author | Christian Urban <christian dot urban at kcl dot ac dot uk> |
Fri, 27 Sep 2013 15:43:25 +0100 | |
changeset 116 | 010ae7288327 |
parent 93 | 4794759139ea |
child 117 | 25999de692b2 |
permissions | -rw-r--r-- |
49 | 1 |
|
54 | 2 |
abstract class Rexp |
49 | 3 |
|
4 |
case object NULL extends Rexp |
|
5 |
case object EMPTY extends Rexp |
|
6 |
case class CHAR(c: Char) extends Rexp |
|
7 |
case class ALT(r1: Rexp, r2: Rexp) extends Rexp |
|
8 |
case class SEQ(r1: Rexp, r2: Rexp) extends Rexp |
|
9 |
case class STAR(r: Rexp) extends Rexp |
|
10 |
||
11 |
// some convenience for typing in regular expressions |
|
12 |
def charlist2rexp(s : List[Char]) : Rexp = s match { |
|
13 |
case Nil => EMPTY |
|
14 |
case c::Nil => CHAR(c) |
|
15 |
case c::s => SEQ(CHAR(c), charlist2rexp(s)) |
|
16 |
} |
|
17 |
implicit def string2rexp(s : String) : Rexp = charlist2rexp(s.toList) |
|
18 |
||
19 |
||
20 |
// nullable function: tests whether the regular |
|
21 |
// expression can recognise the empty string |
|
22 |
def nullable (r: Rexp) : Boolean = r match { |
|
23 |
case NULL => false |
|
24 |
case EMPTY => true |
|
25 |
case CHAR(_) => false |
|
26 |
case ALT(r1, r2) => nullable(r1) || nullable(r2) |
|
27 |
case SEQ(r1, r2) => nullable(r1) && nullable(r2) |
|
28 |
case STAR(_) => true |
|
29 |
} |
|
30 |
||
31 |
// derivative of a regular expression w.r.t. a character |
|
32 |
def der (c: Char, r: Rexp) : Rexp = r match { |
|
33 |
case NULL => NULL |
|
34 |
case EMPTY => NULL |
|
35 |
case CHAR(d) => if (c == d) EMPTY else NULL |
|
36 |
case ALT(r1, r2) => ALT(der(c, r1), der(c, r2)) |
|
37 |
case SEQ(r1, r2) => |
|
38 |
if (nullable(r1)) ALT(SEQ(der(c, r1), r2), der(c, r2)) |
|
39 |
else SEQ(der(c, r1), r2) |
|
40 |
case STAR(r) => SEQ(der(c, r), STAR(r)) |
|
41 |
} |
|
42 |
||
43 |
// derivative w.r.t. a string (iterates der) |
|
44 |
def ders (s: List[Char], r: Rexp) : Rexp = s match { |
|
45 |
case Nil => r |
|
54 | 46 |
case c::s => ders(s, der(c, r)) |
49 | 47 |
} |
48 |
||
49 |
// main matcher function |
|
50 |
def matcher(r: Rexp, s: String) : Boolean = nullable(ders(s.toList, r)) |
|
51 |
||
85
1a4065f965fb
added packages
Christian Urban <christian dot urban at kcl dot ac dot uk>
parents:
59
diff
changeset
|
52 |
//example |
1a4065f965fb
added packages
Christian Urban <christian dot urban at kcl dot ac dot uk>
parents:
59
diff
changeset
|
53 |
//val r = STAR(ALT(SEQ(CHAR('a'), CHAR('b')), CHAR('b'))) |
1a4065f965fb
added packages
Christian Urban <christian dot urban at kcl dot ac dot uk>
parents:
59
diff
changeset
|
54 |
//der('b', r) |
1a4065f965fb
added packages
Christian Urban <christian dot urban at kcl dot ac dot uk>
parents:
59
diff
changeset
|
55 |
//der('b', r) |
49 | 56 |
|
57 |
//one or zero |
|
58 |
def OPT(r: Rexp) = ALT(r, EMPTY) |
|
59 |
||
60 |
//n-times |
|
61 |
def NTIMES(r: Rexp, n: Int) : Rexp = n match { |
|
59 | 62 |
case 0 => EMPTY |
49 | 63 |
case 1 => r |
64 |
case n => SEQ(r, NTIMES(r, n - 1)) |
|
65 |
} |
|
66 |
||
67 |
def RTEST(n: Int) = SEQ(NTIMES(OPT("a"), n), NTIMES("a", n)) |
|
68 |
||
69 |
def time_needed[T](i: Int, code: => T) = { |
|
70 |
val start = System.nanoTime() |
|
71 |
for (j <- 1 to i) code |
|
72 |
val end = System.nanoTime() |
|
73 |
(end - start)/(i * 1.0e9) |
|
74 |
} |
|
75 |
||
54 | 76 |
for (i <- 1 to 29) { |
49 | 77 |
println(i + ": " + "%.5f".format(time_needed(1, matcher(RTEST(i), "a" * i)))) |
78 |
} |
|
79 |
||
80 |