1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
use std::io;
use crate::automaton::Automaton;
use crate::buffer::Buffer;
use crate::dfa::{self, DFA};
use crate::error::Result;
use crate::nfa::{self, NFA};
use crate::packed;
use crate::prefilter::{Prefilter, PrefilterState};
use crate::state_id::StateID;
use crate::Match;
/// An automaton for searching multiple strings in linear time.
///
/// The `AhoCorasick` type supports a few basic ways of constructing an
/// automaton, including
/// [`AhoCorasick::new`](struct.AhoCorasick.html#method.new)
/// and
/// [`AhoCorasick::new_auto_configured`](struct.AhoCorasick.html#method.new_auto_configured).
/// However, there are a fair number of configurable options that can be set
/// by using
/// [`AhoCorasickBuilder`](struct.AhoCorasickBuilder.html)
/// instead. Such options include, but are not limited to, how matches are
/// determined, simple case insensitivity, whether to use a DFA or not and
/// various knobs for controlling the space-vs-time trade offs taken when
/// building the automaton.
///
/// If you aren't sure where to start, try beginning with
/// [`AhoCorasick::new_auto_configured`](struct.AhoCorasick.html#method.new_auto_configured).
///
/// # Resource usage
///
/// Aho-Corasick automatons are always constructed in `O(p)` time, where `p`
/// is the combined length of all patterns being searched. With that said,
/// building an automaton can be fairly costly because of high constant
/// factors, particularly when enabling the
/// [DFA](struct.AhoCorasickBuilder.html#method.dfa)
/// option (which is disabled by default). For this reason, it's generally a
/// good idea to build an automaton once and reuse it as much as possible.
///
/// Aho-Corasick automatons can also use a fair bit of memory. To get a
/// concrete idea of how much memory is being used, try using the
/// [`AhoCorasick::heap_bytes`](struct.AhoCorasick.html#method.heap_bytes)
/// method.
///
/// # Examples
///
/// This example shows how to search for occurrences of multiple patterns
/// simultaneously in a case insensitive fashion. Each match includes the
/// pattern that matched along with the byte offsets of the match.
///
/// ```
/// use aho_corasick::AhoCorasickBuilder;
///
/// let patterns = &["apple", "maple", "snapple"];
/// let haystack = "Nobody likes maple in their apple flavored Snapple.";
///
/// let ac = AhoCorasickBuilder::new()
/// .ascii_case_insensitive(true)
/// .build(patterns);
/// let mut matches = vec![];
/// for mat in ac.find_iter(haystack) {
/// matches.push((mat.pattern(), mat.start(), mat.end()));
/// }
/// assert_eq!(matches, vec![
/// (1, 13, 18),
/// (0, 28, 33),
/// (2, 43, 50),
/// ]);
/// ```
///
/// This example shows how to replace matches with some other string:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let patterns = &["fox", "brown", "quick"];
/// let haystack = "The quick brown fox.";
/// let replace_with = &["sloth", "grey", "slow"];
///
/// let ac = AhoCorasick::new(patterns);
/// let result = ac.replace_all(haystack, replace_with);
/// assert_eq!(result, "The slow grey sloth.");
/// ```
#[derive(Clone, Debug)]
pub struct AhoCorasick<S: StateID = usize> {
imp: Imp<S>,
match_kind: MatchKind,
}
impl AhoCorasick {
/// Create a new Aho-Corasick automaton using the default configuration.
///
/// The default configuration optimizes for less space usage, but at the
/// expense of longer search times. To change the configuration, use
/// [`AhoCorasickBuilder`](struct.AhoCorasickBuilder.html)
/// for fine-grained control, or
/// [`AhoCorasick::new_auto_configured`](struct.AhoCorasick.html#method.new_auto_configured)
/// for automatic configuration if you aren't sure which settings to pick.
///
/// This uses the default
/// [`MatchKind::Standard`](enum.MatchKind.html#variant.Standard)
/// match semantics, which reports a match as soon as it is found. This
/// corresponds to the standard match semantics supported by textbook
/// descriptions of the Aho-Corasick algorithm.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let ac = AhoCorasick::new(&[
/// "foo", "bar", "baz",
/// ]);
/// assert_eq!(Some(1), ac.find("xxx bar xxx").map(|m| m.pattern()));
/// ```
pub fn new<I, P>(patterns: I) -> AhoCorasick
where
I: IntoIterator<Item = P>,
P: AsRef<[u8]>,
{
AhoCorasickBuilder::new().build(patterns)
}
/// Build an Aho-Corasick automaton with an automatically determined
/// configuration.
///
/// Specifically, this requires a slice of patterns instead of an iterator
/// since the configuration is determined by looking at the patterns before
/// constructing the automaton. The idea here is to balance space and time
/// automatically. That is, when searching a small number of patterns, this
/// will attempt to use the fastest possible configuration since the total
/// space required will be small anyway. As the number of patterns grows,
/// this will fall back to slower configurations that use less space.
///
/// If you want auto configuration but with match semantics different from
/// the default `MatchKind::Standard`, then use
/// [`AhoCorasickBuilder::auto_configure`](struct.AhoCorasickBuilder.html#method.auto_configure).
///
/// # Examples
///
/// Basic usage is just like `new`, except you must provide a slice:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let ac = AhoCorasick::new_auto_configured(&[
/// "foo", "bar", "baz",
/// ]);
/// assert_eq!(Some(1), ac.find("xxx bar xxx").map(|m| m.pattern()));
/// ```
pub fn new_auto_configured<B>(patterns: &[B]) -> AhoCorasick
where
B: AsRef<[u8]>,
{
AhoCorasickBuilder::new().auto_configure(patterns).build(patterns)
}
}
impl<S: StateID> AhoCorasick<S> {
/// Returns true if and only if this automaton matches the haystack at any
/// position.
///
/// `haystack` may be any type that is cheaply convertible to a `&[u8]`.
/// This includes, but is not limited to, `String`, `&str`, `Vec<u8>`, and
/// `&[u8]` itself.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let ac = AhoCorasick::new(&[
/// "foo", "bar", "quux", "baz",
/// ]);
/// assert!(ac.is_match("xxx bar xxx"));
/// assert!(!ac.is_match("xxx qux xxx"));
/// ```
pub fn is_match<B: AsRef<[u8]>>(&self, haystack: B) -> bool {
self.earliest_find(haystack).is_some()
}
/// Returns the location of the first detected match in `haystack`.
///
/// This method has the same behavior regardless of the
/// [`MatchKind`](enum.MatchKind.html)
/// of this automaton.
///
/// `haystack` may be any type that is cheaply convertible to a `&[u8]`.
/// This includes, but is not limited to, `String`, `&str`, `Vec<u8>`, and
/// `&[u8]` itself.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let ac = AhoCorasick::new(&[
/// "abc", "b",
/// ]);
/// let mat = ac.earliest_find("abcd").expect("should have match");
/// assert_eq!(1, mat.pattern());
/// assert_eq!((1, 2), (mat.start(), mat.end()));
/// ```
pub fn earliest_find<B: AsRef<[u8]>>(&self, haystack: B) -> Option<Match> {
let mut prestate = PrefilterState::new(self.max_pattern_len());
let mut start = self.imp.start_state();
self.imp.earliest_find_at(
&mut prestate,
haystack.as_ref(),
0,
&mut start,
)
}
/// Returns the location of the first match according to the match
/// semantics that this automaton was constructed with.
///
/// When using `MatchKind::Standard`, this corresponds precisely to the
/// same behavior as
/// [`earliest_find`](struct.AhoCorasick.html#method.earliest_find).
/// Otherwise, match semantics correspond to either
/// [leftmost-first](enum.MatchKind.html#variant.LeftmostFirst)
/// or
/// [leftmost-longest](enum.MatchKind.html#variant.LeftmostLongest).
///
/// `haystack` may be any type that is cheaply convertible to a `&[u8]`.
/// This includes, but is not limited to, `String`, `&str`, `Vec<u8>`, and
/// `&[u8]` itself.
