Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1 | use std::borrow::Cow; |
| 2 | use std::collections::HashMap; |
| 3 | use std::fmt; |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 4 | use std::iter::FusedIterator; |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 5 | use std::ops::{Index, Range}; |
| 6 | use std::str::FromStr; |
| 7 | use std::sync::Arc; |
| 8 | |
| 9 | use find_byte::find_byte; |
| 10 | use syntax; |
| 11 | |
| 12 | use error::Error; |
| 13 | use exec::{Exec, ExecNoSyncStr}; |
| 14 | use expand::expand_str; |
| 15 | use re_builder::unicode::RegexBuilder; |
| 16 | use re_trait::{self, RegularExpression, SubCapturesPosIter}; |
| 17 | |
| 18 | /// Escapes all regular expression meta characters in `text`. |
| 19 | /// |
| 20 | /// The string returned may be safely used as a literal in a regular |
| 21 | /// expression. |
| 22 | pub fn escape(text: &str) -> String { |
| 23 | syntax::escape(text) |
| 24 | } |
| 25 | |
| 26 | /// Match represents a single match of a regex in a haystack. |
| 27 | /// |
| 28 | /// The lifetime parameter `'t` refers to the lifetime of the matched text. |
| 29 | #[derive(Copy, Clone, Debug, Eq, PartialEq)] |
| 30 | pub struct Match<'t> { |
| 31 | text: &'t str, |
| 32 | start: usize, |
| 33 | end: usize, |
| 34 | } |
| 35 | |
| 36 | impl<'t> Match<'t> { |
| 37 | /// Returns the starting byte offset of the match in the haystack. |
| 38 | #[inline] |
| 39 | pub fn start(&self) -> usize { |
| 40 | self.start |
| 41 | } |
| 42 | |
| 43 | /// Returns the ending byte offset of the match in the haystack. |
| 44 | #[inline] |
| 45 | pub fn end(&self) -> usize { |
| 46 | self.end |
| 47 | } |
| 48 | |
| 49 | /// Returns the range over the starting and ending byte offsets of the |
| 50 | /// match in the haystack. |
| 51 | #[inline] |
| 52 | pub fn range(&self) -> Range<usize> { |
| 53 | self.start..self.end |
| 54 | } |
| 55 | |
| 56 | /// Returns the matched text. |
| 57 | #[inline] |
| 58 | pub fn as_str(&self) -> &'t str { |
| 59 | &self.text[self.range()] |
| 60 | } |
| 61 | |
| 62 | /// Creates a new match from the given haystack and byte offsets. |
| 63 | #[inline] |
| 64 | fn new(haystack: &'t str, start: usize, end: usize) -> Match<'t> { |
| 65 | Match { text: haystack, start: start, end: end } |
| 66 | } |
| 67 | } |
| 68 | |
| 69 | impl<'t> From<Match<'t>> for &'t str { |
| 70 | fn from(m: Match<'t>) -> &'t str { |
| 71 | m.as_str() |
| 72 | } |
| 73 | } |
| 74 | |
| 75 | impl<'t> From<Match<'t>> for Range<usize> { |
| 76 | fn from(m: Match<'t>) -> Range<usize> { |
| 77 | m.range() |
| 78 | } |
| 79 | } |
| 80 | |
| 81 | /// A compiled regular expression for matching Unicode strings. |
| 82 | /// |
| 83 | /// It is represented as either a sequence of bytecode instructions (dynamic) |
| 84 | /// or as a specialized Rust function (native). It can be used to search, split |
| 85 | /// or replace text. All searching is done with an implicit `.*?` at the |
| 86 | /// beginning and end of an expression. To force an expression to match the |
| 87 | /// whole string (or a prefix or a suffix), you must use an anchor like `^` or |
| 88 | /// `$` (or `\A` and `\z`). |
| 89 | /// |
| 90 | /// While this crate will handle Unicode strings (whether in the regular |
| 91 | /// expression or in the search text), all positions returned are **byte |
| 92 | /// indices**. Every byte index is guaranteed to be at a Unicode code point |
| 93 | /// boundary. |
| 94 | /// |
| 95 | /// The lifetimes `'r` and `'t` in this crate correspond to the lifetime of a |
| 96 | /// compiled regular expression and text to search, respectively. |
| 97 | /// |
| 98 | /// The only methods that allocate new strings are the string replacement |
| 99 | /// methods. All other methods (searching and splitting) return borrowed |
| 100 | /// pointers into the string given. |
| 101 | /// |
| 102 | /// # Examples |
| 103 | /// |
| 104 | /// Find the location of a US phone number: |
| 105 | /// |
| 106 | /// ```rust |
| 107 | /// # use regex::Regex; |
| 108 | /// let re = Regex::new("[0-9]{3}-[0-9]{3}-[0-9]{4}").unwrap(); |
| 109 | /// let mat = re.find("phone: 111-222-3333").unwrap(); |
| 110 | /// assert_eq!((mat.start(), mat.end()), (7, 19)); |
| 111 | /// ``` |
| 112 | /// |
| 113 | /// # Using the `std::str::pattern` methods with `Regex` |
| 114 | /// |
| 115 | /// > **Note**: This section requires that this crate is compiled with the |
| 116 | /// > `pattern` Cargo feature enabled, which **requires nightly Rust**. |
| 117 | /// |
| 118 | /// Since `Regex` implements `Pattern`, you can use regexes with methods |
| 119 | /// defined on `&str`. For example, `is_match`, `find`, `find_iter` |
| 120 | /// and `split` can be replaced with `str::contains`, `str::find`, |
| 121 | /// `str::match_indices` and `str::split`. |
| 122 | /// |
| 123 | /// Here are some examples: |
| 124 | /// |
| 125 | /// ```rust,ignore |
| 126 | /// # use regex::Regex; |
| 127 | /// let re = Regex::new(r"\d+").unwrap(); |
| 128 | /// let haystack = "a111b222c"; |
| 129 | /// |
| 130 | /// assert!(haystack.contains(&re)); |
| 131 | /// assert_eq!(haystack.find(&re), Some(1)); |
| 132 | /// assert_eq!(haystack.match_indices(&re).collect::<Vec<_>>(), |
| 133 | /// vec![(1, 4), (5, 8)]); |
| 134 | /// assert_eq!(haystack.split(&re).collect::<Vec<_>>(), vec!["a", "b", "c"]); |
| 135 | /// ``` |
| 136 | #[derive(Clone)] |
| 137 | pub struct Regex(Exec); |
| 138 | |
| 139 | impl fmt::Display for Regex { |
| 140 | /// Shows the original regular expression. |
| 141 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
| 142 | write!(f, "{}", self.as_str()) |
| 143 | } |
| 144 | } |
| 145 | |
| 146 | impl fmt::Debug for Regex { |
| 147 | /// Shows the original regular expression. |
| 148 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
| 149 | fmt::Display::fmt(self, f) |
| 150 | } |
| 151 | } |
| 152 | |
| 153 | #[doc(hidden)] |
| 154 | impl From<Exec> for Regex { |
| 155 | fn from(exec: Exec) -> Regex { |
| 156 | Regex(exec) |
| 157 | } |
| 158 | } |
| 159 | |
| 160 | impl FromStr for Regex { |
| 161 | type Err = Error; |
| 162 | |
| 163 | /// Attempts to parse a string into a regular expression |
| 164 | fn from_str(s: &str) -> Result<Regex, Error> { |
| 165 | Regex::new(s) |
| 166 | } |
| 167 | } |
| 168 | |
| 169 | /// Core regular expression methods. |
| 170 | impl Regex { |
| 171 | /// Compiles a regular expression. Once compiled, it can be used repeatedly |
| 172 | /// to search, split or replace text in a string. |
| 173 | /// |
| 174 | /// If an invalid expression is given, then an error is returned. |
| 175 | pub fn new(re: &str) -> Result<Regex, Error> { |
| 176 | RegexBuilder::new(re).build() |
| 177 | } |
| 178 | |
Chih-Hung Hsieh | 849e445 | 2020-10-26 13:16:47 -0700 | [diff] [blame] | 179 | /// Returns true if and only if there is a match for the regex in the |
