1 // Core Part about Regular Expression Matching |
|
2 //============================================= |
|
3 |
|
4 object CW9c { |
|
5 |
|
6 // Regular Expressions |
|
7 abstract class Rexp |
|
8 case object ZERO extends Rexp |
|
9 case object ONE extends Rexp |
|
10 case class CHAR(c: Char) extends Rexp |
|
11 case class ALT(r1: Rexp, r2: Rexp) extends Rexp // alternative |
|
12 case class SEQ(r1: Rexp, r2: Rexp) extends Rexp // sequence |
|
13 case class STAR(r: Rexp) extends Rexp // star |
|
14 |
|
15 |
|
16 // some convenience for typing regular expressions |
|
17 |
|
18 import scala.language.implicitConversions |
|
19 import scala.language.reflectiveCalls |
|
20 |
|
21 def charlist2rexp(s: List[Char]): Rexp = s match { |
|
22 case Nil => ONE |
|
23 case c::Nil => CHAR(c) |
|
24 case c::s => SEQ(CHAR(c), charlist2rexp(s)) |
|
25 } |
|
26 implicit def string2rexp(s: String): Rexp = charlist2rexp(s.toList) |
|
27 |
|
28 implicit def RexpOps (r: Rexp) = new { |
|
29 def | (s: Rexp) = ALT(r, s) |
|
30 def % = STAR(r) |
|
31 def ~ (s: Rexp) = SEQ(r, s) |
|
32 } |
|
33 |
|
34 implicit def stringOps (s: String) = new { |
|
35 def | (r: Rexp) = ALT(s, r) |
|
36 def | (r: String) = ALT(s, r) |
|
37 def % = STAR(s) |
|
38 def ~ (r: Rexp) = SEQ(s, r) |
|
39 def ~ (r: String) = SEQ(s, r) |
|
40 } |
|
41 |
|
42 // (5) Complete the function nullable according to |
|
43 // the definition given in the coursework; this |
|
44 // function checks whether a regular expression |
|
45 // can match the empty string and Returns a boolean |
|
46 // accordingly. |
|
47 |
|
48 def nullable (r: Rexp) : Boolean = { |
|
49 r match { |
|
50 case ZERO => false |
|
51 case ONE => true |
|
52 case CHAR(c) => false |
|
53 case ALT(r1, r2) => (nullable(r1) || nullable(r2)) |
|
54 case SEQ(r1, r2) => (nullable(r1) && nullable(r2)) |
|
55 case STAR(r) => true |
|
56 } |
|
57 } |
|
58 |
|
59 // (6) Complete the function der according to |
|
60 // the definition given in the coursework; this |
|
61 // function calculates the derivative of a |
|
62 // regular expression w.r.t. a character. |
|
63 |
|
64 def der (c: Char, r: Rexp) : Rexp = { |
|
65 r match { |
|
66 case ZERO => ZERO |
|
67 case ONE => ZERO |
|
68 case CHAR(d) => if(d == c) ONE else ZERO |
|
69 case ALT(r1, r2) => ALT(der(c, r1), der(c, r2)) |
|
70 case SEQ(r1, r2) => if(nullable(r1)) { |
|
71 (ALT(SEQ(der(c, r1), r2), der(c, r2))) |
|
72 } else { |
|
73 SEQ(der(c, r1), r2) |
|
74 } |
|
75 case STAR(r) => SEQ(der(c, r), STAR(r)) |
|
76 } |
|
77 } |
|
78 |
|
79 |
|
80 // (7) Complete the simp function according to |
|
81 // the specification given in the coursework; this |
|
82 // function simplifies a regular expression from |
|
83 // the inside out, like you would simplify arithmetic |
|
84 // expressions; however it does not simplify inside |
|
85 // STAR-regular expressions. |
|
86 |
|
87 def simp(r: Rexp) : Rexp = { |
|
88 r match { |
|
89 case STAR(r) => STAR(r) // does not process r star |
|
90 case SEQ(r1, r2) => { |
|
91 val x = (simp(r1), simp(r2)) |
|
92 if(x._1 == ZERO) ZERO else |
|
93 if(x._2 == ZERO) ZERO else |
|
94 if(x._1 == ONE) simp(x._2) else |
|
95 if(x._2 == ONE) simp(x._1) else |
|
96 if(x._1 == x._2) simp(x._2) else |
