experiment.py

Overview

The LoadExperiment class is responsible for initializing and managing simulation settings for different experimental runs based on user-defined arguments and a manifest configuration. It facilitates the setup for various models including EMOD and OpenMalaria. The constructor initializes the LoadExperiment object with basic settings, user configurations, and manifest-related attributes. It also configures the run mode and parameters to vary based on the provided arguments.

Import Statements

import manifest
from utility.helper import *
import warnings

Methods

  1. configure_run_mode()
    Configures settings based on the specified run mode (calibrun, test, productionrun, or simulation).

  2. configure_calibration() Configures calibration settings if the emod_calib_params attribute is enabled.

  3. update(**kwargs)
    Updates the attributes of the LoadExperiment instance with new values provided as keyword arguments.

Attributes

  • SUITEname: The name of the suite for the experiment.
  • exp_name: Name of the experiment.
  • emod_step: Specifies the EMOD step.
  • models: List of models to run.
  • intervention_list: List of interventions to be applied.
  • run_mode: Specifies the run mode of the experiment.
  • emod_calib_params: Calibration parameters for EMOD.
  • malariasimulation_parameter_variation: Indicates whether parameter variation for malariasimulation is enabled.
  • output_type: Type of output expected from the simulation.
  • user: User-related settings.
  • custom_subdir: Subdirectory for user-specific data.
  • hpc: High-Performance Computing settings from the manifest.
  • job_directory_manifest: Directory for job outputs as specified in the manifest.
  • output_directory_manifest: Directory for experiment outputs as specified in the manifest.
  • interp_path: Path for interpolation data.
  • interp_csv: Interpolation CSV file name.
  • burnin_directory: Directory for burn-in outputs (default is None).

Example

import argparse
from manifest import Manifest
from load_experiment import LoadExperiment

# Example of argument parsing
parser = argparse.ArgumentParser(description='Load Experiment Settings')
parser.add_argument('--suite', required=True, help='Suite name for the experiment')
parser.add_argument('--expname', required=True, help='Name of the experiment')
# Add additional arguments as necessary
args = parser.parse_args()

# Load manifest
manifest = Manifest()  # Initialize the manifest with appropriate configurations

# Create LoadExperiment object
experiment = LoadExperiment(args, manifest)

# Accessing attributes
print(experiment.SUITEname)
print(experiment.output_type)