| author | Christian Urban <christian dot urban at kcl dot ac dot uk> | 
| Tue, 18 Oct 2016 20:39:54 +0100 | |
| changeset 456 | 4abd90760ffe | 
| parent 454 | 010237a7dae7 | 
| child 467 | 3fc9b036321d | 
| permissions | -rw-r--r-- | 
| 54 | 1 | abstract class Rexp | 
| 424 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 2 | case object ZERO extends Rexp | 
| 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 3 | case object ONE extends Rexp | 
| 49 | 4 | case class CHAR(c: Char) extends Rexp | 
| 5 | case class ALT(r1: Rexp, r2: Rexp) extends Rexp | |
| 6 | case class SEQ(r1: Rexp, r2: Rexp) extends Rexp | |
| 7 | case class STAR(r: Rexp) extends Rexp | |
| 8 | ||
| 9 | // nullable function: tests whether the regular | |
| 10 | // expression can recognise the empty string | |
| 11 | def nullable (r: Rexp) : Boolean = r match {
 | |
| 424 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 12 | case ZERO => false | 
| 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 13 | case ONE => true | 
| 49 | 14 | case CHAR(_) => false | 
| 15 | case ALT(r1, r2) => nullable(r1) || nullable(r2) | |
| 16 | case SEQ(r1, r2) => nullable(r1) && nullable(r2) | |
| 17 | case STAR(_) => true | |
| 453 
a5525c780b96
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
440diff
changeset | 18 | |
| 49 | 19 | } | 
| 20 | ||
| 21 | // derivative of a regular expression w.r.t. a character | |
| 22 | def der (c: Char, r: Rexp) : Rexp = r match {
 | |
| 424 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 23 | case ZERO => ZERO | 
| 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 24 | case ONE => ZERO | 
| 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 25 | case CHAR(d) => if (c == d) ONE else ZERO | 
| 49 | 26 | case ALT(r1, r2) => ALT(der(c, r1), der(c, r2)) | 
| 27 | case SEQ(r1, r2) => | |
| 28 | if (nullable(r1)) ALT(SEQ(der(c, r1), r2), der(c, r2)) | |
| 29 | else SEQ(der(c, r1), r2) | |
| 440 
e14cd32ad497
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
434diff
changeset | 30 | case STAR(r1) => SEQ(der(c, r1), STAR(r1)) | 
| 49 | 31 | } | 
| 32 | ||
| 33 | // derivative w.r.t. a string (iterates der) | |
| 34 | def ders (s: List[Char], r: Rexp) : Rexp = s match {
 | |
| 35 | case Nil => r | |
| 54 | 36 | case c::s => ders(s, der(c, r)) | 
| 49 | 37 | } | 
| 38 | ||
| 39 | // main matcher function | |
| 261 
24531cfaa36a
updated handouts
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
258diff
changeset | 40 | def matches(r: Rexp, s: String) : Boolean = nullable(ders(s.toList, r)) | 
| 49 | 41 | |
| 424 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 42 | //examples from the homework | 
| 343 
539b2e88f5b9
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
261diff
changeset | 43 | val r = STAR(ALT(SEQ(CHAR('a'), CHAR('b')), CHAR('b')))
 | 
| 
539b2e88f5b9
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
261diff
changeset | 44 | der('a', r)
 | 
| 
539b2e88f5b9
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
261diff
changeset | 45 | der('b', r)
 | 
| 
539b2e88f5b9
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
261diff
changeset | 46 | der('c', r)
 | 
| 49 | 47 | |
| 258 
1e4da6d2490c
updated programs
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
119diff
changeset | 48 | //optional: one or zero times | 
| 424 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 49 | def OPT(r: Rexp) = ALT(r, ONE) | 
| 49 | 50 | |
| 51 | //n-times | |
| 52 | def NTIMES(r: Rexp, n: Int) : Rexp = n match {
 | |
| 424 
1129024b26d5
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
422diff
changeset | 53 | case 0 => ONE | 
| 49 | 54 | case 1 => r | 
| 55 | case n => SEQ(r, NTIMES(r, n - 1)) | |
| 56 | } | |
| 57 | ||
| 258 
1e4da6d2490c
updated programs
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
119diff
changeset | 58 | // the evil regular expression  a?{n} a{n}
 | 
| 415 | 59 | def EVIL1(n: Int) = SEQ(NTIMES(OPT(CHAR('a')), n), NTIMES(CHAR('a'), n))
 | 
| 422 | 60 | |
| 61 | // the evil regular expression (a*)*b | |
| 440 
e14cd32ad497
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
434diff
changeset | 62 | val EVIL2 = SEQ(STAR(STAR(CHAR('a'))), CHAR('b'))
 | 
| 49 | 63 | |
| 258 
1e4da6d2490c
updated programs
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
119diff
changeset | 64 | //for measuring time | 
| 49 | 65 | def time_needed[T](i: Int, code: => T) = {
 | 
| 66 | val start = System.nanoTime() | |
| 67 | for (j <- 1 to i) code | |
| 68 | val end = System.nanoTime() | |
| 69 | (end - start)/(i * 1.0e9) | |
| 70 | } | |
| 71 | ||
| 434 
8664ff87cd77
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
433diff
changeset | 72 | //test: (a?{n}) (a{n})
 | 
| 433 
c08290ee4f1f
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
424diff
changeset | 73 | for (i <- 1 to 20) {
 | 
| 
c08290ee4f1f
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
424diff
changeset | 74 | println(i + ": " + "%.5f".format(time_needed(2, matches(EVIL1(i), "a" * i)))) | 
| 
c08290ee4f1f
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
424diff
changeset | 75 | } | 
| 
c08290ee4f1f
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
424diff
changeset | 76 | |
| 415 | 77 | for (i <- 1 to 20) {
 | 
| 78 | println(i + ": " + "%.5f".format(time_needed(2, matches(EVIL1(i), "a" * i)))) | |
| 49 | 79 | } | 
| 363 
0d6deecdb2eb
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
343diff
changeset | 80 | |
| 434 
8664ff87cd77
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
433diff
changeset | 81 | //test: (a*)* b | 
| 440 
e14cd32ad497
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
434diff
changeset | 82 | for (i <- 1 to 20) {
 | 
| 434 
8664ff87cd77
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
433diff
changeset | 83 | println(i + " " + "%.5f".format(time_needed(2, matches(EVIL2, "a" * i)))) | 
| 
8664ff87cd77
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
433diff
changeset | 84 | } | 
| 
8664ff87cd77
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
433diff
changeset | 85 | |
| 440 
e14cd32ad497
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
434diff
changeset | 86 | for (i <- 1 to 20) {
 | 
| 415 | 87 | println(i + " " + "%.5f".format(time_needed(2, matches(EVIL2, "a" * i)))) | 
| 88 | } | |
| 363 
0d6deecdb2eb
updated
 Christian Urban <christian dot urban at kcl dot ac dot uk> parents: 
343diff
changeset | 89 |