///
/// # Examples
///
/// Basic usage, with standard semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["b", "abc", "abcd"];
/// let haystack = "abcd";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::Standard) // default, not necessary
/// .build(patterns);
/// let mat = ac.find(haystack).expect("should have a match");
/// assert_eq!("b", &haystack[mat.start()..mat.end()]);
/// ```
///
/// Now with leftmost-first semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["b", "abc", "abcd"];
/// let haystack = "abcd";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(patterns);
/// let mat = ac.find(haystack).expect("should have a match");
/// assert_eq!("abc", &haystack[mat.start()..mat.end()]);
/// ```
///
/// And finally, leftmost-longest semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["b", "abc", "abcd"];
/// let haystack = "abcd";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostLongest)
/// .build(patterns);
/// let mat = ac.find(haystack).expect("should have a match");
/// assert_eq!("abcd", &haystack[mat.start()..mat.end()]);
/// ```
pub fn find<B: AsRef<[u8]>>(&self, haystack: B) -> Option<Match> {
let mut prestate = PrefilterState::new(self.max_pattern_len());
self.imp.find_at_no_state(&mut prestate, haystack.as_ref(), 0)
}
/// Returns an iterator of non-overlapping matches, using the match
/// semantics that this automaton was constructed with.
///
/// `haystack` may be any type that is cheaply convertible to a `&[u8]`.
/// This includes, but is not limited to, `String`, `&str`, `Vec<u8>`, and
/// `&[u8]` itself.
///
/// # Examples
///
/// Basic usage, with standard semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = "append the app to the appendage";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::Standard) // default, not necessary
/// .build(patterns);
/// let matches: Vec<usize> = ac
/// .find_iter(haystack)
/// .map(|mat| mat.pattern())
/// .collect();
/// assert_eq!(vec![2, 2, 2], matches);
/// ```
///
/// Now with leftmost-first semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = "append the app to the appendage";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(patterns);
/// let matches: Vec<usize> = ac
/// .find_iter(haystack)
/// .map(|mat| mat.pattern())
/// .collect();
/// assert_eq!(vec![0, 2, 0], matches);
/// ```
///
/// And finally, leftmost-longest semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = "append the app to the appendage";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostLongest)
/// .build(patterns);
/// let matches: Vec<usize> = ac
/// .find_iter(haystack)
/// .map(|mat| mat.pattern())
/// .collect();
/// assert_eq!(vec![0, 2, 1], matches);
/// ```
pub fn find_iter<'a, 'b, B: ?Sized + AsRef<[u8]>>(
&'a self,
haystack: &'b B,
) -> FindIter<'a, 'b, S> {
FindIter::new(self, haystack.as_ref())
}
/// Returns an iterator of overlapping matches in the given `haystack`.
///
/// Overlapping matches can _only_ be detected using
/// `MatchKind::Standard` semantics. If this automaton was constructed with
/// leftmost semantics, then this method will panic. To determine whether
/// this will panic at runtime, use the
/// [`AhoCorasick::supports_overlapping`](struct.AhoCorasick.html#method.supports_overlapping)
/// method.
///
/// `haystack` may be any type that is cheaply convertible to a `&[u8]`.
/// This includes, but is not limited to, `String`, `&str`, `Vec<u8>`, and
/// `&[u8]` itself.
///
/// # Panics
///
/// This panics when `AhoCorasick::supports_overlapping` returns `false`.
/// That is, this panics when this automaton's match semantics are not
/// `MatchKind::Standard`.
///
/// # Examples
///
/// Basic usage, with standard semantics:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = "append the app to the appendage";
///
/// let ac = AhoCorasick::new(patterns);
/// let matches: Vec<usize> = ac
/// .find_overlapping_iter(haystack)
/// .map(|mat| mat.pattern())
/// .collect();
/// assert_eq!(vec![2, 0, 2, 2, 0, 1], matches);
/// ```
pub fn find_overlapping_iter<'a, 'b, B: ?Sized + AsRef<[u8]>>(
&'a self,
haystack: &'b B,
) -> FindOverlappingIter<'a, 'b, S> {
FindOverlappingIter::new(self, haystack.as_ref())
}
/// Replace all matches with a corresponding value in the `replace_with`
/// slice given. Matches correspond to the same matches as reported by
/// [`find_iter`](struct.AhoCorasick.html#method.find_iter).
///
/// Replacements are determined by the index of the matching pattern.
/// For example, if the pattern with index `2` is found, then it is
/// replaced by `replace_with[2]`.
///
/// # Panics
///
/// This panics when `replace_with.len()` does not equal the total number
/// of patterns that are matched by this automaton.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = "append the app to the appendage";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(patterns);
/// let result = ac.replace_all(haystack, &["x", "y", "z"]);
/// assert_eq!("x the z to the xage", result);
/// ```
pub fn replace_all<B>(&self, haystack: &str, replace_with: &[B]) -> String
where
B: AsRef<str>,
{
assert_eq!(
replace_with.len(),
self.pattern_count(),
"replace_all requires a replacement for every pattern \
in the automaton"
);
let mut dst = String::with_capacity(haystack.len());
self.replace_all_with(haystack, &mut dst, |mat, _, dst| {
dst.push_str(replace_with[mat.pattern()].as_ref());
true
});
dst
}
/// Replace all matches using raw bytes with a corresponding value in the
/// `replace_with` slice given. Matches correspond to the same matches as
/// reported by [`find_iter`](struct.AhoCorasick.html#method.find_iter).
///
/// Replacements are determined by the index of the matching pattern.
/// For example, if the pattern with index `2` is found, then it is
/// replaced by `replace_with[2]`.
///
/// # Panics
///
/// This panics when `replace_with.len()` does not equal the total number
/// of patterns that are matched by this automaton.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = b"append the app to the appendage";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(patterns);
/// let result = ac.replace_all_bytes(haystack, &["x", "y", "z"]);
/// assert_eq!(b"x the z to the xage".to_vec(), result);
/// ```
pub fn replace_all_bytes<B>(
&self,
haystack: &[u8],
replace_with: &[B],
) -> Vec<u8>
where
B: AsRef<[u8]>,
{
assert_eq!(
replace_with.len(),
self.pattern_count(),
"replace_all_bytes requires a replacement for every pattern \
in the automaton"
);
let mut dst = Vec::with_capacity(haystack.len());
self.replace_all_with_bytes(haystack, &mut dst, |mat, _, dst| {
dst.extend(replace_with[mat.pattern()].as_ref());
true
});
dst
}
/// Replace all matches using a closure called on each match.
/// Matches correspond to the same matches as reported by
/// [`find_iter`](struct.AhoCorasick.html#method.find_iter).
///
/// The closure accepts three parameters: the match found, the text of
/// the match and a string buffer with which to write the replaced text
/// (if any). If the closure returns `true`, then it continues to the next
/// match. If the closure returns `false`, then searching is stopped.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = "append the app to the appendage";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(patterns);
/// let mut result = String::new();
/// ac.replace_all_with(haystack, &mut result, |mat, _, dst| {
/// dst.push_str(&mat.pattern().to_string());
/// true
/// });
/// assert_eq!("0 the 2 to the 0age", result);
/// ```
///
/// Stopping the replacement by returning `false` (continued from the
/// example above):
///
/// ```
/// # use aho_corasick::{AhoCorasickBuilder, MatchKind};
/// # let patterns = &["append", "appendage", "app"];
/// # let haystack = "append the app to the appendage";
/// # let ac = AhoCorasickBuilder::new()
/// # .match_kind(MatchKind::LeftmostFirst)
/// # .build(patterns);
/// let mut result = String::new();
/// ac.replace_all_with(haystack, &mut result, |mat, _, dst| {
/// dst.push_str(&mat.pattern().to_string());
/// mat.pattern() != 2
/// });
/// assert_eq!("0 the 2 to the appendage", result);
/// ```
pub fn replace_all_with<F>(
&self,
haystack: &str,
dst: &mut String,
mut replace_with: F,
) where
F: FnMut(&Match, &str, &mut String) -> bool,
{
let mut last_match = 0;
for mat in self.find_iter(haystack) {
dst.push_str(&haystack[last_match..mat.start()]);
last_match = mat.end();
if !replace_with(&mat, &haystack[mat.start()..mat.end()], dst) {
break;
};
}
dst.push_str(&haystack[last_match..]);
}
/// Replace all matches using raw bytes with a closure called on each
/// match. Matches correspond to the same matches as reported by
/// [`find_iter`](struct.AhoCorasick.html#method.find_iter).
