scirs2-special 0.4.1

Special functions module for SciRS2 (scirs2-special)
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
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
//! Dirichlet L-functions and Dirichlet Characters
//!
//! This module implements Dirichlet L-functions L(s, χ) = Σ_{n=1}^∞ χ(n)/n^s
//! for Dirichlet characters χ modulo q.
//!
//! Key references:
//! - Davenport, "Multiplicative Number Theory"
//! - Ireland & Rosen, "A Classical Introduction to Modern Number Theory"

use std::f64::consts::PI;

/// A Dirichlet character χ modulo q.
///
/// The character assigns a value χ(n) to each integer n.
/// For the principal character: χ(n)=1 if gcd(n,q)=1, else 0.
/// For real (quadratic) characters: values in {-1, 0, 1}.
#[derive(Debug, Clone)]
pub struct DirichletCharacter {
    /// The modulus of the character
    pub modulus: usize,
    /// Character values for n = 0, 1, ..., modulus-1.
    /// `values[n]` = chi(n mod q) as a rational integer (i64).
    pub values: Vec<i64>,
}

impl DirichletCharacter {
    /// Construct the principal character χ_0 modulo q.
    ///
    /// χ_0(n) = 1 if gcd(n, q) = 1, else 0.
    pub fn principal(modulus: usize) -> Self {
        let values = (0..modulus)
            .map(|n| if gcd(n, modulus) == 1 { 1i64 } else { 0i64 })
            .collect();
        DirichletCharacter { modulus, values }
    }

    /// Construct the Kronecker symbol character (d/·) modulo |d|.
    ///
    /// For a fundamental discriminant d, this gives a primitive real character.
    /// The character is defined as χ(n) = (d/n) if gcd(n, |d|) = 1, else 0.
    ///
    /// # Arguments
    /// * `d` - Discriminant (nonzero integer). Must be ≠ 0.
    pub fn kronecker_symbol(d: i64) -> Self {
        let modulus = d.unsigned_abs() as usize;
        if modulus == 0 {
            // Degenerate: return the trivial modulus-1 character
            return DirichletCharacter {
                modulus: 1,
                values: vec![1],
            };
        }
        let values = (0..modulus)
            .map(|n| {
                // Dirichlet character: 0 when gcd(n, q) > 1
                if gcd(n, modulus) != 1 {
                    0i64
                } else {
                    kronecker(d, n as i64)
                }
            })
            .collect();
        DirichletCharacter { modulus, values }
    }

    /// Evaluate the character at n (extended periodically).
    pub fn eval(&self, n: i64) -> i64 {
        if self.modulus == 0 {
            return 0;
        }
        let idx = n.rem_euclid(self.modulus as i64) as usize;
        self.values[idx]
    }

    /// Check whether the character is primitive.
    ///
    /// A character χ mod q is primitive if it is not induced by a character
    /// of smaller modulus. We check this via the conductor.
    pub fn is_primitive(&self) -> bool {
        self.conductor() == self.modulus
    }

    /// Compute the conductor of the character.
    ///
    /// The conductor is the smallest positive divisor d|q such that χ is
    /// induced by a character mod d.
    pub fn conductor(&self) -> usize {
        let q = self.modulus;
        if q == 1 {
            return 1;
        }
        // Check divisors in increasing order
        let mut divs: Vec<usize> = (1..=q).filter(|&d| q.is_multiple_of(d)).collect();
        divs.sort_unstable();
        for d in divs {
            if d == q {
                return q;
            }
            // Check if χ is induced by character mod d:
            // χ(n) must equal 0 whenever gcd(n,q) > 1, and for n ≡ m (mod d)
            // with gcd(n,q) = 1 = gcd(m,q), we need χ(n) = χ(m).
            if self.is_induced_by(d) {
                return d;
            }
        }
        q
    }

    /// Check if the character is induced by a character of modulus d (d | modulus).
    fn is_induced_by(&self, d: usize) -> bool {
        let q = self.modulus;
        if !q.is_multiple_of(d) {
            return false;
        }
        // For all a, b coprime to q with a ≡ b (mod d), χ(a) = χ(b).
        // Also: if gcd(n, d) > 1 then χ(n) = 0.
        for a in 0..q {
            if gcd(a, q) != 1 {
                continue;
            }
            let chi_a = self.values[a];
            // Check all b ≡ a (mod d) with 0 ≤ b < q
            let mut b = a % d;
            while b < q {
                if gcd(b, q) == 1 && self.values[b] != chi_a {
                    return false;
                }
                b += d;
            }
        }
        true
    }

