peroxide/statistics/

rand.rs

//! Random number generator
//!
//! ## Uniform random number generator
//!
//! * Peroxide uses external [`rand` crate](https://crates.io/crates/rand) to generate random number
//!
//!     ```rust
//!     use rand::prelude::*;
//!
//!     fn main() {
//!         let mut rng = rand::rng();
//!
//!         let a = rng.random_range(0f64..=1f64); // Generate random f64 number ranges from 0 to 1
//!     }
//!     ```
//!
//! * To want more detailed explanation, see [`rand` crate](https://crates.io/crates/rand)
//!
//! ## Piece-wise Rejection Sampling
//!
//!

use rand::prelude::*;
use rand_distr::uniform::SampleUniform;

use crate::statistics::dist::{WeightedUniform, RNG};
#[allow(unused_imports)]
use crate::structure::matrix::*;

/// Small random number generator from seed
///
/// # Examples
/// ```
/// use peroxide::fuga::*;
///
/// fn main() {
///     let mut rng = smallrng_from_seed(42);
///
///     let n = Normal(0f64, 1f64);
///     n.sample_with_rng(&mut rng, 10).print();
/// }
pub fn smallrng_from_seed(seed: u64) -> SmallRng {
    SmallRng::seed_from_u64(seed)
}

/// Std random number generator from seed
///
/// # Examples
/// ```
/// use peroxide::fuga::*;
///
/// fn main() {
///     let mut rng = stdrng_from_seed(42);
///
///     let n = Normal(0f64, 1f64);
///     n.sample_with_rng(&mut rng, 10).print();
/// }
pub fn stdrng_from_seed(seed: u64) -> StdRng {
    StdRng::seed_from_u64(seed)
}

/// Simple uniform random number generator with ThreadRng
///
/// # Examples
/// ```
/// use peroxide::fuga::*;
///
/// let mut rng = rand::rng();
/// println!("{}", rand_num(&mut rng, 1, 7));       // Roll a die
/// println!("{}", rand_num(&mut rng, 0f64, 1f64)); // Uniform [0,1)
/// ```
pub fn rand_num<T>(rng: &mut ThreadRng, start: T, end: T) -> T
where
    T: PartialOrd + SampleUniform + Copy,
{
    rng.random_range(start..=end)
}

// =============================================================================
// Back end utils
// =============================================================================

/// Gaussian random number generator using Marsaglia polar form
pub fn marsaglia_polar(rng: &mut ThreadRng, m: f64, s: f64) -> f64 {
    let mut x1 = 0f64;
    let mut x2 = 0f64;
    let mut _y2 = 0f64;
    let mut w = 0f64;

    while w == 0. || w >= 1. {
        x1 = 2.0 * rng.random_range(0f64..=1f64) - 1.0;
        x2 = 2.0 * rng.random_range(0f64..=1f64) - 1.0;
        w = x1 * x1 + x2 * x2;
    }

    w = (-2.0 * w.ln() / w).sqrt();
    let y1 = x1 * w;
    _y2 = x2 * w;

    return m + y1 * s;
}

// =============================================================================
// Ziggurat Table
// =============================================================================

/// Position of Right-most step
const PARAM_R: f64 = 3.44428647676;

