siderust 0.6.0

High-precision astronomy and satellite mechanics in Rust.
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

// SPDX-License-Identifier: AGPL-3.0-or-later
// Copyright (C) 2026 Vallés Puig, Ramon

//! VSOP87 position / velocity computation (optimized version)
//!
//! Exposes three public helpers:
//! * [`position`]  – only X,Y,Z (AstronomicalUnits)
//! * [`velocity`]  – only Ẋ,Ẏ,Ż (AstronomicalUnits/day)
//! * [`position_velocity`] – both in one pass (≈30 % faster than 2 calls)
//!
//! ## Optimizations applied
//! - Separate specialized functions eliminate inner-loop branching
//! - `mul_add` for FMA precision and potential performance
//! - `#[inline(always)]` on hot paths
//! - SIMD batching via `wide` crate for sin/cos operations

use crate::time::JulianDate;
use wide::f64x4;

/// One VSOP87 coefficient term  _a · cos(b + c·T)_
#[derive(Debug, Clone, Copy)]
pub struct Vsop87 {
    pub a: f64,
    pub b: f64,
    pub c: f64,
}

/// Conversion factor: dT/dt (T in Julian millennia per day)
const DT_DT: f64 = 1.0 / 365_250.0;

// ═══════════════════════════════════════════════════════════════════════════
// SIMD helpers for batched sin/cos using wide crate
// ═══════════════════════════════════════════════════════════════════════════

/// Compute sin and cos for 4 values simultaneously using SIMD.
#[inline(always)]
fn sin_cos_x4(args: f64x4) -> (f64x4, f64x4) {
    // wide's sin/cos use polynomial approximations vectorized across lanes
    (args.sin(), args.cos())
}

// ═══════════════════════════════════════════════════════════════════════════
// Specialized coordinate functions (no branching in hot loops)
// ═══════════════════════════════════════════════════════════════════════════

/// Compute position value only (no derivative).
/// Uses SIMD batching for cos operations.
#[inline(always)]
fn coord_value(series_by_power: &[&[Vsop87]], t: f64) -> f64 {
    let mut t_pow = 1.0_f64;
    let mut value = 0.0_f64;

    for terms in series_by_power.iter() {
        let mut serie_val = 0.0_f64;
        let len = terms.len();

        // Process 4 terms at a time with SIMD
        let chunks = len / 4;
        let remainder = len % 4;

        for i in 0..chunks {
            let base = i * 4;
            let t0 = &terms[base];
            let t1 = &terms[base + 1];
            let t2 = &terms[base + 2];
            let t3 = &terms[base + 3];

            // Compute arguments using mul_add: arg = c * t + b
            let args = f64x4::new([
                t0.c.mul_add(t, t0.b),
                t1.c.mul_add(t, t1.b),
                t2.c.mul_add(t, t2.b),
                t3.c.mul_add(t, t3.b),
            ]);

            let cos_args = args.cos();
            let cos_arr = cos_args.to_array();

            // Accumulate: a * cos(arg)
            serie_val += t0.a * cos_arr[0];
            serie_val += t1.a * cos_arr[1];
            serie_val += t2.a * cos_arr[2];
            serie_val += t3.a * cos_arr[3];
        }

        // Handle remaining terms (0-3) with scalar operations
        for term in &terms[chunks * 4..chunks * 4 + remainder] {
            let arg = term.c.mul_add(t, term.b);
            serie_val += term.a * arg.cos();
        }

        value = t_pow.mul_add(serie_val, value);
        t_pow *= t;
    }

    value
}

/// Compute derivative only (no position value).
/// Uses SIMD batching for sin_cos operations.
#[inline(always)]
fn coord_deriv(series_by_power: &[&[Vsop87]], t: f64) -> f64 {
    let mut t_pow = 1.0_f64;
    let mut t_pow_der = 0.0_f64;
    let mut deriv_t = 0.0_f64;

    for (k, terms) in series_by_power.iter().enumerate() {
        let mut serie_val = 0.0_f64;
        let mut serie_der = 0.0_f64;
        let len = terms.len();

