RandomSec/OpenRefine/refine

#!/usr/bin/env bash

########################################################## 
#               OpenRefine Control System             #
##########################################################

# -------------- utility functions ----------------------

fail () {
   cat <<EOF
ERROR: $1
Type '$0 -h' for usage information.
EOF
exit 1
}

error() {
    echo "Error: $1"
    exit 1
}

warn() {
    echo "Warning: $1"
    exit 0
}

usage() {
    cat <<EOF
Usage: $0 [options] <action>
where [options] include:

  -h print this message and exit
  
  -p <port> the port that OpenRefine will listen to
     default: 3333

  -i <interface> the network interface OpenRefine should bind to
     default: 127.0.0.1
	 
  -H <host> the expected value for the Host header (set to * to disable checks)
     default: <interface>

  -w <path> path to the webapp
     default: main/webapp

  -d <path> path to the data directory
     default: OS dependent

  -m <memory> max memory heap size to use
     default: 1400M

  -v <level> verbosity level [from low to high: error,warn,info,debug,trace]
     default: info

  -x <name=value> additional configuration parameters to pass to OpenRefine
     default: [none]

  -c <path> path to refine.ini file
     default: ./refine.ini
  
  --debug enable JVM debugging (on port 8000)
       
  --jmx enable JMX monitoring (for jconsole and jvisualvm)
  
and <action> is one of

   build ............................... Build OpenRefine      
   run ................................. Run OpenRefine [default]

   test ................................ Run all OpenRefine tests
   server_test ......................... Run only the server tests
   extensions_test ..................... Run only the extensions tests
   ui_tests ............................ Run only the UI tests
   
   broker .............................. Run OpenRefine Broker
   
   broker_appengine_run <id> <ver> ..... Run OpenRefine Broker for Google App Engine in local server
   broker_appengine_upload <id> <ver> .. Upload OpenRefine to Google App Engine
   
   findbugs ............................ Run Findbugs against OpenRefine
   pmd ................................. Run PMD against OpenRefine
   cpd ................................. Run Copy/Paste Detection against OpenRefine
   jslint .............................. Run JSlint against OpenRefine

   whitespace <extension> .............. Normalize whitespace in files with the given extension

   mac_dist <version> .................. Make MacOSX binary distribution
   windows_dist <version> .............. Make Windows binary distribution
   linux_dist <version> ................ Make Linux binary distribution
   dist <version> ...................... Make all distributions

   clean ............................... Clean compiled classes
   distclean ........................... Remove all generated files
                
EOF
    exit 1
}
                
add_option() {
    if [ ! -z "$*" ] ; then
        OPTS+=("$@")
    fi
}
                
load_configs() {
   TEMP_CONFIG=$(mktemp -t refine.XXXXXXX)
   if [ "${TEMP_CONFIG}" = "" ] ; then
       error "Could not create temporary file to load configurations"
   fi
   cat $1 | egrep "^[A-Z]" | sed 's/^\([^=]*\)=\(.*\)$/export \1=(\2)/' > ${TEMP_CONFIG}
   . ${TEMP_CONFIG}
   rm ${TEMP_CONFIG}
}
                    
check_macosx() {
    if [ "$OS" != "macosx" ] ; then
        error "This action can only run on MacOSX"
    fi
}

check_downloaders() {
    CURL="`which curl 2> /dev/null`"
    WGET="`which wget 2> /dev/null`"
    
    if [ -z "$CURL" ] && [ -z "$WGET" ] ; then
        error "We need either 'curl' or 'wget' present in PATH to download external dependencies."
    fi
}

check_unzip() {
    UNZIP="`which unzip 2> /dev/null`"
    
    if [ -z "$UNZIP" ] ; then
        error "We need 'unzip' present in PATH to expand external dependencies."
    fi
}
    
check_python() {
    PYTHON="`which python 2> /dev/null`"
    if [ -z "$PYTHON" ] ; then
        error "This action requires you to have 'python' installed and present in your PATH. You can download it for free at http://www.python.org/"
    fi
    PYTHON_VERSION="`$PYTHON --version 2>&1 | cut -f 2 -d ' ' | cut -f 1,2 -d .`"
    if [ "$PYTHON_VERSION" != "2.6" ] && [ "$PYTHON_VERSION" != "2.7" ]; then
        error "This action requires Python version 2.6.x. or 2.7.x. You can download it for free at http://www.python.org/"
    fi
}

check_pywin32() {
    PYWIN32="`$PYTHON -c 'import win32api' 2>&1`"

    if [ ! -z "$PYWIN32" ] ; then
        error "This action requires you to have 'pywin32' windows extensions for Python installed on your machine. You can download it for free at http://sourceforge.net/projects/pywin32/"
    fi
}