///
/// The closure accepts three parameters: the match found, the text of
/// the match and a byte buffer with which to write the replaced text
/// (if any). If the closure returns `true`, then it continues to the next
/// match. If the closure returns `false`, then searching is stopped.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["append", "appendage", "app"];
/// let haystack = b"append the app to the appendage";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(patterns);
/// let mut result = vec![];
/// ac.replace_all_with_bytes(haystack, &mut result, |mat, _, dst| {
/// dst.extend(mat.pattern().to_string().bytes());
/// true
/// });
/// assert_eq!(b"0 the 2 to the 0age".to_vec(), result);
/// ```
///
/// Stopping the replacement by returning `false` (continued from the
/// example above):
///
/// ```
/// # use aho_corasick::{AhoCorasickBuilder, MatchKind};
/// # let patterns = &["append", "appendage", "app"];
/// # let haystack = b"append the app to the appendage";
/// # let ac = AhoCorasickBuilder::new()
/// # .match_kind(MatchKind::LeftmostFirst)
/// # .build(patterns);
/// let mut result = vec![];
/// ac.replace_all_with_bytes(haystack, &mut result, |mat, _, dst| {
/// dst.extend(mat.pattern().to_string().bytes());
/// mat.pattern() != 2
/// });
/// assert_eq!(b"0 the 2 to the appendage".to_vec(), result);
/// ```
pub fn replace_all_with_bytes<F>(
&self,
haystack: &[u8],
dst: &mut Vec<u8>,
mut replace_with: F,
) where
F: FnMut(&Match, &[u8], &mut Vec<u8>) -> bool,
{
let mut last_match = 0;
for mat in self.find_iter(haystack) {
dst.extend(&haystack[last_match..mat.start()]);
last_match = mat.end();
if !replace_with(&mat, &haystack[mat.start()..mat.end()], dst) {
break;
};
}
dst.extend(&haystack[last_match..]);
}
/// Returns an iterator of non-overlapping matches in the given
/// stream. Matches correspond to the same matches as reported by
/// [`find_iter`](struct.AhoCorasick.html#method.find_iter).
///
/// The matches yielded by this iterator use absolute position offsets in
/// the stream given, where the first byte has index `0`. Matches are
/// yieled until the stream is exhausted.
///
/// Each item yielded by the iterator is an `io::Result<Match>`, where an
/// error is yielded if there was a problem reading from the reader given.
///
/// When searching a stream, an internal buffer is used. Therefore, callers
/// should avoiding providing a buffered reader, if possible.
///
/// Searching a stream requires that the automaton was built with
/// `MatchKind::Standard` semantics. If this automaton was constructed
/// with leftmost semantics, then this method will panic. To determine
/// whether this will panic at runtime, use the
/// [`AhoCorasick::supports_stream`](struct.AhoCorasick.html#method.supports_stream)
/// method.
///
/// # Memory usage
///
/// In general, searching streams will use a constant amount of memory for
/// its internal buffer. The one requirement is that the internal buffer
/// must be at least the size of the longest possible match. In most use
/// cases, the default buffer size will be much larger than any individual
/// match.
///
/// # Panics
///
/// This panics when `AhoCorasick::supports_stream` returns `false`.
/// That is, this panics when this automaton's match semantics are not
/// `MatchKind::Standard`. This restriction may be lifted in the future.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// # fn example() -> Result<(), ::std::io::Error> {
/// let patterns = &["append", "appendage", "app"];
/// let haystack = "append the app to the appendage";
///
/// let ac = AhoCorasick::new(patterns);
/// let mut matches = vec![];
/// for result in ac.stream_find_iter(haystack.as_bytes()) {
/// let mat = result?;
/// matches.push(mat.pattern());
/// }
/// assert_eq!(vec![2, 2, 2], matches);
/// # Ok(()) }; example().unwrap()
/// ```
pub fn stream_find_iter<'a, R: io::Read>(
&'a self,
rdr: R,
) -> StreamFindIter<'a, R, S> {
StreamFindIter::new(self, rdr)
}
/// Search for and replace all matches of this automaton in
/// the given reader, and write the replacements to the given
/// writer. Matches correspond to the same matches as reported by
/// [`find_iter`](struct.AhoCorasick.html#method.find_iter).
///
/// Replacements are determined by the index of the matching pattern.
/// For example, if the pattern with index `2` is found, then it is
/// replaced by `replace_with[2]`.
///
/// After all matches are replaced, the writer is _not_ flushed.
///
/// If there was a problem reading from the given reader or writing to the
/// given writer, then the corresponding `io::Error` is returned and all
/// replacement is stopped.
///
/// When searching a stream, an internal buffer is used. Therefore, callers
/// should avoiding providing a buffered reader, if possible. However,
/// callers may want to provide a buffered writer.
///
/// Searching a stream requires that the automaton was built with
/// `MatchKind::Standard` semantics. If this automaton was constructed
/// with leftmost semantics, then this method will panic. To determine
/// whether this will panic at runtime, use the
/// [`AhoCorasick::supports_stream`](struct.AhoCorasick.html#method.supports_stream)
/// method.
///
/// # Memory usage
///
/// In general, searching streams will use a constant amount of memory for
/// its internal buffer. The one requirement is that the internal buffer
/// must be at least the size of the longest possible match. In most use
/// cases, the default buffer size will be much larger than any individual
/// match.
///
/// # Panics
///
/// This panics when `AhoCorasick::supports_stream` returns `false`.
/// That is, this panics when this automaton's match semantics are not
/// `MatchKind::Standard`. This restriction may be lifted in the future.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// # fn example() -> Result<(), ::std::io::Error> {
/// let patterns = &["fox", "brown", "quick"];
/// let haystack = "The quick brown fox.";
/// let replace_with = &["sloth", "grey", "slow"];
///
/// let ac = AhoCorasick::new(patterns);
/// let mut result = vec![];
/// ac.stream_replace_all(haystack.as_bytes(), &mut result, replace_with)?;
/// assert_eq!(b"The slow grey sloth.".to_vec(), result);
/// # Ok(()) }; example().unwrap()
/// ```
pub fn stream_replace_all<R, W, B>(
&self,
rdr: R,
wtr: W,
replace_with: &[B],
) -> io::Result<()>
where
R: io::Read,
W: io::Write,
B: AsRef<[u8]>,
{
assert_eq!(
replace_with.len(),
self.pattern_count(),
"stream_replace_all requires a replacement for every pattern \
in the automaton"
);
self.stream_replace_all_with(rdr, wtr, |mat, _, wtr| {
wtr.write_all(replace_with[mat.pattern()].as_ref())
})
}
/// Search the given reader and replace all matches of this automaton
/// using the given closure. The result is written to the given
/// writer. Matches correspond to the same matches as reported by
/// [`find_iter`](struct.AhoCorasick.html#method.find_iter).
///
/// The closure accepts three parameters: the match found, the text of
/// the match and the writer with which to write the replaced text (if any).
///
/// After all matches are replaced, the writer is _not_ flushed.
///
/// If there was a problem reading from the given reader or writing to the
/// given writer, then the corresponding `io::Error` is returned and all
/// replacement is stopped.
///
/// When searching a stream, an internal buffer is used. Therefore, callers
/// should avoiding providing a buffered reader, if possible. However,
/// callers may want to provide a buffered writer.
///
/// Searching a stream requires that the automaton was built with
/// `MatchKind::Standard` semantics. If this automaton was constructed
/// with leftmost semantics, then this method will panic. To determine
/// whether this will panic at runtime, use the
/// [`AhoCorasick::supports_stream`](struct.AhoCorasick.html#method.supports_stream)
/// method.
///
/// # Memory usage
///
/// In general, searching streams will use a constant amount of memory for
/// its internal buffer. The one requirement is that the internal buffer
/// must be at least the size of the longest possible match. In most use
/// cases, the default buffer size will be much larger than any individual
/// match.
///
/// # Panics
///
/// This panics when `AhoCorasick::supports_stream` returns `false`.
