| 412 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      1 | // Main Part 3 about Regular Expression Matching
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      2 | //=============================================
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      3 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      4 | object M3 {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      5 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      6 | // Regular Expressions
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      7 | abstract class Rexp
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      8 | case object ZERO extends Rexp
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |      9 | case object ONE extends Rexp
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     10 | case class CHAR(c: Char) extends Rexp
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     11 | case class ALTs(rs: List[Rexp]) extends Rexp      // alternatives 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     12 | case class SEQ(r1: Rexp, r2: Rexp) extends Rexp   // sequence
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     13 | case class STAR(r: Rexp) extends Rexp             // star
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     14 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     15 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     16 | // some convenience for typing regular expressions
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     17 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     18 | //the usual binary choice can be defined in terms of ALTs
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     19 | def ALT(r1: Rexp, r2: Rexp) = ALTs(List(r1, r2))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     20 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     21 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     22 | import scala.language.implicitConversions    
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     23 | import scala.language.reflectiveCalls 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     24 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     25 | def charlist2rexp(s: List[Char]): Rexp = s match {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     26 |   case Nil => ONE
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     27 |   case c::Nil => CHAR(c)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     28 |   case c::s => SEQ(CHAR(c), charlist2rexp(s))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     29 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     30 | implicit def string2rexp(s: String): Rexp = charlist2rexp(s.toList)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     31 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     32 | implicit def RexpOps (r: Rexp) = new {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     33 |   def | (s: Rexp) = ALT(r, s)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     34 |   def % = STAR(r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     35 |   def ~ (s: Rexp) = SEQ(r, s)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     36 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     37 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     38 | implicit def stringOps (s: String) = new {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     39 |   def | (r: Rexp) = ALT(s, r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     40 |   def | (r: String) = ALT(s, r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     41 |   def % = STAR(s)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     42 |   def ~ (r: Rexp) = SEQ(s, r)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     43 |   def ~ (r: String) = SEQ(s, r)
 | 
| 
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 | // (1) Complete the function nullable according to
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     47 | // the definition given in the coursework; this 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     48 | // function checks whether a regular expression
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     49 | // can match the empty string and Returns a boolean
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     50 | // accordingly.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     51 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     52 | def nullable (r: Rexp) : Boolean = ???
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     53 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     54 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     55 | // (2) Complete the function der according to
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     56 | // the definition given in the coursework; this
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     57 | // function calculates the derivative of a 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     58 | // regular expression w.r.t. a character.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     59 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     60 | def der (c: Char, r: Rexp) : Rexp = ???
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     61 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     62 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     63 | // (3) Implement the flatten function flts. It
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     64 | // deletes 0s from a list of regular expressions
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     65 | // and also 'spills out', or flattens, nested 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     66 | // ALTernativeS.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     67 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     68 | def flts(rs: List[Rexp]) : List[Rexp] = ???
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     69 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     70 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     71 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     72 | // (4) Complete the simp function according to
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     73 | // the specification given in the coursework description; 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     74 | // this function simplifies a regular expression from
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     75 | // the inside out, like you would simplify arithmetic 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     76 | // expressions; however it does not simplify inside 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     77 | // STAR-regular expressions. Use the _.distinct and 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     78 | // flts functions.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     79 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     80 | def simp(r: Rexp) : Rexp = ???
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     81 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     82 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     83 | // (5) Complete the two functions below; the first 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     84 | // calculates the derivative w.r.t. a string; the second
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     85 | // is the regular expression matcher taking a regular
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     86 | // expression and a string and checks whether the
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     87 | // string matches the regular expression
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     88 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     89 | def ders (s: List[Char], r: Rexp) : Rexp = ???
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     90 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     91 | def matcher(r: Rexp, s: String): Boolean = ???
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     92 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     93 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     94 | // (6) Complete the size function for regular
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     95 | // expressions according to the specification 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     96 | // given in the coursework.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     97 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |     98 | def size(r: Rexp): Int = ???
 | 
| 
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 | // some testing data
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    102 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    103 | /*
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    104 | matcher(("a" ~ "b") ~ "c", "abc")  // => true
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    105 | matcher(("a" ~ "b") ~ "c", "ab")   // => false
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    106 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    107 | // the supposedly 'evil' regular expression (a*)* b
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    108 | val EVIL = SEQ(STAR(STAR(CHAR('a'))), CHAR('b'))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    109 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    110 | matcher(EVIL, "a" * 1000 ++ "b")   // => true
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    111 | matcher(EVIL, "a" * 1000)          // => false
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    112 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    113 | // size without simplifications
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    114 | size(der('a', der('a', EVIL)))             // => 28
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    115 | size(der('a', der('a', der('a', EVIL))))   // => 58
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    116 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    117 | // size with simplification
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    118 | size(simp(der('a', der('a', EVIL))))           // => 8
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    119 | size(simp(der('a', der('a', der('a', EVIL))))) // => 8
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    120 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    121 | // Python needs around 30 seconds for matching 28 a's with EVIL. 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    122 | // Java 9 and later increase this to an "astonishing" 40000 a's in
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    123 | // 30 seconds.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    124 | //
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    125 | // Lets see how long it really takes to match strings with 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    126 | // 5 Million a's...it should be in the range of a couple
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    127 | // of seconds.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    128 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    129 | def time_needed[T](i: Int, code: => T) = {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    130 |   val start = System.nanoTime()
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    131 |   for (j <- 1 to i) code
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    132 |   val end = System.nanoTime()
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    133 |   "%.5f".format((end - start)/(i * 1.0e9))
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    134 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    135 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    136 | for (i <- 0 to 5000000 by 500000) {
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    137 |   println(s"$i ${time_needed(2, matcher(EVIL, "a" * i))} secs.") 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    138 | }
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    139 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    140 | // another "power" test case 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    141 | simp(Iterator.iterate(ONE:Rexp)(r => SEQ(r, ONE | ONE)).drop(50).next()) == ONE
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    142 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    143 | // the Iterator produces the rexp
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    144 | //
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    145 | //      SEQ(SEQ(SEQ(..., ONE | ONE) , ONE | ONE), ONE | ONE)
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    146 | //
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    147 | //    where SEQ is nested 50 times.
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    148 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    149 | */
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    150 | 
 | 
| 
Christian Urban <christian.urban@kcl.ac.uk> parents: diff
changeset |    151 | }
 |