| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      1 | // A simple matcher for basic regular expressions
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      2 | //
 | 
| 919 |      3 | // Call the testcases with X = {1,2,3}
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      4 | //
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      5 | //   amm re1.sc testX
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      6 | //
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      7 | // or 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      8 | //
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      9 | //   amm re1.sc all
 | 
| 826 |     10 | //
 | 
|  |     11 | 
 | 
| 745 |     12 |  
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     13 | // regular expressions
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     14 | abstract class Rexp
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     15 | case object ZERO extends Rexp                    // matches nothing
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     16 | case object ONE extends Rexp                     // matches an empty string
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     17 | case class CHAR(c: Char) extends Rexp            // matches a character c
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     18 | case class ALT(r1: Rexp, r2: Rexp) extends Rexp  // alternative
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     19 | case class SEQ(r1: Rexp, r2: Rexp) extends Rexp  // sequence
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     20 | case class STAR(r: Rexp) extends Rexp            // star
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     21 | 
 | 
| 919 |     22 | 
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     23 | // nullable function: tests whether a regular 
 | 
| 919 |     24 | // expression can recognise the empty string  
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     25 | def nullable(r: Rexp) : Boolean = r match {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     26 |   case ZERO => false
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     27 |   case ONE => true
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     28 |   case CHAR(_) => false
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     29 |   case ALT(r1, r2) => nullable(r1) || nullable(r2)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     30 |   case SEQ(r1, r2) => nullable(r1) && nullable(r2)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     31 |   case STAR(_) => true
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     32 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     33 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     34 | // the derivative of a regular expression w.r.t. a character
 | 
| 825 |     35 | def der(c: Char, r: Rexp) : Rexp = r match {
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     36 |   case ZERO => ZERO
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     37 |   case ONE => ZERO
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     38 |   case CHAR(d) => if (c == d) ONE else ZERO
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     39 |   case ALT(r1, r2) => ALT(der(c, r1), der(c, r2))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     40 |   case SEQ(r1, r2) => 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     41 |     if (nullable(r1)) ALT(SEQ(der(c, r1), r2), der(c, r2))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     42 |     else SEQ(der(c, r1), r2)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     43 |   case STAR(r1) => SEQ(der(c, r1), STAR(r1))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     44 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     45 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     46 | // the derivative w.r.t. a string (iterates der)
 | 
| 825 |     47 | def ders(s: List[Char], r: Rexp) : Rexp = s match {
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     48 |   case Nil => r
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     49 |   case c::s => ders(s, der(c, r))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     50 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     51 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     52 | // the main matcher function
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     53 | def matcher(r: Rexp, s: String) : Boolean = 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     54 |   nullable(ders(s.toList, r))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     55 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     56 | 
 | 
| 919 |     57 | // some examples from the homework
 | 
| 871 |     58 | 
 | 
| 940 |     59 | val r = SEQ(CHAR('a'), CHAR('c'))
 | 
| 913 |     60 | matcher(r, "ac")
 | 
| 871 |     61 | 
 | 
| 919 |     62 | val r1 = STAR(ALT(SEQ(CHAR('a'), CHAR('b')), CHAR('b')))
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     63 | der('a', r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     64 | der('b', r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     65 | der('c', r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     66 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     67 | val r2 = SEQ(SEQ(CHAR('x'), CHAR('y')), CHAR('z'))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     68 | der('x', r2)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     69 | der('y', der('x', r2))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     70 | der('z', der('y', der('x', r2)))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     71 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     72 | 
 | 
| 919 |     73 | // Test Cases
 | 
|  |     74 | //============
 | 
|  |     75 | 
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     76 | // the optional regular expression (one or zero times)
 | 
| 765 |     77 | def OPT(r: Rexp) = ALT(r, ONE)   // r + 1
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     78 | 
 | 
| 928 |     79 | // the n-times regular expression (explicitly expanded to SEQs)
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     80 | def NTIMES(r: Rexp, n: Int) : Rexp = n match {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     81 |   case 0 => ONE
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     82 |   case 1 => r
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     83 |   case n => SEQ(r, NTIMES(r, n - 1))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     84 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     85 | 