/// That is, this panics when this automaton's match semantics are not
/// `MatchKind::Standard`. This restriction may be lifted in the future.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use std::io::Write;
/// use aho_corasick::AhoCorasick;
///
/// # fn example() -> Result<(), ::std::io::Error> {
/// let patterns = &["fox", "brown", "quick"];
/// let haystack = "The quick brown fox.";
///
/// let ac = AhoCorasick::new(patterns);
/// let mut result = vec![];
/// ac.stream_replace_all_with(
/// haystack.as_bytes(),
/// &mut result,
/// |mat, _, wtr| {
/// wtr.write_all(mat.pattern().to_string().as_bytes())
/// },
/// )?;
/// assert_eq!(b"The 2 1 0.".to_vec(), result);
/// # Ok(()) }; example().unwrap()
/// ```
pub fn stream_replace_all_with<R, W, F>(
&self,
rdr: R,
mut wtr: W,
mut replace_with: F,
) -> io::Result<()>
where
R: io::Read,
W: io::Write,
F: FnMut(&Match, &[u8], &mut W) -> io::Result<()>,
{
let mut it = StreamChunkIter::new(self, rdr);
while let Some(result) = it.next() {
let chunk = result?;
match chunk {
StreamChunk::NonMatch { bytes, .. } => {
wtr.write_all(bytes)?;
}
StreamChunk::Match { bytes, mat } => {
replace_with(&mat, bytes, &mut wtr)?;
}
}
}
Ok(())
}
/// Returns the match kind used by this automaton.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasick, MatchKind};
///
/// let ac = AhoCorasick::new(&[
/// "foo", "bar", "quux", "baz",
/// ]);
/// assert_eq!(&MatchKind::Standard, ac.match_kind());
/// ```
pub fn match_kind(&self) -> &MatchKind {
self.imp.match_kind()
}
/// Returns the length of the longest pattern matched by this automaton.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let ac = AhoCorasick::new(&[
/// "foo", "bar", "quux", "baz",
/// ]);
/// assert_eq!(4, ac.max_pattern_len());
/// ```
pub fn max_pattern_len(&self) -> usize {
self.imp.max_pattern_len()
}
/// Return the total number of patterns matched by this automaton.
///
/// This includes patterns that may never participate in a match. For
/// example, if
/// [`MatchKind::LeftmostFirst`](enum.MatchKind.html#variant.LeftmostFirst)
/// match semantics are used, and the patterns `Sam` and `Samwise` were
/// used to build the automaton, then `Samwise` can never participate in a
/// match because `Sam` will always take priority.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasick;
///
/// let ac = AhoCorasick::new(&[
/// "foo", "bar", "baz",
/// ]);
/// assert_eq!(3, ac.pattern_count());
/// ```
pub fn pattern_count(&self) -> usize {
self.imp.pattern_count()
}
/// Returns true if and only if this automaton supports reporting
/// overlapping matches.
///
/// If this returns false and overlapping matches are requested, then it
/// will result in a panic.
///
/// Since leftmost matching is inherently incompatible with overlapping
/// matches, only
/// [`MatchKind::Standard`](enum.MatchKind.html#variant.Standard)
/// supports overlapping matches. This is unlikely to change in the future.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::Standard)
/// .build(&["foo", "bar", "baz"]);
/// assert!(ac.supports_overlapping());
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(&["foo", "bar", "baz"]);
/// assert!(!ac.supports_overlapping());
/// ```
pub fn supports_overlapping(&self) -> bool {
self.match_kind.supports_overlapping()
}
/// Returns true if and only if this automaton supports stream searching.
///
/// If this returns false and stream searching (or replacing) is attempted,
/// then it will result in a panic.
///
/// Currently, only
/// [`MatchKind::Standard`](enum.MatchKind.html#variant.Standard)
/// supports streaming. This may be expanded in the future.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::Standard)
/// .build(&["foo", "bar", "baz"]);
/// assert!(ac.supports_stream());
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(&["foo", "bar", "baz"]);
/// assert!(!ac.supports_stream());
/// ```
pub fn supports_stream(&self) -> bool {
self.match_kind.supports_stream()
}
/// Returns the approximate total amount of heap used by this automaton, in
/// units of bytes.
///
/// # Examples
///
/// This example shows the difference in heap usage between a few
/// configurations:
///
/// ```ignore
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let ac = AhoCorasickBuilder::new()
/// .dfa(false) // default
/// .build(&["foo", "bar", "baz"]);
/// assert_eq!(10_336, ac.heap_bytes());
///
/// let ac = AhoCorasickBuilder::new()
/// .dfa(false) // default
/// .ascii_case_insensitive(true)
/// .build(&["foo", "bar", "baz"]);
/// assert_eq!(10_384, ac.heap_bytes());
///
/// let ac = AhoCorasickBuilder::new()
/// .dfa(true)
/// .ascii_case_insensitive(true)
/// .build(&["foo", "bar", "baz"]);
/// assert_eq!(1_248, ac.heap_bytes());
/// ```
pub fn heap_bytes(&self) -> usize {
match self.imp {
Imp::NFA(ref nfa) => nfa.heap_bytes(),
Imp::DFA(ref dfa) => dfa.heap_bytes(),
}
}
}
/// The internal implementation of Aho-Corasick, which is either an NFA or
/// a DFA. The NFA is slower but uses less memory. The DFA is faster but uses
/// more memory.
#[derive(Clone, Debug)]
enum Imp<S: StateID> {
NFA(NFA<S>),
DFA(DFA<S>),
}
impl<S: StateID> Imp<S> {
/// Returns the type of match semantics implemented by this automaton.
fn match_kind(&self) -> &MatchKind {
match *self {
Imp::NFA(ref nfa) => nfa.match_kind(),
Imp::DFA(ref dfa) => dfa.match_kind(),
}
}
/// Returns the identifier of the start state.
fn start_state(&self) -> S {
match *self {
Imp::NFA(ref nfa) => nfa.start_state(),
Imp::DFA(ref dfa) => dfa.start_state(),
}
}
/// The length, in bytes, of the longest pattern in this automaton. This
/// information is useful for maintaining correct buffer sizes when
/// searching on streams.
fn max_pattern_len(&self) -> usize {
match *self {
Imp::NFA(ref nfa) => nfa.max_pattern_len(),
Imp::DFA(ref dfa) => dfa.max_pattern_len(),
}
}
/// The total number of patterns added to this automaton. This includes
/// patterns that may never match. The maximum matching pattern that can be
/// reported is exactly one less than this number.
fn pattern_count(&self) -> usize {
match *self {
Imp::NFA(ref nfa) => nfa.pattern_count(),
Imp::DFA(ref dfa) => dfa.pattern_count(),
}
}
/// Returns the prefilter object, if one exists, for the underlying
/// automaton.
fn prefilter(&self) -> Option<&dyn Prefilter> {
match *self {
Imp::NFA(ref nfa) => nfa.prefilter(),
Imp::DFA(ref dfa) => dfa.prefilter(),
}
}
/// Returns true if and only if we should attempt to use a prefilter.
fn use_prefilter(&self) -> bool {
let p = match self.prefilter() {
None => return false,
Some(p) => p,
};
!p.looks_for_non_start_of_match()
}
#[inline(always)]
fn overlapping_find_at(
&self,
prestate: &mut PrefilterState,
haystack: &[u8],
at: usize,
state_id: &mut S,
match_index: &mut usize,
) -> Option<Match> {
match *self {
Imp::NFA(ref nfa) => nfa.overlapping_find_at(
prestate,
haystack,
at,
state_id,
match_index,
),
Imp::DFA(ref dfa) => dfa.overlapping_find_at(
prestate,
haystack,
at,
state_id,
match_index,
),
}
}
#[inline(always)]
fn earliest_find_at(
&self,
prestate: &mut PrefilterState,
haystack: &[u8],
at: usize,
state_id: &mut S,
) -> Option<Match> {
match *self {
Imp::NFA(ref nfa) => {
nfa.earliest_find_at(prestate, haystack, at, state_id)
}
Imp::DFA(ref dfa) => {
dfa.earliest_find_at(prestate, haystack, at, state_id)
}
}
}
#[inline(always)]
fn find_at_no_state(
&self,
prestate: &mut PrefilterState,
haystack: &[u8],
at: usize,
) -> Option<Match> {
match *self {
Imp::NFA(ref nfa) => nfa.find_at_no_state(prestate, haystack, at),
Imp::DFA(ref dfa) => dfa.find_at_no_state(prestate, haystack, at),
}
}
}
/// An iterator of non-overlapping matches in a particular haystack.
///
/// This iterator yields matches according to the
/// [`MatchKind`](enum.MatchKind.html)
/// used by this automaton.
///
/// This iterator is constructed via the
/// [`AhoCorasick::find_iter`](struct.AhoCorasick.html#method.find_iter)
/// method.
///
/// The type variable `S` refers to the representation used for state
/// identifiers. (By default, this is `usize`.)
///
/// The lifetime `'a` refers to the lifetime of the `AhoCorasick` automaton.
///
/// The lifetime `'b` refers to the lifetime of the haystack being searched.
