// Scala Lecture 3

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

// A Web Crawler / Email Harvester

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

//

// the idea is to look for links using the

// regular expression "https?://[^"]*" and for

// email addresses using yet another regex.

import io.Source

import scala.util._

// gets the first 10K of a web-page

def get_page(url: String) : String = {

  Try(Source.fromURL(url)("ISO-8859-1").take(10000).mkString).

    getOrElse { println(s"  Problem with: $url"); ""}

}

// regex for URLs and emails

val http_pattern = """"https?://[^"]*"""".r

val email_pattern = """([a-z0-9_\.-]+)@([\da-z\.-]+)\.([a-z\.]{2,6})""".r

//  val s = "foo bla christian@kcl.ac.uk 1234567"

//  email_pattern.findAllIn(s).toList

// drops the first and last character from a string

def unquote(s: String) = s.drop(1).dropRight(1)

def get_all_URLs(page: String): Set[String] = 

  http_pattern.findAllIn(page).map(unquote).toSet

// naive version of crawl - searches until a given depth,

// visits pages potentially more than once

def crawl(url: String, n: Int) : Set[String] = {

  if (n == 0) Set()

  else {

    println(s"  Visiting: $n $url")

    val page = get_page(url)

    val new_emails = email_pattern.findAllIn(page).toSet

    new_emails ++ 

      (for (u <- get_all_URLs(page).par) yield crawl(u, n - 1)).flatten

  }

}

// some starting URLs for the crawler

val startURL = """https://nms.kcl.ac.uk/christian.urban/"""

crawl(startURL, 2)

// User-defined Datatypes and Pattern Matching

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

abstract class Exp

case class N(n: Int) extends Exp                  // for numbers

case class Plus(e1: Exp, e2: Exp) extends Exp

case class Times(e1: Exp, e2: Exp) extends Exp

def string(e: Exp) : String = e match {

  case N(n) => n.toString

  case Plus(e1, e2) => "(" + string(e1) + " + " + string(e2) + ")" 

  case Times(e1, e2) => "(" + string(e1) + " * " + string(e2) + ")" 

}

val e = Plus(N(9), Times(N(3), N(4)))

println(string(e))

def eval(e: Exp) : Int = e match {

  case N(n) => n

  case Plus(e1, e2) => eval(e1) + eval(e2) 

  case Times(e1, e2) => eval(e1) * eval(e2) 

}

def simp(e: Exp) : Exp = e match {

  case N(n) => N(n)

  case Plus(e1, e2) => (simp(e1), simp(e2)) match {

    case (N(0), e2s) => e2s

    case (e1s, N(0)) => e1s

    case (e1s, e2s) => Plus(e1s, e2s)

  }  

  case Times(e1, e2) => (simp(e1), simp(e2)) match {

    case (N(0), _) => N(0)

    case (_, N(0)) => N(0)

    case (N(1), e2s) => e2s

    case (e1s, N(1)) => e1s

    case (e1s, e2s) => Times(e1s, e2s)

  }  

}

println(eval(e))

val e2 = Times(Plus(N(0), N(1)), Plus(N(0), N(9)))

println(string(e2))

println(string(simp(e2)))

// Tokens and Reverse Polish Notation

abstract class Token

case class T(n: Int) extends Token

case object PL extends Token

case object TI extends Token

def rp(e: Exp) : List[Token] = e match {

  case N(n) => List(T(n))

  case Plus(e1, e2) => rp(e1) ::: rp(e2) ::: List(PL) 

  case Times(e1, e2) => rp(e1) ::: rp(e2) ::: List(TI) 

}

println(string(e2))

println(rp(e2))

def comp(ls: List[Token], st: List[Int]) : Int = (ls, st) match {

  case (Nil, st) => st.head 

  case (T(n)::rest, st) => comp(rest, n::st)

  case (PL::rest, n1::n2::st) => comp(rest, n1 + n2::st)

  case (TI::rest, n1::n2::st) => comp(rest, n1 * n2::st)

}

comp(rp(e), Nil)

def proc(s: String) : Token = s match {

  case  "+" => PL

  case  "*" => TI

  case  _ => T(s.toInt) 