    /// Check if the character is real (values in {-1, 0, 1}).
    pub fn is_real(&self) -> bool {
        self.values.iter().all(|&v| v == -1 || v == 0 || v == 1)
    }
}

// ────────────────────────────────────────────────────────────────────────────
// Number-theoretic helpers
// ────────────────────────────────────────────────────────────────────────────

/// Compute gcd(a, b) using Euclidean algorithm.
pub fn gcd(a: usize, b: usize) -> usize {
    if b == 0 {
        a
    } else {
        gcd(b, a % b)
    }
}

/// Compute the Jacobi symbol (a/n) using quadratic reciprocity.
/// n must be a positive odd integer.
pub fn jacobi_symbol(mut a: i64, mut n: i64) -> i64 {
    debug_assert!(n > 0 && n % 2 == 1);
    let mut result = 1i64;
    a = a.rem_euclid(n);
    while a != 0 {
        while a % 2 == 0 {
            a /= 2;
            let r = n % 8;
            if r == 3 || r == 5 {
                result = -result;
            }
        }
        std::mem::swap(&mut a, &mut n);
        if a % 4 == 3 && n % 4 == 3 {
            result = -result;
        }
        a = a.rem_euclid(n);
    }
    if n == 1 {
        result
    } else {
        0
    }
}

/// Compute the Kronecker symbol (a/n) for any integer n.
///
/// Extends the Jacobi symbol to n=0, n=±1, n=2, and n<0.
pub fn kronecker(a: i64, n: i64) -> i64 {
    if n == 0 {
        return if a.abs() == 1 { 1 } else { 0 };
    }
    if n == 1 {
        return 1;
    }
    if n == -1 {
        return if a < 0 { -1 } else { 1 };
    }
    let mut n = n;
    let mut result = 1i64;
    if n < 0 {
        n = -n;
        if a < 0 {
            result = -result;
        }
    }
    // Factor out powers of 2
    let v2 = n.trailing_zeros() as i64;
    if v2 > 0 {
        n >>= v2;
        let a8 = a.rem_euclid(8);
        let k2 = if a8 == 1 || a8 == 7 { 1i64 } else { -1i64 };
        if v2 % 2 != 0 {
            result *= k2;
        }
    }
    if n == 1 {
        return result;
    }
    // Now n is odd and positive
    result * jacobi_symbol(a, n)
}

// ────────────────────────────────────────────────────────────────────────────
// L-function evaluation
// ────────────────────────────────────────────────────────────────────────────

/// Evaluate the Dirichlet L-function L(s, χ) by partial summation.
///
/// L(s, χ) = Σ_{n=1}^{n_terms} χ(n) / n^s
///
/// This direct partial sum converges for Re(s) > 1. For s near 1 with a
/// non-principal character, the series converges conditionally for Re(s) > 0
/// by the cancellation in χ(n).
///
/// # Arguments
/// * `s` - Real argument (must be > 0 for non-principal, > 1 for principal)
/// * `chi` - Dirichlet character
/// * `n_terms` - Number of terms to sum (default 10000 for ~4-digit accuracy)
pub fn dirichlet_l(s: f64, chi: &DirichletCharacter, n_terms: usize) -> f64 {
    let q = chi.modulus;
    // For efficiency, precompute one period and accumulate in blocks
    if q == 0 {
        return 0.0;
    }

    // Group partial sums by complete periods + remainder
    let full_periods = n_terms / q;
    let remainder = n_terms % q;

    let mut sum = 0.0f64;

    // Complete periods
    for period in 0..full_periods {
        let base = period * q;
        for r in 1..=q {
            let n = base + r;
            let chi_n = chi.eval(n as i64);
            if chi_n != 0 {
                sum += chi_n as f64 / (n as f64).powf(s);
            }
        }
    }
    // Remainder
    let base = full_periods * q;
    for r in 1..=remainder {
        let n = base + r;
        let chi_n = chi.eval(n as i64);
        if chi_n != 0 {
            sum += chi_n as f64 / (n as f64).powf(s);
        }
    }
    sum
}