#[derive(Debug)]
pub struct FindIter<'a, 'b, S: StateID> {
fsm: &'a Imp<S>,
prestate: PrefilterState,
haystack: &'b [u8],
pos: usize,
}
impl<'a, 'b, S: StateID> FindIter<'a, 'b, S> {
fn new(ac: &'a AhoCorasick<S>, haystack: &'b [u8]) -> FindIter<'a, 'b, S> {
let prestate = PrefilterState::new(ac.max_pattern_len());
FindIter { fsm: &ac.imp, prestate, haystack, pos: 0 }
}
}
impl<'a, 'b, S: StateID> Iterator for FindIter<'a, 'b, S> {
type Item = Match;
fn next(&mut self) -> Option<Match> {
if self.pos > self.haystack.len() {
return None;
}
let result = self.fsm.find_at_no_state(
&mut self.prestate,
self.haystack,
self.pos,
);
let mat = match result {
None => return None,
Some(mat) => mat,
};
if mat.end() == self.pos {
// If the automaton can match the empty string and if we found an
// empty match, then we need to forcefully move the position.
self.pos += 1;
} else {
self.pos = mat.end();
}
Some(mat)
}
}
/// An iterator of overlapping matches in a particular haystack.
///
/// This iterator will report all possible matches in a particular haystack,
/// even when the matches overlap.
///
/// This iterator is constructed via the
/// [`AhoCorasick::find_overlapping_iter`](struct.AhoCorasick.html#method.find_overlapping_iter)
/// method.
///
/// The type variable `S` refers to the representation used for state
/// identifiers. (By default, this is `usize`.)
///
/// The lifetime `'a` refers to the lifetime of the `AhoCorasick` automaton.
///
/// The lifetime `'b` refers to the lifetime of the haystack being searched.
#[derive(Debug)]
pub struct FindOverlappingIter<'a, 'b, S: StateID> {
fsm: &'a Imp<S>,
prestate: PrefilterState,
haystack: &'b [u8],
pos: usize,
state_id: S,
match_index: usize,
}
impl<'a, 'b, S: StateID> FindOverlappingIter<'a, 'b, S> {
fn new(
ac: &'a AhoCorasick<S>,
haystack: &'b [u8],
) -> FindOverlappingIter<'a, 'b, S> {
assert!(
ac.supports_overlapping(),
"automaton does not support overlapping searches"
);
let prestate = PrefilterState::new(ac.max_pattern_len());
FindOverlappingIter {
fsm: &ac.imp,
prestate,
haystack,
pos: 0,
state_id: ac.imp.start_state(),
match_index: 0,
}
}
}
impl<'a, 'b, S: StateID> Iterator for FindOverlappingIter<'a, 'b, S> {
type Item = Match;
fn next(&mut self) -> Option<Match> {
let result = self.fsm.overlapping_find_at(
&mut self.prestate,
self.haystack,
self.pos,
&mut self.state_id,
&mut self.match_index,
);
match result {
None => return None,
Some(m) => {
self.pos = m.end();
Some(m)
}
}
}
}
/// An iterator that reports Aho-Corasick matches in a stream.
///
/// This iterator yields elements of type `io::Result<Match>`, where an error
/// is reported if there was a problem reading from the underlying stream.
/// The iterator terminates only when the underlying stream reaches `EOF`.
///
/// This iterator is constructed via the
/// [`AhoCorasick::stream_find_iter`](struct.AhoCorasick.html#method.stream_find_iter)
/// method.
///
/// The type variable `R` refers to the `io::Read` stream that is being read
/// from.
///
/// The type variable `S` refers to the representation used for state
/// identifiers. (By default, this is `usize`.)
///
/// The lifetime `'a` refers to the lifetime of the `AhoCorasick` automaton.
#[derive(Debug)]
pub struct StreamFindIter<'a, R, S: StateID> {
it: StreamChunkIter<'a, R, S>,
}
impl<'a, R: io::Read, S: StateID> StreamFindIter<'a, R, S> {
fn new(ac: &'a AhoCorasick<S>, rdr: R) -> StreamFindIter<'a, R, S> {
StreamFindIter { it: StreamChunkIter::new(ac, rdr) }
}
}
impl<'a, R: io::Read, S: StateID> Iterator for StreamFindIter<'a, R, S> {
type Item = io::Result<Match>;
fn next(&mut self) -> Option<io::Result<Match>> {
loop {
match self.it.next() {
None => return None,
Some(Err(err)) => return Some(Err(err)),
Some(Ok(StreamChunk::NonMatch { .. })) => {}
Some(Ok(StreamChunk::Match { mat, .. })) => {
return Some(Ok(mat));
}
}
}
}
}
/// An iterator over chunks in an underlying reader. Each chunk either
/// corresponds to non-matching bytes or matching bytes, but all bytes from
/// the underlying reader are reported in sequence. There may be an arbitrary
/// number of non-matching chunks before seeing a matching chunk.
///
/// N.B. This does not actually implement Iterator because we need to borrow
/// from the underlying reader. But conceptually, it's still an iterator.
#[derive(Debug)]
struct StreamChunkIter<'a, R, S: StateID> {
/// The AC automaton.
fsm: &'a Imp<S>,
/// State associated with this automaton's prefilter. It is a heuristic
/// for stopping the prefilter if it's deemed ineffective.
prestate: PrefilterState,
/// The source of bytes we read from.
rdr: R,
/// A fixed size buffer. This is what we actually search. There are some
/// invariants around the buffer's size, namely, it must be big enough to
/// contain the longest possible match.
buf: Buffer,
/// The ID of the FSM state we're currently in.
state_id: S,
/// The current position at which to start the next search in `buf`.
search_pos: usize,
/// The absolute position of `search_pos`, where `0` corresponds to the
/// position of the first byte read from `rdr`.
absolute_pos: usize,
/// The ending position of the last StreamChunk that was returned to the
/// caller. This position is used to determine whether we need to emit
/// non-matching bytes before emitting a match.
report_pos: usize,
/// A match that should be reported on the next call.
pending_match: Option<Match>,
/// Enabled only when the automaton can match the empty string. When
/// enabled, we need to execute one final search after consuming the
/// reader to find the trailing empty match.
has_empty_match_at_end: bool,
}
/// A single chunk yielded by the stream chunk iterator.
///
/// The `'r` lifetime refers to the lifetime of the stream chunk iterator.
#[derive(Debug)]
enum StreamChunk<'r> {
/// A chunk that does not contain any matches.
NonMatch { bytes: &'r [u8] },
/// A chunk that precisely contains a match.
Match { bytes: &'r [u8], mat: Match },
}
impl<'a, R: io::Read, S: StateID> StreamChunkIter<'a, R, S> {
fn new(ac: &'a AhoCorasick<S>, rdr: R) -> StreamChunkIter<'a, R, S> {
assert!(
ac.supports_stream(),
"stream searching is only supported for Standard match semantics"
);
let prestate = if ac.imp.use_prefilter() {
PrefilterState::new(ac.max_pattern_len())
} else {
PrefilterState::disabled()
};
let buf = Buffer::new(ac.imp.max_pattern_len());
let state_id = ac.imp.start_state();
StreamChunkIter {
fsm: &ac.imp,
prestate,
rdr,
buf,
state_id,
absolute_pos: 0,
report_pos: 0,
search_pos: 0,
pending_match: None,
has_empty_match_at_end: ac.is_match(""),
}
}
fn next(&mut self) -> Option<io::Result<StreamChunk>> {
loop {
if let Some(mut mat) = self.pending_match.take() {
let bytes = &self.buf.buffer()[mat.start()..mat.end()];
self.report_pos = mat.end();
mat = mat.increment(self.absolute_pos);
return Some(Ok(StreamChunk::Match { bytes, mat }));
}
if self.search_pos >= self.buf.len() {
if let Some(end) = self.unreported() {
let bytes = &self.buf.buffer()[self.report_pos..end];
self.report_pos = end;
return Some(Ok(StreamChunk::NonMatch { bytes }));
}
if self.buf.len() >= self.buf.min_buffer_len() {
// This is the point at which we roll our buffer, which we
// only do if our buffer has at least the minimum amount of
// bytes in it. Before rolling, we update our various
// positions to be consistent with the buffer after it has
// been rolled.
self.report_pos -=
self.buf.len() - self.buf.min_buffer_len();
self.absolute_pos +=
self.search_pos - self.buf.min_buffer_len();
self.search_pos = self.buf.min_buffer_len();
self.buf.roll();
}
match self.buf.fill(&mut self.rdr) {
Err(err) => return Some(Err(err)),
Ok(false) => {
// We've hit EOF, but if there are still some
// unreported bytes remaining, return them now.
