| 1 |
use alloc::string::String; |
| 2 |
|
| 3 |
use regex_automata::{meta, Input, PatternID, PatternSet, PatternSetIter}; |
| 4 |
|
| 5 |
use crate::{bytes::RegexSetBuilder, Error}; |
| 6 |
|
| 7 |
/// Match multiple, possibly overlapping, regexes in a single search. |
| 8 |
/// |
| 9 |
/// A regex set corresponds to the union of zero or more regular expressions. |
| 10 |
/// That is, a regex set will match a haystack when at least one of its |
| 11 |
/// constituent regexes matches. A regex set as its formulated here provides a |
| 12 |
/// touch more power: it will also report *which* regular expressions in the |
| 13 |
/// set match. Indeed, this is the key difference between regex sets and a |
| 14 |
/// single `Regex` with many alternates, since only one alternate can match at |
| 15 |
/// a time. |
| 16 |
/// |
| 17 |
/// For example, consider regular expressions to match email addresses and |
| 18 |
/// domains: `[a-z]+@[a-z]+\.(com|org|net)` and `[a-z]+\.(com|org|net)`. If a |
| 19 |
/// regex set is constructed from those regexes, then searching the haystack |
| 20 |
/// `[email protected]` will report both regexes as matching. Of course, one |
| 21 |
/// could accomplish this by compiling each regex on its own and doing two |
| 22 |
/// searches over the haystack. The key advantage of using a regex set is |
| 23 |
/// that it will report the matching regexes using a *single pass through the |
| 24 |
/// haystack*. If one has hundreds or thousands of regexes to match repeatedly |
| 25 |
/// (like a URL router for a complex web application or a user agent matcher), |
| 26 |
/// then a regex set *can* realize huge performance gains. |
| 27 |
/// |
| 28 |
/// Unlike the top-level [`RegexSet`](crate::RegexSet), this `RegexSet` |
| 29 |
/// searches haystacks with type `&[u8]` instead of `&str`. Consequently, this |
| 30 |
/// `RegexSet` is permitted to match invalid UTF-8. |
| 31 |
/// |
| 32 |
/// # Limitations |
| 33 |
/// |
| 34 |
/// Regex sets are limited to answering the following two questions: |
| 35 |
/// |
| 36 |
/// 1. Does any regex in the set match? |
| 37 |
/// 2. If so, which regexes in the set match? |
| 38 |
/// |
| 39 |
/// As with the main [`Regex`][crate::bytes::Regex] type, it is cheaper to ask |
| 40 |
/// (1) instead of (2) since the matching engines can stop after the first |
| 41 |
/// match is found. |
| 42 |
/// |
| 43 |
/// You cannot directly extract [`Match`][crate::bytes::Match] or |
| 44 |
/// [`Captures`][crate::bytes::Captures] objects from a regex set. If you need |
| 45 |
/// these operations, the recommended approach is to compile each pattern in |
| 46 |
/// the set independently and scan the exact same haystack a second time with |
| 47 |
/// those independently compiled patterns: |
| 48 |
/// |
| 49 |
/// ``` |
| 50 |
/// use regex::bytes::{Regex, RegexSet}; |
| 51 |
/// |
| 52 |
/// let patterns = ["foo", "bar"]; |
| 53 |
/// // Both patterns will match different ranges of this string. |
| 54 |
/// let hay = b"barfoo"; |
| 55 |
/// |
| 56 |
/// // Compile a set matching any of our patterns. |
| 57 |
/// let set = RegexSet::new(patterns).unwrap(); |
| 58 |
/// // Compile each pattern independently. |
| 59 |
/// let regexes: Vec<_> = set |
| 60 |
/// .patterns() |
| 61 |
/// .iter() |
| 62 |
/// .map(|pat| Regex::new(pat).unwrap()) |
| 63 |
/// .collect(); |
| 64 |
/// |
| 65 |
/// // Match against the whole set first and identify the individual |
| 66 |
/// // matching patterns. |
| 67 |
/// let matches: Vec<&[u8]> = set |
| 68 |
/// .matches(hay) |
| 69 |