| 180 | /// string given. |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 181 | /// |
| 182 | /// It is recommended to use this method if all you need to do is test |
| 183 | /// a match, since the underlying matching engine may be able to do less |
| 184 | /// work. |
| 185 | /// |
| 186 | /// # Example |
| 187 | /// |
| 188 | /// Test if some text contains at least one word with exactly 13 |
| 189 | /// Unicode word characters: |
| 190 | /// |
| 191 | /// ```rust |
| 192 | /// # extern crate regex; use regex::Regex; |
| 193 | /// # fn main() { |
| 194 | /// let text = "I categorically deny having triskaidekaphobia."; |
| 195 | /// assert!(Regex::new(r"\b\w{13}\b").unwrap().is_match(text)); |
| 196 | /// # } |
| 197 | /// ``` |
| 198 | pub fn is_match(&self, text: &str) -> bool { |
| 199 | self.is_match_at(text, 0) |
| 200 | } |
| 201 | |
| 202 | /// Returns the start and end byte range of the leftmost-first match in |
| 203 | /// `text`. If no match exists, then `None` is returned. |
| 204 | /// |
| 205 | /// Note that this should only be used if you want to discover the position |
| 206 | /// of the match. Testing the existence of a match is faster if you use |
| 207 | /// `is_match`. |
| 208 | /// |
| 209 | /// # Example |
| 210 | /// |
| 211 | /// Find the start and end location of the first word with exactly 13 |
| 212 | /// Unicode word characters: |
| 213 | /// |
| 214 | /// ```rust |
| 215 | /// # extern crate regex; use regex::Regex; |
| 216 | /// # fn main() { |
| 217 | /// let text = "I categorically deny having triskaidekaphobia."; |
| 218 | /// let mat = Regex::new(r"\b\w{13}\b").unwrap().find(text).unwrap(); |
| 219 | /// assert_eq!(mat.start(), 2); |
| 220 | /// assert_eq!(mat.end(), 15); |
| 221 | /// # } |
| 222 | /// ``` |
| 223 | pub fn find<'t>(&self, text: &'t str) -> Option<Match<'t>> { |
| 224 | self.find_at(text, 0) |
| 225 | } |
| 226 | |
| 227 | /// Returns an iterator for each successive non-overlapping match in |
| 228 | /// `text`, returning the start and end byte indices with respect to |
| 229 | /// `text`. |
| 230 | /// |
| 231 | /// # Example |
| 232 | /// |
| 233 | /// Find the start and end location of every word with exactly 13 Unicode |
| 234 | /// word characters: |
| 235 | /// |
| 236 | /// ```rust |
| 237 | /// # extern crate regex; use regex::Regex; |
| 238 | /// # fn main() { |
| 239 | /// let text = "Retroactively relinquishing remunerations is reprehensible."; |
| 240 | /// for mat in Regex::new(r"\b\w{13}\b").unwrap().find_iter(text) { |
| 241 | /// println!("{:?}", mat); |
| 242 | /// } |
| 243 | /// # } |
| 244 | /// ``` |
| 245 | pub fn find_iter<'r, 't>(&'r self, text: &'t str) -> Matches<'r, 't> { |
| 246 | Matches(self.0.searcher_str().find_iter(text)) |
| 247 | } |
| 248 | |
| 249 | /// Returns the capture groups corresponding to the leftmost-first |
| 250 | /// match in `text`. Capture group `0` always corresponds to the entire |
| 251 | /// match. If no match is found, then `None` is returned. |
| 252 | /// |
| 253 | /// You should only use `captures` if you need access to the location of |
| 254 | /// capturing group matches. Otherwise, `find` is faster for discovering |
| 255 | /// the location of the overall match. |
| 256 | /// |
| 257 | /// # Examples |
| 258 | /// |
| 259 | /// Say you have some text with movie names and their release years, |
| 260 | /// like "'Citizen Kane' (1941)". It'd be nice if we could search for text |
| 261 | /// looking like that, while also extracting the movie name and its release |
| 262 | /// year separately. |
| 263 | /// |
| 264 | /// ```rust |
| 265 | /// # extern crate regex; use regex::Regex; |
| 266 | /// # fn main() { |
| 267 | /// let re = Regex::new(r"'([^']+)'\s+\((\d{4})\)").unwrap(); |
| 268 | /// let text = "Not my favorite movie: 'Citizen Kane' (1941)."; |
| 269 | /// let caps = re.captures(text).unwrap(); |
| 270 | /// assert_eq!(caps.get(1).unwrap().as_str(), "Citizen Kane"); |
| 271 | /// assert_eq!(caps.get(2).unwrap().as_str(), "1941"); |
| 272 | /// assert_eq!(caps.get(0).unwrap().as_str(), "'Citizen Kane' (1941)"); |
| 273 | /// // You can also access the groups by index using the Index notation. |
| 274 | /// // Note that this will panic on an invalid index. |
| 275 | /// assert_eq!(&caps[1], "Citizen Kane"); |
| 276 | /// assert_eq!(&caps[2], "1941"); |
| 277 | /// assert_eq!(&caps[0], "'Citizen Kane' (1941)"); |
| 278 | /// # } |
| 279 | /// ``` |
| 280 | /// |
| 281 | /// Note that the full match is at capture group `0`. Each subsequent |
| 282 | /// capture group is indexed by the order of its opening `(`. |
| 283 | /// |
| 284 | /// We can make this example a bit clearer by using *named* capture groups: |
| 285 | /// |
| 286 | /// ```rust |
| 287 | /// # extern crate regex; use regex::Regex; |
| 288 | /// # fn main() { |
| 289 | /// let re = Regex::new(r"'(?P<title>[^']+)'\s+\((?P<year>\d{4})\)") |
| 290 | /// .unwrap(); |
| 291 | /// let text = "Not my favorite movie: 'Citizen Kane' (1941)."; |
| 292 | /// let caps = re.captures(text).unwrap(); |
| 293 | /// assert_eq!(caps.name("title").unwrap().as_str(), "Citizen Kane"); |
| 294 | /// assert_eq!(caps.name("year").unwrap().as_str(), "1941"); |
| 295 | /// assert_eq!(caps.get(0).unwrap().as_str(), "'Citizen Kane' (1941)"); |
| 296 | /// // You can also access the groups by name using the Index notation. |
| 297 | /// // Note that this will panic on an invalid group name. |
| 298 | /// assert_eq!(&caps["title"], "Citizen Kane"); |
| 299 | /// assert_eq!(&caps["year"], "1941"); |
| 300 | /// assert_eq!(&caps[0], "'Citizen Kane' (1941)"); |
| 301 | /// |
| 302 | /// # } |
| 303 | /// ``` |
| 304 | /// |
| 305 | /// Here we name the capture groups, which we can access with the `name` |
| 306 | /// method or the `Index` notation with a `&str`. Note that the named |
| 307 | /// capture groups are still accessible with `get` or the `Index` notation |
| 308 | /// with a `usize`. |
| 309 | /// |
| 310 | /// The `0`th capture group is always unnamed, so it must always be |
| 311 | /// accessed with `get(0)` or `[0]`. |
| 312 | pub fn captures<'t>(&self, text: &'t str) -> Option<Captures<'t>> { |
| 313 | let mut locs = self.capture_locations(); |
| 314 | self.captures_read_at(&mut locs, text, 0).map(move |_| Captures { |
| 315 | text: text, |
| 316 | locs: locs.0, |
| 317 | named_groups: self.0.capture_name_idx().clone(), |
| 318 | }) |
| 319 | } |
| 320 | |
| 321 | /// Returns an iterator over all the non-overlapping capture groups matched |
| 322 | /// in `text`. This is operationally the same as `find_iter`, except it |
| 323 | /// yields information about capturing group matches. |
| 324 | /// |
| 325 | /// # Example |
| 326 | /// |
| 327 | /// We can use this to find all movie titles and their release years in |
| 328 | /// some text, where the movie is formatted like "'Title' (xxxx)": |
| 329 | /// |
| 330 | /// ```rust |
| 331 | /// # extern crate regex; use regex::Regex; |