check_running() {
    check_downloaders
    URL="http://${REFINE_HOST_INTERNAL}:${REFINE_PORT}/"
    CHECK_STR="<title>OpenRefine</title>"

    if [ "$CURL" ] ; then
        curl -s -S -f $URL > /dev/null 2>&1
        CURL_RETURN=$?
        if [ $CURL_RETURN -eq "7" ] || [ $CURL_RETURN -eq "22" ]  ; then
            NOT_RUNNING="1"
        fi
    elif [ "$WGET" ] ; then
        wget -O - $URL > /dev/null 2>&1
        if [ "$?" = "4" ] ; then
            NOT_RUNNING="1"
        fi
    fi    

    if [ -z "${NOT_RUNNING}" ] ; then
        if [ "$CURL" ] ; then
            RUNNING=`curl -s $URL | grep "$CHECK_STR"`
        elif [ "$WGET" ] ; then
            RUNNING=`wget -q -O - $URL | grep "$CHECK_STR"`
        fi    
        
        if [ -z "${RUNNING}" ] ; then
            error "Something is already running on $URL but doesn't seem to be OpenRefine. Maybe a proxy issue?"
        fi
    else
        RUNNING=""
    fi
}

get_version() {
    VERSION="$1"
    
    if [ -z "$VERSION" ] ; then
        fail "Must specify a version number"
    fi

    NUM_VERSION=`echo $VERSION | sed -E 's/-.*//g'`
    
    if [ "${NUM_VERSION}" = "" ] ; then
        fail "${VERSION} is not a valid version number"
    fi
    
    if [ "`echo "${NUM_VERSION}" | egrep '^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$'`" = "${NUM_VERSION}" ] ; then
        FULL_VERSION="${NUM_VERSION}"
    elif [ "`echo "${NUM_VERSION}" | egrep '^[0-9]+\.[0-9]+\.[0-9]+$'`" = "${NUM_VERSION}" ] ; then
        FULL_VERSION="${NUM_VERSION}.0"
    elif [ "`echo "${NUM_VERSION}" | egrep ''^[0-9]+\.[0-9]+$''`" = "${NUM_VERSION}" ] ; then
        FULL_VERSION="${NUM_VERSION}.0.0"
    elif [ "`echo "${NUM_VERSION}" | egrep '^[0-9]+$'`" = "${NUM_VERSION}" ] ; then
        FULL_VERSION="${NUM_VERSION}.0.0.0"
    else 
        fail "${VERSION} is not a valid version number"
    fi
}

get_revision() {
    if [ -d ".svn" ] ; then
        INFO=`svn info`
        REVISION=`echo $INFO | sed s/^$VERSION-//`
    elif [ -d ".git" ] ; then
        INFO=`git describe`
        REVISION=`echo $INFO`
    REVISION=${REVISION:4}
    else
        error "cannot obtain revision, exiting!"
    fi
}
    
download() {
    URL=$1
    DEST=$2
    
    check_downloaders
    
    if [ "$CURL" ] ; then
        curl -L -o $DEST $URL || exit "Error while downloading $URL"
    elif [ "$WGET" ] ; then
        wget -O $DEST $URL || error "Error while downloading $URL"
    fi
}

tool_download() {
    URL=$1
    FILE=$2
    DIR=$3
    
    cd $REFINE_TOOLS_DIR
        if [ ! -f "$FILE" ] ; then
            download $URL $FILE
        fi
        if [ ! -d "$DIR" ] ; then
            if [ -z "`echo $FILE | sed 's@.*.tar.gz$@@' | sed 's@.*.tgz$@@'`" ] ; then
                tar xzf $FILE || error "Error while expanding $FILE"
            fi
            if [ -z "`echo $FILE | sed 's@.*.zip$@@'`" ] ; then
                check_unzip
                $UNZIP -q $FILE || error "Error while expanding $FILE"
            fi
        fi
    cd ..
}

load_data() {
    FILE=$1
    NAME=$2
    URL="http://${REFINE_HOST_INTERNAL}:${REFINE_PORT}/command/core/create-project-from-upload"
    CURL="`which curl 2> /dev/null`"
    
    if [ -z "$CURL" ] ; then
        error "We need 'curl' present in PATH to upload data to OpenRefine."
    else
        curl -s -F "project-file=@$FILE" -F "project-name=$NAME" $URL > /dev/null || error "Error while uploading $FILE to OpenRefine"
        echo "Loaded $FILE as $NAME"
    fi
}

display() {
    FILE=$1
    if [ "$OS" = "macosx" ] ; then
    open $FILE
    elif [ "$OS" = "linux" ] ; then
    gnome-open $FILE
    else
    notepad $FILE
    fi  

