cdflib 0.4.0

Pure-Rust port of CDFLIB: cumulative distribution functions (CDF) associated with common probability distributions
Documentation
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
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
use thiserror::Error;

use crate::error::SearchError;
use crate::search::{search_bounded_zero, search_monotone, SEARCH_BOUND};
use crate::special::beta_inc;
use crate::special::{beta_log, psi};
use crate::traits::{Continuous, ContinuousCdf, Entropy, Mean, Variance};

/// Β distribution with shape parameters *a* > 0 and *b* > 0.
///
/// Defined over the interval [0 . . 1], with density
/// *f*(*x*; *a*, *b*) = *xᵃ* ⁻ ¹ (1 − *x*)*ᵇ* ⁻ ¹ / Β(*a*, *b*).
///
/// # Example
///
/// ```
/// use cdflib::Beta;
/// use cdflib::traits::ContinuousCdf;
///
/// let b = Beta::new(2.0, 5.0);
///
/// // Pr[X ≤ 0.3]
/// let p = b.cdf(0.3);
///
/// // Compute parameter a given Pr[X ≤ 0.5] = 0.9 and b = 2.0
/// let a = Beta::search_a(0.9, 0.1, 0.5, 2.0).unwrap();
/// ```
#[derive(Debug, Clone, Copy, PartialEq)]
pub struct Beta {
    a: f64,
    b: f64,
}

/// Errors arising from constructing a [`Beta`] or from its parameter searches.
///
/// [`Beta`]: crate::Beta
#[derive(Debug, Clone, Copy, PartialEq, Error)]
pub enum BetaError {
    /// The shape parameter *a* was not strictly positive.
    #[error("shape parameter `a` must be positive, got {0}")]
    ANotPositive(f64),
    /// The shape parameter *a* was not finite.
    #[error("shape parameter `a` must be finite, got {0}")]
    ANotFinite(f64),
    /// The shape parameter *b* was not strictly positive.
    #[error("shape parameter `b` must be positive, got {0}")]
    BNotPositive(f64),
    /// The shape parameter *b* was not finite.
    #[error("shape parameter `b` must be finite, got {0}")]
    BNotFinite(f64),
    /// The argument *x* fell outside [0 . . 1].
    #[error("argument x must be in [0..1], got {0}")]
    XOutOfRange(f64),
    /// The probability *p* fell outside [0 . . 1] (or was non-finite).
    #[error("probability {0} outside [0..1]")]
    PNotInRange(f64),
    /// The probability *q* fell outside [0 . . 1] (or was non-finite).
    #[error("probability {0} outside [0..1]")]
    QNotInRange(f64),
    /// The pair (*p*, *q*) is not complementary (|*p* + *q* − 1| > 3 ε).
    /// Mirrors CDFLIB's `cdfbet` status 3.
    #[error("p ({p}) and q ({q}) are not complementary: |p + q - 1| > 3ε")]
    PQSumNotOne { p: f64, q: f64 },
    /// The internal root-finder failed; see [`SearchError`].
    ///
    /// [`SearchError`]: crate::error::SearchError
    #[error(transparent)]
    Search(#[from] SearchError),
}

impl Beta {
    /// Construct a Β(*a*, *b*) distribution with the given shape parameters.
    ///
    /// # Panics
    ///
    /// Panics if either parameter is invalid; use [`try_new`] for a fallible
    /// variant.
    ///
    /// [`try_new`]: Self::try_new
    #[inline]
    pub fn new(a: f64, b: f64) -> Self {
        Self::try_new(a, b).unwrap()
    }