/// Evaluate L(1, χ) for a non-principal character χ.
///
/// For non-principal characters, L(1,χ) converges. This function uses:
/// - Even real characters: L(1,χ) = -1/q * Σ_{a=1}^{q} χ(a) * log|sin(π a / q)|
/// - Odd real characters: direct partial summation with Euler-Maclaurin acceleration
/// - For non-real characters: direct partial summation
///
/// # Returns
/// `None` if χ is principal (L(1,χ_0) diverges).
pub fn l_function_at_1(chi: &DirichletCharacter) -> Option<f64> {
    let q = chi.modulus;
    // Check for principal character: Σ χ(n) over one period = 0 for non-principal
    let sum_chi: i64 = (0..q).map(|n| chi.eval(n as i64)).sum();
    if sum_chi > 0 && q > 1 {
        return None; // likely principal
    }
    // Also handle q=1 (trivial character)
    if q == 1 {
        return None;
    }

    // Check if chi is even (χ(-1) = 1) or odd (χ(-1) = -1)
    let chi_neg1 = chi.eval(-1);

    if chi.is_real() && chi_neg1 == 1 {
        // Even real character: L(1,χ) = -1/q * Σ_{a=1}^{q-1} χ(a) log|sin(πa/q)|
        let q_f = q as f64;
        let sum: f64 = (1..q)
            .map(|a| {
                let chi_a = chi.eval(a as i64) as f64;
                let sin_val = (PI * a as f64 / q_f).sin().abs();
                if sin_val > 1e-15 {
                    chi_a * sin_val.ln()
                } else {
                    0.0
                }
            })
            .sum();
        Some(-sum / q_f)
    } else if chi.is_real() && chi_neg1 == -1 {
        // Odd real character: use the exact class number formula.
        // L(1,χ) = (πi / (q * τ̄(χ))) * Σ_{a=1}^{q-1} a * χ(a)
        //
        // For a real primitive odd character, τ(χ) = ε i√q (ε ∈ {±1}).
        // Hence τ̄(χ) = -ε i√q.
        // L(1,χ) = πi / (q * (-ε i√q)) * S = π / (q√q * ε) * S  (since i/(-i) = -1... wait)
        // πi / (-ε i q√q) * S = π / (-ε q√q) * S
        //
        // We compute τ(χ) = Σ_{a=1}^q χ(a) * exp(2πia/q) and get the imaginary part ε√q.
        // Then ε = sign(Im(τ)) and L(1,χ) = π * |S| / (q√q) with sign from ε*S.
        //
        // Numerically: compute the real Gauss sum numerator.
        let q_f = q as f64;
        // Compute τ(χ) imaginary part: Im(τ) = Σ χ(a) sin(2πa/q)
        let tau_im: f64 = (1..q)
            .map(|a| chi.eval(a as i64) as f64 * (2.0 * PI * a as f64 / q_f).sin())
            .sum();
        // S = Σ_{a=1}^{q-1} a * χ(a)
        let s: f64 = (1..q).map(|a| a as f64 * chi.eval(a as i64) as f64).sum();
        // L(1,χ) = π * (-S) / (q * tau_im)  (derived from L = πi/(q*τ̄)*S and τ̄=-τ for real χ)
        // Since τ̄ = τ* = Σ χ(a)*exp(-2πia/q) = -Σ χ(a)*exp(2πia/q) for odd real χ = -τ
        // L(1,χ) = πi / (q * (-τ)) * S = -πi * S / (q * τ)
        // τ = i * tau_im (approximately), so:
        // L(1,χ) = -πi * S / (q * i * tau_im) = -π * S / (q * tau_im)
        let l1 = -PI * s / (q_f * tau_im);
        Some(l1)
    } else {
        // Non-real character: direct partial summation
        Some(dirichlet_l(1.0, chi, 100000))
    }
}