}
# ----------------------------------------------------------------------------------------------

build_prepare() {
    if [ ! -d $REFINE_BUILD_DIR ] ; then 
        mkdir $REFINE_BUILD_DIR || error "Error while making directory $REFINE_BUILD_DIR"
    fi
}

dist_prepare() {
    if [ ! -d $REFINE_DIST_DIR ] ; then 
        mkdir $REFINE_DIST_DIR || error "Error while making directory $REFINE_DIST_DIR"
    fi
}

tools_prepare() {
    if [ ! -d $REFINE_TOOLS_DIR ] ; then 
        mkdir $REFINE_TOOLS_DIR || error "Error while making directory $REFINE_TOOLS_DIR"
    fi
}

mvn_prepare() {
    tools_prepare

    MVN_VERSION="3.8.4"
    MVN_URL="https://dlcdn.apache.org/maven/maven-3/${MVN_VERSION}/binaries/apache-maven-${MVN_VERSION}-bin.tar.gz"
    MVN_FILE=`echo $MVN_URL | sed 's|.*/||'`
    MVN_DIR="apache-maven-${MVN_VERSION}"

    MVN="`which mvn 2> /dev/null`"
   
    if [ -z "$MVN" ] ; then
        if [ -z "$MVN_HOME" ] ; then
            cd $REFINE_TOOLS_DIR
                if [ ! -f "$MVN_FILE" ] ; then
                    echo "\nCould not find Maven locally, starting download for Maven ..."
                    download $MVN_URL $MVN_FILE
                fi
                if [ ! -d "$MVN_DIR" ] ; then
                    tar xzf $MVN_FILE -C . || error "Error while expanding $MVN_FILE"
                fi
                export MVN_HOME="`pwd`/$MVN_DIR"
                if [ "$OS" = "windows" ] ; then
                    export MVN_HOME=`cygpath --unix "$MVN_HOME"`
                fi
            cd ..
        fi
        MVN="$MVN_HOME/bin/mvn"
    fi
}

appengine_prepare() {
    if [ -z "$APPENGINE_HOME" ] ; then
        error "You have to have the APPENGINE_HOME environment variable set and pointing to the local installation of the Google AppEngine SDK."
    elif [ ! -f "$APPENGINE_HOME/bin/appcfg.sh" ] ; then
        error "Environment variable APPENGINE_HOME is set to '$APPENGINE_HOME' which doesn't point to a valid Google App Engine SDK."
    fi
    APPENGINE="$APPENGINE_HOME/bin/appcfg.sh"
    APPENGINE_LOCAL="$APPENGINE_HOME/bin/dev_appserver.sh"
    
    ANT_PARAMS="$ANT_PARAMS -Dappengine.sdk.dir=$APPENGINE_HOME"
}
   
virtualenv_prepare() {
    check_python

    VIRTUALENV_DIR="virtualenv-1.9.1"
    VIRTUALENV_FILE="${VIRTUALENV_DIR}.tar.gz"
    VIRTUALENV_URL="http://pypi.python.org/packages/source/v/virtualenv/${VIRTUALENV_FILE}"

    tool_download $VIRTUALENV_URL $VIRTUALENV_FILE $VIRTUALENV_DIR
    
    PYTHON_LOCAL="$REFINE_TOOLS_DIR/python"
    if [ "$OS" = "windows" ] ; then
        PYTHON_LOCAL="${PYTHON_LOCAL}_win"
    fi
    
    if [ ! -d "$PYTHON_LOCAL" ] ; then
        $PYTHON $REFINE_TOOLS_DIR/$VIRTUALENV_DIR/virtualenv.py $PYTHON_LOCAL
    fi
    
    PYTHON_HOME="`pwd`/$PYTHON_LOCAL"

    if [ "$OS" = "windows" ] ; then
        PYTHON="$PYTHON_HOME/Scripts/python.exe"
        PYTHON_INSTALL="$PYTHON_HOME/Scripts/easy_install.exe"
    else 
        PYTHON="$PYTHON_HOME/bin/python"
        PYTHON_INSTALL="$PYTHON_HOME/bin/easy_install"
    fi
}
    
# ----------------------------------------------------------------------------------------------

mvn() {
    mvn_prepare   
    
    "$MVN" $MVN_PARAMS -Dversion="$VERSION" -Dfull_version="$FULL_VERSION"  $1 || error "Error while running maven task '$1'"
}