/// .into_iter() |
| 70 |
/// // Dereference the match index to get the corresponding |
| 71 |
/// // compiled pattern. |
| 72 |
/// .map(|index| ®exes[index]) |
| 73 |
/// // To get match locations or any other info, we then have to search the |
| 74 |
/// // exact same haystack again, using our separately-compiled pattern. |
| 75 |
/// .map(|re| re.find(hay).unwrap().as_bytes()) |
| 76 |
/// .collect(); |
| 77 |
/// |
| 78 |
/// // Matches arrive in the order the constituent patterns were declared, |
| 79 |
/// // not the order they appear in the haystack. |
| 80 |
/// assert_eq!(vec![&b"foo"[..], &b"bar"[..]], matches); |
| 81 |
/// ``` |
| 82 |
/// |
| 83 |
/// # Performance |
| 84 |
/// |
| 85 |
/// A `RegexSet` has the same performance characteristics as `Regex`. Namely, |
| 86 |
/// search takes `O(m * n)` time, where `m` is proportional to the size of the |
| 87 |
/// regex set and `n` is proportional to the length of the haystack. |
| 88 |
/// |
| 89 |
/// # Trait implementations |
| 90 |
/// |
| 91 |
/// The `Default` trait is implemented for `RegexSet`. The default value |
| 92 |
/// is an empty set. An empty set can also be explicitly constructed via |
| 93 |
/// [`RegexSet::empty`]. |
| 94 |
/// |
| 95 |
/// # Example |
| 96 |
/// |
| 97 |
/// This shows how the above two regexes (for matching email addresses and |
| 98 |
/// domains) might work: |
| 99 |
/// |
| 100 |
/// ``` |
| 101 |
/// use regex::bytes::RegexSet; |
| 102 |
/// |
| 103 |
/// let set = RegexSet::new(&[ |
| 104 |
/// r"[a-z]+@[a-z]+\.(com|org|net)", |
| 105 |
/// r"[a-z]+\.(com|org|net)", |
| 106 |
/// ]).unwrap(); |
| 107 |
/// |
| 108 |
/// // Ask whether any regexes in the set match. |
| 109 |
/// assert!(set.is_match(b"[email protected]")); |
| 110 |
/// |
| 111 |
/// // Identify which regexes in the set match. |
| 112 |
/// let matches: Vec<_> = set.matches(b"[email protected]").into_iter().collect(); |
| 113 |
/// assert_eq!(vec![0, 1], matches); |
| 114 |
/// |
| 115 |
/// // Try again, but with a haystack that only matches one of the regexes. |
| 116 |
/// let matches: Vec<_> = set.matches(b"example.com").into_iter().collect(); |
| 117 |
/// assert_eq!(vec![1], matches); |
| 118 |
/// |
| 119 |
/// // Try again, but with a haystack that doesn't match any regex in the set. |
| 120 |
/// let matches: Vec<_> = set.matches(b"example").into_iter().collect(); |
| 121 |
/// assert!(matches.is_empty()); |
| 122 |
/// ``` |
| 123 |
/// |
| 124 |
/// Note that it would be possible to adapt the above example to using `Regex` |
| 125 |
/// with an expression like: |
| 126 |
/// |
| 127 |
/// ```text |
| 128 |
/// (?P<email>[a-z]+@(?P<email_domain>[a-z]+[.](com|org|net)))|(?P<domain>[a-z]+[.](com|org|net)) |
| 129 |
/// ``` |
| 130 |
/// |
| 131 |
/// After a match, one could then inspect the capture groups to figure out |
| 132 |
/// which alternates matched. The problem is that it is hard to make this |
| 133 |
/// approach scale when there are many regexes since the overlap between each |
| 134 |
/// alternate isn't always obvious to reason about. |
| 135 |
#[derive(Clone)] |
| 136 |
pub struct RegexSet { |
| 137 |
pub(crate) meta: meta::Regex, |
| 138 |
pub(crate) patterns: alloc::sync::Arc<[String]>, |
| 139 |
} |
| 140 |
|
| 141 |
impl RegexSet { |
| 142 |
/// Create a new regex set with the given regular expressions. |
| 143 |
/// |
| 144 |
/// This takes an iterator of `S`, where `S` is something that can produce |
| 145 |
/// a `&str`. If any of the strings in the iterator are not valid regular |
| 146 |