| 332 | /// # fn main() { |
| 333 | /// let re = Regex::new(r"'(?P<title>[^']+)'\s+\((?P<year>\d{4})\)") |
| 334 | /// .unwrap(); |
| 335 | /// let text = "'Citizen Kane' (1941), 'The Wizard of Oz' (1939), 'M' (1931)."; |
| 336 | /// for caps in re.captures_iter(text) { |
| 337 | /// println!("Movie: {:?}, Released: {:?}", |
| 338 | /// &caps["title"], &caps["year"]); |
| 339 | /// } |
| 340 | /// // Output: |
| 341 | /// // Movie: Citizen Kane, Released: 1941 |
| 342 | /// // Movie: The Wizard of Oz, Released: 1939 |
| 343 | /// // Movie: M, Released: 1931 |
| 344 | /// # } |
| 345 | /// ``` |
| 346 | pub fn captures_iter<'r, 't>( |
| 347 | &'r self, |
| 348 | text: &'t str, |
| 349 | ) -> CaptureMatches<'r, 't> { |
| 350 | CaptureMatches(self.0.searcher_str().captures_iter(text)) |
| 351 | } |
| 352 | |
| 353 | /// Returns an iterator of substrings of `text` delimited by a match of the |
| 354 | /// regular expression. Namely, each element of the iterator corresponds to |
| 355 | /// text that *isn't* matched by the regular expression. |
| 356 | /// |
| 357 | /// This method will *not* copy the text given. |
| 358 | /// |
| 359 | /// # Example |
| 360 | /// |
| 361 | /// To split a string delimited by arbitrary amounts of spaces or tabs: |
| 362 | /// |
| 363 | /// ```rust |
| 364 | /// # extern crate regex; use regex::Regex; |
| 365 | /// # fn main() { |
| 366 | /// let re = Regex::new(r"[ \t]+").unwrap(); |
| 367 | /// let fields: Vec<&str> = re.split("a b \t c\td e").collect(); |
| 368 | /// assert_eq!(fields, vec!["a", "b", "c", "d", "e"]); |
| 369 | /// # } |
| 370 | /// ``` |
| 371 | pub fn split<'r, 't>(&'r self, text: &'t str) -> Split<'r, 't> { |
| 372 | Split { finder: self.find_iter(text), last: 0 } |
| 373 | } |
| 374 | |
| 375 | /// Returns an iterator of at most `limit` substrings of `text` delimited |
| 376 | /// by a match of the regular expression. (A `limit` of `0` will return no |
| 377 | /// substrings.) Namely, each element of the iterator corresponds to text |
| 378 | /// that *isn't* matched by the regular expression. The remainder of the |
| 379 | /// string that is not split will be the last element in the iterator. |
| 380 | /// |
| 381 | /// This method will *not* copy the text given. |
| 382 | /// |
| 383 | /// # Example |
| 384 | /// |
| 385 | /// Get the first two words in some text: |
| 386 | /// |
| 387 | /// ```rust |
| 388 | /// # extern crate regex; use regex::Regex; |
| 389 | /// # fn main() { |
| 390 | /// let re = Regex::new(r"\W+").unwrap(); |
| 391 | /// let fields: Vec<&str> = re.splitn("Hey! How are you?", 3).collect(); |
| 392 | /// assert_eq!(fields, vec!("Hey", "How", "are you?")); |
| 393 | /// # } |
| 394 | /// ``` |
| 395 | pub fn splitn<'r, 't>( |
| 396 | &'r self, |
| 397 | text: &'t str, |
| 398 | limit: usize, |
| 399 | ) -> SplitN<'r, 't> { |
| 400 | SplitN { splits: self.split(text), n: limit } |
| 401 | } |
| 402 | |
| 403 | /// Replaces the leftmost-first match with the replacement provided. |
| 404 | /// The replacement can be a regular string (where `$N` and `$name` are |
| 405 | /// expanded to match capture groups) or a function that takes the matches' |
| 406 | /// `Captures` and returns the replaced string. |
| 407 | /// |
| 408 | /// If no match is found, then a copy of the string is returned unchanged. |
| 409 | /// |
| 410 | /// # Replacement string syntax |
| 411 | /// |
| 412 | /// All instances of `$name` in the replacement text is replaced with the |
| 413 | /// corresponding capture group `name`. |
| 414 | /// |
| 415 | /// `name` may be an integer corresponding to the index of the |
| 416 | /// capture group (counted by order of opening parenthesis where `0` is the |
| 417 | /// entire match) or it can be a name (consisting of letters, digits or |
| 418 | /// underscores) corresponding to a named capture group. |
| 419 | /// |
| 420 | /// If `name` isn't a valid capture group (whether the name doesn't exist |
| 421 | /// or isn't a valid index), then it is replaced with the empty string. |
| 422 | /// |
| 423 | /// The longest possible name is used. e.g., `$1a` looks up the capture |
| 424 | /// group named `1a` and not the capture group at index `1`. To exert more |
| 425 | /// precise control over the name, use braces, e.g., `${1}a`. |
| 426 | /// |
| 427 | /// To write a literal `$` use `$$`. |
| 428 | /// |
| 429 | /// # Examples |
| 430 | /// |
| 431 | /// Note that this function is polymorphic with respect to the replacement. |
| 432 | /// In typical usage, this can just be a normal string: |
| 433 | /// |
| 434 | /// ```rust |
| 435 | /// # extern crate regex; use regex::Regex; |
| 436 | /// # fn main() { |
| 437 | /// let re = Regex::new("[^01]+").unwrap(); |
| 438 | /// assert_eq!(re.replace("1078910", ""), "1010"); |
| 439 | /// # } |
| 440 | /// ``` |
| 441 | /// |
| 442 | /// But anything satisfying the `Replacer` trait will work. For example, |
| 443 | /// a closure of type `|&Captures| -> String` provides direct access to the |
| 444 | /// captures corresponding to a match. This allows one to access |
| 445 | /// capturing group matches easily: |
| 446 | /// |
| 447 | /// ```rust |
| 448 | /// # extern crate regex; use regex::Regex; |
| 449 | /// # use regex::Captures; fn main() { |
| 450 | /// let re = Regex::new(r"([^,\s]+),\s+(\S+)").unwrap(); |
| 451 | /// let result = re.replace("Springsteen, Bruce", |caps: &Captures| { |
| 452 | /// format!("{} {}", &caps[2], &caps[1]) |
| 453 | /// }); |
| 454 | /// assert_eq!(result, "Bruce Springsteen"); |
| 455 | /// # } |
| 456 | /// ``` |
| 457 | /// |
| 458 | /// But this is a bit cumbersome to use all the time. Instead, a simple |
| 459 | /// syntax is supported that expands `$name` into the corresponding capture |
| 460 | /// group. Here's the last example, but using this expansion technique |
| 461 | /// with named capture groups: |
| 462 | /// |
| 463 | /// ```rust |
| 464 | /// # extern crate regex; use regex::Regex; |
| 465 | /// # fn main() { |
| 466 | /// let re = Regex::new(r"(?P<last>[^,\s]+),\s+(?P<first>\S+)").unwrap(); |
| 467 | /// let result = re.replace("Springsteen, Bruce", "$first $last"); |
| 468 | /// assert_eq!(result, "Bruce Springsteen"); |
| 469 | /// # } |
| 470 | /// ``` |
| 471 | /// |
| 472 | /// Note that using `$2` instead of `$first` or `$1` instead of `$last` |
| 473 | /// would produce the same result. To write a literal `$` use `$$`. |
| 474 | /// |
| 475 | /// Sometimes the replacement string requires use of curly braces to |
| 476 | /// delineate a capture group replacement and surrounding literal text. |
| 477 | /// For example, if we wanted to join two words together with an |
| 478 | /// underscore: |
| 479 | /// |
| 480 | /// ```rust |
| 481 | /// # extern crate regex; use regex::Regex; |
| 482 | /// # fn main() { |
| 483 | /// let re = Regex::new(r"(?P<first>\w+)\s+(?P<second>\w+)").unwrap(); |