# ----------------------------------------------------------------------------------------------

dist() {
    get_version $1
    mvn_prepare
    "$MVN" versions:set -DnewVersion="$VERSION"
    "$MVN" package
}

windows_dist() {
    get_version $1
    mvn_prepare
    "$MVN" versions:set -DnewVersion="$VERSION"
    "$MVN" package -P windows
}

linux_dist() {
    get_version $1
    mvn_prepare
    "$MVN" versions:set -DnewVersion="$VERSION"
    "$MVN" package -P linux
}
 
# Kept just in case someone wants to follow this workflow on a mac,
# but no longer needed as "mvn package" does it directly on both mac and linux.
mac_dist() {
    check_macosx
    dist_prepare
    get_version $1
    get_revision

    appbundler_prepare
    
    ANT_PARAMS="-Dappbundler.dir=${REFINE_TOOLS_DIR}/${APPBUNDLER_DIR}"
    ant mac

    mkdir -p "$REFINE_BUILD_DIR/mac/.background"
    cp graphics/dmg_background/dmg_background.png "$REFINE_BUILD_DIR/mac/.background/dmg_background.png"
    
    SIZE=350
    
    if [ -f "$REFINE_BUILD_DIR/temp_refine.dmg" ] ; then
        rm "$REFINE_BUILD_DIR/temp_refine.dmg"
    fi

# Sign the bundle with a self-signed cert so OS X doesn't frustrate users by making app invisible
    codesign --deep -s "OpenRefine Code Signing" "$REFINE_BUILD_DIR/mac/OpenRefine.app"
    spctl --assess --type execute --verbose=4 "$REFINE_BUILD_DIR/mac/OpenRefine.app"

    TITLE="OpenRefine $VERSION"
    echo "Building MacOSX DMG for $TITLE"
    hdiutil create -srcfolder "$REFINE_BUILD_DIR/mac" -volname "$TITLE" -fs HFS+ -fsargs "-c c=64,a=16,e=16" -format UDRW -size ${SIZE}m "$REFINE_BUILD_DIR/temp_refine.dmg" || error "can't create empty DMG"
    DEVICE=`hdiutil attach -readwrite -noverify -noautoopen "$REFINE_BUILD_DIR/temp_refine.dmg" | egrep '^/dev/' | sed -e "s/^\/dev\///g" -e 1q  | awk '{print $1}'`
    echo $DEVICE
    hdiutil attach "$REFINE_BUILD_DIR/temp_refine.dmg" || error "Can't attach temp DMG"
    
    echo '
       tell application "Finder"
         tell disk "'$TITLE'"
            open
            set current view of container window to icon view
            set toolbar visible of container window to false
            set statusbar visible of container window to false
            set the bounds of container window to {200, 100, 760, 460}
            set theViewOptions to the icon view options of container window
            set arrangement of theViewOptions to not arranged
            set icon size of theViewOptions to 100
            set background picture of theViewOptions to file ".background:dmg_background.png"
            make new alias file at container window to POSIX file "/Applications" with properties {name:"Applications"}
            set position of item "OpenRefine" of container window to {170, 175}
            set position of item "Applications" of container window to {380, 175}
            close
            open               
            update without registering applications
            delay 5            
            eject
         end tell
       end tell
    ' | osascript || error "Error running applescript"
    
    sync
    sync
    sleep 3
    hdiutil detach $DEVICE
    
    if [ -f "$REFINE_DIST_DIR/openrefine-mac-$VERSION.dmg" ] ; then
        rm "$REFINE_DIST_DIR/openrefine-mac-$VERSION.dmg"
    fi
    
    hdiutil convert "$REFINE_BUILD_DIR/temp_refine.dmg" -format UDZO -imagekey zlib-level=9 -o "$REFINE_DIST_DIR/openrefine-mac-$VERSION.dmg" || error "Error compressing DMG"
    hdiutil internet-enable -yes "$REFINE_DIST_DIR/openrefine-mac-$VERSION.dmg" || error "Error internet-enabling DMG"
    
    rm -f "$REFINE_BUILD_DIR/temp_refine.dmg"
}

test() {
    mvn_prepare
    $MVN test
}


ui_tests() {
    download http://okfnlabs.org/reconcile-csv/dist/reconcile-csv-0.1.2.jar ./tools/reconcile-csv-0.1.2.jar
    RECONCILE_SERVER_CMD="$JAVA -Xmx2g -jar ./tools/reconcile-csv-0.1.2.jar  ./main/tests/cypress/cypress/fixtures/csv-reconcile-species.csv scientific_name taxon_id"
    echo "Starting reconcile-csv-0.1.2 ..."
    $RECONCILE_SERVER_CMD 2>&1 &
    RECONCILE_SERVER_PID="$!"
    