/// expressions, then an error is returned. |
| 147 |
/// |
| 148 |
/// # Example |
| 149 |
/// |
| 150 |
/// Create a new regex set from an iterator of strings: |
| 151 |
/// |
| 152 |
/// ``` |
| 153 |
/// use regex::bytes::RegexSet; |
| 154 |
/// |
| 155 |
/// let set = RegexSet::new([r"\w+", r"\d+"]).unwrap(); |
| 156 |
/// assert!(set.is_match(b"foo")); |
| 157 |
/// ``` |
| 158 |
pub fn new<I, S>(exprs: I) -> Result<RegexSet, Error> |
| 159 |
where |
| 160 |
S: AsRef<str>, |
| 161 |
I: IntoIterator<Item = S>, |
| 162 |
{ |
| 163 |
RegexSetBuilder::new(exprs).build() |
| 164 |
} |
| 165 |
|
| 166 |
/// Create a new empty regex set. |
| 167 |
/// |
| 168 |
/// An empty regex never matches anything. |
| 169 |
/// |
| 170 |
/// This is a convenience function for `RegexSet::new([])`, but doesn't |
| 171 |
/// require one to specify the type of the input. |
| 172 |
/// |
| 173 |
/// # Example |
| 174 |
/// |
| 175 |
/// ``` |
| 176 |
/// use regex::bytes::RegexSet; |
| 177 |
/// |
| 178 |
/// let set = RegexSet::empty(); |
| 179 |
/// assert!(set.is_empty()); |
| 180 |
/// // an empty set matches nothing |
| 181 |
/// assert!(!set.is_match(b"")); |
| 182 |
/// ``` |
| 183 |
pub fn empty() -> RegexSet { |
| 184 |
let empty: [&str; 0] = []; |
| 185 |
RegexSetBuilder::new(empty).build().unwrap() |
| 186 |
} |
| 187 |
|
| 188 |
/// Returns true if and only if one of the regexes in this set matches |
| 189 |
/// the haystack given. |
| 190 |
/// |
| 191 |
/// This method should be preferred if you only need to test whether any |
| 192 |
/// of the regexes in the set should match, but don't care about *which* |
| 193 |
/// regexes matched. This is because the underlying matching engine will |
| 194 |
/// quit immediately after seeing the first match instead of continuing to |
| 195 |
/// find all matches. |
| 196 |
/// |
| 197 |
/// Note that as with searches using [`Regex`](crate::bytes::Regex), the |
| 198 |
/// expression is unanchored by default. That is, if the regex does not |
| 199 |
/// start with `^` or `\A`, or end with `$` or `\z`, then it is permitted |
| 200 |
/// to match anywhere in the haystack. |
| 201 |
/// |
| 202 |
/// # Example |
| 203 |
/// |
| 204 |
/// Tests whether a set matches somewhere in a haystack: |
| 205 |
/// |
| 206 |
/// ``` |
| 207 |
/// use regex::bytes::RegexSet; |
| 208 |
/// |
| 209 |
/// let set = RegexSet::new([r"\w+", r"\d+"]).unwrap(); |
| 210 |
/// assert!(set.is_match(b"foo")); |
| 211 |
/// assert!(!set.is_match("☃".as_bytes())); |
| 212 |
/// ``` |
| 213 |
#[inline] |
| 214 |
pub fn is_match(&self, haystack: &[u8]) -> bool { |
| 215 |
self.is_match_at(haystack, 0) |
| 216 |
} |
| 217 |
|
| 218 |
/// Returns true if and only if one of the regexes in this set matches the |
| 219 |
/// haystack given, with the search starting at the offset given. |
| 220 |
/// |
| 221 |
/// The significance of the starting point is that it takes the surrounding |
| 222 |
/// context into consideration. For example, the `\A` anchor can only |
| 223 |
/// match when `start == 0`. |
| 224 |
/// |
| 225 |
/// # Panics |
| 226 |
/// |
| 227 |
/// This panics when `start >= haystack.len() + 1`. |
| 228 |
/// |
| 229 |
/// # Example |
| 230 |
/// |
| 231 |
/// This example shows the significance of `start`. Namely, consider a |
| 232 |
/// haystack `foobar` and a desire to execute a search starting at offset |
| 233 |
/// `3`. You could search a substring explicitly, but then the look-around |
| 234 |
/// assertions won't work correctly. Instead, you can use this method to |
| 235 |