| 484 | /// let result = re.replace("deep fried", "${first}_$second"); |
| 485 | /// assert_eq!(result, "deep_fried"); |
| 486 | /// # } |
| 487 | /// ``` |
| 488 | /// |
| 489 | /// Without the curly braces, the capture group name `first_` would be |
| 490 | /// used, and since it doesn't exist, it would be replaced with the empty |
| 491 | /// string. |
| 492 | /// |
| 493 | /// Finally, sometimes you just want to replace a literal string with no |
| 494 | /// regard for capturing group expansion. This can be done by wrapping a |
| 495 | /// byte string with `NoExpand`: |
| 496 | /// |
| 497 | /// ```rust |
| 498 | /// # extern crate regex; use regex::Regex; |
| 499 | /// # fn main() { |
| 500 | /// use regex::NoExpand; |
| 501 | /// |
| 502 | /// let re = Regex::new(r"(?P<last>[^,\s]+),\s+(\S+)").unwrap(); |
| 503 | /// let result = re.replace("Springsteen, Bruce", NoExpand("$2 $last")); |
| 504 | /// assert_eq!(result, "$2 $last"); |
| 505 | /// # } |
| 506 | /// ``` |
| 507 | pub fn replace<'t, R: Replacer>( |
| 508 | &self, |
| 509 | text: &'t str, |
| 510 | rep: R, |
| 511 | ) -> Cow<'t, str> { |
| 512 | self.replacen(text, 1, rep) |
| 513 | } |
| 514 | |
| 515 | /// Replaces all non-overlapping matches in `text` with the replacement |
| 516 | /// provided. This is the same as calling `replacen` with `limit` set to |
| 517 | /// `0`. |
| 518 | /// |
| 519 | /// See the documentation for `replace` for details on how to access |
| 520 | /// capturing group matches in the replacement string. |
| 521 | pub fn replace_all<'t, R: Replacer>( |
| 522 | &self, |
| 523 | text: &'t str, |
| 524 | rep: R, |
| 525 | ) -> Cow<'t, str> { |
| 526 | self.replacen(text, 0, rep) |
| 527 | } |
| 528 | |
| 529 | /// Replaces at most `limit` non-overlapping matches in `text` with the |
| 530 | /// replacement provided. If `limit` is 0, then all non-overlapping matches |
| 531 | /// are replaced. |
| 532 | /// |
| 533 | /// See the documentation for `replace` for details on how to access |
| 534 | /// capturing group matches in the replacement string. |
| 535 | pub fn replacen<'t, R: Replacer>( |
| 536 | &self, |
| 537 | text: &'t str, |
| 538 | limit: usize, |
| 539 | mut rep: R, |
| 540 | ) -> Cow<'t, str> { |
| 541 | // If we know that the replacement doesn't have any capture expansions, |
| 542 | // then we can fast path. The fast path can make a tremendous |
| 543 | // difference: |
| 544 | // |
| 545 | // 1) We use `find_iter` instead of `captures_iter`. Not asking for |
| 546 | // captures generally makes the regex engines faster. |
| 547 | // 2) We don't need to look up all of the capture groups and do |
| 548 | // replacements inside the replacement string. We just push it |
| 549 | // at each match and be done with it. |
| 550 | if let Some(rep) = rep.no_expansion() { |
| 551 | let mut it = self.find_iter(text).enumerate().peekable(); |
| 552 | if it.peek().is_none() { |
| 553 | return Cow::Borrowed(text); |
| 554 | } |
| 555 | let mut new = String::with_capacity(text.len()); |
| 556 | let mut last_match = 0; |
| 557 | for (i, m) in it { |
| 558 | if limit > 0 && i >= limit { |
| 559 | break; |
| 560 | } |
| 561 | new.push_str(&text[last_match..m.start()]); |
| 562 | new.push_str(&rep); |
| 563 | last_match = m.end(); |
| 564 | } |
| 565 | new.push_str(&text[last_match..]); |
| 566 | return Cow::Owned(new); |
| 567 | } |
| 568 | |
| 569 | // The slower path, which we use if the replacement needs access to |
| 570 | // capture groups. |
| 571 | let mut it = self.captures_iter(text).enumerate().peekable(); |
| 572 | if it.peek().is_none() { |
| 573 | return Cow::Borrowed(text); |
| 574 | } |
| 575 | let mut new = String::with_capacity(text.len()); |
| 576 | let mut last_match = 0; |
| 577 | for (i, cap) in it { |
| 578 | if limit > 0 && i >= limit { |
| 579 | break; |
| 580 | } |
| 581 | // unwrap on 0 is OK because captures only reports matches |
| 582 | let m = cap.get(0).unwrap(); |
| 583 | new.push_str(&text[last_match..m.start()]); |
| 584 | rep.replace_append(&cap, &mut new); |
| 585 | last_match = m.end(); |
| 586 | } |
| 587 | new.push_str(&text[last_match..]); |
| 588 | Cow::Owned(new) |
| 589 | } |
| 590 | } |
| 591 | |
| 592 | /// Advanced or "lower level" search methods. |
| 593 | impl Regex { |
| 594 | /// Returns the end location of a match in the text given. |
| 595 | /// |
| 596 | /// This method may have the same performance characteristics as |
| 597 | /// `is_match`, except it provides an end location for a match. In |
| 598 | /// particular, the location returned *may be shorter* than the proper end |
| 599 | /// of the leftmost-first match. |
| 600 | /// |
| 601 | /// # Example |
| 602 | /// |
| 603 | /// Typically, `a+` would match the entire first sequence of `a` in some |
| 604 | /// text, but `shortest_match` can give up as soon as it sees the first |
| 605 | /// `a`. |
| 606 | /// |
| 607 | /// ```rust |
| 608 | /// # extern crate regex; use regex::Regex; |
| 609 | /// # fn main() { |
| 610 | /// let text = "aaaaa"; |
| 611 | /// let pos = Regex::new(r"a+").unwrap().shortest_match(text); |
| 612 | /// assert_eq!(pos, Some(1)); |
| 613 | /// # } |
| 614 | /// ``` |
| 615 | pub fn shortest_match(&self, text: &str) -> Option<usize> { |
| 616 | self.shortest_match_at(text, 0) |
| 617 | } |
| 618 | |
| 619 | /// Returns the same as shortest_match, but starts the search at the given |
| 620 | /// offset. |
| 621 | /// |
| 622 | /// The significance of the starting point is that it takes the surrounding |
| 623 | /// context into consideration. For example, the `\A` anchor can only |
| 624 | /// match when `start == 0`. |
| 625 | pub fn shortest_match_at( |
| 626 | &self, |
| 627 | text: &str, |
| 628 | start: usize, |
| 629 | ) -> Option<usize> { |
| 630 | self.0.searcher_str().shortest_match_at(text, start) |
| 631 | } |
| 632 | |
| 633 | /// Returns the same as is_match, but starts the search at the given |
| 634 | /// offset. |
| 635 | /// |
| 636 | /// The significance of the starting point is that it takes the surrounding |
| 637 | /// context into consideration. For example, the `\A` anchor can only |
| 638 | /// match when `start == 0`. |
| 639 | pub fn is_match_at(&self, text: &str, start: usize) -> bool { |
| 640 | self.shortest_match_at(text, start).is_some() |
| 641 | } |
| 642 | |
| 643 | /// Returns the same as find, but starts the search at the given |
| 644 | /// offset. |
| 645 | /// |
| 646 | /// The significance of the starting point is that it takes the surrounding |
| 647 | /// context into consideration. For example, the `\A` anchor can only |
| 648 | /// match when `start == 0`. |
| 649 | pub fn find_at<'t>( |
| 650 | &self, |
| 651 | text: &'t str, |
| 652 | start: usize, |
| 653 | ) -> Option<Match<'t>> { |
| 654 | self.0 |
| 655 | .searcher_str() |
| 656 | .find_at(text, start) |
| 657 | .map(|(s, e)| Match::new(text, s, e)) |
| 658 | } |
| 659 | |
| 660 | /// This is like `captures`, but uses |
