progs/matcher/re2.sc
author Christian Urban <christian.urban@kcl.ac.uk>
Fri, 03 Oct 2025 17:07:01 +0100
changeset 999 e719e420cbc7
parent 929 9541e073f2ed
permissions -rw-r--r--
updated
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
929
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
     1
// A Version of the simple matcher with an explicit 
919
53f08d873e09 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 879
diff changeset
     2
// n-times regular expression
53f08d873e09 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 879
diff changeset
     3
//
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
     4
// this keeps the size of the regular expression in the
919
53f08d873e09 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 879
diff changeset
     5
// EVIL1 testcase small
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
     6
//
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
     7
// call the test cases with X = {1,2}
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
     8
//
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
     9
//   amm re2.sc testX
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    10
//
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    11
// or 
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    12
//
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    13
//   amm re2.sc all
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    14
999
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    15
// regular expressions (as enum in Scala 3)
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    16
enum Rexp {
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    17
  case ZERO                     // matches nothing
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    18
  case ONE                      // matches an empty string
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    19
  case CHAR(c: Char)            // matches a character c
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    20
  case ALT(r1: Rexp, r2: Rexp)  // alternative
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    21
  case SEQ(r1: Rexp, r2: Rexp)  // sequence
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    22
  case STAR(r: Rexp)            // star
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    23
  case NTIMES(r: Rexp, n: Int)  // explicit n-times
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    24
}
e719e420cbc7 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 929
diff changeset
    25
import Rexp._
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    26
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    27
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    28
def nullable (r: Rexp) : Boolean = r match {
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    29
  case ZERO => false
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    30
  case ONE => true
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    31
  case CHAR(_) => false
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    32
  case ALT(r1, r2) => nullable(r1) || nullable(r2)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    33
  case SEQ(r1, r2) => nullable(r1) && nullable(r2)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    34
  case STAR(_) => true
769
f9686b22db7e updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 767
diff changeset
    35
  case NTIMES(r, n) => if (n == 0) true else nullable(r)
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    36
}
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    37
825
dca072e2bb7d updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 769
diff changeset
    38
def der(c: Char, r: Rexp) : Rexp = r match {
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    39
  case ZERO => ZERO
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    40
  case ONE => ZERO
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    41
  case CHAR(d) => if (c == d) ONE else ZERO
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    42
  case ALT(r1, r2) => ALT(der(c, r1), der(c, r2))
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    43
  case SEQ(r1, r2) => 
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    44
    if (nullable(r1)) ALT(SEQ(der(c, r1), r2), der(c, r2))
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    45
    else SEQ(der(c, r1), r2)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    46
  case STAR(r1) => SEQ(der(c, r1), STAR(r1))
769
f9686b22db7e updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 767
diff changeset
    47
  case NTIMES(r, n) => 
f9686b22db7e updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 767
diff changeset
    48
    if (n == 0) ZERO else SEQ(der(c, r), NTIMES(r, n - 1))
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    49
}
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    50
825
dca072e2bb7d updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 769
diff changeset
    51
def ders(s: List[Char], r: Rexp) : Rexp = s match {
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    52
  case Nil => r
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    53
  case c::s => ders(s, der(c, r))
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    54
}
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    55
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    56
def matcher(r: Rexp, s: String) : Boolean = 
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    57
  nullable(ders(s.toList, r))
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    58
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    59
919
53f08d873e09 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 879
diff changeset
    60
// Test Cases
53f08d873e09 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 879
diff changeset
    61
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    62
// the optional regular expression: one or zero times
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    63
// this regular expression is still defined in terms of ALT
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    64
def OPT(r: Rexp) = ALT(r, ONE)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    65
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    66
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    67
// evil regular expressions
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    68
def EVIL1(n: Int) = SEQ(NTIMES(OPT(CHAR('a')), n), NTIMES(CHAR('a'), n))
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    69
val EVIL2 = SEQ(STAR(STAR(CHAR('a'))), CHAR('b'))
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    70
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    71
def time_needed[T](i: Int, code: => T) = {
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    72
  val start = System.nanoTime()
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    73
  for (j <- 1 to i) code
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    74
  val end = System.nanoTime()
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    75