 | 
| 765 |     86 | // the evil regular expression  (a?){n} a{n}
 | 
|  |     87 | def EVIL1(n: Int) = 
 | 
|  |     88 |   SEQ(NTIMES(OPT(CHAR('a')), n), NTIMES(CHAR('a'), n))
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     89 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     90 | // the evil regular expression (a*)* b
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     91 | val EVIL2 = SEQ(STAR(STAR(CHAR('a'))), CHAR('b'))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     92 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     93 | // for measuring time
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     94 | def time_needed[T](i: Int, code: => T) = {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     95 |   val start = System.nanoTime()
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     96 |   for (j <- 1 to i) code
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     97 |   val end = System.nanoTime()
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     98 |   (end - start) / (i * 1.0e9)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     99 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    100 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    101 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    102 | // test: (a?{n}) (a{n})
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    103 | def test1() = {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    104 |   println("Test (a?{n}) (a{n})")
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    105 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    106 |   for (i <- 0 to 20 by 2) {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    107 |     println(f"$i: ${time_needed(2, matcher(EVIL1(i), "a" * i))}%.5f")
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    108 |   }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    109 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    110 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    111 | // test: (a*)* b
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    112 | def test2() = {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    113 |   println("Test (a*)* b")
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    114 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    115 |   for (i <- 0 to 20 by 2) {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    116 |     println(f"$i: ${time_needed(2, matcher(EVIL2, "a" * i))}%.5f")
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    117 |   }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    118 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    119 | 
 | 
| 927 |    120 | 
 | 
|  |    121 | 
 | 
|  |    122 | 
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    123 | // the size of a regular expressions - for testing purposes 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    124 | def size(r: Rexp) : Int = r match {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    125 |   case ZERO => 1
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    126 |   case ONE => 1
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    127 |   case CHAR(_) => 1
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    128 |   case ALT(r1, r2) => 1 + size(r1) + size(r2)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    129 |   case SEQ(r1, r2) => 1 + size(r1) + size(r2)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    130 |   case STAR(r) => 1 + size(r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    131 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    132 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    133 | // the expicit expansion in EVIL1(n) increases
 | 
| 769 |    134 | // drastically its size - (a?){n} a{n}
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    135 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    136 | size(EVIL1(1))  // 5
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    137 | size(EVIL1(3))  // 17
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    138 | size(EVIL1(5))  // 29
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    139 | size(EVIL1(7))  // 41
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    140 | size(EVIL1(20)) // 119
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    141 | 
 | 
| 769 |    142 | size(ders(("a" * 20).toList, EVIL1(20))) 
 | 
|  |    143 | 
 | 
|  |    144 | 
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    145 | // given a regular expression and building successive
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    146 | // derivatives might result into bigger and bigger
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    147 | // regular expressions...here is an example for this:
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    148 | 
 | 
| 927 |    149 | 
 | 
|  |    150 | // (a + aa)*
 | 
|  |    151 | val BIG = STAR(ALT(CHAR('a'), SEQ(CHAR('a'), CHAR('a'))))
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    152 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    153 | size(ders("".toList, BIG))              // 13
 | 
| 927 |    154 | size(ders("aa".toList, BIG))            // 51
 | 
|  |    155 | size(ders("aaaa".toList, BIG))          // 112
 | 
|  |    156 | size(ders("aaaaaa".toList, BIG))        // 191
 | 
|  |    157 | size(ders("aaaaaaaa".toList, BIG))      // 288
 | 
|  |    158 | size(ders("aaaaaaaaaa".toList, BIG))    // 403
 | 
|  |    159 | size(ders("aaaaaaaaaaaa".toList, BIG))  // 536
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    160 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    161 | 
 | 
| 927 |    162 | size(ders(("a" * 30).toList, BIG))      // 31010539
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    163 | 
 | 
| 960 |    164 | 
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    165 | def test3() = {
 | 
| 927 |    166 |   println("Test (a + aa)*")
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    167 | 
 | 
| 927 |    168 |   for (i <- 0 to 30 by 5) {
 | 
|  |    169 |     println(f"$i: ${time_needed(2, matcher(BIG, "a" * i))}%.5f")
 | 
| 725 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    170 |   }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    171 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    172 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    173 | @main
 | 
| 825 |    174 | def all() = { test1(); test2() ; test3() } 
 | 
| 960 |    175 | 
 | 
|  |    176 | //all()
 |