plotine-core 0.5.2

Core geometry, color, transforms, scales, and ticks for plotine
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
//! Tick location for linear / log / symlog axes.

use crate::scale::{LinearScale, LogScale, Scale, ScaleKind, SymlogScale};

/// A single major tick with its data value and formatted label.
#[derive(Debug, Clone, PartialEq)]
pub struct Tick {
    pub value: f64,
    pub label: String,
}

impl Tick {
    /// Format a data value with the same rules as the auto locator.
    pub fn from_value(value: f64) -> Self {
        Self {
            value,
            label: format_tick(value),
        }
    }
}

/// Locates aesthetically pleasing ticks inside a scale domain.
#[derive(Debug, Clone, Copy)]
pub struct TickLocator {
    pub target_count: usize,
}

impl Default for TickLocator {
    fn default() -> Self {
        Self { target_count: 6 }
    }
}

impl TickLocator {
    pub fn new(target_count: usize) -> Self {
        Self {
            target_count: target_count.max(2),
        }
    }

    pub fn ticks(&self, scale: ScaleKind) -> Vec<Tick> {
        match scale {
            ScaleKind::Linear(s) => self.ticks_linear(s),
            ScaleKind::Log(s) => self.ticks_log(s),
            ScaleKind::Symlog(s) => self.ticks_symlog(s),
        }
    }

    /// Compute major ticks for a linear scale using a simplified Wilkinson/nice-number approach.
    pub fn ticks_linear(&self, scale: LinearScale) -> Vec<Tick> {
        let (min, max) = scale.domain();
        let span = max - min;
        if !span.is_finite() || span <= 0.0 {
            return vec![Tick {
                value: min,
                label: format_tick(min),
            }];
        }

        // `target_count - 1` ≈ MaxNLocator `nbins` (max intervals). Use the
        // ceiling nice-number path so step is never smaller than `rough`
        // (round-to-nearest was producing 0.5 when nbins required ≥1.0).
        let rough = span / (self.target_count as f64 - 1.0).max(1.0);
        let mut step = nice_number(rough, false);
        if step <= 0.0 || !step.is_finite() {
            return vec![
                Tick {
                    value: min,
                    label: format_tick(min),
                },
                Tick {
                    value: max,
                    label: format_tick(max),
                },
            ];
        }

        let mut values = linear_tick_values(min, max, step);
        // MaxNLocator `min_n_ticks` (default 2): if the ceiling step left too
        // few in-domain ticks (tiny inset axes), step down the nice ladder
        // instead of labeling raw padded endpoints (`0.114` / `0.914`).
        const MIN_N_TICKS: usize = 2;
        let mut guard = 0;
        while values.len() < MIN_N_TICKS && guard < 8 {
            let smaller = next_smaller_nice(step);
            if smaller <= 0.0 || !smaller.is_finite() || smaller >= step * (1.0 - 1e-12) {
                break;
            }
            step = smaller;
            values = linear_tick_values(min, max, step);
            guard += 1;
        }

        if values.is_empty() {
            values.push(min);
            if (max - min).abs() > f64::EPSILON {
                values.push(max);
            }
        }

        let labels = format_aligned_ticks(&values, step);
        values
            .into_iter()
            .zip(labels)
            .map(|(value, label)| Tick { label, value })
            .collect()
    }

    pub fn ticks_log(&self, scale: LogScale) -> Vec<Tick> {
        let (min, max) = scale.domain();
        let mut values = Vec::new();
        let start_exp = min.log10().floor() as i32;
        let end_exp = max.log10().ceil() as i32;
        for exp in start_exp..=end_exp {
            let v = 10f64.powi(exp);
            if v >= min * (1.0 - 1e-12) && v <= max * (1.0 + 1e-12) {
                values.push(v);
            }
        }
        if values.is_empty() {
            values.push(min);
            if (max - min).abs() > f64::EPSILON * min {
                values.push(max);
            }
        }
        values
            .into_iter()
            .map(|value| Tick {
                label: format_log_tick(value),
                value,
            })
            .collect()
    }

    pub fn ticks_symlog(&self, scale: SymlogScale) -> Vec<Tick> {
        // Reuse linear locator on the data domain; good enough for M2.
        match LinearScale::new(scale.min(), scale.max()) {
            Ok(linear) => self.ticks_linear(linear),
            Err(_) => vec![Tick {
                value: 0.0,
                label: "0".into(),
            }],
        }
    }
}

fn linear_tick_values(min: f64, max: f64, step: f64) -> Vec<f64> {
    let start = (min / step).ceil() * step;
    let mut values = Vec::new();
    let mut v = start;
    let mut guard = 0;
    while v <= max + step * 1e-9 && guard < 64 {
        if v >= min - step * 1e-9 {
            // Keep nice step values — do not snap to raw domain endpoints
            // (that produced labels like `5.551e-17` / long floats).
            let mut tick_v = v;
            if tick_v.abs() < step * 1e-10 {
                tick_v = 0.0;
            }
            if values
                .last()
                .map(|last: &f64| (tick_v - *last).abs() > step * 1e-9)
                .unwrap_or(true)
            {
                values.push(tick_v);
            }
        }
        v += step;
        guard += 1;
    }
    values
}

