typst_library/foundations/
decimal.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
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
use std::fmt::{self, Display, Formatter};
use std::hash::{Hash, Hasher};
use std::ops::Neg;
use std::str::FromStr;

use ecow::{eco_format, EcoString};
use rust_decimal::MathematicalOps;
use typst_syntax::{ast, Span, Spanned};

use crate::diag::{warning, At, SourceResult};
use crate::engine::Engine;
use crate::foundations::{cast, func, repr, scope, ty, Repr, Str};
use crate::World;

/// A fixed-point decimal number type.
///
/// This type should be used for precise arithmetic operations on numbers
/// represented in base 10. A typical use case is representing currency.
///
/// # Example
/// ```example
/// Decimal: #(decimal("0.1") + decimal("0.2")) \
/// Float: #(0.1 + 0.2)
/// ```
///
/// # Construction and casts
/// To create a decimal number, use the `{decimal(string)}` constructor, such as
/// in `{decimal("3.141592653")}` **(note the double quotes!)**. This
/// constructor preserves all given fractional digits, provided they are
/// representable as per the limits specified below (otherwise, an error is
/// raised).
///
/// You can also convert any [integer]($int) to a decimal with the
/// `{decimal(int)}` constructor, e.g. `{decimal(59)}`. However, note that
/// constructing a decimal from a [floating-point number]($float), while
/// supported, **is an imprecise conversion and therefore discouraged.** A
/// warning will be raised if Typst detects that there was an accidental `float`
/// to `decimal` cast through its constructor, e.g. if writing `{decimal(3.14)}`
/// (note the lack of double quotes, indicating this is an accidental `float`
/// cast and therefore imprecise). It is recommended to use strings for
/// constant decimal values instead (e.g. `{decimal("3.14")}`).
///
/// The precision of a `float` to `decimal` cast can be slightly improved by
/// rounding the result to 15 digits with [`calc.round`]($calc.round), but there
/// are still no precision guarantees for that kind of conversion.
///
/// # Operations
/// Basic arithmetic operations are supported on two decimals and on pairs of
/// decimals and integers.
///
/// Built-in operations between `float` and `decimal` are not supported in order
/// to guard against accidental loss of precision. They will raise an error
/// instead.
///
/// Certain `calc` functions, such as trigonometric functions and power between
/// two real numbers, are also only supported for `float` (although raising
/// `decimal` to integer exponents is supported). You can opt into potentially
/// imprecise operations with the `{float(decimal)}` constructor, which casts
/// the `decimal` number into a `float`, allowing for operations without
/// precision guarantees.
///
/// # Displaying decimals
/// To display a decimal, simply insert the value into the document. To only
/// display a certain number of digits, [round]($calc.round) the decimal first.
/// Localized formatting of decimals and other numbers is not yet supported, but
/// planned for the future.
///
/// You can convert decimals to strings using the [`str`] constructor. This way,
/// you can post-process the displayed representation, e.g. to replace the
/// period with a comma (as a stand-in for proper built-in localization to
/// languages that use the comma).
///
/// # Precision and limits
/// A `decimal` number has a limit of 28 to 29 significant base-10 digits. This
/// includes the sum of digits before and after the decimal point. As such,
/// numbers with more fractional digits have a smaller range. The maximum and
/// minimum `decimal` numbers have a value of `{79228162514264337593543950335}`
/// and `{-79228162514264337593543950335}` respectively. In contrast with
/// [`float`], this type does not support infinity or NaN, so overflowing or
/// underflowing operations will raise an error.
///
/// Typical operations between `decimal` numbers, such as addition,
/// multiplication, and [power]($calc.pow) to an integer, will be highly precise
/// due to their fixed-point representation. Note, however, that multiplication
/// and division may not preserve all digits in some edge cases: while they are
/// considered precise, digits past the limits specified above are rounded off
/// and lost, so some loss of precision beyond the maximum representable digits
/// is possible. Note that this behavior can be observed not only when dividing,
/// but also when multiplying by numbers between 0 and 1, as both operations can
/// push a number's fractional digits beyond the limits described above, leading
/// to rounding. When those two operations do not surpass the digit limits, they
/// are fully precise.
#[ty(scope, cast)]
#[derive(Debug, Default, Clone, Copy, PartialEq, Eq, PartialOrd, Ord)]
pub struct Decimal(rust_decimal::Decimal);

impl Decimal {
    pub const ZERO: Self = Self(rust_decimal::Decimal::ZERO);
    pub const ONE: Self = Self(rust_decimal::Decimal::ONE);
    pub const MIN: Self = Self(rust_decimal::Decimal::MIN);
    pub const MAX: Self = Self(rust_decimal::Decimal::MAX);