/// specify the start position of a search. |
| 236 |
/// |
| 237 |
/// ``` |
| 238 |
/// use regex::bytes::RegexSet; |
| 239 |
/// |
| 240 |
/// let set = RegexSet::new([r"\bbar\b", r"(?m)^bar$"]).unwrap(); |
| 241 |
/// let hay = b"foobar"; |
| 242 |
/// // We get a match here, but it's probably not intended. |
| 243 |
/// assert!(set.is_match(&hay[3..])); |
| 244 |
/// // No match because the assertions take the context into account. |
| 245 |
/// assert!(!set.is_match_at(hay, 3)); |
| 246 |
/// ``` |
| 247 |
#[inline] |
| 248 |
pub fn is_match_at(&self, haystack: &[u8], start: usize) -> bool { |
| 249 |
self.meta.is_match(Input::new(haystack).span(start..haystack.len())) |
| 250 |
} |
| 251 |
|
| 252 |
/// Returns the set of regexes that match in the given haystack. |
| 253 |
/// |
| 254 |
/// The set returned contains the index of each regex that matches in |
| 255 |
/// the given haystack. The index is in correspondence with the order of |
| 256 |
/// regular expressions given to `RegexSet`'s constructor. |
| 257 |
/// |
| 258 |
/// The set can also be used to iterate over the matched indices. The order |
| 259 |
/// of iteration is always ascending with respect to the matching indices. |
| 260 |
/// |
| 261 |
/// Note that as with searches using [`Regex`](crate::bytes::Regex), the |
| 262 |
/// expression is unanchored by default. That is, if the regex does not |
| 263 |
/// start with `^` or `\A`, or end with `$` or `\z`, then it is permitted |
| 264 |
/// to match anywhere in the haystack. |
| 265 |
/// |
| 266 |
/// # Example |
| 267 |
/// |
| 268 |
/// Tests which regular expressions match the given haystack: |
| 269 |
/// |
| 270 |
/// ``` |
| 271 |
/// use regex::bytes::RegexSet; |
| 272 |
/// |
| 273 |
/// let set = RegexSet::new([ |
| 274 |
/// r"\w+", |
| 275 |
/// r"\d+", |
| 276 |
/// r"\pL+", |
| 277 |
/// r"foo", |
| 278 |
/// r"bar", |
| 279 |
/// r"barfoo", |
| 280 |
/// r"foobar", |
| 281 |
/// ]).unwrap(); |
| 282 |
/// let matches: Vec<_> = set.matches(b"foobar").into_iter().collect(); |
| 283 |
/// assert_eq!(matches, vec![0, 2, 3, 4, 6]); |
| 284 |
/// |
| 285 |
/// // You can also test whether a particular regex matched: |
| 286 |
/// let matches = set.matches(b"foobar"); |
| 287 |
/// assert!(!matches.matched(5)); |
| 288 |
/// assert!(matches.matched(6)); |
| 289 |
/// ``` |
| 290 |
#[inline] |
| 291 |
pub fn matches(&self, haystack: &[u8]) -> SetMatches { |
| 292 |
self.matches_at(haystack, 0) |
| 293 |
} |
| 294 |
|
| 295 |
/// Returns the set of regexes that match in the given haystack. |
| 296 |
/// |
| 297 |
/// The set returned contains the index of each regex that matches in |
| 298 |
/// the given haystack. The index is in correspondence with the order of |
| 299 |
/// regular expressions given to `RegexSet`'s constructor. |
| 300 |
/// |
| 301 |
/// The set can also be used to iterate over the matched indices. The order |
| 302 |
/// of iteration is always ascending with respect to the matching indices. |
| 303 |
/// |
| 304 |
/// The significance of the starting point is that it takes the surrounding |
| 305 |
/// context into consideration. For example, the `\A` anchor can only |
| 306 |
/// match when `start == 0`. |
| 307 |
/// |
| 308 |
/// # Panics |
| 309 |
/// |
| 310 |
/// This panics when `start >= haystack.len() + 1`. |
| 311 |
/// |
| 312 |
/// # Example |
| 313 |
/// |
| 314 |
/// Tests which regular expressions match the given haystack: |
| 315 |
/// |
| 316 |
/// ``` |
| 317 |
/// use regex::bytes::RegexSet; |
| 318 |
/// |
| 319 |
