kyleburton

(defn find-dupes-naieve [inp-seq]
  (reduce (fn [res item]
            (assoc res item (inc (get res item 0))))
          {}
          (map #(second (.split %1 "\t")) inp-seq)))


(defn find-dupes-with-bloom-filter [inp-seq expected-size fp-prob]
  (let [flt (bloom/make-optimal-filter expected-size fp-prob)]
    (reduce (fn [res item]

kyleburton

public class Main {
  public static void main ( String args [] ) {
    int size = 392000000;
    int sampleSize = 20000;
    System.err.println("Start: allocation");
    int [] array = new int[size];
    System.err.println("Done: allocation");
    java.util.Random rnd = new java.util.Random();

    System.err.println("Start: population w/line numbers");

kyleburton

SELECT column FROM table
ORDER BY RANDOM()
LIMIT 20000

kyleburton

(ns clj-etl-utils.sequences)


(def random-sample-seq
     (let [rnd (java.util.Random.)]
       (fn self [[item & population :as population-seq] population-size remaining-samples-needed]
         (if (or (zero? remaining-samples-needed) (empty? population-seq))
           nil
           (if (< (.nextInt rnd population-size) remaining-samples-needed)
             (lazy-cat

kyleburton

#!/usr/bin/env ruby
require File.join(File.dirname(__FILE__),'..','src','main','ruby','util')
require File.join(File.dirname(__FILE__),'deploy-check')

class ClojureServiceDeployer < ServiceUtils

  def run

    unless File.exist? release_jar_file
      unless system("mvn assembly:assembly")

kyleburton

#!/bin/bash

SVC_HOME="/var/lib/the-clj-service/current"

# no args means we fork
if [ -z "${1:-}" ]; then
  export SVC_JAR="$SVC_HOME/the-clj-servicejar"
  if [ -z "$SVC_JAR" ]; then
    echo "Error: SVC_JAR not found"
    exit 1

kyleburton

#!/bin/sh -e

. /lib/lsb/init-functions

PATH=/bin:/usr/bin:/sbin:/usr/sbin
DAEMON=/var/lib/the-clj-service/the-clj-service.sh
NAME=the-clj-service
# any args passed to the process
SERVICES_OPTS="${2:-}"
PIDDIR="/var/run/the-clj-service"

kyleburton

require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', '..', '..','lib','our_chef'))

node[:our_config] ||= {}
node[:our_config][:jvm_opts] ||= {}
node[:our_config][:jvm_opts][:max_heap] ||= "256m"

bash "create user servicerunner" do
  user "root"
  code <<-EOF
    useradd -s /bin/bash -m servicerunner

kyleburton

import scala.io.Source
import scala.xml._

object SbtDepsFromPom {
  def main ( args: Seq[String] ) {
    val pomContents = Source.fromFile(args(0)).mkString
    var xml = XML.loadString(pomContents)
    var deps = (xml \\ "dependency") map { (dep) =>
      List( (dep \\ "groupId")(0).text,
            (dep \\ "artifactId")(0).text,

kyleburton

require 'java'
#require 'clojure-1.0.0.jar'
Dir["#{File.dirname(__FILE__)}/*.jar"].each { |jar| puts "requiring: #{jar}"; require jar }
import "clojure.lang.RT"

class CljHelper
  def initialize *pkgs
    @mappings = {}
    @ns_map  = RT.var "clojure.core", "ns-map"
    @symbol  = RT.var "clojure.core", "symbol"