| 661 | /// [`CaptureLocations`](struct.CaptureLocations.html) |
| 662 | /// instead of |
| 663 | /// [`Captures`](struct.Captures.html) in order to amortize allocations. |
| 664 | /// |
| 665 | /// To create a `CaptureLocations` value, use the |
| 666 | /// `Regex::capture_locations` method. |
| 667 | /// |
| 668 | /// This returns the overall match if this was successful, which is always |
| 669 | /// equivalence to the `0`th capture group. |
| 670 | pub fn captures_read<'t>( |
| 671 | &self, |
| 672 | locs: &mut CaptureLocations, |
| 673 | text: &'t str, |
| 674 | ) -> Option<Match<'t>> { |
| 675 | self.captures_read_at(locs, text, 0) |
| 676 | } |
| 677 | |
| 678 | /// Returns the same as captures, but starts the search at the given |
| 679 | /// offset and populates the capture locations given. |
| 680 | /// |
| 681 | /// The significance of the starting point is that it takes the surrounding |
| 682 | /// context into consideration. For example, the `\A` anchor can only |
| 683 | /// match when `start == 0`. |
| 684 | pub fn captures_read_at<'t>( |
| 685 | &self, |
| 686 | locs: &mut CaptureLocations, |
| 687 | text: &'t str, |
| 688 | start: usize, |
| 689 | ) -> Option<Match<'t>> { |
| 690 | self.0 |
| 691 | .searcher_str() |
| 692 | .captures_read_at(&mut locs.0, text, start) |
| 693 | .map(|(s, e)| Match::new(text, s, e)) |
| 694 | } |
| 695 | |
| 696 | /// An undocumented alias for `captures_read_at`. |
| 697 | /// |
| 698 | /// The `regex-capi` crate previously used this routine, so to avoid |
| 699 | /// breaking that crate, we continue to provide the name as an undocumented |
| 700 | /// alias. |
| 701 | #[doc(hidden)] |
| 702 | pub fn read_captures_at<'t>( |
| 703 | &self, |
| 704 | locs: &mut CaptureLocations, |
| 705 | text: &'t str, |
| 706 | start: usize, |
| 707 | ) -> Option<Match<'t>> { |
| 708 | self.captures_read_at(locs, text, start) |
| 709 | } |
| 710 | } |
| 711 | |
| 712 | /// Auxiliary methods. |
| 713 | impl Regex { |
| 714 | /// Returns the original string of this regex. |
| 715 | pub fn as_str(&self) -> &str { |
| 716 | &self.0.regex_strings()[0] |
| 717 | } |
| 718 | |
| 719 | /// Returns an iterator over the capture names. |
| 720 | pub fn capture_names(&self) -> CaptureNames { |
| 721 | CaptureNames(self.0.capture_names().iter()) |
| 722 | } |
| 723 | |
| 724 | /// Returns the number of captures. |
| 725 | pub fn captures_len(&self) -> usize { |
| 726 | self.0.capture_names().len() |
| 727 | } |
| 728 | |
| 729 | /// Returns an empty set of capture locations that can be reused in |
| 730 | /// multiple calls to `captures_read` or `captures_read_at`. |
| 731 | pub fn capture_locations(&self) -> CaptureLocations { |
| 732 | CaptureLocations(self.0.searcher_str().locations()) |
| 733 | } |
| 734 | |
| 735 | /// An alias for `capture_locations` to preserve backward compatibility. |
| 736 | /// |
| 737 | /// The `regex-capi` crate uses this method, so to avoid breaking that |
| 738 | /// crate, we continue to export it as an undocumented API. |
| 739 | #[doc(hidden)] |
| 740 | pub fn locations(&self) -> CaptureLocations { |
| 741 | CaptureLocations(self.0.searcher_str().locations()) |
| 742 | } |
| 743 | } |
| 744 | |
| 745 | /// An iterator over the names of all possible captures. |
| 746 | /// |
| 747 | /// `None` indicates an unnamed capture; the first element (capture 0, the |
| 748 | /// whole matched region) is always unnamed. |
| 749 | /// |
| 750 | /// `'r` is the lifetime of the compiled regular expression. |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 751 | #[derive(Clone, Debug)] |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 752 | pub struct CaptureNames<'r>(::std::slice::Iter<'r, Option<String>>); |
| 753 | |
| 754 | impl<'r> Iterator for CaptureNames<'r> { |
| 755 | type Item = Option<&'r str>; |
| 756 | |
| 757 | fn next(&mut self) -> Option<Option<&'r str>> { |
| 758 | self.0 |
| 759 | .next() |
| 760 | .as_ref() |
| 761 | .map(|slot| slot.as_ref().map(|name| name.as_ref())) |
| 762 | } |
| 763 | |
| 764 | fn size_hint(&self) -> (usize, Option<usize>) { |
| 765 | self.0.size_hint() |
| 766 | } |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 767 | |
| 768 | fn count(self) -> usize { |
| 769 | self.0.count() |
| 770 | } |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 771 | } |
| 772 | |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 773 | impl<'r> ExactSizeIterator for CaptureNames<'r> {} |
| 774 | |
| 775 | impl<'r> FusedIterator for CaptureNames<'r> {} |
| 776 | |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 777 | /// Yields all substrings delimited by a regular expression match. |
| 778 | /// |
| 779 | /// `'r` is the lifetime of the compiled regular expression and `'t` is the |
| 780 | /// lifetime of the string being split. |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 781 | #[derive(Debug)] |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 782 | pub struct Split<'r, 't> { |
| 783 | finder: Matches<'r, 't>, |
| 784 | last: usize, |
| 785 | } |
| 786 | |
| 787 | impl<'r, 't> Iterator for Split<'r, 't> { |
| 788 | type Item = &'t str; |
| 789 | |
| 790 | fn next(&mut self) -> Option<&'t str> { |
| 791 | let text = self.finder.0.text(); |
| 792 | match self.finder.next() { |
| 793 | None => { |
| 794 | if self.last > text.len() { |
| 795 | None |
| 796 | } else { |
| 797 | let s = &text[self.last..]; |
| 798 | self.last = text.len() + 1; // Next call will return None |
| 799 | Some(s) |
| 800 | } |
| 801 | } |
| 802 | Some(m) => { |
| 803 | let matched = &text[self.last..m.start()]; |
| 804 | self.last = m.end(); |
| 805 | Some(matched) |
| 806 | } |
| 807 | } |
| 808 | } |
| 809 | } |
| 810 | |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 811 | impl<'r, 't> FusedIterator for Split<'r, 't> {} |
| 812 | |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 813 | /// Yields at most `N` substrings delimited by a regular expression match. |
| 814 | /// |
| 815 | /// The last substring will be whatever remains after splitting. |
| 816 | /// |
| 817 | /// `'r` is the lifetime of the compiled regular expression and `'t` is the |
| 818 | /// lifetime of the string being split. |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 819 | #[derive(Debug)] |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 820 | pub struct SplitN<'r, 't> { |
| 821 | splits: Split<'r, 't>, |
| 822 | n: usize, |
| 823 | } |
| 824 | |
| 825 | impl<'r, 't> Iterator for SplitN<'r, 't> { |
| 826 | type Item = &'t str; |
| 827 | |
| 828 | fn next(&mut self) -> Option<&'t str> { |
| 829 | if self.n == 0 { |
| 830 | return None; |
| 831 | } |
| 832 | |
| 833 | self.n -= 1; |
| 834 | if self.n > 0 { |
| 835 | return self.splits.next(); |
| 836 | } |
| 837 | |
| 838 | let text = self.splits.finder.0.text(); |
| 839 | if self.splits.last > text.len() { |
| 840 | // We've already returned all substrings. |