if self.report_pos < self.buf.len() {
let bytes = &self.buf.buffer()[self.report_pos..];
self.report_pos = self.buf.len();
let chunk = StreamChunk::NonMatch { bytes };
return Some(Ok(chunk));
} else {
// We've reported everything, but there might still
// be a match at the very last position.
if !self.has_empty_match_at_end {
return None;
}
// fallthrough for another search to get trailing
// empty matches
self.has_empty_match_at_end = false;
}
}
Ok(true) => {}
}
}
let result = self.fsm.earliest_find_at(
&mut self.prestate,
self.buf.buffer(),
self.search_pos,
&mut self.state_id,
);
match result {
None => {
self.search_pos = self.buf.len();
}
Some(mat) => {
self.state_id = self.fsm.start_state();
if mat.end() == self.search_pos {
// If the automaton can match the empty string and if
// we found an empty match, then we need to forcefully
// move the position.
self.search_pos += 1;
} else {
self.search_pos = mat.end();
}
self.pending_match = Some(mat.clone());
if self.report_pos < mat.start() {
let bytes =
&self.buf.buffer()[self.report_pos..mat.start()];
self.report_pos = mat.start();
let chunk = StreamChunk::NonMatch { bytes };
return Some(Ok(chunk));
}
}
}
}
}
fn unreported(&self) -> Option<usize> {
let end = self.search_pos.saturating_sub(self.buf.min_buffer_len());
if self.report_pos < end {
Some(end)
} else {
None
}
}
}
/// A builder for configuring an Aho-Corasick automaton.
#[derive(Clone, Debug)]
pub struct AhoCorasickBuilder {
nfa_builder: nfa::Builder,
dfa_builder: dfa::Builder,
dfa: bool,
}
impl Default for AhoCorasickBuilder {
fn default() -> AhoCorasickBuilder {
AhoCorasickBuilder::new()
}
}
impl AhoCorasickBuilder {
/// Create a new builder for configuring an Aho-Corasick automaton.
///
/// If you don't need fine grained configuration or aren't sure which knobs
/// to set, try using
/// [`AhoCorasick::new_auto_configured`](struct.AhoCorasick.html#method.new_auto_configured)
/// instead.
pub fn new() -> AhoCorasickBuilder {
AhoCorasickBuilder {
nfa_builder: nfa::Builder::new(),
dfa_builder: dfa::Builder::new(),
dfa: false,
}
}
/// Build an Aho-Corasick automaton using the configuration set on this
/// builder.
///
/// A builder may be reused to create more automatons.
///
/// This method will use the default for representing internal state
/// identifiers, which is `usize`. This guarantees that building the
/// automaton will succeed and is generally a good default, but can make
/// the size of the automaton 2-8 times bigger than it needs to be,
/// depending on your target platform.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasickBuilder;
///
/// let patterns = &["foo", "bar", "baz"];
/// let ac = AhoCorasickBuilder::new()
/// .build(patterns);
/// assert_eq!(Some(1), ac.find("xxx bar xxx").map(|m| m.pattern()));
/// ```
pub fn build<I, P>(&self, patterns: I) -> AhoCorasick
where
I: IntoIterator<Item = P>,
P: AsRef<[u8]>,
{
// The builder only returns an error if the chosen state ID
// representation is too small to fit all of the given patterns. In
// this case, since we fix the representation to usize, it will always
// work because it's impossible to overflow usize since the underlying
// storage would OOM long before that happens.
self.build_with_size::<usize, I, P>(patterns)
.expect("usize state ID type should always work")
}
/// Build an Aho-Corasick automaton using the configuration set on this
/// builder with a specific state identifier representation. This only has
/// an effect when the `dfa` option is enabled.
///
/// Generally, the choices for a state identifier representation are
/// `u8`, `u16`, `u32`, `u64` or `usize`, with `usize` being the default.
/// The advantage of choosing a smaller state identifier representation
/// is that the automaton produced will be smaller. This might be
/// beneficial for just generally using less space, or might even allow it
/// to fit more of the automaton in your CPU's cache, leading to overall
/// better search performance.
///
/// Unlike the standard `build` method, this can report an error if the
/// state identifier representation cannot support the size of the
/// automaton.
///
/// Note that the state identifier representation is determined by the
/// `S` type variable. This requires a type hint of some sort, either
/// by specifying the return type or using the turbofish, e.g.,
/// `build_with_size::<u16, _, _>(...)`.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::{AhoCorasick, AhoCorasickBuilder};
///
/// # fn example() -> Result<(), ::aho_corasick::Error> {
/// let patterns = &["foo", "bar", "baz"];
/// let ac: AhoCorasick<u8> = AhoCorasickBuilder::new()
/// .build_with_size(patterns)?;
/// assert_eq!(Some(1), ac.find("xxx bar xxx").map(|m| m.pattern()));
/// # Ok(()) }; example().unwrap()
/// ```
///
/// Or alternatively, with turbofish:
///
/// ```
/// use aho_corasick::AhoCorasickBuilder;
///
/// # fn example() -> Result<(), ::aho_corasick::Error> {
/// let patterns = &["foo", "bar", "baz"];
/// let ac = AhoCorasickBuilder::new()
/// .build_with_size::<u8, _, _>(patterns)?;
/// assert_eq!(Some(1), ac.find("xxx bar xxx").map(|m| m.pattern()));
/// # Ok(()) }; example().unwrap()
/// ```
pub fn build_with_size<S, I, P>(
&self,
patterns: I,
) -> Result<AhoCorasick<S>>
where
S: StateID,
I: IntoIterator<Item = P>,
P: AsRef<[u8]>,
{
let nfa = self.nfa_builder.build(patterns)?;
let match_kind = nfa.match_kind().clone();
let imp = if self.dfa {
let dfa = self.dfa_builder.build(&nfa)?;
Imp::DFA(dfa)
} else {
Imp::NFA(nfa)
};
Ok(AhoCorasick { imp, match_kind })
}
/// Automatically configure the settings on this builder according to the
/// patterns that will be used to construct the automaton.
///
/// The idea here is to balance space and time automatically. That is, when
/// searching a small number of patterns, this will attempt to use the
/// fastest possible configuration since the total space required will be
/// small anyway. As the number of patterns grows, this will fall back to
/// slower configurations that use less space.
///
/// This is guaranteed to never set `match_kind`, but any other option may
/// be overridden.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasickBuilder;
///
/// let patterns = &["foo", "bar", "baz"];
/// let ac = AhoCorasickBuilder::new()
/// .auto_configure(patterns)
/// .build(patterns);
/// assert_eq!(Some(1), ac.find("xxx bar xxx").map(|m| m.pattern()));
/// ```
pub fn auto_configure<B: AsRef<[u8]>>(
&mut self,
patterns: &[B],
) -> &mut AhoCorasickBuilder {
// N.B. Currently we only use the length of `patterns` to make a
// decision here, and could therefore ask for an `ExactSizeIterator`
// instead. But it's conceivable that we might adapt this to look at
// the total number of bytes, which would requires a second pass.
//
// The logic here is fairly rudimentary at the moment, but probably
// OK. The idea here is to use the fastest thing possible for a small
// number of patterns. That is, a DFA with no byte classes, since byte
// classes require an extra indirection for every byte searched. With a
// moderate number of patterns, we still want a DFA, but save on both
// space and compilation time by enabling byte classes. Finally, fall
// back to the slower but smaller NFA.
if patterns.len() <= 100 {
// N.B. Using byte classes can actually be faster by improving
// locality, but this only really applies for multi-megabyte
// automata (i.e., automata that don't fit in your CPU's cache).
self.dfa(true);
} else if patterns.len() <= 5000 {
self.dfa(true);
}
self
}
/// Set the desired match semantics.
///
/// The default is `MatchKind::Standard`, which corresponds to the match
/// semantics supported by the standard textbook description of the
/// Aho-Corasick algorithm. Namely, matches are reported as soon as they
/// are found. Moreover, this is the only way to get overlapping matches
/// or do stream searching.
///
/// The other kinds of match semantics that are supported are
/// `MatchKind::LeftmostFirst` and `MatchKind::LeftmostLongest`. The former
/// corresponds to the match you would get if you were to try to match
/// each pattern at each position in the haystack in the same order that
/// you give to the automaton. That is, it returns the leftmost match
/// corresponding the earliest pattern given to the automaton. The latter
/// corresponds to finding the longest possible match among all leftmost
/// matches.
