#!/bin/sh
# Giulia Hill Dec. 05 ghill@library.berkeley.edu
# Jan. 06 - Added line count, error reporting, and optional url param.  Janet Garey
# Jan. 06 - Added parsing of params and option to use directory or file for input
#           and functions for modularity. ghill@library


#----------- Functions ---------------------------------
useFile() {
   #open and read the input file
    FILE=$1
    OPTIMO=$2

    if [ "$OPTIMO" != "" ]; then
        ALL=`cat $FILE`
       	ERROR=`java xni.XMLGrammarBuilder -a $URL -i $ALL`
    else 
        exec <$FILE
	while read LINE
	do
	    #read ANSWER </dev/tty  #pause, for testing only
	    LINESREAD=`expr $LINESREAD + 1`
	    echo "processing line #" $LINESREAD ":  " $LINE
	    ERROR=`java xni.XMLGrammarBuilder -a $URL -i $LINE`
	    if [ $ERROR ]; then
		echo "Error code " $ERROR "on line " $LINESREAD 
	    fi
	done
	echo
	echo "Finished."
	echo $LINESREAD "lines processed from file:" $FILE
    fi
    exit
}
###################################
traverse() 
{
  if [ "$OPTIMO" != "" ]; then
    traverseOptimized $1
    exit
  fi

  # Traverse a directory
  cd "$1"

  ls | while read i
  do
    if [ -d "$i" ]; then
      echo "Directory: $i"
      # Calling this as a subshell means that when the called
      # function changes directory, it will not affect our
      # current working directory
      (traverse "$i")
    else 
      validate $i
    fi
  done
}

####################################
validate()
{
    VAL=$1;
    echo "processing file $VAL"
    RET=`java xni.XMLGrammarBuilder -a $URL -i $VAL`
    if [ $RET ]; then
       echo "Error code " $RET
    fi
}
###########################################
traverseOptimized() 
{
  # Traverse a directory
  cd "$1"

ALL=
# Ok why is all the following necessary?... because | i.e. pipe forces
# loops to open a subshell, see solution at:
# http://wooledge.org/mywiki/BashFaq#head-eebf55026903ead4aa7de24ac2ac7e5d80075019
# it drove me nuts..
  ls | (
  while read i
  do
    if [ -f "$i" ]; then
      ALL="$ALL $i"
      export ALL
    fi
  done 
 )
    # I scratched all of that
    java xni.XMLGrammarBuilder -a $URL -i `/bin/ls $1`

}

######################################
helpMenu() {
  echo ""
  echo "usage: batchXerces <option> -s [schema url] -l [log file]"
  echo "       <option>: -f <fileNameOfBatch>"
  echo "                 -d <AbsolutePathdirectoryOfFiles> "
  echo "                    recurses on subdirectories only when not optimized"
  echo "                 -i <nameOfIndividualFile> to validate a single file"
  echo "                 -h this screen"
  echo "                 -l a log file. stdout & stderr default if none specified"
  echo "                 -o optimize run. Caches .xsd but doesn't state progress of run"
  echo "                    it is meaningless when used with -i"
  echo "                 the 'f', 'd', 'i' options are mutually exclusive"
  echo "                 if more than one is used, priority is f-d-i"
  echo "       Lists of schema urls can be entered within quotes. If no url is entered"
  echo "       the defaults are mets.xsd, mods-3-1.xsd, metsrights.xsd, mix.xsd "
  echo ""
  exit;
}
#-------------------- Main body --------------------------------
LOG=
DIREC=
URL="http://www.loc.gov/standards/mets/mets.xsd http://www.loc.gov/standards/mods/v3/mods-3-1.xsd http://cosimo.stanford.edu/sdr/metsrights.xsd http://www.loc.gov/standards/mix/mix.xsd"
FILE=
SINGLE=
OPTIMO= 

if [ "$1" = "" ] || [ "$1" = "-h" ]; then
    helpMenu
fi

while [ $# -gt 0 ]; do
  case "$1" in 
   -f ) FILE=$2 ; shift 2 ;;
   -d ) DIREC=$2 ; shift 2 ;;
   -s ) URL=$2 ; shift 2 ;;
   -l ) LOG=$2 ; shift 2 ;;
   -i ) SINGLE=$2 ; shift 2 ;;
   -o ) OPTIMO="true" ; shift 1 ;;
   * ) echo "Unknow flag $1 " ; helpMenu ;;
  esac
done

# if there is a log file, redirect stdout and stderr in it
if [ "$LOG" != "" ]; then
    exec >$LOG 2>&1
fi

# if both FILE and DIREC and SINGLE are empty there is a problem and bail out
if [ "$FILE" = "" ] && [ "$DIREC" = "" ] && [ "$SINGLE" = "" ]; then
    echo "No batch file nor directory entered; helpMenu"
fi

JAVA_HOME="/usr/local/jdk1.5.0"
LINESREAD=0
ERROR=0

PATH="/home/ghill/Xerces:.:/usr/local/jdk1.5.0/bin:/usr/bin"
export PATH
#unset classpath
CLASSPATH=""
export CLASSPATH

# set all the basic stuff like java libraries 
CLASSPATH="$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar:$JAVA_HOME/lib/htmlconverter.java:$JAVA_HOME/lib/jconsole.jar:$JAVA_HOME/lib/sa-jdi.jar:/home/ghill/java:/usr/local/java_libs/xerces.2.7.1/resolver.jar:/usr/local/java_libs/xerces.2.7.1/xercesImpl.jar:/usr/local/java_libs/xerces.2.7.1/xercesSamples.jar:/usr/local/java_libs/xerces.2.7.1/xml-apis.jar:."
export CLASSPATH

if [ "$FILE" != "" ]; then
    useFile $FILE $OPTIMO
    exit
fi

if [ "$DIREC" != "" ]; then
    traverse $DIREC $OPTIMO
    exit
fi

if [ "$SINGLE" != "" ]; then
    validate $SINGLE ""
    exit
fi