// Simple matcher for basic regular expressions

abstract class Rexp

case object ZERO extends Rexp                    // matches nothing

case object ONE extends Rexp                     // matches the empty string

case class CHAR(c: Char) extends Rexp            // matches a character c

case class ALT(r1: Rexp, r2: Rexp) extends Rexp  // alternative

case class SEQ(r1: Rexp, r2: Rexp) extends Rexp  // sequence

case class STAR(r: Rexp) extends Rexp            // star

// nullable function: tests whether a regular 

// expression can recognise the empty string

def nullable (r: Rexp) : Boolean = r match {

  case ZERO => false

  case ONE => true

  case CHAR(_) => false

  case ALT(r1, r2) => nullable(r1) || nullable(r2)

  case SEQ(r1, r2) => nullable(r1) && nullable(r2)

  case STAR(_) => true

}

// derivative of a regular expression w.r.t. a character

def der (c: Char, r: Rexp) : Rexp = r match {

  case ZERO => ZERO

  case ONE => ZERO

  case CHAR(d) => if (c == d) ONE else ZERO

  case ALT(r1, r2) => ALT(der(c, r1), der(c, r2))

  case SEQ(r1, r2) => 

    if (nullable(r1)) ALT(SEQ(der(c, r1), r2), der(c, r2))

    else SEQ(der(c, r1), r2)

  case STAR(r1) => SEQ(der(c, r1), STAR(r1))

}

// derivative w.r.t. a string (iterates der)

def ders (s: List[Char], r: Rexp) : Rexp = s match {

  case Nil => r

  case c::s => ders(s, der(c, r))

}

// main matcher function

def matches(r: Rexp, s: String) : Boolean = nullable(ders(s.toList, r))

//examples from the homework

val r = STAR(ALT(SEQ(CHAR('a'), CHAR('b')), CHAR('b')))

der('a', r)

der('b', r)

der('c', r)

val r2 = SEQ(SEQ(CHAR('x'), CHAR('y')), CHAR('z'))

der('x', r2)

der('y', der('x', r2))

der('z', der('y', der('x', r2)))

//optional regular expression (one or zero times)

def OPT(r: Rexp) = ALT(r, ONE)

//n-times regular expression (explicitly expanded)

def NTIMES(r: Rexp, n: Int) : Rexp = n match {

  case 0 => ONE

  case 1 => r

  case n => SEQ(r, NTIMES(r, n - 1))

}

// Test Cases

// the evil regular expression  a?{n} a{n}

def EVIL1(n: Int) = SEQ(NTIMES(OPT(CHAR('a')), n), NTIMES(CHAR('a'), n))

// the evil regular expression (a*)*b

val EVIL2 = SEQ(STAR(STAR(CHAR('a'))), CHAR('b'))

//for measuring time

def time_needed[T](i: Int, code: => T) = {

  val start = System.nanoTime()

  for (j <- 1 to i) code

  val end = System.nanoTime()

  (end - start)/(i * 1.0e9)

}

//test: (a?{n}) (a{n})

println("Test (a?{n}) (a{n})")

for (i <- 1 to 20) {

  println(i + ": " + "%.5f".format(time_needed(2, matches(EVIL1(i), "a" * i))))

}

for (i <- 1 to 20) {

  println(i + ": " + "%.5f".format(time_needed(2, matches(EVIL1(i), "a" * i))))

}

//test: (a*)* b

println("Test (a*)* b")

for (i <- 1 to 20) {

  println(i + " " + "%.5f".format(time_needed(2, matches(EVIL2, "a" * i))))

}

for (i <- 1 to 20) {

  println(i + " " + "%.5f".format(time_needed(2, matches(EVIL2, "a" * i))))

}

// size of a regular expressions - for testing purposes 

def size(r: Rexp) : Int = r match {

  case ZERO => 1

  case ONE => 1

  case CHAR(_) => 1

  case ALT(r1, r2) => 1 + size(r1) + size(r2)

  case SEQ(r1, r2) => 1 + size(r1) + size(r2)

  case STAR(r) => 1 + size(r)

}

// the expicit expansion in EVIL1(n) increases

// drastically its size

size(EVIL1(1))  // 5

size(EVIL1(3))  // 17

size(EVIL1(5))  // 29

size(EVIL1(7))  // 41

// given a regular expression and building successive

// derivatives might result into bigger and bigger

// regular expressions...here is an example for this:

val BIG_aux = STAR(ALT(CHAR('a'), CHAR('b')))

val BIG = SEQ(BIG_aux, SEQ(CHAR('a'),SEQ(CHAR('b'), BIG_aux)))

size(ders("".toList, BIG))              // 13

size(ders("ab".toList, BIG))            // 51

size(ders("abab".toList, BIG))          // 112

size(ders("ababab".toList, BIG))        // 191

size(ders("abababab".toList, BIG))      // 288

size(ders("ababababab".toList, BIG))    // 403

size(ders("abababababab".toList, BIG))  // 536

size(ders(("ab" * 200).toList, BIG))    // 366808

for (i <- 1 to 21) {

  println(i + " " + "%.5f".format(time_needed(2, matches(BIG, "ab" * i))))

}