/// Compute the generalized Bernoulli number B_{k, χ}.
///
/// B_{k, χ} = q^{k-1} Σ_{a=1}^q χ(a) B_k(a/q)
///
/// where B_k(x) is the Bernoulli polynomial.
///
/// # Arguments
/// * `k` - Non-negative integer
/// * `chi` - Dirichlet character
pub fn generalized_bernoulli(k: usize, chi: &DirichletCharacter) -> f64 {
    let q = chi.modulus as f64;
    let q_pow = q.powi(k as i32 - 1);
    let sum: f64 = (1..=chi.modulus)
        .map(|a| {
            let chi_a = chi.eval(a as i64) as f64;
            let x = a as f64 / q;
            chi_a * bernoulli_poly(k, x)
        })
        .sum();
    q_pow * sum
}

/// Evaluate the Bernoulli polynomial B_k(x).
///
/// Uses the explicit formula B_k(x) = Σ_{j=0}^k C(k,j) B_j x^{k-j}
/// where B_j are Bernoulli numbers.
pub fn bernoulli_poly(k: usize, x: f64) -> f64 {
    let berns = bernoulli_numbers(k);
    let mut result = 0.0f64;
    let mut binom = 1.0f64;
    for j in 0..=k {
        result += binom * berns[j] * x.powi((k - j) as i32);
        // Update binom = C(k, j+1)
        if j < k {
            binom *= (k - j) as f64 / (j + 1) as f64;
        }
    }
    result
}

/// Compute Bernoulli numbers B_0, B_1, ..., B_n.
pub fn bernoulli_numbers(n: usize) -> Vec<f64> {
    let mut b = vec![0.0f64; n + 1];
    b[0] = 1.0;
    if n == 0 {
        return b;
    }
    b[1] = -0.5;
    // Use the recurrence: B_n = -1/(n+1) * Σ_{k=0}^{n-1} C(n+1,k) B_k
    for m in 2..=n {
        let mut s = 0.0f64;
        let mut binom = 1.0f64;
        for k in 0..m {
            s += binom * b[k];
            binom *= (m + 1 - k) as f64 / (k + 1) as f64;
        }
        b[m] = -s / (m + 1) as f64;
    }
    b
}

/// Evaluate the completed L-function Λ(s, χ) satisfying the functional equation.
///
/// For a primitive character χ mod q, the functional equation is:
/// Λ(s, χ) = (q/π)^{s/2} Γ((s+a)/2) L(s, χ)
/// where a = 0 if χ(-1) = 1, a = 1 if χ(-1) = -1.
///
/// The functional equation: Λ(s, χ) = ε(χ) * Λ(1-s, χ̄)
/// where ε(χ) = τ(χ) / (i^a * sqrt(q)) is the root number.
///
/// This implementation approximates L(s, χ) via partial sums; the completion
/// factor is applied analytically.
pub fn l_function_complete(s: f64, chi: &DirichletCharacter) -> f64 {
    let q = chi.modulus as f64;
    let a = if chi.eval(-1) == 1 { 0.0f64 } else { 1.0f64 };

    // Gamma factor: Γ((s+a)/2)
    let gamma_arg = (s + a) / 2.0;
    let gamma_factor = gamma_function(gamma_arg);

    // L(s, χ) via partial sums
    let l_val = dirichlet_l(s, chi, 50000);

    // Completion factor: (q/π)^{s/2}
    let completion = (q / PI).powf(s / 2.0);

    completion * gamma_factor * l_val
}

/// Simple gamma function approximation via Stirling / Lanczos.
fn gamma_function(x: f64) -> f64 {
    if x <= 0.0 {
        return f64::INFINITY;
    }
    if x < 0.5 {
        return PI / ((PI * x).sin() * gamma_function(1.0 - x));
    }
    // Lanczos approximation (g=7, n=9)
    const G: f64 = 7.0;
    const C: [f64; 9] = [
        0.99999999999980993,
        676.5203681218851,
        -1259.1392167224028,
        771.32342877765313,
        -176.61502916214059,
        12.507343278686905,
        -0.13857109526572012,
        9.9843695780195716e-6,
        1.5056327351493116e-7,
    ];
    let x = x - 1.0;
    let mut s = C[0];
    for (i, &c) in C[1..].iter().enumerate() {
        s += c / (x + (i + 1) as f64);
    }
    let t = x + G + 0.5;
    (2.0 * PI).sqrt() * t.powf(x + 0.5) * (-t).exp() * s
}