    get_revision

    CYPRESS_RECORD=0

    if [ -z "$CYPRESS_BROWSER" ] ; then
        CYPRESS_BROWSER="electron"
    fi
    
    if [ ! -z "$CYPRESS_PROJECT_ID" ] && [ ! -z "$CYPRESS_RECORD_KEY" ] ; then
        CYPRESS_RECORD=1
        echo "Tests will be recorded in Cypress Dashboard"
    elif [ ! -z "$CYPRESS_PROJECT_ID" ] && [ -z "$CYPRESS_RECORD_KEY" ] ; then
        fail "Found a Cypress project id but no record key"
    fi

    REFINE_DATA_DIR="${TMPDIR:=/tmp}/openrefine-tests"
    
    add_option "-Drefine.headless=true"
    add_option "-Drefine.autoreload=false"
    add_option "-Dbutterfly.autoreload=false"
    
    run fork > /dev/null
    
    echo "Waiting for OpenRefine to load..."
    sleep 5
    check_running
    if [ -z "$RUNNING" ] ; then
        sleep 10
    fi
    echo "... proceed with the tests."
    echo ""

    echo "Starting Cypress..."
    # Cypress needs a unique group id
    # We're hashing the list of files to generate such Group Id
    CYPRESS_GROUP=$(echo $CYPRESS_BROWSER$CYPRESS_SPECS | shasum)
    CYPRESS_RUN_CMD="yarn --cwd ./main/tests/cypress run cypress run --spec "$CYPRESS_SPECS" --browser $CYPRESS_BROWSER --group "$CYPRESS_GROUP" --headless --quiet --reporter list --env DISABLE_PROJECT_CLEANUP=1,OPENREFINE_URL=http://$REFINE_HOST_INTERNAL:$REFINE_PORT"
    if [ "$CYPRESS_RECORD" = "1" ] ; then
        # if tests are recorded, project id is added to env vars, and --record flag is added to the cmd-line
        export CYPRESS_PROJECT_ID=$CYPRESS_PROJECT_ID
        CYPRESS_RUN_CMD="$CYPRESS_RUN_CMD --record --key $CYPRESS_RECORD_KEY --ci-build-id=$CYPRESS_CI_BUILD_ID --tag $CYPRESS_BROWSER"
    fi
    export MOZ_FORCE_DISABLE_E10S=1
    echo $CYPRESS_RUN_CMD
    $CYPRESS_RUN_CMD

    if [ "$?" = "0" ] ; then
        UI_TEST_SUCCESS="1"
    else
        UI_TEST_SUCCESS="0"
    fi

    if [ "$CYPRESS_RECORD" = "1" ] ; then
        echo "You can review tests on Cypress.io: https://dashboard.cypress.io/projects/$CYPRESS_PROJECT_ID/runs"
    fi
    
    echo ""
    echo "Killing OpenRefine"
    /bin/kill -9 $REFINE_PID
    echo "Killing Reconciliation Server"
    /bin/kill -9 $RECONCILE_SERVER_PID
    echo "Cleaning up"
    rm -rf "$REFINE_DATA_DIR"

    if [ "$UI_TEST_SUCCESS" = "0" ] ; then
        error "The UI test suite failed."
    fi
}


server_test() {
    mvn_prepare
    $MVN test -f main
}

extensions_test() {
    mvn_prepare
    $MVN test -f extensions
}

run() {
    FORK=$1
    
    check_running
    
    if [ "$RUNNING" ] ; then
        warn "OpenRefine is already running."
    fi
    
    if [ ! -d $REFINE_CLASSES_DIR ] ; then
        IS_JAR=`ls $REFINE_LIB_DIR | grep openrefine`
        if [ -z "$IS_JAR" ] ; then
            mvn_prepare
            $MVN process-resources
            $MVN compile test-compile
            echo ""
        fi
    fi

    if [ -d $REFINE_CLASSES_DIR ] ; then
        add_option "-Drefine.autoreload=true" "-Dbutterfly.autoreload=true"
    fi

    if [ "$OS" = "macosx" ] ; then
        add_option '-Xdock:icon=graphics/icon/openrefine.icns'
    fi
    
    if [ "$REFINE_DATA_DIR" ] ; then
        add_option "-Drefine.data_dir=$REFINE_DATA_DIR"
    fi

    if [ "$REFINE_WEBAPP" ] ; then
        add_option "-Drefine.webapp=$REFINE_WEBAPP"
    fi
                    
    if [ "$REFINE_PORT" ] ; then
        add_option "-Drefine.port=$REFINE_PORT"
    fi
    
    if [ "$REFINE_INTERFACE" ] ; then
        add_option "-Drefine.interface=$REFINE_INTERFACE"
    fi
    
    if [ "$REFINE_HOST" ] ; then
        add_option "-Drefine.host=$REFINE_HOST"
    fi

    if [ "$REFINE_AUTOSAVE_PERIOD" ] ; then
        add_option "-Drefine.autosave=$REFINE_AUTOSAVE_PERIOD"
    fi

