adic 0.5.1

Arithmetic and rootfinding for p-adic numbers
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
use std::{
    iter::{once, repeat},
    ops,
};
use num::traits::{Inv, Pow};
use itertools::Itertools;

use crate::{
    divisible::{Composite, Divisible, Prime},
    error::{AdicError, AdicResult},
    normed::Valuation,
    traits::{AdicPrimitive, CanApproximate, HasApproximateDigits, HasDigits, PrimedFrom},
    ZAdic,
};
use super::PowAdic;


#[derive(Debug, Clone, PartialEq, Eq)]
/// Adic composite digits that can be added and multiplied
pub struct MAdicDigits {
    b: Composite,
    d: Vec<u32>,
}


#[allow(dead_code)]
impl MAdicDigits {

    pub fn new(b: Composite, d: Vec<u32>) -> Self {
        MAdicDigits { b, d }
    }

    pub fn from_p_adic<A>(b: &Composite, ap: PowAdic<A>) -> AdicResult<Self>
    where A: AdicPrimitive, PowAdic<A>: HasApproximateDigits<DigitIndex = usize> {

        let p = ap.p();
        if !b.has_prime(p) {
            return Err(AdicError::IllDefined(format!(
                "MAdicDigits composite needs to include the prime: {} not in {}",
                ap.p(), b
            )));
        }

        let b_pow = b.prime_power(p);
        let matched_ap = PowAdic::new(ap.adic, b_pow.power());

        let Valuation::Finite(certainty) = matched_ap.certainty() else {
            return Err(AdicError::InappropriatePrecision(
                "No adic number uncertainty found; infinite n-adic digits".to_string()
            ));
        };

        // Let's use Horner's method to convert from radix p to radix n
        // 1213_6 = (( 1*6 + 2 )*6 + 1)*6 + 3 => 1*6 + 2 = 10)*6 = 60) + 1) = 61)*6 = 446) + 3 = 451_8
        //
        // Note: There's some efficiency to be gained here, e.g. avoiding cloning the Composite.
        // Consider doing the digit calculation manually, outside of MAdicDigits.
        let old_base = Self::new(b.clone(), vec![b_pow.value32()]);
        let mut nadic_ap = Self::zero(b.clone(), certainty);
        for d in matched_ap.digits().collect::<Vec<_>>().into_iter().rev() {
            nadic_ap = nadic_ap * old_base.clone() + Self::new(b.clone(), vec![d]);
        }

        // Multiply by the correct idempotent
        // This is the product of all prime idempotents EXCEPT the one for p
        let all_but_p_idempotent = Self::idempotent_excluding(p, b, certainty)?;

        Ok((nadic_ap * all_but_p_idempotent).into_truncation(certainty))

    }


    pub fn base(&self) -> &Composite {
        &self.b
    }

    // TODO: Impl HasDigits
    pub fn digits(&self) -> impl Iterator<Item=u32> + use<'_> {
        self.d.iter().copied()
    }

    pub fn len(&self) -> usize {
        self.d.len()
    }

    pub fn truncation(&self, n: usize) -> Self {
        self.clone().into_truncation(n)
    }

    pub fn into_truncation(self, n: usize) -> Self {
        let mut d = self.d;
        d.truncate(n);
        Self::new(self.b, d)
    }

    pub fn zero(b: Composite, prec: usize) -> Self {
        MAdicDigits { b, d: vec![0; prec] }
    }

    pub fn one(b: Composite, prec: usize) -> Self {
        MAdicDigits { b, d: once(1).chain(repeat(0)).take(prec).collect() }
    }


    // There are idempotents in multi-adics, zero-like and one-like numbers.
    // These satisfy the property T^2 = T.
    // For pure p-adics or p^n-adics, only 0 and 1 satisfy these.
    // For mixed n-adics, there are more.

    // For multi-adics, you can convert down from n-adic to p-adic, a surjective (onto) function.
    // E.g. the 10-adic 523._10 converts to 3 + 2*10 + 5*100 = 3 + 4*5 + 4*5^3 = 4043._5.
    // But again, this is SURJECTIVE, and e.g. multiple numbers convert to 0.
    // In particular, the number T_5 = lim_n->inf 5^(2^n) is a valid 10-adic but converts to the 5-adic 0.
    // This particular number also converts to the 2-adic 1, since 5^(2^n) -> 1 mod 2^m (Fermat/Euler?)
    // And subtracting T_5 from 1 gives the other idempotent: T_2 = 1 - T_5.