/// Tabulated values for the height of the Ziggurat levels
const YTAB: [f64; 128] = [
    1.,
    0.963598623011,
    0.936280813353,
    0.913041104253,
    0.892278506696,
    0.873239356919,
    0.855496407634,
    0.838778928349,
    0.822902083699,
    0.807732738234,
    0.793171045519,
    0.779139726505,
    0.765577436082,
    0.752434456248,
    0.739669787677,
    0.727249120285,
    0.715143377413,
    0.703327646455,
    0.691780377035,
    0.68048276891,
    0.669418297233,
    0.65857233912,
    0.647931876189,
    0.637485254896,
    0.62722199145,
    0.617132611532,
    0.607208517467,
    0.597441877296,
    0.587825531465,
    0.578352913803,
    0.569017984198,
    0.559815170911,
    0.550739320877,
    0.541785656682,
    0.532949739145,
    0.524227434628,
    0.515614886373,
    0.507108489253,
    0.498704867478,
    0.490400854812,
    0.482193476986,
    0.47407993601,
    0.466057596125,
    0.458123971214,
    0.450276713467,
    0.442513603171,
    0.434832539473,
    0.427231532022,
    0.419708693379,
    0.41226223212,
    0.404890446548,
    0.397591718955,
    0.390364510382,
    0.383207355816,
    0.376118859788,
    0.369097692334,
    0.362142585282,
    0.355252328834,
    0.348425768415,
    0.341661801776,
    0.334959376311,
    0.328317486588,
    0.321735172063,
    0.31521151497,
    0.308745638367,
    0.302336704338,
    0.29598391232,
    0.289686497571,
    0.283443729739,
    0.27725491156,
    0.271119377649,
    0.265036493387,
    0.259005653912,
    0.253026283183,
    0.247097833139,
    0.241219782932,
    0.235391638239,
    0.229612930649,
    0.223883217122,
    0.218202079518,
    0.212569124201,
    0.206983981709,
    0.201446306496,
    0.195955776745,
    0.190512094256,
    0.185114984406,
    0.179764196185,
    0.174459502324,
    0.169200699492,
    0.1639876086,
    0.158820075195,
    0.153697969964,
    0.148621189348,
    0.143589656295,
    0.138603321143,
    0.133662162669,
    0.128766189309,
    0.123915440582,
    0.119109988745,
    0.114349940703,
    0.10963544023,
    0.104966670533,
    0.100343857232,
    0.0957672718266,
    0.0912372357329,
    0.0867541250127,
    0.082318375932,
    0.0779304915295,
    0.0735910494266,
    0.0693007111742,
    0.065060233529,
    0.0608704821745,
    0.056732448584,
    0.05264727098,
    0.0486162607163,
    0.0446409359769,
    0.0407230655415,
    0.0368647267386,
    0.0330683839378,
    0.0293369977411,
    0.0256741818288,
    0.0220844372634,
    0.0185735200577,
    0.0151490552854,
    0.0118216532614,
    0.00860719483079,
    0.00553245272614,
    0.00265435214565,
];

/// Tabulated values for 2^24 times `x[i] / x[i+1]`
/// Used to accept for `U*x[i+1] <= x[i]` without any floating point operations
const KTAB: [u32; 128] = [
    0, 12590644, 14272653, 14988939, 15384584, 15635009, 15807561, 15933577, 16029594, 16105155,
    16166147, 16216399, 16258508, 16294295, 16325078, 16351831, 16375291, 16396026, 16414479,
    16431002, 16445880, 16459343, 16471578, 16482744, 16492970, 16502368, 16511031, 16519039,
    16526459, 16533352, 16539769, 16545755, 16551348, 16556584, 16561493, 16566101, 16570433,
    16574511, 16578353, 16581977, 16585398, 16588629, 16591685, 16594575, 16597311, 16599901,
    16602354, 16604679, 16606881, 16608968, 16610945, 16612818, 16614592, 16616272, 16617861,
    16619363, 16620782, 16622121, 16623383, 16624570, 16625685, 16626730, 16627708, 16628619,
    16629465, 16630248, 16630969, 16631628, 16632228, 16632768, 16633248, 16633671, 16634034,
    16634340, 16634586, 16634774, 16634903, 16634972, 16634980, 16634926, 16634810, 16634628,
    16634381, 16634066, 16633680, 16633222, 16632688, 16632075, 16631380, 16630598, 16629726,
    16628757, 16627686, 16626507, 16625212, 16623794, 16622243, 16620548, 16618698, 16616679,
    16614476, 16612071, 16609444, 16606571, 16603425, 16599973, 16596178, 16591995, 16587369,
    16582237, 16576520, 16570120, 16562917, 16554758, 16545450, 16534739, 16522287, 16507638,
    16490152, 16468907, 16442518, 16408804, 16364095, 16301683, 16207738, 16047994, 15704248,
    15472926,
];

