘bliss’, etc.).

  • run_parallel (bool): Whether to run in parallel.

  • plot_roc_pr_curves (bool): Whether to plot ROC and PR curves.

  • save_predictions (bool): Whether to save predictions.

  • save_path (str): Path to save predictions.

  • model_directory (str): Directory to load models from if not provided.

  • attractor_tool (str): Tool for attractor analysis.

  • attractor_type (str): Type of attractor analysis.

  • cores (int): Number of CPU cores to use.