|
97 SEQ(simp(x._1), simp(x._2)) |
|
98 } |
|
99 case ALT(r1, r2) => { |
|
100 val x = (simp(r1), simp(r2)) |
|
101 if(x._1 == ZERO) simp(x._2) else |
|
102 if(x._2 == ZERO) simp(x._1) else |
|
103 if(x._1 == x._2) simp(x._2) else |
|
104 ALT(simp(x._1), simp(x._2)) |
|
105 } |
|
106 case r => r // if single regex, return it |
|
107 } |
|
108 } |
|
109 |
|
110 |
|
111 // (8) Complete the two functions below; the first |
|
112 // calculates the derivative w.r.t. a string; the second |
|
113 // is the regular expression matcher taking a regular |
|
114 // expression and a string and checks whether the |
|
115 // string matches the regular expression |
|
116 |
|
117 def ders (s: List[Char], r: Rexp) : Rexp = { |
|
118 s match { |
|
119 case Nil => r |
|
120 case c :: cs => ders(cs, simp(der(c,r))) |
|
121 } |
|
122 } |
|
123 |
|
124 def matcher(r: Rexp, s: String): Boolean = { |
|
125 val listOfCharacters = s.toList |
|
126 val result = ders(listOfCharacters, r) |
|
127 nullable(result) |
|
128 } |
|
129 |
|
130 |
|
131 // (9) Complete the size function for regular |
|
132 // expressions according to the specification |
|
133 // given in the coursework. |
|
134 |
|
135 def size(r: Rexp): Int = { |
|
136 r match { |
|
137 case ZERO => 1 |
|
138 case ONE => 1 |
|
139 case CHAR(c) => 1 |
|
140 case ALT(r1, r2) => 1 + size(r1) + size(r2) |
|
141 case SEQ(r1, r2) => 1 + size(r1) + size(r2) |
|
142 case STAR(r) => 1 + size(r) |
|
143 } |
|
144 } |
|
145 |
|
146 // some testing data |
|
147 |
|
148 /* |
|
149 matcher(("a" ~ "b") ~ "c", "abc") // => true |
|
150 matcher(("a" ~ "b") ~ "c", "ab") // => false |
|
151 |
|
152 // the supposedly 'evil' regular expression (a*)* b |
|
153 // val EVIL = SEQ(STAR(STAR(CHAR('a'))), CHAR('b')) |
|
154 |
|
155 matcher(EVIL, "a" * 1000 ++ "b") // => true |
|
156 matcher(EVIL, "a" * 1000) // => false |
|
157 |
|
158 // size without simplifications |
|
159 size(der('a', der('a', EVIL))) // => 28 |
|
160 size(der('a', der('a', der('a', EVIL)))) // => 58 |
|
161 |
|
162 // size with simplification |
|
163 size(simp(der('a', der('a', EVIL)))) // => 8 |
|
164 size(simp(der('a', der('a', der('a', EVIL))))) // => 8 |
|
165 |
|
166 // Python needs around 30 seconds for matching 28 a's with EVIL. |
|
167 // Java 9 and later increase this to an "astonishing" 40000 a's in |
|
168 // 30 seconds. |
|
169 // |
|
170 // Lets see how long it really takes to match strings with |
|
171 // 5 Million a's...it should be in the range of a couple |
|
172 // of seconds. |
|
173 |
|
174 def time_needed[T](i: Int, code: => T) = { |
|
175 val start = System.nanoTime() |
|
176 for (j <- 1 to i) code |
|
177 val end = System.nanoTime() |
|
178 (end - start)/(i * 1.0e9) |
|
179 } |
|
180 |
|
181 for (i <- 0 to 5000000 by 500000) { |
|
182 println(i + " " + "%.5f".format(time_needed(2, matcher(EVIL, "a" * i)))) |
|
183 } |
|
184 |
|
185 // another "power" test case |
|
186 simp(Iterator.iterate(ONE:Rexp)(r => SEQ(r, ONE | ONE)).drop(50).next) == ONE |
|
187 |
|
188 // the Iterator produces the rexp |
|
189 // |
|
190 // SEQ(SEQ(SEQ(..., ONE | ONE) , ONE | ONE), ONE | ONE) |
|
191 // |
|
192 // where SEQ is nested 50 times. |
|
193 |
|
194 */ |
|
195 |
|
196 } |
|