///
/// For more details on match semantics, see the
/// [documentation for `MatchKind`](enum.MatchKind.html).
///
/// # Examples
///
/// In these examples, we demonstrate the differences between match
/// semantics for a particular set of patterns in a specific order:
/// `b`, `abc`, `abcd`.
///
/// Standard semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["b", "abc", "abcd"];
/// let haystack = "abcd";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::Standard) // default, not necessary
/// .build(patterns);
/// let mat = ac.find(haystack).expect("should have a match");
/// assert_eq!("b", &haystack[mat.start()..mat.end()]);
/// ```
///
/// Leftmost-first semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["b", "abc", "abcd"];
/// let haystack = "abcd";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostFirst)
/// .build(patterns);
/// let mat = ac.find(haystack).expect("should have a match");
/// assert_eq!("abc", &haystack[mat.start()..mat.end()]);
/// ```
///
/// Leftmost-longest semantics:
///
/// ```
/// use aho_corasick::{AhoCorasickBuilder, MatchKind};
///
/// let patterns = &["b", "abc", "abcd"];
/// let haystack = "abcd";
///
/// let ac = AhoCorasickBuilder::new()
/// .match_kind(MatchKind::LeftmostLongest)
/// .build(patterns);
/// let mat = ac.find(haystack).expect("should have a match");
/// assert_eq!("abcd", &haystack[mat.start()..mat.end()]);
/// ```
pub fn match_kind(&mut self, kind: MatchKind) -> &mut AhoCorasickBuilder {
self.nfa_builder.match_kind(kind);
self
}
/// Enable anchored mode, which requires all matches to start at the
/// first position in a haystack.
///
/// This option is disabled by default.
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasickBuilder;
///
/// let patterns = &["foo", "bar"];
/// let haystack = "foobar";
///
/// let ac = AhoCorasickBuilder::new()
/// .anchored(true)
/// .build(patterns);
/// assert_eq!(1, ac.find_iter(haystack).count());
/// ```
///
/// When searching for overlapping matches, all matches that start at
/// the beginning of a haystack will be reported:
///
/// ```
/// use aho_corasick::AhoCorasickBuilder;
///
/// let patterns = &["foo", "foofoo"];
/// let haystack = "foofoo";
///
/// let ac = AhoCorasickBuilder::new()
/// .anchored(true)
/// .build(patterns);
/// assert_eq!(2, ac.find_overlapping_iter(haystack).count());
/// // A non-anchored search would return 3 matches.
/// ```
pub fn anchored(&mut self, yes: bool) -> &mut AhoCorasickBuilder {
self.nfa_builder.anchored(yes);
self
}
/// Enable ASCII-aware case insensitive matching.
///
/// When this option is enabled, searching will be performed without
/// respect to case for ASCII letters (`a-z` and `A-Z`) only.
///
/// Enabling this option does not change the search algorithm, but it may
/// increase the size of the automaton.
///
/// **NOTE:** It is unlikely that support for Unicode case folding will
/// be added in the future. The ASCII case works via a simple hack to the
/// underlying automaton, but full Unicode handling requires a fair bit of
/// sophistication. If you do need Unicode handling, you might consider
/// using the [`regex` crate](https://docs.rs/regex) or the lower level
/// [`regex-automata` crate](https://docs.rs/regex-automata).
///
/// # Examples
///
/// Basic usage:
///
/// ```
/// use aho_corasick::AhoCorasickBuilder;
///
/// let patterns = &["FOO", "bAr", "BaZ"];
/// let haystack = "foo bar baz";
///
/// let ac = AhoCorasickBuilder::new()
/// .ascii_case_insensitive(true)
/// .build(patterns);
/// assert_eq!(3, ac.find_iter(haystack).count());
/// ```
pub fn ascii_case_insensitive(
&mut self,
yes: bool,
) -> &mut AhoCorasickBuilder {
self.nfa_builder.ascii_case_insensitive(yes);
self
}
/// Set the limit on how many NFA states use a dense representation for
/// their transitions.
///
/// A dense representation uses more space, but supports faster access to
/// transitions at search time. Thus, this setting permits the control of a
/// space vs time trade off when using the NFA variant of Aho-Corasick.
///
/// This limit is expressed in terms of the depth of a state, i.e., the
/// number of transitions from the starting state of the NFA. The idea is
/// that most of the time searching will be spent near the starting state
/// of the automaton, so states near the start state should use a dense
/// representation. States further away from the start state would then use
/// a sparse representation, which uses less space but is slower to access
/// transitions at search time.
///
/// By default, this is set to a low but non-zero number.
///
/// This setting has no effect if the `dfa` option is enabled.
pub fn dense_depth(&mut self, depth: usize) -> &mut AhoCorasickBuilder {
self.nfa_builder.dense_depth(depth);
self
}
/// Compile the standard Aho-Corasick automaton into a deterministic finite
/// automaton (DFA).
///
/// When this is disabled (which is the default), then a non-deterministic
/// finite automaton (NFA) is used instead.
///
/// The main benefit to a DFA is that it can execute searches more quickly
/// than a NFA (perhaps 2-4 times as fast). The main drawback is that the
/// DFA uses more space and can take much longer to build.
///
/// Enabling this option does not change the time complexity for
/// constructing the Aho-Corasick automaton (which is `O(p)` where
/// `p` is the total number of patterns being compiled). Enabling this
/// option does however reduce the time complexity of non-overlapping
/// searches from `O(n + p)` to `O(n)`, where `n` is the length of the
/// haystack.
///
/// In general, it's a good idea to enable this if you're searching a
/// small number of fairly short patterns (~1000), or if you want the
/// fastest possible search without regard to compilation time or space
/// usage.
pub fn dfa(&mut self, yes: bool) -> &mut AhoCorasickBuilder {
self.dfa = yes;
self
}
/// Enable heuristic prefilter optimizations.
///
/// When enabled, searching will attempt to quickly skip to match
/// candidates using specialized literal search routines. A prefilter
/// cannot always be used, and is generally treated as a heuristic. It
/// can be useful to disable this if the prefilter is observed to be
/// sub-optimal for a particular workload.
///
/// This is enabled by default.
pub fn prefilter(&mut self, yes: bool) -> &mut AhoCorasickBuilder {
self.nfa_builder.prefilter(yes);
self
}
/// Shrink the size of the transition alphabet by mapping bytes to their
/// equivalence classes. This only has an effect when the `dfa` option is
/// enabled.
///
/// When enabled, each a DFA will use a map from all possible bytes
/// to their corresponding equivalence class. Each equivalence class
/// represents a set of bytes that does not discriminate between a match
/// and a non-match in the DFA. For example, the patterns `bar` and `baz`
/// have at least five equivalence classes: singleton sets of `b`, `a`, `r`
/// and `z`, and a final set that contains every other byte.
///
/// The advantage of this map is that the size of the transition table can
/// be reduced drastically from `#states * 256 * sizeof(id)` to
/// `#states * k * sizeof(id)` where `k` is the number of equivalence
/// classes. As a result, total space usage can decrease substantially.
/// Moreover, since a smaller alphabet is used, compilation becomes faster
/// as well.
///
/// The disadvantage of this map is that every byte searched must be
/// passed through this map before it can be used to determine the next
/// transition. This has a small match time performance cost. However, if
/// the DFA is otherwise very large without byte classes, then using byte
/// classes can greatly improve memory locality and thus lead to better
/// overall performance.
///
/// This option is enabled by default.
#[deprecated(
since = "0.7.16",
note = "not carrying its weight, will be always enabled, see: https://github.com/BurntSushi/aho-corasick/issues/57"
)]
pub fn byte_classes(&mut self, yes: bool) -> &mut AhoCorasickBuilder {
self.dfa_builder.byte_classes(yes);
self
}
/// Premultiply state identifiers in the transition table. This only has
/// an effect when the `dfa` option is enabled.
///
/// When enabled, state identifiers are premultiplied to point to their
/// corresponding row in the transition table. That is, given the `i`th
/// state, its corresponding premultiplied identifier is `i * k` where `k`
/// is the alphabet size of the automaton. (The alphabet size is at most
/// 256, but is in practice smaller if byte classes is enabled.)
///
/// When state identifiers are not premultiplied, then the identifier of
/// the `i`th state is `i`.