/// Tabulated values of `2^{-24} * x[i]`
const WTAB: [f64; 128] = [
    1.62318314817e-08,
    2.16291505214e-08,
    2.54246305087e-08,
    2.84579525938e-08,
    3.10340022482e-08,
    3.33011726243e-08,
    3.53439060345e-08,
    3.72152672658e-08,
    3.8950989572e-08,
    4.05763964764e-08,
    4.21101548915e-08,
    4.35664624904e-08,
    4.49563968336e-08,
    4.62887864029e-08,
    4.75707945735e-08,
    4.88083237257e-08,
    5.00063025384e-08,
    5.11688950428e-08,
    5.22996558616e-08,
    5.34016475624e-08,
    5.44775307871e-08,
    5.55296344581e-08,
    5.65600111659e-08,
    5.75704813695e-08,
    5.85626690412e-08,
    5.95380306862e-08,
    6.04978791776e-08,
    6.14434034901e-08,
    6.23756851626e-08,
    6.32957121259e-08,
    6.42043903937e-08,
    6.51025540077e-08,
    6.59909735447e-08,
    6.68703634341e-08,
    6.77413882848e-08,
    6.8604668381e-08,
    6.94607844804e-08,
    7.03102820203e-08,
    7.11536748229e-08,
    7.1991448372e-08,
    7.2824062723e-08,
    7.36519550992e-08,
    7.44755422158e-08,
    7.52952223703e-08,
    7.61113773308e-08,
    7.69243740467e-08,
    7.77345662086e-08,
    7.85422956743e-08,
    7.93478937793e-08,
    8.01516825471e-08,
    8.09539758128e-08,
    8.17550802699e-08,
    8.25552964535e-08,
    8.33549196661e-08,
    8.41542408569e-08,
    8.49535474601e-08,
    8.57531242006e-08,
    8.65532538723e-08,
    8.73542180955e-08,
    8.8156298059e-08,
    8.89597752521e-08,
    8.97649321908e-08,
    9.05720531451e-08,
    9.138142487e-08,
    9.21933373471e-08,
    9.30080845407e-08,
    9.38259651738e-08,
    9.46472835298e-08,
    9.54723502847e-08,
    9.63014833769e-08,
    9.71350089201e-08,
    9.79732621669e-08,
    9.88165885297e-08,
    9.96653446693e-08,
    1.00519899658e-07,
    1.0138063623e-07,
    1.02247952126e-07,
    1.03122261554e-07,
    1.04003996769e-07,
    1.04893609795e-07,
    1.05791574313e-07,
    1.06698387725e-07,
    1.07614573423e-07,
    1.08540683296e-07,
    1.09477300508e-07,
    1.1042504257e-07,
    1.11384564771e-07,
    1.12356564007e-07,
    1.13341783071e-07,
    1.14341015475e-07,
    1.15355110887e-07,
    1.16384981291e-07,
    1.17431607977e-07,
    1.18496049514e-07,
    1.19579450872e-07,
    1.20683053909e-07,
    1.21808209468e-07,
    1.2295639141e-07,
    1.24129212952e-07,
    1.25328445797e-07,
    1.26556042658e-07,
    1.27814163916e-07,
    1.29105209375e-07,
    1.30431856341e-07,
    1.31797105598e-07,
    1.3320433736e-07,
    1.34657379914e-07,
    1.36160594606e-07,
    1.37718982103e-07,
    1.39338316679e-07,
    1.41025317971e-07,
    1.42787873535e-07,
    1.44635331499e-07,
    1.4657889173e-07,
    1.48632138436e-07,
    1.50811780719e-07,
    1.53138707402e-07,
    1.55639532047e-07,
    1.58348931426e-07,
    1.61313325908e-07,
    1.64596952856e-07,
    1.68292495203e-07,
    1.72541128694e-07,
    1.77574279496e-07,
    1.83813550477e-07,
    1.92166040885e-07,
    2.05295471952e-07,
    2.22600839893e-07,
];

/// Gaussian random numbers using the Ziggurat Method
///
/// The code is based on a [C implementation][1] by Jochen Voss.
///
/// [1]: https://www.seehuhn.de/pages/ziggurat.html
#[allow(unused_assignments)]
pub fn ziggurat(rng: &mut ThreadRng, sigma: f64) -> f64 {
    let (mut u, mut i, mut sign, mut j) = (0u32, 0usize, 0u32, 0u32);
    let mut x = 0f64;
    let mut y = 0f64;

    loop {
        u = rand_num(rng, u32::MIN, u32::MAX);
        i = (u & 0x0000007F) as usize; // 7 bit to choose the step
        sign = u & 0x00000080; // 1 bit for the sign
        j = u >> 8; // 24 bit for the x-value