    CLASSPATH="$REFINE_CLASSES_DIR${SEP}$REFINE_LIB_DIR/*"

    RUN_CMD=("$JAVA" -cp "$CLASSPATH" "${OPTS[@]}" "com.google.refine.Refine")

    echo ""

    if [ -z "$FORK" ] ; then
        exec "${RUN_CMD[@]}"
    else
        "${RUN_CMD[@]}" &
        REFINE_PID="$!"
    fi
}
    
broker_build() {
    build_prepare
    get_revision    
    # TODO migrate to Maven
    $MVN prepare_broker
}

broker_run() {
    FORK=$1

    if [ ! -d "broker/core/WEB-INF/lib" ] ; then
        broker_build
        echo ""
    fi

    if [ -d $REFINE_CLASSES_DIR ] ; then
        add_option "-Drefine.autoreload=true" "-Dbutterfly.autoreload=true"
        add_option "-Drefine.development=true"
    fi

    add_option "-Drefine.webapp=broker/core"
    add_option "-Drefine.headless=true"

    add_option "-Drefine.port=$REFINE_PORT"
    add_option "-Drefine.host=0.0.0.0"
                    
    LIB_PATHS=`cat server/classpath.txt`
    CLASSPATH="$REFINE_CLASSES_DIR${SEP}$LIB_PATHS"

    RUN_CMD="$JAVA -cp $CLASSPATH $OPTS com.google.refine.Refine"

    #echo "$RUN_CMD"
    #echo ""
  
    echo "Starting OpenRefine Broker at 'http://0.0.0.0:${REFINE_PORT}/'"
    echo ""

    if [ -z "$FORK" ] ; then
        exec $RUN_CMD
    else
        $RUN_CMD &
        REFINE_PID="$!"
    fi
    
}

broker_appengine_build() {
    appengine_prepare
    get_revision
    
    MVN_PARAMS="-Dappengine.sdk.dir=${APPENGINE_HOME}"

    if [ "$1" ] ; then
        MVN_PARAMS="$MVN_PARAMS -Dappengine.app_id=$1"
    fi
    
    if [ "$2" ] ; then
        MVN_PARAMS="$MVN_PARAMS -Dappengine.version=$2"
    fi
        
    # TODO migrate this to Maven
    $MVN prepare_broker_appengine
}
       
broker_appengine_upload() {
    broker_appengine_build $1 $2
    "$APPENGINE" update "$REFINE_BUILD_DIR/broker/appengine"
}

broker_appengine_run() {
    broker_appengine_build $1 $2
    "$APPENGINE_LOCAL" "$REFINE_BUILD_DIR/broker/appengine"
}
    
findbugs() {
    findbugs_prepare
    
    MVN_PARAMS="-Dfindbugs.dir=${REFINE_TOOLS_DIR}/${FINDBUGS_DIR}"
    ant findbugs
    
    display "$REFINE_BUILD_DIR/reports/findbugs.html"
}    

pmd() {
    pmd_prepare
    
    ANT_PARAMS="-Dpmd.dir=${REFINE_TOOLS_DIR}/${PMD_DIR}"
    ant pmd
    
    display "$REFINE_BUILD_DIR/reports/pmd.html"
}    

cpd() {
    pmd_prepare
    
    ANT_PARAMS="-Dpmd.dir=${REFINE_TOOLS_DIR}/${PMD_DIR}"
    ant cpd

    display "$REFINE_BUILD_DIR/reports/cpd.txt"
}    

jslint() {
    jslint_prepare

    ANT_PARAMS="-Djslint.dir=${REFINE_TOOLS_DIR}/${JSLINT_DIR}"
    ant jslint

    display "$REFINE_BUILD_DIR/reports/jslint.txt"
}

whitespace() {
    [ $# -gt 0 ] || usage
    
    for i in `find . -name *.$1`; do
        # expand tabs to spaces
        expand -t 4 < $i > $i.1
        
