// A version of the simple matcher with simplification 

// of derivatives

//

// this keeps the regular expressions (often) small, which

// is good for the run-time

// 

// call the test cases with X = {1,2,3,4}

//

//   amm re3.sc testX

//

// or 

//

//   amm re3.sc all

// the nullable function: tests whether the 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

  case NTIMES(r, i) => if (i == 0) true else nullable(r)

}

// the 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))

  case NTIMES(r, i) => 

    if (i == 0) ZERO else SEQ(der(c, r), NTIMES(r, i - 1))

}

// simplification

def simp(r: Rexp) : Rexp = r match {

  case ALT(r1, r2) => (simp(r1), simp(r2)) match {

    case (ZERO, r2s) => r2s

    case (r1s, ZERO) => r1s

    case (r1s, r2s) => if (r1s == r2s) r1s else ALT (r1s, r2s)

  }

  case SEQ(r1, r2) =>  (simp(r1), simp(r2)) match {

    case (ZERO, _) => ZERO

    case (_, ZERO) => ZERO

    case (ONE, r2s) => r2s

    case (r1s, ONE) => r1s

    case (r1s, r2s) => SEQ(r1s, r2s)

  }

  case r => r

}

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

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

  case Nil => r

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

}

// the main matcher function

def matcher(r: Rexp, s: String) : Boolean = 

  nullable(ders(s.toList, r))

// Test Cases

//============

// optional is still just defined

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

// evil regular expressions: (a?){n} a{n}  and (a*)* b

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

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

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)

}

@arg(doc = "Test (a?{n}) (a{n})")

@main

def test1() = {

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

  for (i <- 0 to 9000 by 1000) {

    println(f"$i: ${time_needed(3, matcher(EVIL1(i), "a" * i))}%.5f")

  }

}  

@arg(doc = "Test (a*)* b")

@main

def test2() = {

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

  for (i <- 0 to 6000000 by 500000) {

    println(f"$i: ${time_needed(3, matcher(EVIL2, "a" * i))}%.5f")

  }

}

// 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)

  case NTIMES(r, _) => 1 + size(r)

}

// now the size of the derivatives grows 

// much, much slower and actually maxes out

// at size 8

size(ders("".toList, EVIL2))      // 5

size(ders("a".toList, EVIL2))     // 8

size(ders("aa".toList, EVIL2))    // 8

size(ders("aaa".toList, EVIL2))   // 8

size(ders("aaaa".toList, EVIL2))  // 8

size(ders("aaaaa".toList, EVIL2)) // 8

size(ders(("a" * 20000).toList, EVIL2)) // 8

@arg(doc = "All tests.")

@main

def all() = { test1(); test2() } 

// PS:

//

// If you want to dig deeper into the topic, you can have 

// a look at these examples which still explode. They

// need an even more aggressive simplification.

// test: (a + aa)*

val EVIL3 = STAR(ALT(CHAR('a'), SEQ(CHAR('a'), CHAR('a'))))

@arg(doc = "Test EVIL3")

@main

def test3() = {

  println("Test (a + aa)*")

  for (i <- 0 to 30 by 5) {

    println(f"$i: ${time_needed(1, matcher(EVIL3, "a" * i))}%.5f")

  }

}

// test: (1 + a + aa)*

val EVIL4 = STAR(ALT(ONE, ALT(CHAR('a'), SEQ(CHAR('a'), CHAR('a')))))

@arg(doc = "Test EVIL4")

@main

def test4() = {

  println("Test (1 + a + aa)*")

  for (i <- 0 to 30 by 5) {

    println(f"$i: ${time_needed(1, matcher(EVIL4, "a" * i))}%.5f")

  }

}

@main

def test5() = {

  println("Test (abcdef){n}")

  val re = SEQ(CHAR('a'), SEQ(CHAR('b'), SEQ(CHAR('c'), SEQ(CHAR('d'), SEQ(CHAR('e'), CHAR('f'))))))

  for (i <- 0 to 40000 by 5000) {

    println(f"$i: ${time_needed(1, matcher(NTIMES(re, i), "abcdef" * i))}%.5f")

  }

}

@arg(doc = "Tests that show not all is hunky-dory, but a solution leads too far afield.")

@main

def fail() = { test3(); test4(); test5() } 

// simplification

def simp2(r: Rexp) : Rexp = r match {

  case ALT(r1, r2) => (simp2(r1), simp2(r2)) match {

    case (ZERO, r2s) => r2s

    case (r1s, ZERO) => r1s

    //case (r1s, r2s) => if (r1s == r2s) r1s else ALT (r1s, r2s)

    case (r1s, r2s) => ALT(r1s, r2s)

  }

  case SEQ(r1, r2) =>  (simp2(r1), simp2(r2)) match {

    case (ZERO, _) => ZERO

    case (_, ZERO) => ZERO

    //case (ONE, r2s) => r2s

    //case (r1s, ONE) => r1s

    case (r1s, r2s) => SEQ(r1s, r2s)

  }

  case r => r

}

// some convenience for typing in regular expressions

import scala.language.implicitConversions

def charlist2rexp(s : List[Char]): Rexp = s match {

  case Nil => ONE

  case c::Nil => CHAR(c)

  case c::s => SEQ(CHAR(c), charlist2rexp(s))

}

given Conversion[String, Rexp] = (s => charlist2rexp(s.toList))

val HELLO : Rexp = "hello"

extension (r: Rexp) {

  def | (s: Rexp) = ALT(r, s)

  def % = STAR(r)

  def ~ (s: Rexp) = SEQ(r, s)

}

val re = (ONE | "a" | "ab") ~ ("c" | "bc") ~ ("d" | "e")

simp2(der('d', der('c', der('b', der('a', re)))))