        x = j as f64 * WTAB[i];
        if j < KTAB[i] {
            break;
        }

        if i < 127 {
            let y0 = YTAB[i];
            let y1 = YTAB[i + 1];
            y = y1 + (y0 - y1) * rand_num(rng, 0f64, 1f64);
        } else {
            x = PARAM_R - (1.0 - rand_num(rng, 0f64, 1f64).ln()) / PARAM_R;
            y = (-PARAM_R * (x - 0.5 * PARAM_R)).exp() * rand_num(rng, 0f64, 1f64);
        }

        if y < (-0.5 * x * x).exp() {
            break;
        }
    }

    if sign != 0 {
        sigma * x
    } else {
        -sigma * x
    }
}

// =============================================================================
// Rejection Sampling
// =============================================================================
/// Piecewise Rejection Sampling
///
/// # Arguments
/// * `f` - Function to sample (unnormalized function is allowed)
/// * `n` - Number of samples
/// * `(a, b)` - Range of sampling
/// * `m` - Number of pieces
/// * `eps` - Epsilon for max pooling
///
/// # Examples
/// ```
/// use peroxide::fuga::*;
///
/// fn main() -> Result<(), Box<dyn Error>> {
///     let f = |x: f64| {
///         if (0f64..=2f64).contains(&x) {
///             -(x - 1f64).powi(2) + 1f64
///         } else {
///             0f64
///         }
///     };
///
///     let samples = prs(f, 1000, (-1f64, 3f64), 200, 1e-4)?;
///     samples.mean().print(); // near 1
///
///     Ok(())
/// }
pub fn prs<F>(f: F, n: usize, (a, b): (f64, f64), m: usize, eps: f64) -> anyhow::Result<Vec<f64>>
where
    F: Fn(f64) -> f64 + Copy,
{
    let mut rng = rand::rng();

    let mut result = vec![0f64; n];

    let w = WeightedUniform::from_max_pool_1d(f, (a, b), m, eps)?;

    let mut initial_x = w.sample(n);
    let mut left_num = n;

    while left_num > 0 {
        for &x in initial_x.iter() {
            let weight = w.weight_at(x);
            if weight <= 0f64 {
                continue;
            } else {
                let y = rng.random_range(0f64..=weight);

                if y <= f(x) {
                    result[n - left_num] = x;
                    left_num -= 1;
                    if left_num == 0 {
                        return Ok(result);
                    }
                }
            }
        }
        initial_x = w.sample(left_num);
    }
    panic!("Error: failed to generate {} samples", n);
}

/// Piecewise Rejection Sampling with specific Rng
///
/// # Arguments
/// * `f` - Function to sample (unnormalized function is allowed)
/// * `n` - Number of samples
/// * `(a, b)` - Range of sampling
/// * `m` - Number of pieces
/// * `eps` - Epsilon for max pooling
/// * `rng` - Random number generator
///
/// # Examples
/// ```
/// use peroxide::fuga::*;
///
/// fn main() -> Result<(), Box<dyn Error>> {
///     let mut rng = smallrng_from_seed(42);
///     let f = |x: f64| {
///         if (0f64..=2f64).contains(&x) {
///             -(x - 1f64).powi(2) + 1f64
///         } else {
///             0f64
///         }
///     };
///
///     let samples = prs_with_rng(f, 1000, (-1f64, 3f64), 200, 1e-4, &mut rng)?;
///     assert!((samples.mean() - 1f64).abs() < 1e-1);
///
///     Ok(())
/// }
pub fn prs_with_rng<F, R: Rng + Clone>(
    f: F,
    n: usize,
    (a, b): (f64, f64),
    m: usize,
    eps: f64,
    rng: &mut R,
) -> anyhow::Result<Vec<f64>>
where
    F: Fn(f64) -> f64 + Copy,
{
    let mut result = vec![0f64; n];

    let w = WeightedUniform::from_max_pool_1d(f, (a, b), m, eps)?;

    let mut initial_x = w.sample_with_rng(rng, n);
    let mut left_num = n;

    while left_num > 0 {
        for &x in initial_x.iter() {
            let weight = w.weight_at(x);
            if weight <= 0f64 {
                continue;
            } else {
                let y = rng.random_range(0f64..=weight);

                if y <= f(x) {
                    result[n - left_num] = x;
                    left_num -= 1;
                    if left_num == 0 {
                        return Ok(result);
                    }
                }
            }
        }
        initial_x = w.sample_with_rng(rng, left_num);
    }
    panic!("Error: failed to generate {} samples", n);
}