        # convert DOS to UNIX newlines
        tr -d '\r' < $i.1 > $i.2
            
        rm $i $i.1
        mv $i.2 $i
    done
}   
    
checkJavaMajorVersion() {
  java_ver=$("$JAVA" -version 2>&1 | grep version | cut -d ' ' -f 3 | tr -d \")
  # Java 6, 7, 8 starts with 1.x
  if [ ${java_ver:0:2} == "1." ] ; then
    major=`echo ${java_ver} | sed -E 's/1\.([0-9])[0-9_.]{2,6}/\1/g'`
  else
    # Java 9+ starts with x using semver versioning
    major=`echo ${java_ver} | sed -E 's/([0-9]+)(-ea|(\.[0-9]+)*)/\1/g'`
  fi
  if (( ${major} < 8 )); then
    error "OpenRefine requires Java version 8 or later. If you have multiple versions of Java installed, please set the environment variable JAVA_HOME to the correct version."
  fi
  if (( ${major} > 12 )); then
    echo "WARNING: OpenRefine is not tested and not recommended for use with Java versions greater than 12."
  fi
}

# -------------------------- script -----------------------------
    
# ----- Normalize the current directory -------------------------

cd `dirname $0`

# ----- Default values ------------------------------------------

OPTS=()

# ---- OS-specific support --------------------------------------

SYSTEM=`uname`

case "$SYSTEM" in
  CYGWIN*) OS="windows" ;;
  Darwin*) OS="macosx"  ;;
  Linux*)  OS="linux"   ;;
  *) OS="other" ;;
esac

SEP=":" 
if [ "$OS" = "windows" ] ; then
    SEP=";" 
fi


# ----- Make sure there is an appropriate java environment is available -------------

if [ "$OS" = "macosx" ] ; then
    if [ -z "$JAVA_HOME" ] ; then
    # We need want recent Java because we're bundling JRE - may want to warn and force developer to set JAVA_HOME
    # The /usr/libexec/java_home utility may be tied to the Java prefs app, so could go away when Apple removes it
        export JAVA_HOME=$(/usr/libexec/java_home)
    fi
fi

if [ "$JAVA_HOME" ] ; then
    JAVA="$JAVA_HOME/bin/java"
else 
    JAVA="`which java 2> /dev/null`"
fi

if [ ! -x "$JAVA" ] ; then
    error "Could not find the 'java' executable at '$JAVA', are you sure your JAVA_HOME environment variable is pointing to a proper java installation?"
fi

checkJavaMajorVersion

# ----- Parse the command line args ------------------------------------------

while [ $# -ne 0 ] ; do
  case "$1" in
	-h) usage;;
    -p) shift; REFINE_PORT="$1"; shift; continue;;
    -H) shift; REFINE_HOST="$1"; shift; continue;;
    -i) shift; REFINE_INTERFACE="$1"; shift; continue;;
    -w) shift; REFINE_WEBAPP="$1"; shift; continue;;
    -d) shift; REFINE_DATA_DIR="$1"; shift; continue;;
    -m) shift; REFINE_MEMORY="$1"; REFINE_MIN_MEMORY="$1"; shift; continue;;
    -k) shift; REFINE_GOOGLE_API_KEY="$1"; shift; continue;;
    -v) shift; REFINE_VERBOSITY="$1"; shift; continue;;
    -x) shift; REFINE_EXTRA_OPTS="$1"; shift; continue;;
    -c) shift; REFINE_INI_PATH="$1"; shift; continue;;
    --debug) shift; add_option '-Xdebug' '-Xrunjdwp:transport=dt_socket,address=8000,server=y,suspend=n'; continue;;
    --jmx) shift; add_option '-Dcom.sun.management.jmxremote'; continue;;
    -*) fail "Invalid option: $1";;
    *) break;;
  esac
done

if [ $# -ne 0 ] ; then
    ACTION=$1; shift
fi

if [ -z "$ACTION" ] ; then
    ACTION="run"
fi

# ----- Load configurations -------------------------------------

if [ -z $REFINE_INI_PATH ]; then
  REFINE_INI_PATH=refine.ini
fi

echo "Using ${REFINE_INI_PATH} for configuration"
load_configs $REFINE_INI_PATH

# ----- Verify and Set Required Environment Variables -------------------------

if [ -z "$JAVA_OPTIONS" ] ; then
  JAVA_OPTIONS=""
fi
add_option "${JAVA_OPTIONS[@]}"

if [ -z "$REFINE_MEMORY" ] ; then
    REFINE_MEMORY="1024M"
fi
if [ -z "$REFINE_MIN_MEMORY" ] ; then
    REFINE_MIN_MEMORY="256M"
fi
add_option "-Xms$REFINE_MIN_MEMORY" "-Xmx$REFINE_MEMORY" "-Drefine.memory=$REFINE_MEMORY"

freeRam=UNKNOWN
if [ "$OS" = "macosx" ] ; then
    freeRam=$(top -l 1 | grep PhysMem | awk '{print $6}' | tr -d M)
elif [ "$OS" = "linux" ] ; then
    freeRam=$(free -m | grep -oP '\d+' | head -n 1)
fi

