1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374
// Copyright 2021 Developers of the Rand project.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// https://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or https://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
//! The Zeta and related distributions.
use num_traits::Float;
use crate::{Distribution, Standard};
use rand::{Rng, distributions::OpenClosed01};
use core::fmt;
/// Samples integers according to the [zeta distribution].
///
/// The zeta distribution is a limit of the [`Zipf`] distribution. Sometimes it
/// is called one of the following: discrete Pareto, Riemann-Zeta, Zipf, or
/// Zipf–Estoup distribution.
///
/// It has the density function `f(k) = k^(-a) / C(a)` for `k >= 1`, where `a`
/// is the parameter and `C(a)` is the Riemann zeta function.
///
/// # Example
/// ```
/// use rand::prelude::*;
/// use rand_distr::Zeta;
///
/// let val: f64 = thread_rng().sample(Zeta::new(1.5).unwrap());
/// println!("{}", val);
/// ```
///
/// # Remarks
///
/// The zeta distribution has no upper limit. Sampled values may be infinite.
/// In particular, a value of infinity might be returned for the following
/// reasons:
/// 1. it is the best representation in the type `F` of the actual sample.
/// 2. to prevent infinite loops for very small `a`.
///
/// # Implementation details
///
/// We are using the algorithm from [Non-Uniform Random Variate Generation],
/// Section 6.1, page 551.
///
/// [zeta distribution]: https://en.wikipedia.org/wiki/Zeta_distribution
/// [Non-Uniform Random Variate Generation]: https://doi.org/10.1007/978-1-4613-8643-8
#[derive(Clone, Copy, Debug)]
pub struct Zeta<F>
where F: Float, Standard: Distribution<F>, OpenClosed01: Distribution<F>
{
a_minus_1: F,
b: F,
}
/// Error type returned from `Zeta::new`.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum ZetaError {
/// `a <= 1` or `nan`.
ATooSmall,
}
impl fmt::Display for ZetaError {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(match self {
ZetaError::ATooSmall => "a <= 1 or is NaN in Zeta distribution",
})
}
}
#[cfg(feature = "std")]
#[cfg_attr(doc_cfg, doc(cfg(feature = "std")))]
impl std::error::Error for ZetaError {}
impl<F> Zeta<F>
where F: Float, Standard: Distribution<F>, OpenClosed01: Distribution<F>
{
/// Construct a new `Zeta` distribution with given `a` parameter.
#[inline]
pub fn new(a: F) -> Result<Zeta<F>, ZetaError> {
if !(a > F::one()) {
return Err(ZetaError::ATooSmall);
}
let a_minus_1 = a - F::one();
let two = F::one() + F::one();
Ok(Zeta {
a_minus_1,
b: two.powf(a_minus_1),
})
}
}
impl<F> Distribution<F> for Zeta<F>
where F: Float, Standard: Distribution<F>, OpenClosed01: Distribution<F>
{
#[inline]
fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> F {
loop {
let u = rng.sample(OpenClosed01);
let x = u.powf(-F::one() / self.a_minus_1).floor();
debug_assert!(x >= F::one());
if x.is_infinite() {
// For sufficiently small `a`, `x` will always be infinite,
// which is rejected, resulting in an infinite loop. We avoid
// this by always returning infinity instead.
return x;
}
let t = (F::one() + F::one() / x).powf(self.a_minus_1);
let v = rng.sample(Standard);
if v * x * (t - F::one()) * self.b <= t * (self.b - F::one()) {
return x;
}
}
}
}
/// Samples integers according to the Zipf distribution.
///
/// The samples follow Zipf's law: The frequency of each sample from a finite
/// set of size `n` is inversely proportional to a power of its frequency rank
/// (with exponent `s`).
///
/// For large `n`, this converges to the [`Zeta`] distribution.
///
/// For `s = 0`, this becomes a uniform distribution.
///
/// # Example
/// ```
/// use rand::prelude::*;
/// use rand_distr::Zipf;
///
/// let val: f64 = thread_rng().sample(Zipf::new(10, 1.5).unwrap());
/// println!("{}", val);
/// ```
///
/// # Implementation details
///
/// Implemented via [rejection sampling](https://en.wikipedia.org/wiki/Rejection_sampling),
/// due to Jason Crease[1].
///
/// [1]: https://jasoncrease.medium.com/rejection-sampling-the-zipf-distribution-6b359792cffa
#[derive(Clone, Copy, Debug)]
pub struct Zipf<F>
where F: Float, Standard: Distribution<F> {
n: F,
s: F,
t: F,
q: F,
}
/// Error type returned from `Zipf::new`.
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum ZipfError {
/// `s < 0` or `nan`.
STooSmall,
/// `n < 1`.
NTooSmall,
}
impl fmt::Display for ZipfError {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(match self {
ZipfError::STooSmall => "s < 0 or is NaN in Zipf distribution",
ZipfError::NTooSmall => "n < 1 in Zipf distribution",
})
}
}
#[cfg(feature = "std")]
#[cfg_attr(doc_cfg, doc(cfg(feature = "std")))]
impl std::error::Error for ZipfError {}
impl<F> Zipf<F>
where F: Float, Standard: Distribution<F> {
/// Construct a new `Zipf` distribution for a set with `n` elements and a
/// frequency rank exponent `s`.