/// let set = RegexSet::new([r"\bbar\b", r"(?m)^bar$"]).unwrap(); |
| 320 |
/// let hay = b"foobar"; |
| 321 |
/// // We get matches here, but it's probably not intended. |
| 322 |
/// let matches: Vec<_> = set.matches(&hay[3..]).into_iter().collect(); |
| 323 |
/// assert_eq!(matches, vec![0, 1]); |
| 324 |
/// // No matches because the assertions take the context into account. |
| 325 |
/// let matches: Vec<_> = set.matches_at(hay, 3).into_iter().collect(); |
| 326 |
/// assert_eq!(matches, vec![]); |
| 327 |
/// ``` |
| 328 |
#[inline] |
| 329 |
pub fn matches_at(&self, haystack: &[u8], start: usize) -> SetMatches { |
| 330 |
let input = Input::new(haystack).span(start..haystack.len()); |
| 331 |
let mut patset = PatternSet::new(self.meta.pattern_len()); |
| 332 |
self.meta.which_overlapping_matches(&input, &mut patset); |
| 333 |
SetMatches(patset) |
| 334 |
} |
| 335 |
|
| 336 |
/// Returns the same as matches, but starts the search at the given |
| 337 |
/// offset and stores the matches into the slice given. |
| 338 |
/// |
| 339 |
/// The significance of the starting point is that it takes the surrounding |
| 340 |
/// context into consideration. For example, the `\A` anchor can only |
| 341 |
/// match when `start == 0`. |
| 342 |
/// |
| 343 |
/// `matches` must have a length that is at least the number of regexes |
| 344 |
/// in this set. |
| 345 |
/// |
| 346 |
/// This method returns true if and only if at least one member of |
| 347 |
/// `matches` is true after executing the set against `haystack`. |
| 348 |
#[doc(hidden)] |
| 349 |
#[inline] |
| 350 |
pub fn matches_read_at( |
| 351 |
&self, |
| 352 |
matches: &mut [bool], |
| 353 |
haystack: &[u8], |
| 354 |
start: usize, |
| 355 |
) -> bool { |
| 356 |
// This is pretty dumb. We should try to fix this, but the |
| 357 |
// regex-automata API doesn't provide a way to store matches in an |
| 358 |
// arbitrary &mut [bool]. Thankfully, this API is is doc(hidden) and |
| 359 |
// thus not public... But regex-capi currently uses it. We should |
| 360 |
// fix regex-capi to use a PatternSet, maybe? Not sure... PatternSet |
| 361 |
// is in regex-automata, not regex. So maybe we should just accept a |
| 362 |
// 'SetMatches', which is basically just a newtype around PatternSet. |
| 363 |
let mut patset = PatternSet::new(self.meta.pattern_len()); |
| 364 |
let mut input = Input::new(haystack); |
| 365 |
input.set_start(start); |
| 366 |
self.meta.which_overlapping_matches(&input, &mut patset); |
| 367 |
for pid in patset.iter() { |
| 368 |
matches[pid] = true; |
| 369 |
} |
| 370 |
!patset.is_empty() |
| 371 |
} |
| 372 |
|
| 373 |
/// An alias for `matches_read_at` to preserve backward compatibility. |
| 374 |
/// |
| 375 |
/// The `regex-capi` crate used this method, so to avoid breaking that |
| 376 |
/// crate, we continue to export it as an undocumented API. |
| 377 |
#[doc(hidden)] |
| 378 |
#[inline] |
| 379 |
pub fn read_matches_at( |
| 380 |
&self, |
| 381 |
matches: &mut [bool], |
| 382 |
haystack: &[u8], |
| 383 |
start: usize, |
| 384 |
) -> bool { |
| 385 |
self.matches_read_at(matches, haystack, start) |
| 386 |
} |
| 387 |
|
| 388 |
/// Returns the total number of regexes in this set. |
| 389 |
/// |
| 390 |
/// # Example |
| 391 |
/// |
| 392 |
/// ``` |
| 393 |
/// use regex::bytes::RegexSet; |
| 394 |
/// |
| 395 |
/// assert_eq!(0, RegexSet::empty().len()); |
| 396 |
/// assert_eq!(1, RegexSet::new([r"[0-9]"]).unwrap().len()); |
| 397 |
/// assert_eq!(2, RegexSet::new([r"[0-9]", r"[a-z]"]).unwrap().len()); |
| 398 |
/// ``` |
| 399 |