}

comp("1 2 + 4 * 5 + 3 +".split(" ").toList.map(proc), Nil)

def string(e: Exp) : String = e match {

  case N(n) => n.toString

  case Plus(e1, e2) => "(" + string(e1) + " + " + string(e2) + ")"

  case Times(e1, e2) => "(" + string(e1) + " * " + string(e2) + ")"

}

val e = Plus(N(9), Times(N(3), N(4)))

println(string(e))

def eval(e: Exp) : Int = e match {

  case N(n) => n

  case Plus(e1, e2) => eval(e1) + eval(e2)

  case Times(e1, e2) => eval(e1) * eval(e2)

}

eval(e)

def simp(e: Exp) : Exp = e match {

  case N(n) => N(n)

  case Plus(e1, e2) => (simp(e1), simp(e2)) match {

    case (N(0), e2s) => e2s

    case (e1s, N(0)) => e1s

    case (e1s, e2s) => Plus(e1s, e2s) 

  }

  case Times(e1, e2) => (simp(e1), simp(e2)) match {

    case (N(0), e2s) => N(0)

    case (e1s, N(0)) => N(0)

    case (N(1), e2s) => e2s

    case (e1s, N(1)) => e1s

    case (e1s, e2s) => Times(e1s, e2s) 

  }

}

val e2 = Times(Plus(N(0), N(1)), Plus(N(0), N(9)))

println(string(e2))

println(string(simp(e2)))

// Token and Reverse Polish Notation

abstract class Token

case class T(n: Int) extends Token

case object PL extends Token

case object TI extends Token

def rp(e: Exp) : List[Token] = e match {

  case N(n) => List(T(n))

  case Plus(e1, e2) => rp(e1) ::: rp(e2) ::: List(PL)

  case Times(e1, e2) => rp(e1) ::: rp(e2) ::: List(TI)

}

def comp(ts: List[Token], stk: List[Int]) : Int = (ts, stk) match {

  case (Nil, st) => st.head

  case (T(n)::rest, st) => comp(rest, n::st)

  case (PL::rest, n1::n2::st) => comp(rest, n1 + n2::st)

  case (TI::rest, n1::n2::st) => comp(rest, n1 * n2::st)

}

def exp(ts: List[Token], st: List[Exp]) : Exp = (ts, st) match {

  case (Nil, st) => st.head

  case (T(n)::rest, st) => exp(rest, N(n)::st)

  case (PL::rest, n1::n2::st) => exp(rest, Plus(n2, n1)::st)

  case (TI::rest, n1::n2::st) => exp(rest, Times(n2, n1)::st)

}

exp(toks(e2), Nil)

def proc(s: String) = s match {

  case "+" => PL

  case "*" => TI

  case n => T(n.toInt)

}

string(exp("1 2 + 4 * 5 + 3 +".split(" ").toList.map(proc), Nil))

// Tail recursion

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

def fact(n: Long): Long = 

  if (n == 0) 1 else n * fact(n - 1)

def factB(n: BigInt): BigInt = 

  if (n == 0) 1 else n * factB(n - 1)

factB(100000)

fact(10)              //ok

fact(10000)           // produces a stackoverflow

def factT(n: BigInt, acc: BigInt): BigInt =

  if (n == 0) acc else factT(n - 1, n * acc)

factT(10, 1)

println(factT(100000, 1))

// there is a flag for ensuring a function is tail recursive

import scala.annotation.tailrec

@tailrec

def factT(n: BigInt, acc: BigInt): BigInt =

  if (n == 0) acc else factT(n - 1, n * acc)

// for tail-recursive functions the Scala compiler

// generates loop-like code, which does not need

// to allocate stack-space in each recursive

// call; Scala can do this only for tail-recursive

// functions

// Jumping Towers

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

// the first n prefixes of xs

// for 1 => include xs

def moves(xs: List[Int], n: Int) : List[List[Int]] = (xs, n) match {

  case (Nil, _) => Nil

  case (xs, 0) => Nil

  case (x::xs, n) => (x::xs) :: moves(xs, n - 1)

