emath/
lib.rs

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
//! Opinionated 2D math library for building GUIs.
//!
//! Includes vectors, positions, rectangles etc.
//!
//! Conventions (unless otherwise specified):
//!
//! * All angles are in radians
//! * X+ is right and Y+ is down.
//! * (0,0) is left top.
//! * Dimension order is always `x y`
//!
//! ## Integrating with other math libraries.
//! `emath` does not strive to become a general purpose or all-powerful math library.
//!
//! For that, use something else ([`glam`](https://docs.rs/glam), [`nalgebra`](https://docs.rs/nalgebra), …)
//! and enable the `mint` feature flag in `emath` to enable implicit conversion to/from `emath`.
//!
//! ## Feature flags
#![cfg_attr(feature = "document-features", doc = document_features::document_features!())]
//!

#![allow(clippy::float_cmp)]

use std::ops::{Add, Div, Mul, RangeInclusive, Sub};

// ----------------------------------------------------------------------------

pub mod align;
pub mod easing;
mod history;
mod numeric;
mod ordered_float;
mod pos2;
mod range;
mod rect;
mod rect_transform;
mod rot2;
pub mod smart_aim;
mod ts_transform;
mod vec2;
mod vec2b;

pub use self::{
    align::{Align, Align2},
    history::History,
    numeric::*,
    ordered_float::*,
    pos2::*,
    range::Rangef,
    rect::*,
    rect_transform::*,
    rot2::*,
    ts_transform::*,
    vec2::*,
    vec2b::*,
};

// ----------------------------------------------------------------------------

/// Helper trait to implement [`lerp`] and [`remap`].
pub trait One {
    const ONE: Self;
}

impl One for f32 {
    const ONE: Self = 1.0;
}

impl One for f64 {
    const ONE: Self = 1.0;
}

/// Helper trait to implement [`lerp`] and [`remap`].
pub trait Real:
    Copy
    + PartialEq
    + PartialOrd
    + One
    + Add<Self, Output = Self>
    + Sub<Self, Output = Self>
    + Mul<Self, Output = Self>
    + Div<Self, Output = Self>
{
}

impl Real for f32 {}

impl Real for f64 {}

// ----------------------------------------------------------------------------

/// Linear interpolation.
///
/// ```
/// # use emath::lerp;
/// assert_eq!(lerp(1.0..=5.0, 0.0), 1.0);
/// assert_eq!(lerp(1.0..=5.0, 0.5), 3.0);
/// assert_eq!(lerp(1.0..=5.0, 1.0), 5.0);
/// assert_eq!(lerp(1.0..=5.0, 2.0), 9.0);
/// ```
#[inline(always)]
pub fn lerp<R, T>(range: impl Into<RangeInclusive<R>>, t: T) -> R
where
    T: Real + Mul<R, Output = R>,
    R: Copy + Add<R, Output = R>,
{
    let range = range.into();
    (T::ONE - t) * *range.start() + t * *range.end()
}

/// Where in the range is this value? Returns 0-1 if within the range.
///
/// Returns <0 if before and >1 if after.
///
/// Returns `None` if the input range is zero-width.
///
/// ```
/// # use emath::inverse_lerp;
/// assert_eq!(inverse_lerp(1.0..=5.0, 1.0), Some(0.0));
/// assert_eq!(inverse_lerp(1.0..=5.0, 3.0), Some(0.5));
/// assert_eq!(inverse_lerp(1.0..=5.0, 5.0), Some(1.0));
/// assert_eq!(inverse_lerp(1.0..=5.0, 9.0), Some(2.0));
/// assert_eq!(inverse_lerp(1.0..=1.0, 3.0), None);
/// ```
#[inline]
pub fn inverse_lerp<R>(range: RangeInclusive<R>, value: R) -> Option<R>
where
    R: Copy + PartialEq + Sub<R, Output = R> + Div<R, Output = R>,
{
    let min = *range.start();
    let max = *range.end();
    if min == max {
        None
    } else {
        Some((value - min) / (max - min))
    }
}

/// Linearly remap a value from one range to another,
/// so that when `x == from.start()` returns `to.start()`
/// and when `x == from.end()` returns `to.end()`.
pub fn remap<T>(x: T, from: impl Into<RangeInclusive<T>>, to: impl Into<RangeInclusive<T>>) -> T
where
    T: Real,
{
    let from = from.into();
    let to = to.into();
    debug_assert!(from.start() != from.end());
    let t = (x - *from.start()) / (*from.end() - *from.start());
    lerp(to, t)
}