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

    #[test]
    fn test_gcd() {
        assert_eq!(gcd(12, 8), 4);
        assert_eq!(gcd(7, 5), 1);
        assert_eq!(gcd(0, 5), 5);
    }

    #[test]
    fn test_kronecker() {
        // Legendre symbol (5/3) = -1 (5 ≡ 2 mod 3, 2 is not QR mod 3)
        assert_eq!(kronecker(5, 3), -1);
        // (1/n) = 1 for all n
        assert_eq!(kronecker(1, 5), 1);
        assert_eq!(kronecker(1, 7), 1);
        // (-1/2) = Kronecker extension
        assert_eq!(kronecker(-1, 2), 1); // (-1)^{(2^2-1)/8} = (-1)^0 = no wait, Kronecker(-1,2) = (-1)^{(2^2-1)/8} = 1
    }

    #[test]
    fn test_principal_character() {
        let chi = DirichletCharacter::principal(4);
        assert_eq!(chi.eval(1), 1);
        assert_eq!(chi.eval(2), 0);
        assert_eq!(chi.eval(3), 1);
        assert_eq!(chi.eval(4), 0);
    }

    #[test]
    fn test_kronecker_character_neg4() {
        // Kronecker symbol (-4/n) gives the non-principal character mod 4
        // chi(-4/1)=1, chi(-4/2)=0, chi(-4/3)=-1, chi(-4/4)=0
        let chi = DirichletCharacter::kronecker_symbol(-4);
        assert_eq!(chi.eval(1), 1);
        assert_eq!(chi.eval(3), -1);
        assert_eq!(chi.eval(2), 0);
    }

    #[test]
    fn test_dirichlet_l_principal_mod4() {
        // L(2, χ_0 mod 4) = Σ_{n odd} 1/n^2 = π²/8 ≈ 1.2337
        let chi = DirichletCharacter::principal(4);
        let l2 = dirichlet_l(2.0, &chi, 100000);
        let expected = std::f64::consts::PI * std::f64::consts::PI / 8.0;
        assert_relative_eq!(l2, expected, epsilon = 1e-3);
    }

    #[test]
    fn test_l_function_at_1_legendre_neg4() {
        // χ(-4/·): L(1, χ) = π/4 ≈ 0.7854 (Leibniz formula)
        let chi = DirichletCharacter::kronecker_symbol(-4);
        let l1 = l_function_at_1(&chi).expect("should compute L(1,chi)");
        let expected = std::f64::consts::PI / 4.0;
        assert_relative_eq!(l1, expected, epsilon = 1e-6);
    }

    #[test]
    fn test_bernoulli_numbers() {
        let b = bernoulli_numbers(6);
        // B_0=1, B_1=-1/2, B_2=1/6, B_3=0, B_4=-1/30
        assert_relative_eq!(b[0], 1.0, epsilon = 1e-12);
        assert_relative_eq!(b[1], -0.5, epsilon = 1e-12);
        assert_relative_eq!(b[2], 1.0 / 6.0, epsilon = 1e-12);
        assert_relative_eq!(b[3], 0.0, epsilon = 1e-12);
        assert_relative_eq!(b[4], -1.0 / 30.0, epsilon = 1e-12);
    }

    #[test]
    fn test_bernoulli_poly_b1() {
        // B_1(x) = x - 1/2
        let b1 = bernoulli_poly(1, 0.3);
        assert_relative_eq!(b1, 0.3 - 0.5, epsilon = 1e-12);
    }

    #[test]
    fn test_generalized_bernoulli_chi0() {
        // For the principal character mod 1 (trivial), B_{k,χ_0} = B_k
        let chi = DirichletCharacter::principal(1);
        let b2 = generalized_bernoulli(2, &chi);
        // B_{2, χ_0 mod 1} = 1^1 * chi(1)*B_2(1) = B_2(1) = 1/6 - 1 + 1 = 1/6
        assert!(b2.is_finite());
    }

    #[test]
    fn test_conductor_principal() {
        let chi = DirichletCharacter::principal(6);
        // Principal character mod 6 has conductor... let's just verify it divides 6
        let cond = chi.conductor();
        assert!(6 % cond == 0);
    }

    #[test]
    fn test_character_real() {
        let chi = DirichletCharacter::kronecker_symbol(-4);
        assert!(chi.is_real());
    }
}