    /// Whether this decimal value is zero.
    pub const fn is_zero(self) -> bool {
        self.0.is_zero()
    }

    /// Whether this decimal value is negative.
    pub const fn is_negative(self) -> bool {
        self.0.is_sign_negative()
    }

    /// Whether this decimal has fractional part equal to zero (is an integer).
    pub fn is_integer(self) -> bool {
        self.0.is_integer()
    }

    /// Computes the absolute value of this decimal.
    pub fn abs(self) -> Self {
        Self(self.0.abs())
    }

    /// Computes the largest integer less than or equal to this decimal.
    ///
    /// A decimal is returned as this may not be within `i64`'s range of
    /// values.
    pub fn floor(self) -> Self {
        Self(self.0.floor())
    }

    /// Computes the smallest integer greater than or equal to this decimal.
    ///
    /// A decimal is returned as this may not be within `i64`'s range of
    /// values.
    pub fn ceil(self) -> Self {
        Self(self.0.ceil())
    }

    /// Returns the integer part of this decimal.
    pub fn trunc(self) -> Self {
        Self(self.0.trunc())
    }

    /// Returns the fractional part of this decimal (with the integer part set
    /// to zero).
    pub fn fract(self) -> Self {
        Self(self.0.fract())
    }

    /// Rounds this decimal up to the specified amount of digits with the
    /// traditional rounding rules, using the "midpoint away from zero"
    /// strategy (6.5 -> 7, -6.5 -> -7).
    ///
    /// If given a negative amount of digits, rounds to integer digits instead
    /// with the same rounding strategy. For example, rounding to -3 digits
    /// will turn 34567.89 into 35000.00 and -34567.89 into -35000.00.
    ///
    /// Note that this can return `None` when using negative digits where the
    /// rounded number would overflow the available range for decimals.
    pub fn round(self, digits: i32) -> Option<Self> {
        // Positive digits can be handled by just rounding with rust_decimal.
        if let Ok(positive_digits) = u32::try_from(digits) {
            return Some(Self(self.0.round_dp_with_strategy(
                positive_digits,
                rust_decimal::RoundingStrategy::MidpointAwayFromZero,
            )));
        }

        // We received negative digits, so we round to integer digits.
        let mut num = self.0;
        let old_scale = num.scale();
        let digits = -digits as u32;

        let (Ok(_), Some(ten_to_digits)) = (
            // Same as dividing by 10^digits.
            num.set_scale(old_scale + digits),
            rust_decimal::Decimal::TEN.checked_powi(digits as i64),
        ) else {
            // Scaling more than any possible amount of integer digits.
            let mut zero = rust_decimal::Decimal::ZERO;
            zero.set_sign_negative(self.is_negative());
            return Some(Self(zero));
        };

        // Round to this integer digit.
        num = num.round_dp_with_strategy(
            0,
            rust_decimal::RoundingStrategy::MidpointAwayFromZero,
        );

        // Multiply by 10^digits again, which can overflow and fail.
        num.checked_mul(ten_to_digits).map(Self)
    }

    /// Attempts to add two decimals.
    ///
    /// Returns `None` on overflow or underflow.
    pub fn checked_add(self, other: Self) -> Option<Self> {
        self.0.checked_add(other.0).map(Self)
    }

    /// Attempts to subtract a decimal from another.
    ///
    /// Returns `None` on overflow or underflow.
    pub fn checked_sub(self, other: Self) -> Option<Self> {
        self.0.checked_sub(other.0).map(Self)
    }

    /// Attempts to multiply two decimals.
    ///
    /// Returns `None` on overflow or underflow.
    pub fn checked_mul(self, other: Self) -> Option<Self> {
        self.0.checked_mul(other.0).map(Self)
    }

    /// Attempts to divide two decimals.
    ///
    /// Returns `None` if `other` is zero, as well as on overflow or underflow.
    pub fn checked_div(self, other: Self) -> Option<Self> {
        self.0.checked_div(other.0).map(Self)
    }

    /// Attempts to obtain the quotient of Euclidean division between two
    /// decimals. Implemented similarly to [`f64::div_euclid`].
    ///
    /// The returned quotient is truncated and adjusted if the remainder was
    /// negative.
    ///
    /// Returns `None` if `other` is zero, as well as on overflow or underflow.
    pub fn checked_div_euclid(self, other: Self) -> Option<Self> {
        let q = self.0.checked_div(other.0)?.trunc();
        if self
            .0
            .checked_rem(other.0)
            .as_ref()
            .is_some_and(rust_decimal::Decimal::is_sign_negative)
        {
            return if other.0.is_sign_positive() {
                q.checked_sub(rust_decimal::Decimal::ONE).map(Self)
            } else {
                q.checked_add(rust_decimal::Decimal::ONE).map(Self)
            };
        }
        Some(Self(q))
    }

