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
//! Word wrapping algorithms.
//!
//! After a text has been broken into words (or [`Fragment`]s), one
//! now has to decide how to break the fragments into lines. The
//! simplest algorithm for this is implemented by
//! [`wrap_first_fit()`]: it uses no look-ahead and simply adds
//! fragments to the line as long as they fit. However, this can lead
//! to poor line breaks if a large fragment almost-but-not-quite fits
//! on a line. When that happens, the fragment is moved to the next
//! line and it will leave behind a large gap.
//!
//! A more advanced algorithm, implemented by [`wrap_optimal_fit()`],
//! will take this into account. The optimal-fit algorithm considers
//! all possible line breaks and will attempt to minimize the gaps
//! left behind by overly short lines.
//!
//! While both algorithms run in linear time, the first-fit algorithm
//! is about 4 times faster than the optimal-fit algorithm.
#[cfg(feature = "smawk")]
mod optimal_fit;
#[cfg(feature = "smawk")]
pub use optimal_fit::{wrap_optimal_fit, OverflowError, Penalties};
use crate::core::{Fragment, Word};
/// Describes how to wrap words into lines.
///
/// The simplest approach is to wrap words one word at a time and
/// accept the first way of wrapping which fit
/// ([`WrapAlgorithm::FirstFit`]). If the `smawk` Cargo feature is
/// enabled, a more complex algorithm is available which will look at
/// an entire paragraph at a time in order to find optimal line breaks
/// ([`WrapAlgorithm::OptimalFit`]).
#[derive(Clone, Copy)]
pub enum WrapAlgorithm {
/// Wrap words using a fast and simple algorithm.
///
/// This algorithm uses no look-ahead when finding line breaks.
/// Implemented by [`wrap_first_fit()`], please see that function
/// for details and examples.
FirstFit,
/// Wrap words using an advanced algorithm with look-ahead.
///
/// This wrapping algorithm considers the entire paragraph to find
/// optimal line breaks. When wrapping text, "penalties" are
/// assigned to line breaks based on the gaps left at the end of
/// lines. See [`Penalties`] for details.
///
/// The underlying wrapping algorithm is implemented by
/// [`wrap_optimal_fit()`], please see that function for examples.
///
/// **Note:** Only available when the `smawk` Cargo feature is
/// enabled.
#[cfg(feature = "smawk")]
OptimalFit(Penalties),
/// Custom wrapping function.
///
/// Use this if you want to implement your own wrapping algorithm.
/// The function can freely decide how to turn a slice of
/// [`Word`]s into lines.
///
/// # Example
///
/// ```
/// use textwrap::core::Word;
/// use textwrap::{wrap, Options, WrapAlgorithm};
///
/// fn stair<'a, 'b>(words: &'b [Word<'a>], _: &'b [usize]) -> Vec<&'b [Word<'a>]> {
/// let mut lines = Vec::new();
/// let mut step = 1;
/// let mut start_idx = 0;
/// while start_idx + step <= words.len() {
/// lines.push(&words[start_idx .. start_idx+step]);
/// start_idx += step;
/// step += 1;
/// }
/// lines
/// }
///
/// let options = Options::new(10).wrap_algorithm(WrapAlgorithm::Custom(stair));
/// assert_eq!(wrap("First, second, third, fourth, fifth, sixth", options),
/// vec!["First,",
/// "second, third,",
/// "fourth, fifth, sixth"]);
/// ```
Custom(for<'a, 'b> fn(words: &'b [Word<'a>], line_widths: &'b [usize]) -> Vec<&'b [Word<'a>]>),
}
impl PartialEq for WrapAlgorithm {
/// Compare two wrap algorithms.
///
/// ```
/// use textwrap::WrapAlgorithm;
///
/// assert_eq!(WrapAlgorithm::FirstFit, WrapAlgorithm::FirstFit);
/// #[cfg(feature = "smawk")] {
/// assert_eq!(WrapAlgorithm::new_optimal_fit(), WrapAlgorithm::new_optimal_fit());
/// }
/// ```
///
/// Note that `WrapAlgorithm::Custom` values never compare equal:
///
/// ```
/// use textwrap::WrapAlgorithm;
///
/// assert_ne!(WrapAlgorithm::Custom(|words, line_widths| vec![words]),
/// WrapAlgorithm::Custom(|words, line_widths| vec![words]));
/// ```
fn eq(&self, other: &Self) -> bool {
match (self, other) {
(WrapAlgorithm::FirstFit, WrapAlgorithm::FirstFit) => true,
#[cfg(feature = "smawk")]
(WrapAlgorithm::OptimalFit(a), WrapAlgorithm::OptimalFit(b)) => a == b,
(_, _) => false,
}
}
}
impl std::fmt::Debug for WrapAlgorithm {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
match self {
WrapAlgorithm::FirstFit => f.write_str("FirstFit"),
#[cfg(feature = "smawk")]
WrapAlgorithm::OptimalFit(penalties) => write!(f, "OptimalFit({:?})", penalties),
WrapAlgorithm::Custom(_) => f.write_str("Custom(...)"),
}
}
}
impl WrapAlgorithm {
/// Create new wrap algorithm.