///
/// The advantage of premultiplying state identifiers is that is saves a
/// multiplication instruction per byte when searching with a DFA. This has
/// been observed to lead to a 20% performance benefit in micro-benchmarks.
///
/// The primary disadvantage of premultiplying state identifiers is
/// that they require a larger integer size to represent. For example,
/// if the DFA has 200 states, then its premultiplied form requires 16
/// bits to represent every possible state identifier, where as its
/// non-premultiplied form only requires 8 bits.
///
/// This option is enabled by default.
#[deprecated(
since = "0.7.16",
note = "not carrying its weight, will be always enabled, see: https://github.com/BurntSushi/aho-corasick/issues/57"
)]
pub fn premultiply(&mut self, yes: bool) -> &mut AhoCorasickBuilder {
self.dfa_builder.premultiply(yes);
self
}
}
/// A knob for controlling the match semantics of an Aho-Corasick automaton.
///
/// There are two generally different ways that Aho-Corasick automatons can
/// report matches. The first way is the "standard" approach that results from
/// implementing most textbook explanations of Aho-Corasick. The second way is
/// to report only the leftmost non-overlapping matches. The leftmost approach
/// is in turn split into two different ways of resolving ambiguous matches:
/// leftmost-first and leftmost-longest.
///
/// The `Standard` match kind is the default and is the only one that supports
/// overlapping matches and stream searching. (Trying to find overlapping
/// or streaming matches using leftmost match semantics will result in a
/// panic.) The `Standard` match kind will report matches as they are seen.
/// When searching for overlapping matches, then all possible matches are
/// reported. When searching for non-overlapping matches, the first match seen
/// is reported. For example, for non-overlapping matches, given the patterns
/// `abcd` and `b` and the subject string `abcdef`, only a match for `b` is
/// reported since it is detected first. The `abcd` match is never reported
/// since it overlaps with the `b` match.
///
/// In contrast, the leftmost match kind always prefers the leftmost match
/// among all possible matches. Given the same example as above with `abcd` and
/// `b` as patterns and `abcdef` as the subject string, the leftmost match is
/// `abcd` since it begins before the `b` match, even though the `b` match is
/// detected before the `abcd` match. In this case, the `b` match is not
/// reported at all since it overlaps with the `abcd` match.
///
/// The difference between leftmost-first and leftmost-longest is in how they
/// resolve ambiguous matches when there are multiple leftmost matches to
/// choose from. Leftmost-first always chooses the pattern that was provided
/// earliest, where as leftmost-longest always chooses the longest matching
/// pattern. For example, given the patterns `a` and `ab` and the subject
/// string `ab`, the leftmost-first match is `a` but the leftmost-longest match
/// is `ab`. Conversely, if the patterns were given in reverse order, i.e.,
/// `ab` and `a`, then both the leftmost-first and leftmost-longest matches
/// would be `ab`. Stated differently, the leftmost-first match depends on the
/// order in which the patterns were given to the Aho-Corasick automaton.
/// Because of that, when leftmost-first matching is used, if a pattern `A`
/// that appears before a pattern `B` is a prefix of `B`, then it is impossible
/// to ever observe a match of `B`.
///
/// If you're not sure which match kind to pick, then stick with the standard
/// kind, which is the default. In particular, if you need overlapping or
/// streaming matches, then you _must_ use the standard kind. The leftmost
/// kinds are useful in specific circumstances. For example, leftmost-first can
/// be very useful as a way to implement match priority based on the order of
/// patterns given and leftmost-longest can be useful for dictionary searching
/// such that only the longest matching words are reported.
///
/// # Relationship with regular expression alternations
///
/// Understanding match semantics can be a little tricky, and one easy way
/// to conceptualize non-overlapping matches from an Aho-Corasick automaton
/// is to think about them as a simple alternation of literals in a regular
/// expression. For example, let's say we wanted to match the strings
/// `Sam` and `Samwise`, which would turn into the regex `Sam|Samwise`. It
/// turns out that regular expression engines have two different ways of
/// matching this alternation. The first way, leftmost-longest, is commonly
/// found in POSIX compatible implementations of regular expressions (such as
/// `grep`). The second way, leftmost-first, is commonly found in backtracking
/// implementations such as Perl. (Some regex engines, such as RE2 and Rust's
/// regex engine do not use backtracking, but still implement leftmost-first
/// semantics in an effort to match the behavior of dominant backtracking
/// regex engines such as those found in Perl, Ruby, Python, Javascript and
/// PHP.)
///
/// That is, when matching `Sam|Samwise` against `Samwise`, a POSIX regex
/// will match `Samwise` because it is the longest possible match, but a
/// Perl-like regex will match `Sam` since it appears earlier in the
/// alternation. Indeed, the regex `Sam|Samwise` in a Perl-like regex engine
/// will never match `Samwise` since `Sam` will always have higher priority.
/// Conversely, matching the regex `Samwise|Sam` against `Samwise` will lead to
/// a match of `Samwise` in both POSIX and Perl-like regexes since `Samwise` is
/// still longest match, but it also appears earlier than `Sam`.
///
/// The "standard" match semantics of Aho-Corasick generally don't correspond
/// to the match semantics of any large group of regex implementations, so
/// there's no direct analogy that can be made here. Standard match semantics
/// are generally useful for overlapping matches, or if you just want to see
/// matches as they are detected.
///
/// The main conclusion to draw from this section is that the match semantics
/// can be tweaked to precisely match either Perl-like regex alternations or
/// POSIX regex alternations.
#[derive(Clone, Copy, Debug, Eq, PartialEq)]
pub enum MatchKind {
/// Use standard match semantics, which support overlapping matches. When
/// used with non-overlapping matches, matches are reported as they are
/// seen.
Standard,
/// Use leftmost-first match semantics, which reports leftmost matches.
/// When there are multiple possible leftmost matches, the match
/// corresponding to the pattern that appeared earlier when constructing
/// the automaton is reported.
///
/// This does **not** support overlapping matches or stream searching. If
/// this match kind is used, attempting to find overlapping matches or
/// stream matches will panic.
LeftmostFirst,
/// Use leftmost-longest match semantics, which reports leftmost matches.
/// When there are multiple possible leftmost matches, the longest match
/// is chosen.
///
/// This does **not** support overlapping matches or stream searching. If
/// this match kind is used, attempting to find overlapping matches or
/// stream matches will panic.
LeftmostLongest,
/// Hints that destructuring should not be exhaustive.
///
/// This enum may grow additional variants, so this makes sure clients
/// don't count on exhaustive matching. (Otherwise, adding a new variant
/// could break existing code.)
#[doc(hidden)]
__Nonexhaustive,
}
/// The default match kind is `MatchKind::Standard`.
impl Default for MatchKind {
fn default() -> MatchKind {
MatchKind::Standard
}
}
impl MatchKind {
fn supports_overlapping(&self) -> bool {
self.is_standard()
}
fn supports_stream(&self) -> bool {
// TODO: It may be possible to support this. It's hard.
//
// See: https://github.com/rust-lang/regex/issues/425#issuecomment-471367838
self.is_standard()
}
pub(crate) fn is_standard(&self) -> bool {
*self == MatchKind::Standard
}
pub(crate) fn is_leftmost(&self) -> bool {
*self == MatchKind::LeftmostFirst
|| *self == MatchKind::LeftmostLongest
}
pub(crate) fn is_leftmost_first(&self) -> bool {
*self == MatchKind::LeftmostFirst
}
/// Convert this match kind into a packed match kind. If this match kind
/// corresponds to standard semantics, then this returns None, since
/// packed searching does not support standard semantics.
pub(crate) fn as_packed(&self) -> Option<packed::MatchKind> {
match *self {
MatchKind::Standard => None,
MatchKind::LeftmostFirst => Some(packed::MatchKind::LeftmostFirst),
MatchKind::LeftmostLongest => {
Some(packed::MatchKind::LeftmostLongest)
}
MatchKind::__Nonexhaustive => unreachable!(),
}
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn oibits() {
use std::panic::{RefUnwindSafe, UnwindSafe};
fn assert_send<T: Send>() {}
fn assert_sync<T: Sync>() {}
fn assert_unwind_safe<T: RefUnwindSafe + UnwindSafe>() {}
assert_send::<AhoCorasick>();
assert_sync::<AhoCorasick>();
assert_unwind_safe::<AhoCorasick>();
assert_send::<AhoCorasickBuilder>();
assert_sync::<AhoCorasickBuilder>();
assert_unwind_safe::<AhoCorasickBuilder>();
}
}