/// Next-smaller AutoLocator step on the `[1, 2, 2.5, 5, 10]` ladder.
fn next_smaller_nice(step: f64) -> f64 {
    if step <= 0.0 || !step.is_finite() {
        return step;
    }
    let exp = step.log10().floor();
    let base = 10f64.powf(exp);
    let frac = step / base;
    let (next_frac, next_exp) = if (frac - 10.0).abs() < 1e-9 {
        (5.0, exp)
    } else if (frac - 5.0).abs() < 1e-9 {
        (2.5, exp)
    } else if (frac - 2.5).abs() < 1e-9 {
        (2.0, exp)
    } else if (frac - 2.0).abs() < 1e-9 {
        (1.0, exp)
    } else if (frac - 1.0).abs() < 1e-9 {
        (5.0, exp - 1.0)
    } else {
        // Non-canonical step: halve, then re-nice.
        return nice_number(step * 0.5, true);
    };
    next_frac * 10f64.powf(next_exp)
}

fn nice_number(x: f64, round: bool) -> f64 {
    if x <= 0.0 || !x.is_finite() {
        return 1.0;
    }
    let exp = x.log10().floor();
    let frac = x / 10f64.powf(exp);
    // `round=true` (step from rough span/nbins): AutoLocator-ish `[1, 2, 5, 10]`.
    // Thresholds bias toward *larger* steps (fewer ticks): rough=0.33 → 0.5,
    // rough=3 → 5 (not 2). Keep 2.5 only in the `round=false` ceiling path.
    let nice_frac = if round {
        if frac < 1.5 {
            1.0
        } else if frac < 2.25 {
            2.0
        } else if frac < 7.0 {
            5.0
        } else {
            10.0
        }
    } else if frac <= 1.0 {
        1.0
    } else if frac <= 2.0 {
        2.0
    } else if frac <= 2.5 {
        2.5
    } else if frac <= 5.0 {
        5.0
    } else {
        10.0
    };
    nice_frac * 10f64.powf(exp)
}

fn format_tick(value: f64) -> String {
    format_tick_with_step(value, 1.0)
}

/// Build ticks from explicit values with axis-wide decimal alignment.
pub fn ticks_from_values(values: &[f64]) -> Vec<Tick> {
    if values.is_empty() {
        return Vec::new();
    }
    let mut step = f64::INFINITY;
    for w in values.windows(2) {
        let d = (w[1] - w[0]).abs();
        if d.is_finite() && d > 0.0 {
            step = step.min(d);
        }
    }
    if !step.is_finite() {
        step = 1.0;
    }
    let labels = format_aligned_ticks(values, step);
    values
        .iter()
        .copied()
        .zip(labels)
        .map(|(value, label)| Tick { value, label })
        .collect()
}

/// Hard cap on fractional digits for auto tick labels (avoids float-noise
/// labels like `0.913580` on tiny inset axes).
pub const MAX_TICK_DECIMALS: usize = 3;

/// Format every tick on an axis to the same decimal width.
///
/// Decimals = max(step precision, fractional digits present in any value),
/// capped at [`MAX_TICK_DECIMALS`], so `1 / 0.75 / 0.5` becomes
/// `1.00 / 0.75 / 0.50`. Integer-only axes stay compact (`0`, `1`, `2`).
pub fn format_aligned_ticks(values: &[f64], step: f64) -> Vec<String> {
    let mut decimals = step_decimals(step);
    for &v in values {
        decimals = decimals.max(value_frac_decimals(v));
    }
    decimals = decimals.min(MAX_TICK_DECIMALS);
    values
        .iter()
        .map(|&v| format_tick_fixed(v, decimals))
        .collect()
}

/// Format a tick using the locator step (matplotlib `ScalarFormatter`-ish).
///
/// Fractional steps keep trailing decimals (`0.0`, `1.0` for step `0.5`);
/// integer steps stay compact (`0`, `1`, `2`).
fn format_tick_with_step(value: f64, step: f64) -> String {
    format_tick_fixed(value, step_decimals(step))
}

fn value_frac_decimals(value: f64) -> usize {
    if !value.is_finite() {
        return 0;
    }
    let v = if value.abs() < 1e-12 { 0.0 } else { value };
    if (v - v.round()).abs() < 1e-9 * v.abs().max(1.0) {
        return 0;
    }
    if !(1e-3..1e4).contains(&v.abs()) {
        return 0;
    }
    let s = format!("{v:.6}");
    let Some(frac) = s.split('.').nth(1) else {
        return 0;
    };
    frac.trim_end_matches('0').len().min(MAX_TICK_DECIMALS)
}