        // Process 4 terms at a time with SIMD
        let chunks = len / 4;
        let remainder = len % 4;

        for i in 0..chunks {
            let base = i * 4;
            let t0 = &terms[base];
            let t1 = &terms[base + 1];
            let t2 = &terms[base + 2];
            let t3 = &terms[base + 3];

            let args = f64x4::new([
                t0.c.mul_add(t, t0.b),
                t1.c.mul_add(t, t1.b),
                t2.c.mul_add(t, t2.b),
                t3.c.mul_add(t, t3.b),
            ]);

            let (sin_args, cos_args) = sin_cos_x4(args);
            let sin_arr = sin_args.to_array();
            let cos_arr = cos_args.to_array();

            // serie_val += a * cos(arg)
            // serie_der += -a * c * sin(arg)
            serie_val += t0.a * cos_arr[0];
            serie_val += t1.a * cos_arr[1];
            serie_val += t2.a * cos_arr[2];
            serie_val += t3.a * cos_arr[3];

            serie_der -= t0.a * t0.c * sin_arr[0];
            serie_der -= t1.a * t1.c * sin_arr[1];
            serie_der -= t2.a * t2.c * sin_arr[2];
            serie_der -= t3.a * t3.c * sin_arr[3];
        }

        // Handle remaining terms with scalar operations
        for term in &terms[chunks * 4..chunks * 4 + remainder] {
            let arg = term.c.mul_add(t, term.b);
            let (sin_arg, cos_arg) = arg.sin_cos();
            serie_val += term.a * cos_arg;
            serie_der -= term.a * term.c * sin_arg;
        }

        deriv_t = t_pow.mul_add(serie_der, deriv_t);
        deriv_t = t_pow_der.mul_add(serie_val, deriv_t);

        t_pow_der = (k as f64 + 1.0) * t_pow;
        t_pow *= t;
    }

    deriv_t * DT_DT
}

/// Compute both position value and derivative in one pass.
/// Uses SIMD batching for sin_cos operations.
#[inline(always)]
fn coord_both(series_by_power: &[&[Vsop87]], t: f64) -> (f64, f64) {
    let mut t_pow = 1.0_f64;
    let mut t_pow_der = 0.0_f64;
    let mut value = 0.0_f64;
    let mut deriv_t = 0.0_f64;

    for (k, terms) in series_by_power.iter().enumerate() {
        let mut serie_val = 0.0_f64;
        let mut serie_der = 0.0_f64;
        let len = terms.len();

        // Process 4 terms at a time with SIMD
        let chunks = len / 4;
        let remainder = len % 4;

        for i in 0..chunks {
            let base = i * 4;
            let t0 = &terms[base];
            let t1 = &terms[base + 1];
            let t2 = &terms[base + 2];
            let t3 = &terms[base + 3];

            let args = f64x4::new([
                t0.c.mul_add(t, t0.b),
                t1.c.mul_add(t, t1.b),
                t2.c.mul_add(t, t2.b),
                t3.c.mul_add(t, t3.b),
            ]);

            let (sin_args, cos_args) = sin_cos_x4(args);
            let sin_arr = sin_args.to_array();
            let cos_arr = cos_args.to_array();

            serie_val += t0.a * cos_arr[0];
            serie_val += t1.a * cos_arr[1];
            serie_val += t2.a * cos_arr[2];
            serie_val += t3.a * cos_arr[3];

            serie_der -= t0.a * t0.c * sin_arr[0];
            serie_der -= t1.a * t1.c * sin_arr[1];
            serie_der -= t2.a * t2.c * sin_arr[2];
            serie_der -= t3.a * t3.c * sin_arr[3];
        }

