kaja47

// you need Scala 2.10 to run this

val text = ???

// we are interested in bi-grams
val n = 2

// map from ngrams to seq of following words
val map = collection.mutable.Map[Seq[String], Seq[String]]() withDefaultValue Seq()

kaja47

<?php

// PHP implementation of http://norvig.com/spell-correct.html in 49 lines of code
// based on http://soundofemotion.com/spellcorrect.txt which is much longer

function words($text) {
  return preg_split("~[^a-z]+~", $text, null, PREG_SPLIT_NO_EMPTY);
}

function train($features) {

kaja47

import collection.immutable.BitSet

case class Matrix(w: Int, h: Int, bitmap: BitSet) {
  def reset(cells: Seq[(Int, Int)]): Matrix =
    copy(bitmap = BitSet() ++ (cells map (pos _).tupled))

  def pos(x: Int, y: Int): Int = x + y * w

  def at(x: Int, y: Int): Boolean = bitmap(pos(x, y))

kaja47

class A { 
  def callThis = println("YOU MANIACS! YOU BLEW IT UP! OH, DAMN YOU! GODDAMN YOU ALL TO HELL!")
}
val a = new A

a.getClass.getMethod("callThis").invoke(a) // reflection
a.asInstanceOf[{ def callThis }].callThis  // duck-typed reflection

kaja47

<?php

trait WantsAuthenticator {
  function injectAuthenticator(Authenticator $authenticator) {
    $this->authenticator = $authenticator;  
  }  
}
trait WantsCache {
  function injectCache(Cache $cache) {
    $this->cache = $cache;  

kaja47

import javax.imageio.ImageIO
import java.io.File
import java.awt.image.BufferedImage
import java.awt.{ RenderingHints, AlphaComposite }

// Content-aware image cropping with ChunkyPNG
// https://gist.github.com/a54cd41137b678935c91

final class Cropper(file: String) {
  val image = ImageIO.read(new File(file))

kaja47

import java.util.regex.Pattern

def add(numbers: String, delimiters: Seq[String] = Seq(",")): Int = {
  val delimRegex = "^//(.+?)\n(.*)$".r
  val multiDelim = "^//(\\[.+?\\])\n(.*)$".r
  numbers match {
    case "" => 0
    case multiDelim(dels, ns) => add(ns, dels.init.tail split "\\]\\[")
    case delimRegex(del, ns) => add(ns, Seq(del))
    case ns =>

kaja47

-- InnoDB table posts (uid, chan, board, threadId, id, ....)
-- primary (uid)
-- index (chan, board, threadId, id)

-- slow as fuck (Using where; Using index)
select distinct board
from posts
where chan = '4chan'

-- fast (Using where; Using index for group-by)

kaja47

-- rather slow
select * from posts
where text rlike '[[:<:]](pic|picture|image|photo)[[:>:]].*related'

-- much much faster
select * from posts
where text  like '%related%'
  and text rlike '[[:<:]](pic|picture|image|photo)[[:>:]].*related'

kaja47

// http://blog.greweb.fr/2012/11/play-framework-enumerator-outputstream/
import java.io.{ BufferedWriter, OutputStreamWriter }

val enumerator = Enumerator.outputStream { os =>
  val writer = new BufferedWriter(new OutputStreamWriter(os, encoding), 4096)
  template.writeTo(writer)
  writer.close()
}
Ok.stream(enumerator andThen Enumerator.eof).withHeaders(
  "Content-Type" -> "text/html"