///
/// For large `n`, rounding may occur to fit the number into the float type.
#[inline]
pub fn new(n: u64, s: F) -> Result<Zipf<F>, ZipfError> {
if !(s >= F::zero()) {
return Err(ZipfError::STooSmall);
}
if n < 1 {
return Err(ZipfError::NTooSmall);
}
let n = F::from(n).unwrap(); // This does not fail.
let q = if s != F::one() {
// Make sure to calculate the division only once.
F::one() / (F::one() - s)
} else {
// This value is never used.
F::zero()
};
let t = if s != F::one() {
(n.powf(F::one() - s) - s) * q
} else {
F::one() + n.ln()
};
debug_assert!(t > F::zero());
Ok(Zipf {
n, s, t, q
})
}
/// Inverse cumulative density function
#[inline]
fn inv_cdf(&self, p: F) -> F {
let one = F::one();
let pt = p * self.t;
if pt <= one {
pt
} else if self.s != one {
(pt * (one - self.s) + self.s).powf(self.q)
} else {
(pt - one).exp()
}
}
}
impl<F> Distribution<F> for Zipf<F>
where F: Float, Standard: Distribution<F>
{
#[inline]
fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> F {
let one = F::one();
loop {
let inv_b = self.inv_cdf(rng.sample(Standard));
let x = (inv_b + one).floor();
let mut ratio = x.powf(-self.s);
if x > one {
ratio = ratio * inv_b.powf(self.s)
};
let y = rng.sample(Standard);
if y < ratio {
return x;
}
}
}
}
#[cfg(test)]
mod tests {
use super::*;
fn test_samples<F: Float + core::fmt::Debug, D: Distribution<F>>(
distr: D, zero: F, expected: &[F],
) {
let mut rng = crate::test::rng(213);
let mut buf = [zero; 4];
for x in &mut buf {
*x = rng.sample(&distr);
}
assert_eq!(buf, expected);
}
#[test]
#[should_panic]
fn zeta_invalid() {
Zeta::new(1.).unwrap();
}
#[test]
#[should_panic]
fn zeta_nan() {
Zeta::new(core::f64::NAN).unwrap();
}
#[test]
fn zeta_sample() {
let a = 2.0;
let d = Zeta::new(a).unwrap();
let mut rng = crate::test::rng(1);
for _ in 0..1000 {
let r = d.sample(&mut rng);
assert!(r >= 1.);
}
}
#[test]
fn zeta_small_a() {
let a = 1. + 1e-15;
let d = Zeta::new(a).unwrap();
let mut rng = crate::test::rng(2);
for _ in 0..1000 {
let r = d.sample(&mut rng);
assert!(r >= 1.);
}
}
#[test]
fn zeta_value_stability() {
test_samples(Zeta::new(1.5).unwrap(), 0f32, &[
1.0, 2.0, 1.0, 1.0,
]);
test_samples(Zeta::new(2.0).unwrap(), 0f64, &[
2.0, 1.0, 1.0, 1.0,
]);
}
#[test]
#[should_panic]
fn zipf_s_too_small() {
Zipf::new(10, -1.).unwrap();
}
#[test]
#[should_panic]
fn zipf_n_too_small() {
Zipf::new(0, 1.).unwrap();
}
#[test]
#[should_panic]
fn zipf_nan() {
Zipf::new(10, core::f64::NAN).unwrap();
}
#[test]
fn zipf_sample() {
let d = Zipf::new(10, 0.5).unwrap();
let mut rng = crate::test::rng(2);
for _ in 0..1000 {
let r = d.sample(&mut rng);
assert!(r >= 1.);
}
}
#[test]
fn zipf_sample_s_1() {
let d = Zipf::new(10, 1.).unwrap();
let mut rng = crate::test::rng(2);
for _ in 0..1000 {
let r = d.sample(&mut rng);
assert!(r >= 1.);
}
}
#[test]
fn zipf_sample_s_0() {
let d = Zipf::new(10, 0.).unwrap();
let mut rng = crate::test::rng(2);
for _ in 0..1000 {
let r = d.sample(&mut rng);
assert!(r >= 1.);
}
// TODO: verify that this is a uniform distribution
}
#[test]
fn zipf_sample_large_n() {
let d = Zipf::new(core::u64::MAX, 1.5).unwrap();
let mut rng = crate::test::rng(2);
for _ in 0..1000 {
let r = d.sample(&mut rng);
assert!(r >= 1.);
}
// TODO: verify that this is a zeta distribution
}
#[test]
fn zipf_value_stability() {
test_samples(Zipf::new(10, 0.5).unwrap(), 0f32, &[
10.0, 2.0, 6.0, 7.0
]);
test_samples(Zipf::new(10, 2.0).unwrap(), 0f64, &[
1.0, 2.0, 3.0, 2.0
]);
}
}