| 841 | None |
| 842 | } else { |
| 843 | // self.n == 0, so future calls will return None immediately |
| 844 | Some(&text[self.splits.last..]) |
| 845 | } |
| 846 | } |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 847 | |
| 848 | fn size_hint(&self) -> (usize, Option<usize>) { |
| 849 | (0, Some(self.n)) |
| 850 | } |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 851 | } |
| 852 | |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 853 | impl<'r, 't> FusedIterator for SplitN<'r, 't> {} |
| 854 | |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 855 | /// CaptureLocations is a low level representation of the raw offsets of each |
| 856 | /// submatch. |
| 857 | /// |
| 858 | /// You can think of this as a lower level |
| 859 | /// [`Captures`](struct.Captures.html), where this type does not support |
| 860 | /// named capturing groups directly and it does not borrow the text that these |
| 861 | /// offsets were matched on. |
| 862 | /// |
| 863 | /// Primarily, this type is useful when using the lower level `Regex` APIs |
| 864 | /// such as `read_captures`, which permits amortizing the allocation in which |
| 865 | /// capture match locations are stored. |
| 866 | /// |
| 867 | /// In order to build a value of this type, you'll need to call the |
| 868 | /// `capture_locations` method on the `Regex` being used to execute the search. |
| 869 | /// The value returned can then be reused in subsequent searches. |
| 870 | #[derive(Clone, Debug)] |
| 871 | pub struct CaptureLocations(re_trait::Locations); |
| 872 | |
| 873 | /// A type alias for `CaptureLocations` for backwards compatibility. |
| 874 | /// |
| 875 | /// Previously, we exported `CaptureLocations` as `Locations` in an |
| 876 | /// undocumented API. To prevent breaking that code (e.g., in `regex-capi`), |
| 877 | /// we continue re-exporting the same undocumented API. |
| 878 | #[doc(hidden)] |
| 879 | pub type Locations = CaptureLocations; |
| 880 | |
| 881 | impl CaptureLocations { |
| 882 | /// Returns the start and end positions of the Nth capture group. Returns |
| 883 | /// `None` if `i` is not a valid capture group or if the capture group did |
| 884 | /// not match anything. The positions returned are *always* byte indices |
| 885 | /// with respect to the original string matched. |
| 886 | #[inline] |
| 887 | pub fn get(&self, i: usize) -> Option<(usize, usize)> { |
| 888 | self.0.pos(i) |
| 889 | } |
| 890 | |
| 891 | /// Returns the total number of capturing groups. |
| 892 | /// |
| 893 | /// This is always at least `1` since every regex has at least `1` |
| 894 | /// capturing group that corresponds to the entire match. |
| 895 | #[inline] |
| 896 | pub fn len(&self) -> usize { |
| 897 | self.0.len() |
| 898 | } |
| 899 | |
| 900 | /// An alias for the `get` method for backwards compatibility. |
| 901 | /// |
| 902 | /// Previously, we exported `get` as `pos` in an undocumented API. To |
| 903 | /// prevent breaking that code (e.g., in `regex-capi`), we continue |
| 904 | /// re-exporting the same undocumented API. |
| 905 | #[doc(hidden)] |
| 906 | #[inline] |
| 907 | pub fn pos(&self, i: usize) -> Option<(usize, usize)> { |
| 908 | self.get(i) |
| 909 | } |
| 910 | } |
| 911 | |
| 912 | /// Captures represents a group of captured strings for a single match. |
| 913 | /// |
| 914 | /// The 0th capture always corresponds to the entire match. Each subsequent |
| 915 | /// index corresponds to the next capture group in the regex. If a capture |
| 916 | /// group is named, then the matched string is *also* available via the `name` |
| 917 | /// method. (Note that the 0th capture is always unnamed and so must be |
| 918 | /// accessed with the `get` method.) |
| 919 | /// |
| 920 | /// Positions returned from a capture group are always byte indices. |
| 921 | /// |
| 922 | /// `'t` is the lifetime of the matched text. |
| 923 | pub struct Captures<'t> { |
| 924 | text: &'t str, |
| 925 | locs: re_trait::Locations, |
| 926 | named_groups: Arc<HashMap<String, usize>>, |
| 927 | } |
| 928 | |
| 929 | impl<'t> Captures<'t> { |
| 930 | /// Returns the match associated with the capture group at index `i`. If |
| 931 | /// `i` does not correspond to a capture group, or if the capture group |
| 932 | /// did not participate in the match, then `None` is returned. |
| 933 | /// |
| 934 | /// # Examples |
| 935 | /// |
| 936 | /// Get the text of the match with a default of an empty string if this |
| 937 | /// group didn't participate in the match: |
| 938 | /// |
| 939 | /// ```rust |
| 940 | /// # use regex::Regex; |
| 941 | /// let re = Regex::new(r"[a-z]+(?:([0-9]+)|([A-Z]+))").unwrap(); |
| 942 | /// let caps = re.captures("abc123").unwrap(); |
| 943 | /// |
| 944 | /// let text1 = caps.get(1).map_or("", |m| m.as_str()); |
| 945 | /// let text2 = caps.get(2).map_or("", |m| m.as_str()); |
| 946 | /// assert_eq!(text1, "123"); |
| 947 | /// assert_eq!(text2, ""); |
| 948 | /// ``` |
| 949 | pub fn get(&self, i: usize) -> Option<Match<'t>> { |
| 950 | self.locs.pos(i).map(|(s, e)| Match::new(self.text, s, e)) |
| 951 | } |
| 952 | |
| 953 | /// Returns the match for the capture group named `name`. If `name` isn't a |
| 954 | /// valid capture group or didn't match anything, then `None` is returned. |
| 955 | pub fn name(&self, name: &str) -> Option<Match<'t>> { |
| 956 | self.named_groups.get(name).and_then(|&i| self.get(i)) |
| 957 | } |
| 958 | |
| 959 | /// An iterator that yields all capturing matches in the order in which |
| 960 | /// they appear in the regex. If a particular capture group didn't |
| 961 | /// participate in the match, then `None` is yielded for that capture. |
| 962 | /// |
| 963 | /// The first match always corresponds to the overall match of the regex. |
| 964 | pub fn iter<'c>(&'c self) -> SubCaptureMatches<'c, 't> { |
| 965 | SubCaptureMatches { caps: self, it: self.locs.iter() } |
| 966 | } |
| 967 | |
| 968 | /// Expands all instances of `$name` in `replacement` to the corresponding |
| 969 | /// capture group `name`, and writes them to the `dst` buffer given. |
| 970 | /// |
Chih-Hung Hsieh | 849e445 | 2020-10-26 13:16:47 -0700 | [diff] [blame] | 971 | /// `name` may be an integer corresponding to the index of the capture |
| 972 | /// group (counted by order of opening parenthesis where `0` is the |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 973 | /// entire match) or it can be a name (consisting of letters, digits or |
| 974 | /// underscores) corresponding to a named capture group. |
| 975 | /// |
| 976 | /// If `name` isn't a valid capture group (whether the name doesn't exist |
| 977 | /// or isn't a valid index), then it is replaced with the empty string. |
| 978 | /// |
Chih-Hung Hsieh | 849e445 | 2020-10-26 13:16:47 -0700 | [diff] [blame] | 979 | /// The longest possible name consisting of the characters `[_0-9A-Za-z]` |