    /// Fallible counterpart of [`new`](Self::new) returning a [`BetaError`]
    /// instead of panicking.
    ///
    /// Returns [`ANotPositive`], [`ANotFinite`], [`BNotPositive`], or
    /// [`BNotFinite`] if either parameter fails its validity check.
    ///
    /// [`ANotPositive`]: BetaError::ANotPositive
    /// [`ANotFinite`]: BetaError::ANotFinite
    /// [`BNotPositive`]: BetaError::BNotPositive
    /// [`BNotFinite`]: BetaError::BNotFinite
    #[inline]
    pub fn try_new(a: f64, b: f64) -> Result<Self, BetaError> {
        if !a.is_finite() {
            return Err(BetaError::ANotFinite(a));
        }
        if a <= 0.0 {
            return Err(BetaError::ANotPositive(a));
        }
        if !b.is_finite() {
            return Err(BetaError::BNotFinite(b));
        }
        if b <= 0.0 {
            return Err(BetaError::BNotPositive(b));
        }
        Ok(Self { a, b })
    }

    /// Returns the shape parameter *a*.
    #[inline]
    pub const fn a(&self) -> f64 {
        self.a
    }

    /// Returns the shape parameter *b*.
    #[inline]
    pub const fn b(&self) -> f64 {
        self.b
    }

    /// Returns the shape parameter *a* satisfying Pr[*X* ≤ *x*] = *p*.
    ///
    /// CDFLIB's `cdfbet` with `which = 3`. Caller passes both *p* and
    /// *q* = 1 − *p*; consistency is enforced within 3 ε.
    #[inline]
    pub fn search_a(p: f64, q: f64, x: f64, b: f64) -> Result<f64, BetaError> {
        check_pq(p, q)?;
        if !(0.0..=1.0).contains(&x) {
            return Err(BetaError::XOutOfRange(x));
        }
        if !b.is_finite() {
            return Err(BetaError::BNotFinite(b));
        }
        if b <= 0.0 {
            return Err(BetaError::BNotPositive(b));
        }
        let f = |a: f64| {
            let (cum, ccum) = beta_inc(a, b, x, 1.0 - x);
            if p <= q {
                cum - p
            } else {
                ccum - q
            }
        };
        // I_x(a, b) is decreasing in a (more weight near 1 when a grows).
        // Match cdfbet's which=3: range (zero, inf), start = 5.0;
        // mirror Fortran's cum-p if p<=q else ccum-q precision pivot.
        Ok(search_monotone(
            0.0,
            SEARCH_BOUND,
            5.0,
            0.0,
            SEARCH_BOUND,
            f,
        )?)
    }

    /// Returns the shape parameter *b* satisfying Pr[*X* ≤ *x*] = *p*.
    ///
    /// CDFLIB's `cdfbet` with `which = 4`. Caller passes both *p* and
    /// *q* = 1 − *p*; consistency is enforced within 3 ε.
    #[inline]
    pub fn search_b(p: f64, q: f64, x: f64, a: f64) -> Result<f64, BetaError> {
        check_pq(p, q)?;
        if !(0.0..=1.0).contains(&x) {
            return Err(BetaError::XOutOfRange(x));
        }
        if !a.is_finite() {
            return Err(BetaError::ANotFinite(a));
        }
        if a <= 0.0 {
            return Err(BetaError::ANotPositive(a));
        }
        let f = |b: f64| {
            let (cum, ccum) = beta_inc(a, b, x, 1.0 - x);
            if p <= q {
                cum - p
            } else {
                ccum - q
            }
        };
        // I_x(a, b) is increasing in b. Match cdfbet's which=4 setup and
        // precision pivot.
        Ok(search_monotone(
            0.0,
            SEARCH_BOUND,
            5.0,
            0.0,
            SEARCH_BOUND,
            f,
        )?)
    }
}

#[inline]
fn check_p(p: f64) -> Result<(), BetaError> {
    if !(0.0..=1.0).contains(&p) || !p.is_finite() {
        Err(BetaError::PNotInRange(p))
    } else {
        Ok(())
    }
}