fn format_tick_fixed(value: f64, decimals: usize) -> String {
    if !value.is_finite() {
        return format!("{value}");
    }
    // Floating-point dust around zero must not become `5.551e-17`.
    if value == 0.0 || value.abs() < 1e-12 {
        return if decimals == 0 {
            "0".to_string()
        } else {
            format!("{:.*}", decimals, 0.0)
        };
    }
    let abs = value.abs();
    let body = if !(1e-3..1e4).contains(&abs) {
        // Keep sign in scientific form; rare on plot axes.
        return format!("{value:.3e}");
    } else if decimals > 0 {
        format!("{:.*}", decimals, abs)
    } else {
        let near_int = (value - value.round()).abs() < 1e-9 * abs.max(1.0);
        if near_int {
            format!("{:.0}", abs.round())
        } else if abs >= 100.0 {
            format!("{abs:.0}")
        } else if abs >= 10.0 {
            format!("{abs:.1}")
                .trim_end_matches('0')
                .trim_end_matches('.')
                .to_string()
        } else {
            format!("{abs:.2}")
                .trim_end_matches('0')
                .trim_end_matches('.')
                .to_string()
        }
    };
    // Matplotlib `ScalarFormatter` uses Unicode minus (U+2212).
    if value < 0.0 {
        format!("\u{2212}{body}")
    } else {
        body
    }
}

fn step_decimals(step: f64) -> usize {
    let s = step.abs();
    if !s.is_finite() || s >= 0.999 {
        return 0;
    }
    // Enough decimals for the step to format cleanly (0.5→1, 0.25→2).
    for d in 1..=MAX_TICK_DECIMALS {
        let scaled = s * 10f64.powi(d as i32);
        if (scaled - scaled.round()).abs() < 1e-8 * scaled.max(1.0) {
            return d;
        }
    }
    MAX_TICK_DECIMALS
}

fn format_log_tick(value: f64) -> String {
    if value <= 0.0 {
        return format_tick(value);
    }
    let exp = value.log10();
    if (exp - exp.round()).abs() < 1e-9 {
        // Matplotlib `LogFormatterMathtext` / `LogFormatterSciNotation`: `$10^{n}$`.
        let e = exp.round() as i32;
        format!("$10^{{{e}}}$")
    } else {
        format_tick(value)
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::scale::{LinearScale, LogScale, ScaleKind};
    use proptest::prelude::*;

    #[test]
    fn produces_ticks_inside_domain() {
        let scale = ScaleKind::Linear(LinearScale::new(0.0, 1.0).unwrap());
        let ticks = TickLocator::default().ticks(scale);
        assert!(ticks.len() >= 2);
        for t in &ticks {
            assert!(t.value >= -1e-9 && t.value <= 1.0 + 1e-9);
        }
    }

    #[test]
    fn formats_near_zero_as_zero() {
        assert_eq!(format_tick(0.0), "0");
        assert_eq!(format_tick(5.551115123125783e-17), "0");
        assert_eq!(format_tick(-1e-13), "0");
    }

    #[test]
    fn fractional_step_keeps_decimals() {
        assert_eq!(format_tick_with_step(0.0, 0.5), "0.0");
        assert_eq!(format_tick_with_step(1.0, 0.5), "1.0");
        assert_eq!(format_tick_with_step(0.5, 0.5), "0.5");
        assert_eq!(format_tick_with_step(0.0, 0.25), "0.00");
        assert_eq!(format_tick_with_step(0.25, 0.25), "0.25");
        assert_eq!(format_tick_with_step(-0.5, 0.5), "\u{2212}0.5");
    }

    #[test]
    fn aligned_decimals_pad_axis() {
        let labels = format_aligned_ticks(&[1.0, 0.75, 0.5, 0.25], 0.25);
        assert_eq!(labels, vec!["1.00", "0.75", "0.50", "0.25"]);
        let labels = format_aligned_ticks(&[0.0, 1.0, 2.0], 1.0);
        assert_eq!(labels, vec!["0", "1", "2"]);
        let ticks = ticks_from_values(&[1.0, 0.5, 0.25]);
        assert_eq!(
            ticks.iter().map(|t| t.label.as_str()).collect::<Vec<_>>(),
            vec!["1.00", "0.50", "0.25"]
        );
    }

    #[test]
    fn aligned_decimals_cap_at_three() {
        // Float-noise endpoints must not produce 6-digit inset labels.
        let labels = format_aligned_ticks(&[0.113639482, 0.913580247], 0.1);
        for lab in &labels {
            let frac = lab.split('.').nth(1).unwrap_or("");
            assert!(
                frac.len() <= MAX_TICK_DECIMALS,
                "label {lab} exceeds {MAX_TICK_DECIMALS} decimals"
            );
        }
    }