}

moves(List(5,1,0), 1)

moves(List(5,1,0), 2)

moves(List(5,1,0), 5)

// checks whether a jump tour exists at all

def search(xs: List[Int]) : Boolean = xs match {

  case Nil => true

  case (x::xs) =>

    if (xs.length < x) true else moves(xs, x).exists(search(_))

}

search(List(5,3,2,5,1,1))

search(List(3,5,1,0,0,0,1))

search(List(3,5,1,0,0,0,0,1))

search(List(3,5,1,0,0,0,1,1))

search(List(3,5,1))

search(List(5,1,1))

search(Nil)

search(List(1))

search(List(5,1,1))

search(List(3,5,1,0,0,0,0,0,0,0,0,1))

// generates *all* jump tours

//    if we are only interested in the shortes one, we could

//    shortcircut the calculation and only return List(x) in

//    case where xs.length < x, because no tour can be shorter

//    than 1

// 

def jumps(xs: List[Int]) : List[List[Int]] = xs match {

  case Nil => Nil

  case (x::xs) => {

    val children = moves(xs, x)

    val results = children.map((cs) => jumps(cs).map(x :: _)).flatten

    if (xs.length < x) List(x) :: results else results

  }

}

println(jumps(List(5,3,2,5,1,1)).minBy(_.length))

jumps(List(3,5,1,2,1,2,1))

jumps(List(3,5,1,2,3,4,1))

jumps(List(3,5,1,0,0,0,1))

jumps(List(3,5,1))

jumps(List(5,1,1))

jumps(Nil)

jumps(List(1))

jumps(List(5,1,2))

moves(List(1,2), 5)

jumps(List(1,5,1,2))

jumps(List(3,5,1,0,0,0,0,0,0,0,0,1))

jumps(List(5,3,2,5,1,1)).minBy(_.length)

jumps(List(1,3,5,8,9,2,6,7,6,8,9)).minBy(_.length)

jumps(List(1,3,6,1,0,9)).minBy(_.length)

jumps(List(2,3,1,1,2,4,2,0,1,1)).minBy(_.length)

// Sudoku 

//========

// THE POINT OF THIS CODE IS NOT TO BE SUPER

// EFFICIENT AND FAST, just explaining exhaustive

// depth-first search

val game0 = """.14.6.3..

              |62...4..9

              |.8..5.6..

              |.6.2....3

              |.7..1..5.

              |5....9.6.

              |..6.2..3.

              |1..5...92

              |..7.9.41.""".stripMargin.replaceAll("\\n", "")

type Pos = (Int, Int)

val EmptyValue = '.'

val MaxValue = 9

val allValues = "123456789".toList

val indexes = (0 to 8).toList

def empty(game: String) = game.indexOf(EmptyValue)

def isDone(game: String) = empty(game) == -1 

def emptyPosition(game: String) = 

  (empty(game) % MaxValue, empty(game) / MaxValue)

def get_row(game: String, y: Int) = 

  indexes.map(col => game(y * MaxValue + col))

def get_col(game: String, x: Int) = 

  indexes.map(row => game(x + row * MaxValue))

def get_box(game: String, pos: Pos): List[Char] = {

    def base(p: Int): Int = (p / 3) * 3

    val x0 = base(pos._1)

    val y0 = base(pos._2)

    val ys = (y0 until y0 + 3).toList

    (x0 until x0 + 3).toList.flatMap(x => ys.map(y => game(x + y * MaxValue)))

}

//get_row(game0, 0)

//get_row(game0, 1)

//get_col(game0, 0)

//get_box(game0, (3, 1))

// this is not mutable!!

def update(game: String, pos: Int, value: Char): String = 

  game.updated(pos, value)

def toAvoid(game: String, pos: Pos): List[Char] = 

  (get_col(game, pos._1) ++ get_row(game, pos._2) ++ get_box(game, pos))

def candidates(game: String, pos: Pos): List[Char] = 

  allValues.diff(toAvoid(game, pos))

//candidates(game0, (0,0))