    // T_5 = ...59918212890625_10 -> 0._5 & 1._2
    // T_2 = ...40081787109376_10 -> 1._5 & 0._2

    // Note that these have the properties: T_2 * T_5 = 0; T_2 + T_5 = 1

    // In general, we want numbers that behave like 0 for one p-adic and 1 for the others.
    // This way, we get the idempotents:
    // - 0, 1
    // - T_p, for distinct prime p in n
    // - \prod_p T_p excluding the full product, since that gives 0

    // Or perhaps another way to think of it, the product of any number of T_p from the primes:
    // - The empty product gives 1
    // - The single product gives each "prime idempotent" separately, T_p
    // - The composite products multiply the prime idempotents once each, T_p1 * T_p2 * T_p3
    // - The full product gives 0

    // So we just need to calculate the prime idempotents and then multiply them together judiciously.
    // The prime idempotents should just be lim_n->inf (p1)^((p2*p3*...)^n)

    // Or more properly, if the adic has base b = p0^k0 p1^k1 p2^k2 ...
    // T_p0 = lim_N->inf (p0^k0)^carmichael( (p1^k1 p2^k2 ...)^N )


    /// The product of all idempotents except `T_p ~ p^carmichael((n / p^k)^infinity)`.
    pub fn idempotent_excluding<P>(p: P, c: &Composite, precision: usize) -> AdicResult<MAdicDigits>
    where P: Into<Prime> {

        // Actually, carmichael modpow is pretty inefficient.
        // Instead, here's a decently efficient algorithm using p-adic (b/p)^-n * (b/p)^n

        let p = p.into();
        let p_pow = c.prime_power(p).power();
        if p_pow == 0 {
            return Err(AdicError::IllDefined(format!("Prime {p} does not exist in composite {c}")));
        }
        let pow_prec = precision * usize::try_from(p_pow)?;

        let bval = c.value32();
        let b_wo_pp = Composite::new(c.prime_powers().filter(|qq| qq.p() != p)).value32();

        // Idea:
        //  Calculate the p-adic (b/p)^(-precision)
        //  Multiply (b/p) back in as an MAdic

        // z = (b/p)
        let z = ZAdic::primed_from(p, b_wo_pp).into_approximation(pow_prec);
        // r = p-adic z^(-1)
        let r = z.inv();
        // rpow = p-power-adic z^(-precision)
        let rpow = r.pow(precision.try_into()?);
        let rpow = PowAdic::new(rpow, p_pow);

        let mut a = Vec::new();
        for d in rpow.digits() {
            a.push(d);
            let mut carry = 0;
            for da in &mut a {
                let new_d = b_wo_pp * (*da) + carry;
                carry = new_d / bval;
                *da = new_d % bval;
            }
        }

        Ok(Self::new(c.clone(), a))

    }

    /// The prime idempotent `T_p ~ p^carmichael((n / p^k)^infinity)`.
    pub fn prime_idempotent<P>(p: P, c: &Composite, precision: usize) -> AdicResult<MAdicDigits>
    where P: Into<Prime> {

        let p = p.into();
        if Composite::from(p) == *c {
            // If the Composite is the same as the single Prime, the idempotent is zero
            return Ok(Self::zero(c.clone(), precision));
        }

        // Prime idempotent is 1 - idempotent without prime
        let nadic = Self::idempotent_excluding(p, c, precision)?;
        let mut a = nadic.digits();
        let first = a.next();
        let Some(f) = first else {
            return Ok(Self::zero(c.clone(), 0));
        };
        if [0, 1].contains(&f) {
            return Err(AdicError::Severe("idempotent should not start with 0 or 1; something is wrong".to_string()));
        }

        let mut b = vec![];
        b.push(c.value32() + 1 - f);
        for d in a {
            b.push(c.m1() - d);
        }

        Ok(Self::new(c.clone(), b))