    #[test]
    fn rough_three_ceils_to_five() {
        // Ceiling path (used by TickLocator): never shrink below rough.
        assert!((nice_number(3.0, false) - 5.0).abs() < 1e-12);
        assert!((nice_number(0.33, false) - 0.5).abs() < 1e-12);
        assert!((nice_number(0.61, false) - 1.0).abs() < 1e-12);
    }

    #[test]
    fn colorbar_compressed_x_matches_mpl_maxn() {
        // Stock mpl + colorbar: get_tick_space ≈ 7–8 → targets≈8–9.
        // Flooring targets to 11 densified coolwarm to step 2.5 / "0.0" labels.
        let scale = ScaleKind::Linear(LinearScale::new(-0.5, 23.5).unwrap());
        let ticks = TickLocator::new(8).ticks(scale);
        let labels: Vec<&str> = ticks.iter().map(|t| t.label.as_str()).collect();
        assert_eq!(labels, ["0", "5", "10", "15", "20"]);

        // Short tripcolor span: mpl uses 0.5 (not 0.25) when axes are narrow.
        let scale = ScaleKind::Linear(LinearScale::new(-0.1, 2.1).unwrap());
        let ticks = TickLocator::new(8).ticks(scale);
        let step = ticks[1].value - ticks[0].value;
        assert!((step - 0.5).abs() < 1e-12, "expected step 0.5, got {step}");
        assert_eq!(ticks[0].label, "0.0");
    }

    #[test]
    fn locator_respects_max_intervals() {
        // span=5.5, nbins≈8 → targets=9 → must not land on step 0.5.
        let scale = ScaleKind::Linear(LinearScale::new(0.0, 5.5).unwrap());
        let ticks = TickLocator::new(9).ticks(scale);
        let step = ticks[1].value - ticks[0].value;
        assert!(
            step >= 0.99,
            "expected step≥1, got {step} from {} ticks",
            ticks.len()
        );
    }

    #[test]
    fn tiny_inset_span_uses_nice_steps_not_endpoints() {
        // Nested inset y after 5% margins — ceiling step 1.0 would be empty.
        let scale = ScaleKind::Linear(LinearScale::new(0.113639482, 0.913580247).unwrap());
        let ticks = TickLocator::new(2).ticks(scale);
        assert!(ticks.len() >= 2);
        let vals: Vec<f64> = ticks.iter().map(|t| t.value).collect();
        assert!(
            vals.iter().any(|v| (*v - 0.5).abs() < 1e-9),
            "expected a 0.5 tick, got {vals:?}"
        );
        for t in &ticks {
            assert!(
                !t.label.contains("0.114") && !t.label.contains("0.913"),
                "raw endpoint label {}",
                t.label
            );
        }
    }

    #[test]
    fn log_ticks_are_powers_of_ten() {
        let scale = ScaleKind::Log(LogScale::new(1.0, 1000.0).unwrap());
        let ticks = TickLocator::default().ticks(scale);
        assert!(ticks.iter().any(|t| (t.value - 1.0).abs() < 1e-12));
        assert!(ticks.iter().any(|t| (t.value - 10.0).abs() < 1e-12));
        assert!(ticks.iter().any(|t| (t.value - 100.0).abs() < 1e-12));
        assert!(ticks.iter().any(|t| (t.value - 1000.0).abs() < 1e-12));
        let ten = ticks
            .iter()
            .find(|t| (t.value - 10.0).abs() < 1e-12)
            .unwrap();
        assert_eq!(ten.label, "$10^{1}$");
        let tenth = TickLocator::default().ticks(ScaleKind::Log(LogScale::new(0.1, 10.0).unwrap()));
        let lab = tenth
            .iter()
            .find(|t| (t.value - 0.1).abs() < 1e-12)
            .unwrap();
        assert_eq!(lab.label, "$10^{-1}$");
    }

    #[test]
    fn handles_large_range() {
        let scale = ScaleKind::Linear(LinearScale::new(0.0, 1e6).unwrap());
        let ticks = TickLocator::new(5).ticks(scale);
        assert!(ticks.len() >= 2);
        assert!(ticks.len() <= 12);
    }

    proptest! {
        #[test]
        fn ticks_never_panic(a in -1e6f64..1e6, b in -1e6f64..1e6) {
            prop_assume!((a - b).abs() > 1e-6);
            let (min, max) = if a < b { (a, b) } else { (b, a) };
            let scale = ScaleKind::Linear(LinearScale::new(min, max).unwrap());
            let ticks = TickLocator::default().ticks(scale);
            prop_assert!(!ticks.is_empty());
            prop_assert!(ticks.len() <= 64);
        }
    }
}