Developing pangadfs plugins

The best way to understand how to create a pangadfs plugin is to look at the structure and code of the existing pangadfs-showdown plugin. The default plugins for pangadfs will not gracefully handle "Showdown" (a/k/a "Captain Mode") contests on DraftKings. This plugin extends pangadfs to optimizes Showdown lineups.

Plugin Namespaces

The first step is to determine which plugin namespaces are required for your plugin. pangadfs-showdown provides classes for the following namespaces:

• pangadfs.pospool: ShowdownPospool

The default pospool assumes a classic contest, where players are split into position groups along with a flex. The simplest way to handle showdown contests is to assume every player is a FLEX and then handle the 1.5x multiplier later.

• pangadfs.populate: ShowdownPopulate

Populate works differently because there is no need for a separate FLEX draw. Instead, we just need to a random, unique draw of players to fill out a lineup.

• pangadfs.fitness: ShowdownFitness

This handles the 1.5x multiplier for the first player (at index 0) when summing lineup points.

• pangadfs.validate: ShowdownSalaryValidate

This handles the 1.5x multiplier for the first player (at index 0) when summing lineup salary.

Loading Plugins

It is advisable to provide an app that shows the proper configuration for your plugin. For pangadfs-showdown, the configuration is as follows:

    # setup context dict for configuration
    ctx = {
        'ga_settings': {
            'crossover_method': 'uniform',
            'csvpth': Path(__file__).parent / 'pool.csv',
            'elite_divisor': 10,
            'elite_method': 'fittest',
            'mutation_rate': .10,
            'n_generations': 20,
            'points_column': 'proj',
            'population_size': 5000,
            'position_column': 'pos',
            'salary_column': 'salary',
            'select_method': 'diverse',
            'stop_criteria': 10,
            'verbose': True
        },

        'site_settings': {
            'lineup_size': 6,
            'posfilter': 2.0,
            'salary_cap': 50000
        }
    }

    # setup plugins
    plugin_names = {
      'pool': 'pool_default',
      'pospool': 'pospool_showdown',
      'populate': 'populate_showdown',
      'optimize': 'optimize_default',
      'crossover': 'crossover_default',
      'mutate': 'mutate_default',
      'fitness': 'fitness_showdown',
      'select': 'select_default'
    }

    dmgrs = {ns: DriverManager(namespace=f'pangadfs.{ns}', name=pn, invoke_on_load=True)
             for ns, pn in plugin_names.items()}

    # when have multiple validators, need to use NamedExtensionManager so they all run
    validate_names = ['salary_validate_showdown', 'validate_duplicates']
    emgrs = {'validate': NamedExtensionManager('pangadfs.validate', names=validate_names, invoke_on_load=True)}

    ga = GeneticAlgorithm(ctx=ctx, driver_managers=dmgrs, extension_managers=emgrs)