#!/usr/bin/env bash

HADOOP_HOME=${HADOOP_HOME:?HADOOP_HOME is not set}

script_path=`dirname $0`

# Clean up output from prior runs
for dir in "/tmp/input" "/tmp/output" "/tmp/final_output"; do
  if [[ -d $dir ]] ; then
    rm -fr $dir
  fi
done

mkdir /tmp/input

user=$1
host=$2
database=$3
pass=$4

mysql -u $user --password=$pass --host=$host --batch -e \
      "SELECT path_id, product_id FROM recent_items" \
      $database > /tmp/input/recent_items.tsv

jarfile="$HADOOP_HOME/mapred/contrib/streaming/hadoop-0.21.0-streaming.jar"

"$HADOOP_HOME/bin/hadoop" jar \
  $jarfile \
  -input /tmp/input \
  -output /tmp/output \
  -mapper cat \
  -reducer $script_path/reduce.rb

cat /tmp/output/part* > /tmp/output/related.tsv

"$HADOOP_HOME/bin/hadoop" jar \
  $jarfile \
  -input /tmp/output/related.tsv \
  -output /tmp/final_output \
  -mapper cat \
  -reducer $script_path/reduce2.rb

cat /tmp/final_output/part* > /tmp/final_output/viewed_products.tsv

mysqlimport --local --compress -u $user --host=$host \
            --columns=source_product_id,target_product_id,count \
            --replace $database \
            /tmp/final_output/viewed_products.tsv