#[inline] |
| 400 |
pub fn len(&self) -> usize { |
| 401 |
self.meta.pattern_len() |
| 402 |
} |
| 403 |
|
| 404 |
/// Returns `true` if this set contains no regexes. |
| 405 |
/// |
| 406 |
/// # Example |
| 407 |
/// |
| 408 |
/// ``` |
| 409 |
/// use regex::bytes::RegexSet; |
| 410 |
/// |
| 411 |
/// assert!(RegexSet::empty().is_empty()); |
| 412 |
/// assert!(!RegexSet::new([r"[0-9]"]).unwrap().is_empty()); |
| 413 |
/// ``` |
| 414 |
#[inline] |
| 415 |
pub fn is_empty(&self) -> bool { |
| 416 |
self.meta.pattern_len() == 0 |
| 417 |
} |
| 418 |
|
| 419 |
/// Returns the regex patterns that this regex set was constructed from. |
| 420 |
/// |
| 421 |
/// This function can be used to determine the pattern for a match. The |
| 422 |
/// slice returned has exactly as many patterns givens to this regex set, |
| 423 |
/// and the order of the slice is the same as the order of the patterns |
| 424 |
/// provided to the set. |
| 425 |
/// |
| 426 |
/// # Example |
| 427 |
/// |
| 428 |
/// ``` |
| 429 |
/// use regex::bytes::RegexSet; |
| 430 |
/// |
| 431 |
/// let set = RegexSet::new(&[ |
| 432 |
/// r"\w+", |
| 433 |
/// r"\d+", |
| 434 |
/// r"\pL+", |
| 435 |
/// r"foo", |
| 436 |
/// r"bar", |
| 437 |
/// r"barfoo", |
| 438 |
/// r"foobar", |
| 439 |
/// ]).unwrap(); |
| 440 |
/// let matches: Vec<_> = set |
| 441 |
/// .matches(b"foobar") |
| 442 |
/// .into_iter() |
| 443 |
/// .map(|index| &set.patterns()[index]) |
| 444 |
/// .collect(); |
| 445 |
/// assert_eq!(matches, vec![r"\w+", r"\pL+", r"foo", r"bar", r"foobar"]); |
| 446 |
/// ``` |
| 447 |
#[inline] |
| 448 |
pub fn patterns(&self) -> &[String] { |
| 449 |
&self.patterns |
| 450 |
} |
| 451 |
} |
| 452 |
|
| 453 |
impl Default for RegexSet { |
| 454 |
fn default() -> Self { |
| 455 |
RegexSet::empty() |
| 456 |
} |
| 457 |
} |
| 458 |
|
| 459 |
/// A set of matches returned by a regex set. |
| 460 |
/// |
| 461 |
/// Values of this type are constructed by [`RegexSet::matches`]. |
| 462 |
#[derive(Clone, Debug)] |
| 463 |
pub struct SetMatches(PatternSet); |
| 464 |
|
| 465 |
impl SetMatches { |
| 466 |
/// Whether this set contains any matches. |
| 467 |
/// |
| 468 |
/// # Example |
| 469 |
/// |
| 470 |
/// ``` |
| 471 |
/// use regex::bytes::RegexSet; |
| 472 |
/// |
| 473 |
/// let set = RegexSet::new(&[ |
| 474 |
/// r"[a-z]+@[a-z]+\.(com|org|net)", |
| 475 |
/// r"[a-z]+\.(com|org|net)", |
| 476 |
/// ]).unwrap(); |
| 477 |
/// let matches = set.matches(b"[email protected]"); |
| 478 |
/// assert!(matches.matched_any()); |
| 479 |
/// ``` |
| 480 |
#[inline] |
| 481 |
pub fn matched_any(&self) -> bool { |
| 482 |
!self.0.is_empty() |
| 483 |
} |
| 484 |
|
| 485 |
/// Whether the regex at the given index matched. |
| 486 |
/// |
| 487 |
/// The index for a regex is determined by its insertion order upon the |
| 488 |
/// initial construction of a `RegexSet`, starting at `0`. |
| 489 |
/// |
| 490 |
/// # Panics |
| 491 |
/// |
| 492 |
/// If `index` is greater than or equal to the number of regexes in the |
| 493 |
/// original set that produced these matches. Equivalently, when `index` |
| 494 |
/// is greater than or equal to [`SetMatches::len`]. |
| 495 |
/// |
| 496 |
/// # Example |
| 497 |
/// |
| 498 |
/// ``` |
| 499 |
/// use regex::bytes::RegexSet; |
| 500 |
/// |
| 501 |
/// let set = RegexSet::new([ |
| 502 |
/// r"[a-z]+@[a-z]+\.(com|org|net)", |
| 503 |
/// r"[a-z]+\.(com|org|net)", |
| 504 |
/// ]).unwrap(); |
| 505 |
/// let matches = set.matches(b"example.com"); |
| 506 |
/// assert!(!matches.matched(0)); |
| 507 |