#[inline]
fn check_q(q: f64) -> Result<(), BetaError> {
    if !(0.0..=1.0).contains(&q) || !q.is_finite() {
        Err(BetaError::QNotInRange(q))
    } else {
        Ok(())
    }
}

#[inline]
fn check_pq(p: f64, q: f64) -> Result<(), BetaError> {
    check_p(p)?;
    check_q(q)?;
    if (p + q - 1.0).abs() > 3.0 * f64::EPSILON {
        return Err(BetaError::PQSumNotOne { p, q });
    }
    Ok(())
}

impl ContinuousCdf for Beta {
    type Error = BetaError;

    #[inline]
    fn cdf(&self, x: f64) -> f64 {
        if x <= 0.0 {
            return 0.0;
        }
        if x >= 1.0 {
            return 1.0;
        }
        let (cum, _) = beta_inc(self.a, self.b, x, 1.0 - x);
        cum
    }

    #[inline]
    fn ccdf(&self, x: f64) -> f64 {
        if x <= 0.0 {
            return 1.0;
        }
        if x >= 1.0 {
            return 0.0;
        }
        let (_, ccum) = beta_inc(self.a, self.b, x, 1.0 - x);
        ccum
    }

    #[inline]
    fn inverse_cdf(&self, p: f64) -> Result<f64, BetaError> {
        check_p(p)?;
        if p == 0.0 {
            return Ok(0.0);
        }
        if p == 1.0 {
            return Ok(1.0);
        }
        let a = self.a;
        let b = self.b;
        let q = 1.0 - p;
        // F90 cdfbet which=2 drives dzror directly on x when p<=q and on y
        // when p>q, keeping x+y=1 exactly throughout the search.
        if p <= q {
            let f = |x: f64| {
                let (cum, _) = beta_inc(a, b, x, 1.0 - x);
                cum - p
            };
            Ok(search_bounded_zero(0.0, 1.0, f)?)
        } else {
            let f = |y: f64| {
                let (_, ccum) = beta_inc(a, b, 1.0 - y, y);
                ccum - q
            };
            let y = search_bounded_zero(0.0, 1.0, f)?;
            Ok(1.0 - y)
        }
    }
}

impl Beta {
    /// Returns the quantile *x* such that [ccdf]\(*x*\) = *q*.
    ///
    /// Mirrors CDFLIB's `cdfbet` with `which = 2`, using the same
    /// `cum - p` / `ccum - q` pivot as the Fortran routine.
    ///
    /// [ccdf]: crate::traits::ContinuousCdf::ccdf
    #[inline]
    pub fn inverse_ccdf(&self, q: f64) -> Result<f64, BetaError> {
        check_q(q)?;
        if q == 1.0 {
            return Ok(0.0);
        }
        if q == 0.0 {
            return Ok(1.0);
        }
        let a = self.a;
        let b = self.b;
        let p = 1.0 - q;
        // F90 cdfbet which=2 (cdflib.f90:2713-2745) switches the search
        // variable: dzror on x with cum-p when p<=q, dzror on y with
        // ccum-q when p>q, keeping y = 1-x (or x = 1-y) updated each
        // iteration. The variable-switch preserves precision when the
        // small tail is near the right edge.
        if p <= q {
            let f = |x: f64| {
                let (cum, _) = beta_inc(a, b, x, 1.0 - x);
                cum - p
            };
            Ok(search_bounded_zero(0.0, 1.0, f)?)
        } else {
            let f = |y: f64| {
                let (_, ccum) = beta_inc(a, b, 1.0 - y, y);
                ccum - q
            };
            let y = search_bounded_zero(0.0, 1.0, f)?;
            Ok(1.0 - y)
        }
    }
}

impl Continuous for Beta {
    #[inline]
    fn pdf(&self, x: f64) -> f64 {
        if x <= 0.0 || x >= 1.0 {
            return 0.0;
        }
        self.ln_pdf(x).exp()
    }
    #[inline]
    fn ln_pdf(&self, x: f64) -> f64 {
        if x <= 0.0 || x >= 1.0 {
            return f64::NEG_INFINITY;
        }
        (self.a - 1.0) * x.ln() + (self.b - 1.0) * (1.0 - x).ln() - beta_log(self.a, self.b)
    }
}