    /// Attempts to obtain the remainder of Euclidean division between two
    /// decimals. Implemented similarly to [`f64::rem_euclid`].
    ///
    /// The returned decimal `r` is non-negative within the range
    /// `0.0 <= r < other.abs()`.
    ///
    /// Returns `None` if `other` is zero, as well as on overflow or underflow.
    pub fn checked_rem_euclid(self, other: Self) -> Option<Self> {
        let r = self.0.checked_rem(other.0)?;
        Some(Self(if r.is_sign_negative() { r.checked_add(other.0.abs())? } else { r }))
    }

    /// Attempts to calculate the remainder of the division of two decimals.
    ///
    /// Returns `None` if `other` is zero, as well as on overflow or underflow.
    pub fn checked_rem(self, other: Self) -> Option<Self> {
        self.0.checked_rem(other.0).map(Self)
    }

    /// Attempts to take one decimal to the power of an integer.
    ///
    /// Returns `None` for invalid operands, as well as on overflow or
    /// underflow.
    pub fn checked_powi(self, other: i64) -> Option<Self> {
        self.0.checked_powi(other).map(Self)
    }
}

#[scope]
impl Decimal {
    /// Converts a value to a `decimal`.
    ///
    /// It is recommended to use a string to construct the decimal number, or an
    /// [integer]($int) (if desired). The string must contain a number in the
    /// format `{"3.14159"}` (or `{"-3.141519"}` for negative numbers). The
    /// fractional digits are fully preserved; if that's not possible due to the
    /// limit of significant digits (around 28 to 29) having been reached, an
    /// error is raised as the given decimal number wouldn't be representable.
    ///
    /// While this constructor can be used with [floating-point numbers]($float)
    /// to cast them to `decimal`, doing so is **discouraged** as **this cast is
    /// inherently imprecise.** It is easy to accidentally perform this cast by
    /// writing `{decimal(1.234)}` (note the lack of double quotes), which is
    /// why Typst will emit a warning in that case. Please write
    /// `{decimal("1.234")}` instead for that particular case (initialization of
    /// a constant decimal). Also note that floats that are NaN or infinite
    /// cannot be cast to decimals and will raise an error.
    ///
    /// ```example
    /// #decimal("1.222222222222222")
    /// ```
    #[func(constructor)]
    pub fn construct(
        engine: &mut Engine,
        /// The value that should be converted to a decimal.
        value: Spanned<ToDecimal>,
    ) -> SourceResult<Decimal> {
        match value.v {
            ToDecimal::Str(str) => Self::from_str(&str.replace(repr::MINUS_SIGN, "-"))
                .map_err(|_| eco_format!("invalid decimal: {str}"))
                .at(value.span),
            ToDecimal::Int(int) => Ok(Self::from(int)),
            ToDecimal::Float(float) => {
                warn_on_float_literal(engine, value.span);
                Self::try_from(float)
                    .map_err(|_| {
                        eco_format!(
                            "float is not a valid decimal: {}",
                            repr::format_float(float, None, true, "")
                        )
                    })
                    .at(value.span)
            }
            ToDecimal::Decimal(decimal) => Ok(decimal),
        }
    }
}

/// Emits a warning when a decimal is constructed from a float literal.
fn warn_on_float_literal(engine: &mut Engine, span: Span) -> Option<()> {
    let id = span.id()?;
    let source = engine.world.source(id).ok()?;
    let node = source.find(span)?;
    if node.is::<ast::Float>() {
        engine.sink.warn(warning!(
            span,
            "creating a decimal using imprecise float literal";
            hint: "use a string in the decimal constructor to avoid loss \
                   of precision: `decimal({})`",
            node.text().repr()
        ));
    }
    Some(())
}

impl FromStr for Decimal {
    type Err = rust_decimal::Error;

    fn from_str(s: &str) -> Result<Self, Self::Err> {
        rust_decimal::Decimal::from_str_exact(s).map(Self)
    }
}

impl From<i64> for Decimal {
    fn from(value: i64) -> Self {
        Self(rust_decimal::Decimal::from(value))
    }
}

impl TryFrom<f64> for Decimal {
    type Error = ();