| 980 | /// is used. e.g., `$1a` looks up the capture group named `1a` and not the |
| 981 | /// capture group at index `1`. To exert more precise control over the |
| 982 | /// name, or to refer to a capture group name that uses characters outside |
| 983 | /// of `[_0-9A-Za-z]`, use braces, e.g., `${1}a` or `${foo[bar].baz}`. When |
| 984 | /// using braces, any sequence of characters is permitted. If the sequence |
| 985 | /// does not refer to a capture group name in the corresponding regex, then |
| 986 | /// it is replaced with an empty string. |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 987 | /// |
| 988 | /// To write a literal `$` use `$$`. |
| 989 | pub fn expand(&self, replacement: &str, dst: &mut String) { |
| 990 | expand_str(self, replacement, dst) |
| 991 | } |
| 992 | |
| 993 | /// Returns the number of captured groups. |
| 994 | /// |
| 995 | /// This is always at least `1`, since every regex has at least one capture |
| 996 | /// group that corresponds to the full match. |
| 997 | #[inline] |
| 998 | pub fn len(&self) -> usize { |
| 999 | self.locs.len() |
| 1000 | } |
| 1001 | } |
| 1002 | |
| 1003 | impl<'t> fmt::Debug for Captures<'t> { |
| 1004 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
| 1005 | f.debug_tuple("Captures").field(&CapturesDebug(self)).finish() |
| 1006 | } |
| 1007 | } |
| 1008 | |
| 1009 | struct CapturesDebug<'c, 't: 'c>(&'c Captures<'t>); |
| 1010 | |
| 1011 | impl<'c, 't> fmt::Debug for CapturesDebug<'c, 't> { |
| 1012 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
| 1013 | // We'd like to show something nice here, even if it means an |
| 1014 | // allocation to build a reverse index. |
| 1015 | let slot_to_name: HashMap<&usize, &String> = |
| 1016 | self.0.named_groups.iter().map(|(a, b)| (b, a)).collect(); |
| 1017 | let mut map = f.debug_map(); |
| 1018 | for (slot, m) in self.0.locs.iter().enumerate() { |
| 1019 | let m = m.map(|(s, e)| &self.0.text[s..e]); |
| 1020 | if let Some(name) = slot_to_name.get(&slot) { |
| 1021 | map.entry(&name, &m); |
| 1022 | } else { |
| 1023 | map.entry(&slot, &m); |
| 1024 | } |
| 1025 | } |
| 1026 | map.finish() |
| 1027 | } |
| 1028 | } |
| 1029 | |
| 1030 | /// Get a group by index. |
| 1031 | /// |
| 1032 | /// `'t` is the lifetime of the matched text. |
| 1033 | /// |
| 1034 | /// The text can't outlive the `Captures` object if this method is |
| 1035 | /// used, because of how `Index` is defined (normally `a[i]` is part |
| 1036 | /// of `a` and can't outlive it); to do that, use `get()` instead. |
| 1037 | /// |
| 1038 | /// # Panics |
| 1039 | /// |
| 1040 | /// If there is no group at the given index. |
| 1041 | impl<'t> Index<usize> for Captures<'t> { |
| 1042 | type Output = str; |
| 1043 | |
| 1044 | fn index(&self, i: usize) -> &str { |
| 1045 | self.get(i) |
| 1046 | .map(|m| m.as_str()) |
| 1047 | .unwrap_or_else(|| panic!("no group at index '{}'", i)) |
| 1048 | } |
| 1049 | } |
| 1050 | |
| 1051 | /// Get a group by name. |
| 1052 | /// |
| 1053 | /// `'t` is the lifetime of the matched text and `'i` is the lifetime |
| 1054 | /// of the group name (the index). |
| 1055 | /// |
| 1056 | /// The text can't outlive the `Captures` object if this method is |
| 1057 | /// used, because of how `Index` is defined (normally `a[i]` is part |
| 1058 | /// of `a` and can't outlive it); to do that, use `name` instead. |
| 1059 | /// |
| 1060 | /// # Panics |
| 1061 | /// |
| 1062 | /// If there is no group named by the given value. |
| 1063 | impl<'t, 'i> Index<&'i str> for Captures<'t> { |
| 1064 | type Output = str; |
| 1065 | |
| 1066 | fn index<'a>(&'a self, name: &'i str) -> &'a str { |
| 1067 | self.name(name) |
| 1068 | .map(|m| m.as_str()) |
| 1069 | .unwrap_or_else(|| panic!("no group named '{}'", name)) |
| 1070 | } |
| 1071 | } |
| 1072 | |
| 1073 | /// An iterator that yields all capturing matches in the order in which they |
| 1074 | /// appear in the regex. |
| 1075 | /// |
| 1076 | /// If a particular capture group didn't participate in the match, then `None` |
| 1077 | /// is yielded for that capture. The first match always corresponds to the |
| 1078 | /// overall match of the regex. |
| 1079 | /// |
| 1080 | /// The lifetime `'c` corresponds to the lifetime of the `Captures` value, and |
| 1081 | /// the lifetime `'t` corresponds to the originally matched text. |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 1082 | #[derive(Clone, Debug)] |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1083 | pub struct SubCaptureMatches<'c, 't: 'c> { |
| 1084 | caps: &'c Captures<'t>, |
| 1085 | it: SubCapturesPosIter<'c>, |
| 1086 | } |
| 1087 | |
| 1088 | impl<'c, 't> Iterator for SubCaptureMatches<'c, 't> { |
| 1089 | type Item = Option<Match<'t>>; |
| 1090 | |
| 1091 | fn next(&mut self) -> Option<Option<Match<'t>>> { |
| 1092 | self.it |
| 1093 | .next() |
| 1094 | .map(|cap| cap.map(|(s, e)| Match::new(self.caps.text, s, e))) |
| 1095 | } |
| 1096 | } |
| 1097 | |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 1098 | impl<'c, 't> FusedIterator for SubCaptureMatches<'c, 't> {} |
| 1099 | |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1100 | /// An iterator that yields all non-overlapping capture groups matching a |
| 1101 | /// particular regular expression. |
| 1102 | /// |
| 1103 | /// The iterator stops when no more matches can be found. |
| 1104 | /// |
| 1105 | /// `'r` is the lifetime of the compiled regular expression and `'t` is the |
| 1106 | /// lifetime of the matched string. |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 1107 | #[derive(Debug)] |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1108 | pub struct CaptureMatches<'r, 't>( |
| 1109 | re_trait::CaptureMatches<'t, ExecNoSyncStr<'r>>, |
| 1110 | ); |
| 1111 | |
| 1112 | impl<'r, 't> Iterator for CaptureMatches<'r, 't> { |
| 1113 | type Item = Captures<'t>; |
| 1114 | |
| 1115 | fn next(&mut self) -> Option<Captures<'t>> { |
| 1116 | self.0.next().map(|locs| Captures { |
| 1117 | text: self.0.text(), |
| 1118 | locs: locs, |
| 1119 | named_groups: self.0.regex().capture_name_idx().clone(), |
| 1120 | }) |
| 1121 | } |
| 1122 | } |
| 1123 | |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 1124 | impl<'r, 't> FusedIterator for CaptureMatches<'r, 't> {} |
| 1125 | |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1126 | /// An iterator over all non-overlapping matches for a particular string. |
| 1127 | /// |
| 1128 | /// The iterator yields a `Match` value. The iterator stops when no more |
| 1129 | /// matches can be found. |
| 1130 | /// |
| 1131 | /// `'r` is the lifetime of the compiled regular expression and `'t` is the |
| 1132 | /// lifetime of the matched string. |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 1133 | #[derive(Debug)] |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1134 | pub struct Matches<'r, 't>(re_trait::Matches<'t, ExecNoSyncStr<'r>>); |
| 1135 | |