impl Mean for Beta {
    #[inline]
    fn mean(&self) -> f64 {
        self.a / (self.a + self.b)
    }
}

impl Variance for Beta {
    #[inline]
    fn variance(&self) -> f64 {
        let s = self.a + self.b;
        self.a * self.b / (s * s * (s + 1.0))
    }
}

impl Entropy for Beta {
    #[inline]
    fn entropy(&self) -> f64 {
        // H = ln Β(a,b) - (a-1)ψ(a) - (b-1)ψ(b) + (a+b-2)ψ(a+b)
        beta_log(self.a, self.b) - (self.a - 1.0) * psi(self.a) - (self.b - 1.0) * psi(self.b)
            + (self.a + self.b - 2.0) * psi(self.a + self.b)
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn rejects_invalid_parameters() {
        assert!(matches!(
            Beta::try_new(0.0, 1.0),
            Err(BetaError::ANotPositive(0.0))
        ));
        assert!(matches!(
            Beta::try_new(1.0, 0.0),
            Err(BetaError::BNotPositive(0.0))
        ));
        assert!(matches!(
            Beta::try_new(f64::NAN, 1.0),
            Err(BetaError::ANotFinite(_))
        ));
        assert!(matches!(
            Beta::try_new(1.0, f64::INFINITY),
            Err(BetaError::BNotFinite(_))
        ));
    }

    #[test]
    fn inverse_boundaries_and_density_edges() {
        let d = Beta::new(2.0, 3.0);
        assert_eq!(d.cdf(0.0), 0.0);
        assert_eq!(d.cdf(1.0), 1.0);
        assert_eq!(d.ccdf(0.0), 1.0);
        assert_eq!(d.ccdf(1.0), 0.0);
        assert_eq!(d.inverse_cdf(0.0).unwrap(), 0.0);
        assert_eq!(d.inverse_cdf(1.0).unwrap(), 1.0);
        assert_eq!(d.inverse_ccdf(1.0).unwrap(), 0.0);
        assert_eq!(d.inverse_ccdf(0.0).unwrap(), 1.0);
        assert_eq!(d.pdf(0.0), 0.0);
        assert_eq!(d.pdf(1.0), 0.0);
        assert_eq!(d.ln_pdf(0.0), f64::NEG_INFINITY);
        assert_eq!(d.ln_pdf(1.0), f64::NEG_INFINITY);
        assert!(d.pdf(0.4).is_finite());
        assert!(d.ln_pdf(0.4).is_finite());
        assert!(d.inverse_ccdf(0.4).unwrap().is_finite());
        assert!(d.mean().is_finite());
        assert!(d.variance().is_finite());
        assert!(d.entropy().is_finite());
    }

    #[test]
    fn search_parameter_rejects_invalid_inputs() {
        assert!(matches!(
            Beta::search_a(-0.1, 1.1, 0.5, 2.0),
            Err(BetaError::PNotInRange(-0.1))
        ));
        assert!(matches!(
            Beta::search_a(0.5, 0.5, 0.5, 0.0),
            Err(BetaError::BNotPositive(0.0))
        ));
        assert!(matches!(
            Beta::search_b(0.5, 0.5, 0.5, 0.0),
            Err(BetaError::ANotPositive(0.0))
        ));
        assert!(matches!(
            Beta::search_a(0.5, 0.5, 1.5, 2.0),
            Err(BetaError::XOutOfRange(1.5))
        ));
        assert!(matches!(
            Beta::search_b(0.5, 0.5, -0.1, 2.0),
            Err(BetaError::XOutOfRange(x)) if x == -0.1
        ));
    }
}