/// Like [`remap`], but also clamps the value so that the returned value is always in the `to` range.
pub fn remap_clamp<T>(
    x: T,
    from: impl Into<RangeInclusive<T>>,
    to: impl Into<RangeInclusive<T>>,
) -> T
where
    T: Real,
{
    let from = from.into();
    let to = to.into();
    if from.end() < from.start() {
        return remap_clamp(x, *from.end()..=*from.start(), *to.end()..=*to.start());
    }
    if x <= *from.start() {
        *to.start()
    } else if *from.end() <= x {
        *to.end()
    } else {
        debug_assert!(from.start() != from.end());
        let t = (x - *from.start()) / (*from.end() - *from.start());
        // Ensure no numerical inaccuracies sneak in:
        if T::ONE <= t {
            *to.end()
        } else {
            lerp(to, t)
        }
    }
}

/// Round a value to the given number of decimal places.
pub fn round_to_decimals(value: f64, decimal_places: usize) -> f64 {
    // This is a stupid way of doing this, but stupid works.
    format!("{value:.decimal_places$}").parse().unwrap_or(value)
}

pub fn format_with_minimum_decimals(value: f64, decimals: usize) -> String {
    format_with_decimals_in_range(value, decimals..=6)
}

/// Use as few decimals as possible to show the value accurately, but within the given range.
///
/// Decimals are counted after the decimal point.
pub fn format_with_decimals_in_range(value: f64, decimal_range: RangeInclusive<usize>) -> String {
    let min_decimals = *decimal_range.start();
    let max_decimals = *decimal_range.end();
    debug_assert!(min_decimals <= max_decimals);
    debug_assert!(max_decimals < 100);
    let max_decimals = max_decimals.min(16);
    let min_decimals = min_decimals.min(max_decimals);

    if min_decimals < max_decimals {
        // Ugly/slow way of doing this. TODO(emilk): clean up precision.
        for decimals in min_decimals..max_decimals {
            let text = format!("{value:.decimals$}");
            let epsilon = 16.0 * f32::EPSILON; // margin large enough to handle most peoples round-tripping needs
            if almost_equal(text.parse::<f32>().unwrap(), value as f32, epsilon) {
                // Enough precision to show the value accurately - good!
                return text;
            }
        }
        // The value has more precision than we expected.
        // Probably the value was set not by the slider, but from outside.
        // In any case: show the full value
    }
    format!("{value:.max_decimals$}")
}

/// Return true when arguments are the same within some rounding error.
///
/// For instance `almost_equal(x, x.to_degrees().to_radians(), f32::EPSILON)` should hold true for all x.
/// The `epsilon`  can be `f32::EPSILON` to handle simple transforms (like degrees -> radians)
/// but should be higher to handle more complex transformations.
pub fn almost_equal(a: f32, b: f32, epsilon: f32) -> bool {
    if a == b {
        true // handle infinites
    } else {
        let abs_max = a.abs().max(b.abs());
        abs_max <= epsilon || ((a - b).abs() / abs_max) <= epsilon
    }
}

#[allow(clippy::approx_constant)]
#[test]
fn test_format() {
    assert_eq!(format_with_minimum_decimals(1_234_567.0, 0), "1234567");
    assert_eq!(format_with_minimum_decimals(1_234_567.0, 1), "1234567.0");
    assert_eq!(format_with_minimum_decimals(3.14, 2), "3.14");
    assert_eq!(format_with_minimum_decimals(3.14, 3), "3.140");
    assert_eq!(
        format_with_minimum_decimals(std::f64::consts::PI, 2),
        "3.14159"
    );
}

#[test]
fn test_almost_equal() {
    for &x in &[
        0.0_f32,
        f32::MIN_POSITIVE,
        1e-20,
        1e-10,
        f32::EPSILON,
        0.1,
        0.99,
        1.0,
        1.001,
        1e10,
        f32::MAX / 100.0,
        // f32::MAX, // overflows in rad<->deg test
        f32::INFINITY,
    ] {
        for &x in &[-x, x] {
            for roundtrip in &[
                |x: f32| x.to_degrees().to_radians(),
                |x: f32| x.to_radians().to_degrees(),
            ] {
                let epsilon = f32::EPSILON;
                assert!(
                    almost_equal(x, roundtrip(x), epsilon),
                    "{} vs {}",
                    x,
                    roundtrip(x)
                );
            }
        }
    }
}

#[test]
fn test_remap() {
    assert_eq!(remap_clamp(1.0, 0.0..=1.0, 0.0..=16.0), 16.0);
    assert_eq!(remap_clamp(1.0, 1.0..=0.0, 16.0..=0.0), 16.0);
    assert_eq!(remap_clamp(0.5, 1.0..=0.0, 16.0..=0.0), 8.0);
}

// ----------------------------------------------------------------------------

/// Extends `f32`, [`Vec2`] etc with `at_least` and `at_most` as aliases for `max` and `min`.
pub trait NumExt {
    /// More readable version of `self.max(lower_limit)`
    #[must_use]
    fn at_least(self, lower_limit: Self) -> Self;