| 1136 | impl<'r, 't> Iterator for Matches<'r, 't> { |
| 1137 | type Item = Match<'t>; |
| 1138 | |
| 1139 | fn next(&mut self) -> Option<Match<'t>> { |
| 1140 | let text = self.0.text(); |
| 1141 | self.0.next().map(|(s, e)| Match::new(text, s, e)) |
| 1142 | } |
| 1143 | } |
| 1144 | |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 1145 | impl<'r, 't> FusedIterator for Matches<'r, 't> {} |
| 1146 | |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1147 | /// Replacer describes types that can be used to replace matches in a string. |
| 1148 | /// |
| 1149 | /// In general, users of this crate shouldn't need to implement this trait, |
Elliott Hughes | ffb6030 | 2021-04-01 17:11:40 -0700 | [diff] [blame^] | 1150 | /// since implementations are already provided for `&str` along with other |
| 1151 | /// variants of string types and `FnMut(&Captures) -> String` (or any |
| 1152 | /// `FnMut(&Captures) -> T` where `T: AsRef<str>`), which covers most use cases. |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1153 | pub trait Replacer { |
| 1154 | /// Appends text to `dst` to replace the current match. |
| 1155 | /// |
| 1156 | /// The current match is represented by `caps`, which is guaranteed to |
| 1157 | /// have a match at capture group `0`. |
| 1158 | /// |
| 1159 | /// For example, a no-op replacement would be |
Chih-Hung Hsieh | 849e445 | 2020-10-26 13:16:47 -0700 | [diff] [blame] | 1160 | /// `dst.push_str(caps.get(0).unwrap().as_str())`. |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1161 | fn replace_append(&mut self, caps: &Captures, dst: &mut String); |
| 1162 | |
| 1163 | /// Return a fixed unchanging replacement string. |
| 1164 | /// |
| 1165 | /// When doing replacements, if access to `Captures` is not needed (e.g., |
| 1166 | /// the replacement byte string does not need `$` expansion), then it can |
| 1167 | /// be beneficial to avoid finding sub-captures. |
| 1168 | /// |
| 1169 | /// In general, this is called once for every call to `replacen`. |
| 1170 | fn no_expansion<'r>(&'r mut self) -> Option<Cow<'r, str>> { |
| 1171 | None |
| 1172 | } |
| 1173 | |
| 1174 | /// Return a `Replacer` that borrows and wraps this `Replacer`. |
| 1175 | /// |
| 1176 | /// This is useful when you want to take a generic `Replacer` (which might |
| 1177 | /// not be cloneable) and use it without consuming it, so it can be used |
| 1178 | /// more than once. |
| 1179 | /// |
| 1180 | /// # Example |
| 1181 | /// |
| 1182 | /// ``` |
| 1183 | /// use regex::{Regex, Replacer}; |
| 1184 | /// |
| 1185 | /// fn replace_all_twice<R: Replacer>( |
| 1186 | /// re: Regex, |
| 1187 | /// src: &str, |
| 1188 | /// mut rep: R, |
| 1189 | /// ) -> String { |
| 1190 | /// let dst = re.replace_all(src, rep.by_ref()); |
| 1191 | /// let dst = re.replace_all(&dst, rep.by_ref()); |
| 1192 | /// dst.into_owned() |
| 1193 | /// } |
| 1194 | /// ``` |
| 1195 | fn by_ref<'r>(&'r mut self) -> ReplacerRef<'r, Self> { |
| 1196 | ReplacerRef(self) |
| 1197 | } |
| 1198 | } |
| 1199 | |
| 1200 | /// By-reference adaptor for a `Replacer` |
| 1201 | /// |
| 1202 | /// Returned by [`Replacer::by_ref`](trait.Replacer.html#method.by_ref). |
| 1203 | #[derive(Debug)] |
| 1204 | pub struct ReplacerRef<'a, R: ?Sized + 'a>(&'a mut R); |
| 1205 | |
| 1206 | impl<'a, R: Replacer + ?Sized + 'a> Replacer for ReplacerRef<'a, R> { |
| 1207 | fn replace_append(&mut self, caps: &Captures, dst: &mut String) { |
| 1208 | self.0.replace_append(caps, dst) |
| 1209 | } |
| 1210 | fn no_expansion(&mut self) -> Option<Cow<str>> { |
| 1211 | self.0.no_expansion() |
| 1212 | } |
| 1213 | } |
| 1214 | |
| 1215 | impl<'a> Replacer for &'a str { |
| 1216 | fn replace_append(&mut self, caps: &Captures, dst: &mut String) { |
| 1217 | caps.expand(*self, dst); |
| 1218 | } |
| 1219 | |
| 1220 | fn no_expansion(&mut self) -> Option<Cow<str>> { |
Elliott Hughes | ffb6030 | 2021-04-01 17:11:40 -0700 | [diff] [blame^] | 1221 | no_expansion(self) |
| 1222 | } |
| 1223 | } |
| 1224 | |
| 1225 | impl<'a> Replacer for &'a String { |
| 1226 | fn replace_append(&mut self, caps: &Captures, dst: &mut String) { |
| 1227 | self.as_str().replace_append(caps, dst) |
| 1228 | } |
| 1229 | |
| 1230 | fn no_expansion(&mut self) -> Option<Cow<str>> { |
| 1231 | no_expansion(self) |
| 1232 | } |
| 1233 | } |
| 1234 | |
| 1235 | impl Replacer for String { |
| 1236 | fn replace_append(&mut self, caps: &Captures, dst: &mut String) { |
| 1237 | self.as_str().replace_append(caps, dst) |
| 1238 | } |
| 1239 | |
| 1240 | fn no_expansion(&mut self) -> Option<Cow<str>> { |
| 1241 | no_expansion(self) |
| 1242 | } |
| 1243 | } |
| 1244 | |
| 1245 | impl<'a> Replacer for Cow<'a, str> { |
| 1246 | fn replace_append(&mut self, caps: &Captures, dst: &mut String) { |
| 1247 | self.as_ref().replace_append(caps, dst) |
| 1248 | } |
| 1249 | |
| 1250 | fn no_expansion(&mut self) -> Option<Cow<str>> { |
| 1251 | no_expansion(self) |
| 1252 | } |
| 1253 | } |
| 1254 | |
| 1255 | impl<'a> Replacer for &'a Cow<'a, str> { |
| 1256 | fn replace_append(&mut self, caps: &Captures, dst: &mut String) { |
| 1257 | self.as_ref().replace_append(caps, dst) |
| 1258 | } |
| 1259 | |
| 1260 | fn no_expansion(&mut self) -> Option<Cow<str>> { |
| 1261 | no_expansion(self) |
| 1262 | } |
| 1263 | } |
| 1264 | |
| 1265 | fn no_expansion<T: AsRef<str>>(t: &T) -> Option<Cow<str>> { |
| 1266 | let s = t.as_ref(); |
| 1267 | match find_byte(b'$', s.as_bytes()) { |
| 1268 | Some(_) => None, |
| 1269 | None => Some(Cow::Borrowed(s)), |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1270 | } |
| 1271 | } |
| 1272 | |
| 1273 | impl<F, T> Replacer for F |
| 1274 | where |
| 1275 | F: FnMut(&Captures) -> T, |
| 1276 | T: AsRef<str>, |
| 1277 | { |
| 1278 | fn replace_append(&mut self, caps: &Captures, dst: &mut String) { |
| 1279 | dst.push_str((*self)(caps).as_ref()); |
| 1280 | } |
| 1281 | } |
| 1282 | |
| 1283 | /// `NoExpand` indicates literal string replacement. |
| 1284 | /// |
| 1285 | /// It can be used with `replace` and `replace_all` to do a literal string |
| 1286 | /// replacement without expanding `$name` to their corresponding capture |
| 1287 | /// groups. This can be both convenient (to avoid escaping `$`, for example) |
| 1288 | /// and performant (since capture groups don't need to be found). |
| 1289 | /// |
| 1290 | /// `'t` is the lifetime of the literal text. |
Haibo Huang | 47619dd | 2021-01-08 17:05:43 -0800 | [diff] [blame] | 1291 | #[derive(Clone, Debug)] |
Chih-Hung Hsieh | e42c505 | 2020-04-16 10:44:21 -0700 | [diff] [blame] | 1292 | pub struct NoExpand<'t>(pub &'t str); |
| 1293 | |
| 1294 | impl<'t> Replacer for NoExpand<'t> { |
| 1295 | fn replace_append(&mut self, _: &Captures, dst: &mut String) { |
| 1296 | dst.push_str(self.0); |
| 1297 | } |
| 1298 | |
| 1299 | fn no_expansion(&mut self) -> Option<Cow<str>> { |
| 1300 | Some(Cow::Borrowed(self.0)) |
| 1301 | } |
| 1302 | } |