Base Parameter

baccarat.params.Param

Overview

The Param class is an abstract base class that defines the interface for all parameter types. It implements Python's descriptor protocol to automatically generate new random values when accessed within a Simulator instance.

Class Definition

class Param(ABC):
    def __get__(self, instance, owner):
        # Generate a new value each time the parameter is accessed
        return self.generate(rng=instance.rng, num_generated=instance.num_samples)

    @abstractmethod
    def generate(self, rng: np.random.Generator, num_generated: int) -> np.ndarray:
        """
        Generate random values using the given random number generator.

        Args:
            rng: The NumPy random number generator to use
            num_generated: Number of values to generate

        Returns:
            np.ndarray: Array of generated values
        """

Descriptor Protocol

The Param class uses Python's descriptor protocol to provide dynamic behavior when parameters are accessed as instance attributes in a Simulator subclass:

1. When a parameter is accessed (e.g., self.x in a Simulator subclass), Python calls the __get__ method
2. The __get__ method calls the generate method with the simulator's random number generator and sample count
3. The generate method returns an array of random values according to the parameter's specific distribution

Custom Parameter Types

To create a custom parameter type, subclass Param and implement the generate method:

import numpy as np
from baccarat.params import Param

class ExponentialParam(Param):
    def __init__(self, scale=1.0):
        self.scale = scale

    def generate(self, rng: np.random.Generator, num_generated: int) -> np.ndarray:
        return rng.exponential(scale=self.scale, size=num_generated)