/// assert!(matches.matched(1)); |
| 508 |
/// ``` |
| 509 |
#[inline] |
| 510 |
pub fn matched(&self, index: usize) -> bool { |
| 511 |
self.0.contains(PatternID::new_unchecked(index)) |
| 512 |
} |
| 513 |
|
| 514 |
/// The total number of regexes in the set that created these matches. |
| 515 |
/// |
| 516 |
/// **WARNING:** This always returns the same value as [`RegexSet::len`]. |
| 517 |
/// In particular, it does *not* return the number of elements yielded by |
| 518 |
/// [`SetMatches::iter`]. The only way to determine the total number of |
| 519 |
/// matched regexes is to iterate over them. |
| 520 |
/// |
| 521 |
/// # Example |
| 522 |
/// |
| 523 |
/// Notice that this method returns the total number of regexes in the |
| 524 |
/// original set, and *not* the total number of regexes that matched. |
| 525 |
/// |
| 526 |
/// ``` |
| 527 |
/// use regex::bytes::RegexSet; |
| 528 |
/// |
| 529 |
/// let set = RegexSet::new([ |
| 530 |
/// r"[a-z]+@[a-z]+\.(com|org|net)", |
| 531 |
/// r"[a-z]+\.(com|org|net)", |
| 532 |
/// ]).unwrap(); |
| 533 |
/// let matches = set.matches(b"example.com"); |
| 534 |
/// // Total number of patterns that matched. |
| 535 |
/// assert_eq!(1, matches.iter().count()); |
| 536 |
/// // Total number of patterns in the set. |
| 537 |
/// assert_eq!(2, matches.len()); |
| 538 |
/// ``` |
| 539 |
#[inline] |
| 540 |
pub fn len(&self) -> usize { |
| 541 |
self.0.capacity() |
| 542 |
} |
| 543 |
|
| 544 |
/// Returns an iterator over the indices of the regexes that matched. |
| 545 |
/// |
| 546 |
/// This will always produces matches in ascending order, where the index |
| 547 |
/// yielded corresponds to the index of the regex that matched with respect |
| 548 |
/// to its position when initially building the set. |
| 549 |
/// |
| 550 |
/// # Example |
| 551 |
/// |
| 552 |
/// ``` |
| 553 |
/// use regex::bytes::RegexSet; |
| 554 |
/// |
| 555 |
/// let set = RegexSet::new([ |
| 556 |
/// r"[0-9]", |
| 557 |
/// r"[a-z]", |
| 558 |
/// r"[A-Z]", |
| 559 |
/// r"\p{Greek}", |
| 560 |
/// ]).unwrap(); |
| 561 |
/// let hay = "βa1".as_bytes(); |
| 562 |
/// let matches: Vec<_> = set.matches(hay).iter().collect(); |
| 563 |
/// assert_eq!(matches, vec![0, 1, 3]); |
| 564 |
/// ``` |
| 565 |
/// |
| 566 |
/// Note that `SetMatches` also implemnets the `IntoIterator` trait, so |
| 567 |
/// this method is not always needed. For example: |
| 568 |
/// |
| 569 |
/// ``` |
| 570 |
/// use regex::bytes::RegexSet; |
| 571 |
/// |
| 572 |
/// let set = RegexSet::new([ |
| 573 |
/// r"[0-9]", |
| 574 |
/// r"[a-z]", |
| 575 |
/// r"[A-Z]", |
| 576 |
/// r"\p{Greek}", |
| 577 |
/// ]).unwrap(); |
| 578 |
/// let hay = "βa1".as_bytes(); |
| 579 |
/// let mut matches = vec![]; |
| 580 |
/// for index in set.matches(hay) { |
| 581 |
/// matches.push(index); |
| 582 |
/// } |
| 583 |
/// assert_eq!(matches, vec![0, 1, 3]); |
| 584 |
/// ``` |
| 585 |
#[inline] |
| 586 |
pub fn iter(&self) -> SetMatchesIter<'_> { |
| 587 |
SetMatchesIter(self.0.iter()) |
| 588 |
} |
| 589 |
} |
| 590 |
|
| 591 |
impl IntoIterator for SetMatches { |
| 592 |
type IntoIter = SetMatchesIntoIter; |
| 593 |
type Item = usize; |
| 594 |
|
| 595 |
fn into_iter(self) -> Self::IntoIter { |
| 596 |
let it = 0..self.0.capacity(); |
| 597 |
SetMatchesIntoIter { patset: self.0, it } |
| 598 |
} |
| 599 |
} |
| 600 |
|
| 601 |
impl<'a> IntoIterator for &'a SetMatches { |
| 602 |
type IntoIter = SetMatchesIter<'a>; |
| 603 |
type Item = usize; |
| 604 |
|
| 605 |
fn into_iter(self) -> Self::IntoIter { |
| 606 |