///
/// The best wrapping algorithm is used by default, i.e.,
/// [`WrapAlgorithm::OptimalFit`] if available, otherwise
/// [`WrapAlgorithm::FirstFit`].
pub const fn new() -> Self {
#[cfg(not(feature = "smawk"))]
{
WrapAlgorithm::FirstFit
}
#[cfg(feature = "smawk")]
{
WrapAlgorithm::new_optimal_fit()
}
}
/// New [`WrapAlgorithm::OptimalFit`] with default penalties. This
/// works well for monospace text.
///
/// **Note:** Only available when the `smawk` Cargo feature is
/// enabled.
#[cfg(feature = "smawk")]
pub const fn new_optimal_fit() -> Self {
WrapAlgorithm::OptimalFit(Penalties::new())
}
/// Wrap words according to line widths.
///
/// The `line_widths` slice gives the target line width for each
/// line (the last slice element is repeated as necessary). This
/// can be used to implement hanging indentation.
#[inline]
pub fn wrap<'a, 'b>(
&self,
words: &'b [Word<'a>],
line_widths: &'b [usize],
) -> Vec<&'b [Word<'a>]> {
// Every integer up to 2u64.pow(f64::MANTISSA_DIGITS) = 2**53
// = 9_007_199_254_740_992 can be represented without loss by
// a f64. Larger line widths will be rounded to the nearest
// representable number.
let f64_line_widths = line_widths.iter().map(|w| *w as f64).collect::<Vec<_>>();
match self {
WrapAlgorithm::FirstFit => wrap_first_fit(words, &f64_line_widths),
#[cfg(feature = "smawk")]
WrapAlgorithm::OptimalFit(penalties) => {
// The computation cannot overflow when the line
// widths are restricted to usize.
wrap_optimal_fit(words, &f64_line_widths, penalties).unwrap()
}
WrapAlgorithm::Custom(func) => func(words, line_widths),
}
}
}
impl Default for WrapAlgorithm {
fn default() -> Self {
WrapAlgorithm::new()
}
}
/// Wrap abstract fragments into lines with a first-fit algorithm.
///
/// The `line_widths` slice gives the target line width for each line
/// (the last slice element is repeated as necessary). This can be
/// used to implement hanging indentation.
///
/// The fragments must already have been split into the desired
/// widths, this function will not (and cannot) attempt to split them
/// further when arranging them into lines.
///
/// # First-Fit Algorithm
///
/// This implements a simple “greedy” algorithm: accumulate fragments
/// one by one and when a fragment no longer fits, start a new line.
/// There is no look-ahead, we simply take first fit of the fragments
/// we find.
///
/// While fast and predictable, this algorithm can produce poor line
/// breaks when a long fragment is moved to a new line, leaving behind
/// a large gap:
///
/// ```
/// use textwrap::core::Word;
/// use textwrap::wrap_algorithms::wrap_first_fit;
/// use textwrap::WordSeparator;
///
/// // Helper to convert wrapped lines to a Vec<String>.
/// fn lines_to_strings(lines: Vec<&[Word<'_>]>) -> Vec<String> {
/// lines.iter().map(|line| {
/// line.iter().map(|word| &**word).collect::<Vec<_>>().join(" ")
/// }).collect::<Vec<_>>()
/// }
///
/// let text = "These few words will unfortunately not wrap nicely.";
/// let words = WordSeparator::AsciiSpace.find_words(text).collect::<Vec<_>>();
/// assert_eq!(lines_to_strings(wrap_first_fit(&words, &[15.0])),
/// vec!["These few words",
/// "will", // <-- short line
/// "unfortunately",
/// "not wrap",
/// "nicely."]);
///
/// // We can avoid the short line if we look ahead:
/// #[cfg(feature = "smawk")]
/// use textwrap::wrap_algorithms::{wrap_optimal_fit, Penalties};
/// #[cfg(feature = "smawk")]
/// assert_eq!(lines_to_strings(wrap_optimal_fit(&words, &[15.0], &Penalties::new()).unwrap()),
/// vec!["These few",
/// "words will",
/// "unfortunately",
/// "not wrap",
/// "nicely."]);
/// ```
///
/// The [`wrap_optimal_fit()`] function was used above to get better
/// line breaks. It uses an advanced algorithm which tries to avoid
/// short lines. This function is about 4 times faster than
/// [`wrap_optimal_fit()`].
///
/// # Examples
///
/// Imagine you're building a house site and you have a number of
/// tasks you need to execute. Things like pour foundation, complete
/// framing, install plumbing, electric cabling, install insulation.
///
/// The construction workers can only work during daytime, so they
/// need to pack up everything at night. Because they need to secure
/// their tools and move machines back to the garage, this process
/// takes much more time than the time it would take them to simply
/// switch to another task.
///
/// You would like to make a list of tasks to execute every day based
/// on your estimates. You can model this with a program like this:
///
/// ```
/// use textwrap::core::{Fragment, Word};
/// use textwrap::wrap_algorithms::wrap_first_fit;
///
/// #[derive(Debug)]
/// struct Task<'a> {
/// name: &'a str,
/// hours: f64, // Time needed to complete task.