    /// More readable version of `self.min(upper_limit)`
    #[must_use]
    fn at_most(self, upper_limit: Self) -> Self;
}

macro_rules! impl_num_ext {
    ($t: ty) => {
        impl NumExt for $t {
            #[inline(always)]
            fn at_least(self, lower_limit: Self) -> Self {
                self.max(lower_limit)
            }

            #[inline(always)]
            fn at_most(self, upper_limit: Self) -> Self {
                self.min(upper_limit)
            }
        }
    };
}

impl_num_ext!(u8);
impl_num_ext!(u16);
impl_num_ext!(u32);
impl_num_ext!(u64);
impl_num_ext!(u128);
impl_num_ext!(usize);
impl_num_ext!(i8);
impl_num_ext!(i16);
impl_num_ext!(i32);
impl_num_ext!(i64);
impl_num_ext!(i128);
impl_num_ext!(isize);
impl_num_ext!(f32);
impl_num_ext!(f64);
impl_num_ext!(Vec2);
impl_num_ext!(Pos2);

// ----------------------------------------------------------------------------

/// Wrap angle to `[-PI, PI]` range.
pub fn normalized_angle(mut angle: f32) -> f32 {
    use std::f32::consts::{PI, TAU};
    angle %= TAU;
    if angle > PI {
        angle -= TAU;
    } else if angle < -PI {
        angle += TAU;
    }
    angle
}

#[test]
fn test_normalized_angle() {
    macro_rules! almost_eq {
        ($left: expr, $right: expr) => {
            let left = $left;
            let right = $right;
            assert!((left - right).abs() < 1e-6, "{} != {}", left, right);
        };
    }

    use std::f32::consts::TAU;
    almost_eq!(normalized_angle(-3.0 * TAU), 0.0);
    almost_eq!(normalized_angle(-2.3 * TAU), -0.3 * TAU);
    almost_eq!(normalized_angle(-TAU), 0.0);
    almost_eq!(normalized_angle(0.0), 0.0);
    almost_eq!(normalized_angle(TAU), 0.0);
    almost_eq!(normalized_angle(2.7 * TAU), -0.3 * TAU);
}

// ----------------------------------------------------------------------------

/// Calculate a lerp-factor for exponential smoothing using a time step.
///
/// * `exponential_smooth_factor(0.90, 1.0, dt)`: reach 90% in 1.0 seconds
/// * `exponential_smooth_factor(0.50, 0.2, dt)`: reach 50% in 0.2 seconds
///
/// Example:
/// ```
/// # use emath::{lerp, exponential_smooth_factor};
/// # let (mut smoothed_value, target_value, dt) = (0.0_f32, 1.0_f32, 0.01_f32);
/// let t = exponential_smooth_factor(0.90, 0.2, dt); // reach 90% in 0.2 seconds
/// smoothed_value = lerp(smoothed_value..=target_value, t);
/// ```
pub fn exponential_smooth_factor(
    reach_this_fraction: f32,
    in_this_many_seconds: f32,
    dt: f32,
) -> f32 {
    1.0 - (1.0 - reach_this_fraction).powf(dt / in_this_many_seconds)
}

/// If you have a value animating over time,
/// how much towards its target do you need to move it this frame?
///
/// You only need to store the start time and target value in order to animate using this function.
///
/// ``` rs
/// struct Animation {
///     current_value: f32,
///
///     animation_time_span: (f64, f64),
///     target_value: f32,
/// }
///
/// impl Animation {
///     fn update(&mut self, now: f64, dt: f32) {
///         let t = interpolation_factor(self.animation_time_span, now, dt, ease_in_ease_out);
///         self.current_value = emath::lerp(self.current_value..=self.target_value, t);
///     }
/// }
/// ```
pub fn interpolation_factor(
    (start_time, end_time): (f64, f64),
    current_time: f64,
    dt: f32,
    easing: impl Fn(f32) -> f32,
) -> f32 {
    let animation_duration = (end_time - start_time) as f32;
    let prev_time = current_time - dt as f64;
    let prev_t = easing((prev_time - start_time) as f32 / animation_duration);
    let end_t = easing((current_time - start_time) as f32 / animation_duration);
    if end_t < 1.0 {
        (end_t - prev_t) / (1.0 - prev_t)
    } else {
        1.0
    }
}

/// Ease in, ease out.
///
/// `f(0) = 0, f'(0) = 0, f(1) = 1, f'(1) = 0`.
#[inline]
pub fn ease_in_ease_out(t: f32) -> f32 {
    let t = t.clamp(0.0, 1.0);
    (3.0 * t * t - 2.0 * t * t * t).clamp(0.0, 1.0)
}