echo "-------------------------------------------------------------------------------------------------"
echo You have "$freeRam"M of free memory. 
echo Your current configuration is set to use $REFINE_MEMORY of memory.
echo OpenRefine can run better when given more memory. Read our FAQ on how to allocate more memory here:
echo https://docs.openrefine.org/manual/installing\#increasing-memory-allocation
echo "-------------------------------------------------------------------------------------------------"

if [ -z "$REFINE_MAX_FORM_CONTENT_SIZE" ] ; then
    REFINE_MAX_FORM_CONTENT_SIZE="1048576"
fi
add_option "-Drefine.max_form_content_size=$REFINE_MAX_FORM_CONTENT_SIZE"

if [ -z "$REFINE_PORT" ] ; then
    REFINE_PORT="3333"
fi

if [ -z "$REFINE_INTERFACE" ] ; then
    REFINE_INTERFACE="127.0.0.1"
fi

if [ -z "$REFINE_HOST" ] ; then
    if [ "$REFINE_INTERFACE" = "0.0.0.0" ] ; then
        REFINE_HOST='*'
    else
        REFINE_HOST="$REFINE_INTERFACE"
    fi
fi

if [ "$REFINE_HOST" = '*' ] ; then
    echo No host specified while binding to interface 0.0.0.0, guessing localhost.
    REFINE_HOST_INTERNAL="localhost"
else
    REFINE_HOST_INTERNAL="$REFINE_HOST"
fi

if [ -z "$REFINE_WEBAPP" ] ; then
    REFINE_WEBAPP="main/webapp"
fi

if [ -z "$REFINE_TEST_DIR" ] ; then
    REFINE_TEST_DIR="main/tests"
fi

if [ -z "$REFINE_CLASSES_DIR" ] ; then
    REFINE_CLASSES_DIR="server/classes"
fi

if [ -z "$REFINE_LIB_DIR" ] ; then
    REFINE_LIB_DIR="server/target/lib"
fi

if [ -z "$REFINE_BUILD_DIR" ] ; then
    REFINE_BUILD_DIR="build"
fi

if [ -z "$REFINE_TOOLS_DIR" ] ; then
    REFINE_TOOLS_DIR="tools"
fi

if [ -z "$REFINE_DIST_DIR" ] ; then
    REFINE_DIST_DIR="dist"
fi

if [ -z "$REFINE_VERBOSITY" ] ; then
    REFINE_VERBOSITY="info"
fi
add_option "-Drefine.verbosity=$REFINE_VERBOSITY"

if [ ! -z "$REFINE_EXTRA_OPTS" ] ; then
    add_option "-D$REFINE_EXTRA_OPTS"
fi

if [ -z "$JYTHONPATH" ] ; then
    JYTHONPATH="$REFINE_WEBAPP/WEB-INF/lib/jython"
else
    JYTHONPATH="$REFINE_WEBAPP/WEB-INF/lib/jython${SEP}$JYTHONPATH"
fi

add_option "-Dpython.path=$JYTHONPATH"
add_option "-Dpython.cachedir=$HOME/.local/share/google/refine/cachedir"

if [ ! -z "$GDATA_CLIENT_ID" ] ; then
    if [ ! -z "$GDATA_CLIENT_SECRET" ] ; then
        if [ ! -z "$GDATA_API_KEY" ] ; then
            add_option "-Dext.gdata.clientid=$GDATA_CLIENT_ID" "-Dext.gdata.clientsecret=$GDATA_CLIENT_SECRET" "-Dext.gdata.apikey=$GDATA_API_KEY"
        fi
    fi
fi

# ----- Respond to the action given --------------------------------------------
case "$ACTION" in
  build) build_prepare; mvn process-resources; mvn compile\ test-compile;;
  clean) mvn clean;;
  whitespace) whitespace $1;;
  distclean) mvn distclean;;
  test) test $1;;  
  tests) test $1;;  
  ui_tests) ui_tests;;    
  server_test) server_test $1;;  
  server_tests) server_test $1;;  
  extensions_test) extensions_test $1;;
  extensions_tests) extensions_test $1;;
  findbugs) findbugs;;  
  pmd) pmd;;  
  cpd) cpd;;  
  jslint) jslint;;  
  run) run;;  
  broker) broker_run;;
  broker_appengine_run) broker_appengine_run $1 $2;;
  broker_appengine_upload) broker_appengine_upload $1 $2;;
  mac_dist) mac_dist $1;;
  windows_dist) windows_dist $1;;
  linux_dist) linux_dist $1;;
  dist) dist $1;;
  *) usage; ;;
esac

# ----------- end of file --------------------