/// sweep: f64, // Time needed for a quick sweep after task during the day.
/// cleanup: f64, // Time needed for full cleanup if day ends with this task.
/// }
///
/// impl Fragment for Task<'_> {
/// fn width(&self) -> f64 { self.hours }
/// fn whitespace_width(&self) -> f64 { self.sweep }
/// fn penalty_width(&self) -> f64 { self.cleanup }
/// }
///
/// // The morning tasks
/// let tasks = vec![
/// Task { name: "Foundation", hours: 4.0, sweep: 2.0, cleanup: 3.0 },
/// Task { name: "Framing", hours: 3.0, sweep: 1.0, cleanup: 2.0 },
/// Task { name: "Plumbing", hours: 2.0, sweep: 2.0, cleanup: 2.0 },
/// Task { name: "Electrical", hours: 2.0, sweep: 1.0, cleanup: 2.0 },
/// Task { name: "Insulation", hours: 2.0, sweep: 1.0, cleanup: 2.0 },
/// Task { name: "Drywall", hours: 3.0, sweep: 1.0, cleanup: 2.0 },
/// Task { name: "Floors", hours: 3.0, sweep: 1.0, cleanup: 2.0 },
/// Task { name: "Countertops", hours: 1.0, sweep: 1.0, cleanup: 2.0 },
/// Task { name: "Bathrooms", hours: 2.0, sweep: 1.0, cleanup: 2.0 },
/// ];
///
/// // Fill tasks into days, taking `day_length` into account. The
/// // output shows the hours worked per day along with the names of
/// // the tasks for that day.
/// fn assign_days<'a>(tasks: &[Task<'a>], day_length: f64) -> Vec<(f64, Vec<&'a str>)> {
/// let mut days = Vec::new();
/// // Assign tasks to days. The assignment is a vector of slices,
/// // with a slice per day.
/// let assigned_days: Vec<&[Task<'a>]> = wrap_first_fit(&tasks, &[day_length]);
/// for day in assigned_days.iter() {
/// let last = day.last().unwrap();
/// let work_hours: f64 = day.iter().map(|t| t.hours + t.sweep).sum();
/// let names = day.iter().map(|t| t.name).collect::<Vec<_>>();
/// days.push((work_hours - last.sweep + last.cleanup, names));
/// }
/// days
/// }
///
/// // With a single crew working 8 hours a day:
/// assert_eq!(
/// assign_days(&tasks, 8.0),
/// [
/// (7.0, vec!["Foundation"]),
/// (8.0, vec!["Framing", "Plumbing"]),
/// (7.0, vec!["Electrical", "Insulation"]),
/// (5.0, vec!["Drywall"]),
/// (7.0, vec!["Floors", "Countertops"]),
/// (4.0, vec!["Bathrooms"]),
/// ]
/// );
///
/// // With two crews working in shifts, 16 hours a day:
/// assert_eq!(
/// assign_days(&tasks, 16.0),
/// [
/// (14.0, vec!["Foundation", "Framing", "Plumbing"]),
/// (15.0, vec!["Electrical", "Insulation", "Drywall", "Floors"]),
/// (6.0, vec!["Countertops", "Bathrooms"]),
/// ]
/// );
/// ```
///
/// Apologies to anyone who actually knows how to build a house and
/// knows how long each step takes :-)
pub fn wrap_first_fit<'a, T: Fragment>(
fragments: &'a [T],
line_widths: &[f64],
) -> Vec<&'a [T]> {
// The final line width is used for all remaining lines.
let default_line_width = line_widths.last().copied().unwrap_or(0.0);
let mut lines = Vec::new();
let mut start = 0;
let mut width = 0.0;
for (idx, fragment) in fragments.iter().enumerate() {
let line_width = line_widths
.get(lines.len())
.copied()
.unwrap_or(default_line_width);
if width + fragment.width() + fragment.penalty_width() > line_width && idx > start {
lines.push(&fragments[start..idx]);
start = idx;
width = 0.0;
}
width += fragment.width() + fragment.whitespace_width();
}
lines.push(&fragments[start..]);
lines
}
#[cfg(test)]
mod tests {
use super::*;
#[derive(Debug, PartialEq)]
struct Word(f64);
#[rustfmt::skip]
impl Fragment for Word {
fn width(&self) -> f64 { self.0 }
fn whitespace_width(&self) -> f64 { 1.0 }
fn penalty_width(&self) -> f64 { 0.0 }
}
#[test]
fn wrap_string_longer_than_f64() {
let words = vec![
Word(1e307),
Word(2e307),
Word(3e307),
Word(4e307),
Word(5e307),
Word(6e307),
];
// Wrap at just under f64::MAX (~19e307). The tiny
// whitespace_widths disappear because of loss of precision.
assert_eq!(
wrap_first_fit(&words, &[15e307]),
&[
vec![
Word(1e307),
Word(2e307),
Word(3e307),
Word(4e307),
Word(5e307)
],
vec![Word(6e307)]
]
);
}
}