num::categorical_sampling

Defined in header <stochastic/concepts.hpp>

template <class S, class G>
concept categorical_sampling =
    std::uniform_random_bit_generator<G> && requires(S &sampler, G &rng) {
      { sampler(rng) } -> std::convertible_to<idx>;
    };

Sampler drawing an index from a fixed distribution.

\[ P(X = k) = \frac{w_k}{\sum_j w_j} \]

The weights are non-negative and need not sum to one. Whether a particular set satisfies that is a property of the values, checked by num::stochastic::debug::verify_weights.