    }

}


impl ops::Add for MAdicDigits {
    type Output = Self;
    fn add(self, rhs: Self) -> Self::Output {

        // Add the like digits one-by-one
        // Then reduce each to the [0, base) range and handle the carry

        assert_eq!(self.b, rhs.b, "Cannot add MAdicDigits with different bases");
        let bval = self.b.value32();

        let la = self.d.len();
        let lb = rhs.d.len();

        let mut summed_digits = Vec::with_capacity(std::cmp::max(la, lb) + 1);
        summed_digits.extend(
            self.d.iter().copied()
                .zip_longest(rhs.d.iter().copied())
                .map(|lr| lr.reduce(|l, r| l+r))
                .collect::<Vec<_>>()
        );

        let mut carry = 0;
        for digit in &mut summed_digits {
            let bigger_digit = *digit + carry;
            *digit = bigger_digit % bval;
            carry = bigger_digit / bval;
        }
        while carry > 0 {
            summed_digits.push(carry % bval);
            carry = carry / bval;
        }

        MAdicDigits::new(self.b.clone(), summed_digits)

    }
}


impl ops::Mul for MAdicDigits {
    type Output = Self;
    fn mul(self, rhs: Self) -> Self::Output {

        // Turn b around and "drag it across" longer to create the multiplied digits one-by-one
        // Then reduce each to the [0, base) range and handle the carry

        assert_eq!(self.b, rhs.b, "Cannot mul MAdicDigits with different bases");
        let bval = self.b.value32();

        let la = self.d.len();
        let lb = rhs.d.len();
        if la * lb == 0 {
            return MAdicDigits { b: self.b.clone(), d: vec![] };
        }
        let lt = la + lb - 1;

        // Performance critical here!
        let a_digits = self.d.clone();
        let mut rev_b_digits = rhs.d.clone();
        rev_b_digits.reverse();
        let mut summed_digits = Vec::with_capacity(lt + 1);
        summed_digits.extend((0..lt).map(|digit_place| {
            let (a_skip, b_skip) = if (digit_place >= lb) {
                (digit_place + 1 - lb, 0)
            } else {
                (0, lb - digit_place - 1)
            };
            let mut d = 0;
            for (&ds, &dr) in a_digits[a_skip..].iter().zip(rev_b_digits[b_skip..].iter()) {
                d += ds * dr;
            }
            d
        }));

        let mut carry = 0;
        for digit in &mut summed_digits {
            let bigger_digit = *digit + carry;
            *digit = bigger_digit % bval;
            carry = bigger_digit / bval;
        }
        while carry > 0 {
            summed_digits.push(carry % bval);
            carry = carry / bval;
        }

        MAdicDigits::new(self.b.clone(), summed_digits)

    }
}



#[cfg(test)]
mod tests {

    use crate::divisible::Composite;

    use super::MAdicDigits;


    #[test]
    fn two_primes() {

        let b = 10;
        let c = Composite::try_from(b).unwrap();
        let prec = 10;

        let two_to_inf = MAdicDigits::new(c.clone(), vec![6, 7, 3, 9, 0, 1, 7, 8, 7, 1]);
        let five_to_inf = MAdicDigits::new(c.clone(), vec![5, 2, 6, 0, 9, 8, 2, 1, 2, 8]);
        assert_eq!(Ok(two_to_inf.clone()), MAdicDigits::prime_idempotent(2, &c, prec));
        assert_eq!(Ok(five_to_inf.clone()), MAdicDigits::prime_idempotent(5, &c, prec));
        assert_eq!(Ok(two_to_inf.clone()), MAdicDigits::idempotent_excluding(5, &c, prec));
        assert_eq!(Ok(five_to_inf.clone()), MAdicDigits::idempotent_excluding(2, &c, prec));

        let zero = MAdicDigits::zero(c.clone(), prec);
        assert_eq!(zero, (two_to_inf * five_to_inf).into_truncation(prec));

