behemoth

#!/bin/bash
#
# shortcuts for behemoth commands
# assumes that mvn clean install has been called succesfully

# if no args specified, show usage
if [ $# = 0 ]; then
  echo "Usage: behemoth COMMAND"
  echo "where COMMAND is one of:"
  echo "  reader"
  echo "  exporter"
  echo "  filter"
  echo "  importer"
  echo "  gate"
  echo "  tika"
  echo "  uima"
  echo "  mahout"
  echo "  solr"
  echo "  language-id"
  echo "Most commands print help when invoked w/o parameters."
  exit 1
fi

# get arguments
COMMAND=$1
shift

# figure out which class to run
if [ "$COMMAND" = "reader" ] ; then
  CLASS=com.digitalpebble.behemoth.util.CorpusReader
  MODULE=core
elif [ "$COMMAND" = "filter" ] ; then
  CLASS=com.digitalpebble.behemoth.util.CorpusFilter
  MODULE=core
elif [ "$COMMAND" = "importer" ] ; then
  CLASS=com.digitalpebble.behemoth.util.CorpusGenerator
  MODULE=core
elif [ "$COMMAND" = "exporter" ] ; then
  CLASS=com.digitalpebble.behemoth.util.ContentExtractor
  MODULE=core
elif [ "$COMMAND" = "gate" ] ; then
  CLASS=com.digitalpebble.behemoth.gate.GATEDriver
  MODULE=gate
elif [ "$COMMAND" = "tika" ] ; then
  CLASS=com.digitalpebble.behemoth.tika.TikaDriver
  MODULE=tika
elif [ "$COMMAND" = "uima" ] ; then
  CLASS=com.digitalpebble.behemoth.uima.UIMADriver
  MODULE=uima
elif [ "$COMMAND" = "mahout" ] ; then
  CLASS=com.digitalpebble.behemoth.mahout.SparseVectorsFromBehemoth
  MODULE=mahout
elif [ "$COMMAND" = "solr" ] ; then
  CLASS=com.digitalpebble.behemoth.solr.SOLRIndexerJob
  MODULE=solr
elif [ "$COMMAND" = "language-id" ] ; then
  CLASS=com.digitalpebble.behemoth.languageidentification.LanguageIdDriver
  MODULE=language-id
else
   echo "unknown command"
   exit -1;
fi

BEHE_JOB=$(dirname $0)/$MODULE/target/behemoth-$MODULE-1.1-SNAPSHOT-job.jar
EXEC_CALL="hadoop jar $BEHE_JOB"

# check that hadoop can be found on the path
if [ $(which hadoop | wc -l ) -eq 0 ]; then
    echo "Can't find Hadoop executable. Add HADOOP_HOME/bin to the path"
    exit -1;
fi

# run it
exec $EXEC_CALL $CLASS "$@"