  (end - start) / (i * 1.0e9)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    76
}
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    77
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    78
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    79
826
b0352633bf48 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 825
diff changeset
    80
@arg(doc = "Test (a?{n}) (a{n})")
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    81
@main
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    82
def test1() = {
742
b5b5583a3a08 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 725
diff changeset
    83
  println("Test (a?{n}) (a{n})")
b5b5583a3a08 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 725
diff changeset
    84
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    85
  for (i <- 0 to 1000 by 100) {
769
f9686b22db7e updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 767
diff changeset
    86
    println(f"$i: ${time_needed(1, matcher(EVIL1(i), "a" * i))}%.5f")
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    87
  }
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    88
}  
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    89
742
b5b5583a3a08 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 725
diff changeset
    90
826
b0352633bf48 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 825
diff changeset
    91
@arg(doc = "Test (a*)* b")
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    92
@main
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    93
def test2() = {
742
b5b5583a3a08 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 725
diff changeset
    94
  println("Test (a*)* b")
b5b5583a3a08 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 725
diff changeset
    95
929
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
    96
  for (i <- 0 to 20 by 2) {
769
f9686b22db7e updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 767
diff changeset
    97
    println(f"$i: ${time_needed(1, matcher(EVIL2, "a" * i))}%.5f")
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    98
  }
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
    99
}
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   100
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   101
// the size of a regular expressions - for testing purposes 
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   102
def size(r: Rexp) : Int = r match {
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   103
  case ZERO => 1
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   104
  case ONE => 1
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   105
  case CHAR(_) => 1
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   106
  case ALT(r1, r2) => 1 + size(r1) + size(r2)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   107
  case SEQ(r1, r2) => 1 + size(r1) + size(r2)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   108
  case STAR(r) => 1 + size(r)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   109
  case NTIMES(r, _) => 1 + size(r)
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   110
}
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   111
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   112
// EVIL1(n) has now a constant size, no matter
929
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
   113
// what n is
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   114
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   115
size(EVIL1(1))  // 7
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   116
size(EVIL1(3))  // 7
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   117
size(EVIL1(5))  // 7
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   118
size(EVIL1(7))  // 7
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   119
size(EVIL1(20)) // 7
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   120
929
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
   121
// also the derivative only grows much more 
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
   122
// moderately and actually maxes out at size 211
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
   123
// (for n = 5)
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   124
size(ders("".toList, EVIL1(5)))       // 7
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   125
size(ders("a".toList, EVIL1(5)))      // 16
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   126
size(ders("aa".toList, EVIL1(5)))     // 35
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   127
size(ders("aaa".toList, EVIL1(5)))    // 59
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   128
size(ders("aaaa".toList, EVIL1(5)))   // 88
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   129
size(ders("aaaaa".toList, EVIL1(5)))  // 122
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   130
size(ders("aaaaaa".toList, EVIL1(5))) // 151
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   131
929
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
   132
size(ders(("a" * 20).toList, EVIL1(5)))   // 211
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
   133
size(ders(("a" * 200).toList, EVIL1(5)))  // 211
9541e073f2ed updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 919
diff changeset
   134
size(ders(("a" * 2000).toList, EVIL1(5))) // 211
769
f9686b22db7e updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 767
diff changeset
   135
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   136
// but the size of the derivatives can still grow 
769
f9686b22db7e updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 767
diff changeset
   137
// quite dramatically in case of EVIL2:  (a*)* b
725
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   138
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   139
size(ders("".toList, EVIL2))       // 5
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   140
size(ders("a".toList, EVIL2))      // 12
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   141
size(ders("aa".toList, EVIL2))     // 28
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   142
size(ders("aaa".toList, EVIL2))    // 58
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   143
size(ders("aaaa".toList, EVIL2))   // 116
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   144
size(ders("aaaaa".toList, EVIL2))  // 230
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   145
size(ders("aaaaaa".toList, EVIL2)) // 456
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   146
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   147
size(ders(("a" * 20).toList, EVIL2)) // 7340068
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   148
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   149
f345e89895f5 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents:
diff changeset
   150
@main
825
dca072e2bb7d updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 769
diff changeset
   151
def all() = { test1(); test2() } 
826
b0352633bf48 updated
Christian Urban <christian.urban@kcl.ac.uk>
parents: 825
diff changeset
   152