        // Handle remaining terms with scalar operations
        for term in &terms[chunks * 4..chunks * 4 + remainder] {
            let arg = term.c.mul_add(t, term.b);
            let (sin_arg, cos_arg) = arg.sin_cos();
            serie_val += term.a * cos_arg;
            serie_der -= term.a * term.c * sin_arg;
        }

        value = t_pow.mul_add(serie_val, value);
        deriv_t = t_pow.mul_add(serie_der, deriv_t);
        deriv_t = t_pow_der.mul_add(serie_val, deriv_t);

        t_pow_der = (k as f64 + 1.0) * t_pow;
        t_pow *= t;
    }

    (value, deriv_t * DT_DT)
}

// ═══════════════════════════════════════════════════════════════════════════
// Public façade
// ═══════════════════════════════════════════════════════════════════════════

pub fn position(
    jd: JulianDate,
    x_series: &[&[Vsop87]],
    y_series: &[&[Vsop87]],
    z_series: &[&[Vsop87]],
) -> (f64, f64, f64) {
    let t = JulianDate::tt_to_tdb(jd).julian_millennias().value();

    let x = coord_value(x_series, t);
    let y = coord_value(y_series, t);
    let z = coord_value(z_series, t);

    (x, y, z)
}

pub fn velocity(
    jd: JulianDate,
    x_series: &[&[Vsop87]],
    y_series: &[&[Vsop87]],
    z_series: &[&[Vsop87]],
) -> (f64, f64, f64) {
    let t = JulianDate::tt_to_tdb(jd).julian_millennias().value();

    let xdot = coord_deriv(x_series, t);
    let ydot = coord_deriv(y_series, t);
    let zdot = coord_deriv(z_series, t);

    (xdot, ydot, zdot)
}

/// Position **and** velocity in a single pass (≈30 % faster than calling
/// the two previous helpers).
pub fn position_velocity(
    jd: JulianDate,
    x_series: &[&[Vsop87]],
    y_series: &[&[Vsop87]],
    z_series: &[&[Vsop87]],
) -> ((f64, f64, f64), (f64, f64, f64)) {
    let t = JulianDate::tt_to_tdb(jd).julian_millennias().value();

    let (x, xdot) = coord_both(x_series, t);
    let (y, ydot) = coord_both(y_series, t);
    let (z, zdot) = coord_both(z_series, t);

    ((x, y, z), (xdot, ydot, zdot))
}

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

    const X0: [Vsop87; 1] = [Vsop87 {
        a: 1.0,
        b: 0.0,
        c: 0.0,
    }];
    const X1: [Vsop87; 1] = [Vsop87 {
        a: 2.0,
        b: 0.0,
        c: 0.0,
    }];
    const Y0: [Vsop87; 1] = [Vsop87 {
        a: 0.0,
        b: 0.0,
        c: 0.0,
    }];
    const Y1: [Vsop87; 0] = [];
    const Y2: [Vsop87; 1] = [Vsop87 {
        a: 3.0,
        b: 0.0,
        c: 0.0,
    }];
    const Z0: [Vsop87; 1] = [Vsop87 {
        a: 4.0,
        b: 0.0,
        c: 0.0,
    }];

    #[test]
    fn test_position_velocity() {
        // 100 years after J2000 -> T = 0.1
        let jd = JulianDate::new(2451545.0 + 36525.0);
        let pos = position(jd, &[&X0, &X1], &[&Y0, &Y1, &Y2], &[&Z0]);
        assert!((pos.0 - 1.2).abs() < 1e-12);
        assert!((pos.1 - 0.03).abs() < 1e-12);
        assert!((pos.2 - 4.0).abs() < 1e-12);

        let vel = velocity(jd, &[&X0, &X1], &[&Y0, &Y1, &Y2], &[&Z0]);
        let dt = 1.0 / 365_250.0;
        assert!((vel.0 - 2.0 * dt).abs() < 1e-12);
        assert!((vel.1 - 0.6 * dt).abs() < 1e-12);
        assert!((vel.2 - 0.0).abs() < 1e-12);

        let (p2, v2) = position_velocity(jd, &[&X0, &X1], &[&Y0, &Y1, &Y2], &[&Z0]);
        assert!((p2.0 - pos.0).abs() < 1e-12);
        assert!((p2.1 - pos.1).abs() < 1e-12);
        assert!((p2.2 - pos.2).abs() < 1e-12);
        assert!((v2.0 - vel.0).abs() < 1e-12);
        assert!((v2.1 - vel.1).abs() < 1e-12);
        assert!((v2.2 - vel.2).abs() < 1e-12);
    }
}