self.iter() |
| 607 |
} |
| 608 |
} |
| 609 |
|
| 610 |
/// An owned iterator over the set of matches from a regex set. |
| 611 |
/// |
| 612 |
/// This will always produces matches in ascending order of index, where the |
| 613 |
/// index corresponds to the index of the regex that matched with respect to |
| 614 |
/// its position when initially building the set. |
| 615 |
/// |
| 616 |
/// This iterator is created by calling `SetMatches::into_iter` via the |
| 617 |
/// `IntoIterator` trait. This is automatically done in `for` loops. |
| 618 |
/// |
| 619 |
/// # Example |
| 620 |
/// |
| 621 |
/// ``` |
| 622 |
/// use regex::bytes::RegexSet; |
| 623 |
/// |
| 624 |
/// let set = RegexSet::new([ |
| 625 |
/// r"[0-9]", |
| 626 |
/// r"[a-z]", |
| 627 |
/// r"[A-Z]", |
| 628 |
/// r"\p{Greek}", |
| 629 |
/// ]).unwrap(); |
| 630 |
/// let hay = "βa1".as_bytes(); |
| 631 |
/// let mut matches = vec![]; |
| 632 |
/// for index in set.matches(hay) { |
| 633 |
/// matches.push(index); |
| 634 |
/// } |
| 635 |
/// assert_eq!(matches, vec![0, 1, 3]); |
| 636 |
/// ``` |
| 637 |
#[derive(Debug)] |
| 638 |
pub struct SetMatchesIntoIter { |
| 639 |
patset: PatternSet, |
| 640 |
it: core::ops::Range<usize>, |
| 641 |
} |
| 642 |
|
| 643 |
impl Iterator for SetMatchesIntoIter { |
| 644 |
type Item = usize; |
| 645 |
|
| 646 |
fn next(&mut self) -> Option<usize> { |
| 647 |
loop { |
| 648 |
let id = self.it.next()?; |
| 649 |
if self.patset.contains(PatternID::new_unchecked(id)) { |
| 650 |
return Some(id); |
| 651 |
} |
| 652 |
} |
| 653 |
} |
| 654 |
|
| 655 |
fn size_hint(&self) -> (usize, Option<usize>) { |
| 656 |
self.it.size_hint() |
| 657 |
} |
| 658 |
} |
| 659 |
|
| 660 |
impl DoubleEndedIterator for SetMatchesIntoIter { |
| 661 |
fn next_back(&mut self) -> Option<usize> { |
| 662 |
loop { |
| 663 |
let id = self.it.next_back()?; |
| 664 |
if self.patset.contains(PatternID::new_unchecked(id)) { |
| 665 |
return Some(id); |
| 666 |
} |
| 667 |
} |
| 668 |
} |
| 669 |
} |
| 670 |
|
| 671 |
impl core::iter::FusedIterator for SetMatchesIntoIter {} |
| 672 |
|
| 673 |
/// A borrowed iterator over the set of matches from a regex set. |
| 674 |
/// |
| 675 |
/// The lifetime `'a` refers to the lifetime of the [`SetMatches`] value that |
| 676 |
/// created this iterator. |
| 677 |
/// |
| 678 |
/// This will always produces matches in ascending order, where the index |
| 679 |
/// corresponds to the index of the regex that matched with respect to its |
| 680 |
/// position when initially building the set. |
| 681 |
/// |
| 682 |
/// This iterator is created by the [`SetMatches::iter`] method. |
| 683 |
#[derive(Clone, Debug)] |
| 684 |
pub struct SetMatchesIter<'a>(PatternSetIter<'a>); |
| 685 |
|
| 686 |
impl<'a> Iterator for SetMatchesIter<'a> { |
| 687 |
type Item = usize; |
| 688 |
|
| 689 |
fn next(&mut self) -> Option<usize> { |
| 690 |
self.0.next().map(|pid| pid.as_usize()) |
| 691 |
} |
| 692 |
|
| 693 |
fn size_hint(&self) -> (usize, Option<usize>) { |
| 694 |
self.0.size_hint() |
| 695 |
} |
| 696 |
} |
| 697 |
|
| 698 |
impl<'a> DoubleEndedIterator for SetMatchesIter<'a> { |
| 699 |
fn next_back(&mut self) -> Option<usize> { |
| 700 |
self.0.next_back().map(|pid| pid.as_usize()) |
| 701 |
} |
| 702 |
} |
| 703 |
|
| 704 |
impl<'a> core::iter::FusedIterator for SetMatchesIter<'a> {} |
| 705 |
|
| 706 |
impl core::fmt::Debug for RegexSet { |
| 707 |
fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result { |
| 708 |
write!(f, "RegexSet({:?})", self.patterns()) |
| 709 |
} |
| 710 |
} |