#enums #randomness #ergonomics

random-constructible

Provides a trait for creating random instances of enums with weighted probabilities

8 releases (breaking)

new 0.9.0 Nov 21, 2024
0.8.0 Nov 21, 2024
0.7.0 Nov 21, 2024
0.6.0 Nov 21, 2024
0.1.0 Nov 16, 2024

#28 in #randomness

Download history 234/week @ 2024-11-13

234 downloads per month
Used in 2 crates

MIT/Apache

27KB
335 lines

random-constructible

Crates.io Documentation

random-constructible is a Rust crate that provides traits and macros to facilitate the random generation of primitive types and enums, with support for custom probability distributions. It simplifies the process of creating random instances of your types, especially when dealing with enums that require weighted random selection.

Features

  • Random Generation for Primitive Types: Automatically implements random generation for all primitive integer and floating-point types.
  • Random Enums with Custom Probabilities: Easily define how your enums should be randomly generated, including specifying custom weights for each variant.
  • Uniform Random Generation: Support for uniform random generation across all variants.
  • Extensible Probability Maps: Create and use custom probability maps for more complex random generation scenarios.

Installation

Add the following to your Cargo.toml:

[dependencies]
random-constructible = "0.6.0"

and include it in your crate:

use random_constructible::*;

Getting Started

Random Generation for Primitive Types

The crate automatically provides implementations for all primitive integer and floating-point types. You can generate random values using:

use random_constructible::RandConstruct;

let random_u32: u32 = u32::random();
let random_f64: f64 = f64::random();

Random Enums

To enable random generation for your enums, implement the RandConstructEnum trait. You can do this manually or use the provided macros for convenience.

Manual Implementation

use random_constructible::{RandConstructEnum, RandConstruct};

#[derive(Default, Copy, Clone, Debug, PartialEq, Eq, Hash)]
enum MyEnum {
    #[default]
    VariantA,
    VariantB,
    VariantC,
}

impl RandConstructEnum for MyEnum {
    fn all_variants() -> Vec<Self> {
        vec![Self::VariantA, Self::VariantB, Self::VariantC]
    }

    fn default_weight(&self) -> f64 {
        match self {
            Self::VariantA => 1.0,
            Self::VariantB => 2.0,
            Self::VariantC => 3.0,
        }
    }

    fn create_default_probability_map() -> std::sync::Arc<std::collections::HashMap<Self, f64>> {
        use std::collections::HashMap;
        let mut map = HashMap::new();
        for variant in Self::all_variants() {
            map.insert(variant, variant.default_weight());
        }
        std::sync::Arc::new(map)
    }
}

Using the rand_construct_env! Macro

Alternatively, you can use the rand_construct_env! macro to define the probability map:

use random_constructible::{RandConstructEnumWithEnv, rand_construct_env};

struct DefaultProvider;

rand_construct_env!(DefaultProvider => MyEnum {
    VariantA => 1.0,
    VariantB => 2.0,
    VariantC => 3.0,
});

Sampling Random Variants

Once you've implemented RandConstructEnum for your enum, you can generate random variants:

use random_constructible::RandConstruct;

let random_variant = MyEnum::random(); // Uses default weights
let uniform_variant = MyEnum::uniform(); // Uniform probability

To sample using a custom provider:

use random_constructible::RandConstructEnumWithEnv;

let random_variant = MyEnum::sample_from_provider::<DefaultProvider, _>(&mut rand::thread_rng());

Examples

Complete Example

use random_constructible::{RandConstructEnum, RandConstructEnumWithEnv, rand_construct_env, RandConstruct};
use rand::Rng;

// Define your enum
#[derive(Default, Copy, Clone, Debug, PartialEq, Eq, Hash)]
enum Color {
    #[default]
    Red,
    Green,
    Blue,
}

// Implement RandConstructEnum
impl RandConstructEnum for Color {
    fn all_variants() -> Vec<Self> {
        vec![Self::Red, Self::Green, Self::Blue]
    }

    fn default_weight(&self) -> f64 {
        match self {
            Self::Red => 1.0,
            Self::Green => 1.0,
            Self::Blue => 1.0,
        }
    }

    fn create_default_probability_map() -> std::sync::Arc<std::collections::HashMap<Self, f64>> {
        use std::collections::HashMap;
        let mut map = HashMap::new();
        for variant in Self::all_variants() {
            map.insert(variant, variant.default_weight());
        }
        std::sync::Arc::new(map)
    }
}

// Define a custom probability provider
struct CustomColorProvider;

rand_construct_env!(CustomColorProvider => Color {
    Red => 0.5,
    Green => 0.3,
    Blue => 0.2,
});

fn main() {
    // Random variant using default weights
    let random_color = Color::random();
    println!("Random Color: {:?}", random_color);

    // Random variant using custom probabilities
    let mut rng = rand::thread_rng();
    let random_color = Color::sample_from_provider::<CustomColorProvider, _>(&mut rng);
    println!("Custom Random Color: {:?}", random_color);
}

Testing Randomness

The crate also provides utilities for testing the distribution of your random generation:

use random_constructible::{RandConstructEnum, RandConstruct};
use std::collections::HashMap;

fn main() {
    let mut counts = HashMap::new();
    for _ in 0..10000 {
        let variant = Color::random();
        *counts.entry(variant).or_insert(0) += 1;
    }

    for (variant, count) in counts {
        println!("{:?}: {}", variant, count);
    }
}

License

This project is licensed under the MIT License - see the LICENSE file for details.


Note: This README is generated based on the crate's code and is meant to help you get started with random-constructible. For more detailed information, please refer to the documentation.

Dependencies

~360–510KB