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

#!/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

(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

SELECT column FROM table
ORDER BY RANDOM()
LIMIT 20000

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

(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

(sequences/random-sample-seq (ds/read-lines "phone-nums.txt")
                             2000000
                             200)

kyleburton

(defn count-area-codes [inp-seq]
  (reduce (fn [m line]
            (let [phnum         (second (.split line "\t"))
                  [_ area-code] (first (re-seq #"\((\d+)\)" phnum))]
              (assoc m area-code (inc (get m area-code 0)))))
          {}
          inp-seq))

  (apply
   merge-with +

kyleburton

(reduce
   (fn [res counts]
     (merge-with + res counts))
   (pmap (fn [[start end]]
           (count-area-codes 
             (io/read-lines-from-file-segment inp-file start end)))
         (partition 2 1
           (io/byte-partitions-at-line-boundaries
             inp-file
             (* 1024 1024)))))

kyleburton

(defun krb-rename-file-visited-by-this-buffer (new-name)
  (interactive (list (read-string (format "Rename [%s] to: " buffer-file-name))))
  (let ((file-name buffer-file-name))
    (if (not (string-match "/" new-name))
        (setq new-name (format "%s%s" (file-name-directory file-name)
                               new-name)))
    (message "rename-file: %s to %s" file-name new-name)
    (rename-file file-name new-name)
    (kill-buffer (buffer-name))
    (find-file new-name)))