    /// Attempts to convert a Decimal to a float.
    ///
    /// This can fail if the float is infinite or NaN, or otherwise cannot be
    /// represented by a decimal number.
    fn try_from(value: f64) -> Result<Self, Self::Error> {
        rust_decimal::Decimal::from_f64_retain(value).map(Self).ok_or(())
    }
}

impl TryFrom<Decimal> for f64 {
    type Error = rust_decimal::Error;

    /// Attempts to convert a Decimal to a float.
    ///
    /// This should in principle be infallible according to the implementation,
    /// but we mirror the decimal implementation's API either way.
    fn try_from(value: Decimal) -> Result<Self, Self::Error> {
        value.0.try_into()
    }
}

impl TryFrom<Decimal> for i64 {
    type Error = rust_decimal::Error;

    /// Attempts to convert a Decimal to an integer.
    ///
    /// Returns an error if the decimal has a fractional part, or if there
    /// would be overflow or underflow.
    fn try_from(value: Decimal) -> Result<Self, Self::Error> {
        value.0.try_into()
    }
}

impl Display for Decimal {
    fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
        if self.0.is_sign_negative() {
            f.write_str(repr::MINUS_SIGN)?;
        }
        self.0.abs().fmt(f)
    }
}

impl Repr for Decimal {
    fn repr(&self) -> EcoString {
        eco_format!("decimal({})", eco_format!("{}", self.0).repr())
    }
}

impl Neg for Decimal {
    type Output = Self;

    fn neg(self) -> Self {
        Self(-self.0)
    }
}

impl Hash for Decimal {
    fn hash<H: Hasher>(&self, state: &mut H) {
        // `rust_decimal`'s Hash implementation normalizes decimals before
        // hashing them. This means decimals with different scales but
        // equivalent value not only compare equal but also hash equally. Here,
        // we hash all bytes explicitly to ensure the scale is also considered.
        // This means that 123.314 == 123.31400, but 123.314.hash() !=
        // 123.31400.hash().
        //
        // Note that this implies that equal decimals can have different hashes,
        // which might generate problems with certain data structures, such as
        // HashSet and HashMap.
        self.0.serialize().hash(state);
    }
}

/// A value that can be cast to a decimal.
pub enum ToDecimal {
    /// A decimal to be converted to itself.
    Decimal(Decimal),
    /// A string with the decimal's representation.
    Str(EcoString),
    /// An integer to be converted to the equivalent decimal.
    Int(i64),
    /// A float to be converted to the equivalent decimal.
    Float(f64),
}

cast! {
    ToDecimal,
    v: Decimal => Self::Decimal(v),
    v: i64 => Self::Int(v),
    v: bool => Self::Int(v as i64),
    v: f64 => Self::Float(v),
    v: Str => Self::Str(EcoString::from(v)),
}

#[cfg(test)]
mod tests {
    use std::str::FromStr;

    use typst_utils::hash128;

    use super::Decimal;

    #[test]
    fn test_decimals_with_equal_scales_hash_identically() {
        let a = Decimal::from_str("3.14").unwrap();
        let b = Decimal::from_str("3.14").unwrap();
        assert_eq!(a, b);
        assert_eq!(hash128(&a), hash128(&b));
    }

    #[test]
    fn test_decimals_with_different_scales_hash_differently() {
        let a = Decimal::from_str("3.140").unwrap();
        let b = Decimal::from_str("3.14000").unwrap();
        assert_eq!(a, b);
        assert_ne!(hash128(&a), hash128(&b));
    }

    #[track_caller]
    fn test_round(value: &str, digits: i32, expected: &str) {
        assert_eq!(
            Decimal::from_str(value).unwrap().round(digits),
            Some(Decimal::from_str(expected).unwrap()),
        );
    }

    #[test]
    fn test_decimal_positive_round() {
        test_round("312.55553", 0, "313.00000");
        test_round("312.55553", 3, "312.556");
        test_round("312.5555300000", 3, "312.556");
        test_round("-312.55553", 3, "-312.556");
        test_round("312.55553", 28, "312.55553");
        test_round("312.55553", 2341, "312.55553");
        test_round("-312.55553", 2341, "-312.55553");
    }

    #[test]
    fn test_decimal_negative_round() {
        test_round("4596.55553", -1, "4600");
        test_round("4596.555530000000", -1, "4600");
        test_round("-4596.55553", -3, "-5000");
        test_round("4596.55553", -28, "0");
        test_round("-4596.55553", -2341, "0");
        assert_eq!(Decimal::MAX.round(-1), None);
    }
}