    }

    #[test]
    fn three_primes() {

        let b = 30;
        let c = Composite::try_from(b).unwrap();
        let prec = 10;

        let two_to_inf = MAdicDigits::new(c.clone(), vec![16, 22, 3, 28, 15, 14, 26, 14, 23, 2]);
        let three_to_inf = MAdicDigits::new(c.clone(), vec![21, 26, 28, 12, 5, 22, 29, 26, 27, 5]);
        let five_to_inf = MAdicDigits::new(c.clone(), vec![25, 10, 27, 18, 8, 23, 3, 18, 8, 21]);
        assert_eq!(Ok(two_to_inf.clone()), MAdicDigits::prime_idempotent(2, &c, prec));
        assert_eq!(Ok(three_to_inf.clone()), MAdicDigits::prime_idempotent(3, &c, prec));
        assert_eq!(Ok(five_to_inf.clone()), MAdicDigits::prime_idempotent(5, &c, prec));

        let six_to_inf = MAdicDigits::new(c.clone(), vec![6, 19, 2, 11, 21, 6, 26, 11, 21, 8]);
        let ten_to_inf = MAdicDigits::new(c.clone(), vec![10, 3, 1, 17, 24, 7, 0, 3, 2, 24]);
        let fifteen_to_inf = MAdicDigits::new(c.clone(), vec![15, 7, 26, 1, 14, 15, 3, 15, 6, 27]);
        assert_eq!(Ok(six_to_inf.clone()), MAdicDigits::idempotent_excluding(5, &c, prec));
        assert_eq!(six_to_inf, (two_to_inf.clone() * three_to_inf.clone()).into_truncation(prec));
        assert_eq!(Ok(ten_to_inf.clone()), MAdicDigits::idempotent_excluding(3, &c, prec));
        assert_eq!(ten_to_inf, (two_to_inf.clone() * five_to_inf.clone()).into_truncation(prec));
        assert_eq!(Ok(fifteen_to_inf.clone()), MAdicDigits::idempotent_excluding(2, &c, prec));
        assert_eq!(fifteen_to_inf, (three_to_inf.clone() * five_to_inf.clone()).into_truncation(prec));

        assert_eq!(
            MAdicDigits::one(c.clone(), prec),
            (
                MAdicDigits::prime_idempotent(2, &c, prec).unwrap() +
                MAdicDigits::idempotent_excluding(2, &c, prec).unwrap()
            ).into_truncation(prec)
        );
        assert_eq!(
            MAdicDigits::one(c.clone(), prec),
            (
                MAdicDigits::prime_idempotent(3, &c, prec).unwrap() +
                MAdicDigits::idempotent_excluding(3, &c, prec).unwrap()
            ).into_truncation(prec)
        );
        assert_eq!(
            MAdicDigits::one(c.clone(), prec),
            (
                MAdicDigits::prime_idempotent(5, &c, prec).unwrap() +
                MAdicDigits::idempotent_excluding(5, &c, prec).unwrap()
            ).into_truncation(prec)
        );

        assert_eq!(
            MAdicDigits::zero(c.clone(), prec),
            (two_to_inf * three_to_inf * five_to_inf).into_truncation(prec)
        );

    }

    #[test]
    fn prime_power() {

        let b = 20;
        let c = Composite::try_from(b).unwrap();
        let prec = 10;

        let two_to_inf = MAdicDigits::new(c.clone(), vec![16, 8, 13, 8, 9, 18, 17, 18, 6, 12]);
        let five_to_inf = MAdicDigits::new(c.clone(), vec![5, 11, 6, 11, 10, 1, 2, 1, 13, 7]);
        assert_eq!(Ok(two_to_inf), MAdicDigits::idempotent_excluding(5, &c, prec));
        assert_eq!(Ok(five_to_inf), MAdicDigits::idempotent_excluding(2, &c, prec));

    }

    #[test]
    fn big_idempotent() {
        let b = 10;
        let c = Composite::try_from(b).unwrap();
        assert_eq!(10, MAdicDigits::prime_idempotent(2, &c, 10).unwrap().len());
        assert_eq!(100, MAdicDigits::prime_idempotent(2, &c, 100).unwrap().len());
        assert_eq!(1000, MAdicDigits::prime_idempotent(2, &c, 1000).unwrap().len());
        // Too slow:
        // assert_eq!(10000, MAdicDigits::prime_idempotent(2, &c, 10000).unwrap().len());
    }

}