d2t

#!/bin/bash

usage() {
    echo ""
    echo "docx2tex"
    echo ""
    echo "Usage: d2t [options ...] <docx file>"
    echo ""
    echo "Options:"
    echo "   -o    path to custom output directory"
    echo "   -c    path to custom word2tex configuration file"
    echo "   -p    generate PDF with pdflatex"
    echo "   -m    choose MathType source (ole|wmf|ole+wmf)"
    echo "   -r    omit page and cross-references and labels"
    echo "   -i    select lists detection mode (indent|role|none)"
    echo "   -t    choose table model (tabularx|tabular|htmltabs)"
    echo "   -l    suppress table grid lines"
    echo "   -x    custom XSLT stylesheet for Hub processing (after evolve-hub)"
    echo "   -e    custom XSLT stylesheet for evolve-hub (should import xsl/evolve-hub-driver.xsl)"
    echo "   -d    debug mode"
    echo "   -h    heap (main memory) that Java is allowed to use (for ex., 4096m)"
    1>&2; exit 1;
}
# print out error step and exit code
function exitonerror {
    echo "Errors encountered while running $2. Please see $LOG for details."
    exit 1
}
# readlink -f is unavailable on Mac OS X
function real_dir() {
    SOURCE="$1"
    while [ -h "$SOURCE" ]; do # resolve $SOURCE until the file is no longer a symlink
	DIR="$( cd -P "$( dirname "$SOURCE" )" && pwd )"
	SOURCE="$(readlink "$SOURCE")"
	[[ $SOURCE != /* ]] && SOURCE="$DIR/$SOURCE" # if $SOURCE was a relative symlink, we need to resolve it relative to the path where the symlink file was located
    done
    echo "$( cd -P "$( dirname "$SOURCE" )" && pwd  )"
}

# cygwin check
cygwin=false;
case "`uname`" in
    CYGWIN*) cygwin=true;
esac

# script directory
PWD="$( pwd)"
DIR="$( real_dir "${BASH_SOURCE[0]}" )"
CALABASH=$DIR/calabash/calabash.sh

# invoke pdflatex
PDFLATEX=pdflatex
MTEFSOURCE="ole"
TABLEMODEL=""
TABLEGRID=""
LISTMODE=""
PDF=no
REFS=yes
DEBUG=no

# specify options
while getopts ":o:c:m:t:i:x:e:plrdh:" opt; do
    case "${opt}" in
	o)
	    OUT_DIR=${OPTARG}
	    ;;
	c)
	    CONF=${OPTARG}
	    ;;
  m)
      MTEFSOURCE=${OPTARG}
      ;;
	x)
	    CUSTOMXSL=${OPTARG}
	    ;;
	e)
	    CUSTOMDRIVER=${OPTARG}
	    ;;
	t)
	    TABLEMODEL=${OPTARG}
	    ;;
  i)
	    LISTMODE=${OPTARG}
	    ;;
	p)
	    PDF=yes
	    ;;
	l)
	    TABLEGRID=no
	    ;;
	r)
	    REFS=no
	    ;;
	d)
	    DEBUG=yes
	    ;;
	h)
	    HEAP=${OPTARG}
	    ;;
	\?)
	    echo "Invalid option: -$OPTARG" >&2
	    usage
	    ;;
	:)
	    echo "Option -$OPTARG requires an argument." >&2
	    usage
	    ;;
    esac
done
shift $((OPTIND-1))
# check if argument for file is set
if [ -z "$1" ]; then
    usage
fi

# custom evolve-hub driver given (as URI, may also be relative to this directory)
if ! [ -z "$CUSTOMDRIVER" ]; then
    INPUTCUSTOMDRIVER="-i custom-evolve-hub-driver=$CUSTOMDRIVER"
fi

# file as 1st argument
FILE="$1"
if ! [[ $FILE =~ \.docx$ ]]; then
    echo "Only docx files are supported."
    exit 1
fi

# set basename
BASENAME="$(basename "$(basename "$FILE" .docx)" .docm)"
BASENAME="${BASENAME// /_}"

# output directory path
if [[ -z "$OUT_DIR" ]]; then
    OUT_DIR="$(real_dir "$FILE")"
else
    if [[ "$OUT_DIR" != /* ]]; then	
	OUT_DIR="$PWD/$OUT_DIR"
    fi
fi

mkdir -p "$OUT_DIR"

# path to configuration file
if [[ -z "$CONF" ]]; then
    CONF="$DIR/conf/conf.csv"
else
    if [[ "$CONF" != /* ]]; then	
	CONF="$PWD/$CONF"
    fi
fi

# logging
LOG="$OUT_DIR/$BASENAME.d2t.log"

# remove log from previous runs
if [ -e "$LOG" ]; then
    rm "$LOG"
fi

# Java max. amount of main memory
if [ -z $HEAP ]; then
    HEAP=1400m
fi


# debugging
DEBUG_DIR="$OUT_DIR/$BASENAME.debug"

# pdflatex outdir path
PDFLATEX_OUT="$OUT_DIR"

# make absolute paths
if $cygwin; then
    FILE="$(cygpath -ma "$FILE")"
    OUT_DIR="$(cygpath -ma "$OUT_DIR")"
    DIR="$(cygpath -ma "$DIR")"
    CONF="$(cygpath -ma "$CONF")"
    DEBUG_DIR_URI=file:/$(cygpath -ma "$DEBUG_DIR" )
else
    DEBUG_DIR_URI="file:$DEBUG_DIR"
fi

DEBUG_DIR_URI="${DEBUG_DIR_URI// /%20}"
OUT_DIR_PSEUDOURI="${OUT_DIR// /%20}"

FILE_CP="$OUT_DIR/$BASENAME.docx"
cp "$FILE" "$FILE_CP" 2>&1 2>>"$LOG"

# check if file exists
if [ ! -f "$FILE" ]; then
    echo "Error: input file not found: $FILE"
    usage
fi

echo "starting docx2tex"

if [ "$DEBUG" = "yes" ]; then
    echo "debug mode: $DEBUG"
    echo "storing debug files to $DEBUG_DIR"
    echo ""
    echo "Parameters"
    echo "  workdir: $DIR"
    echo "  outdir: $OUT_DIR"
    echo "  file: $FILE"
    echo "  config: $CONF"
    echo "  MathType source: $MTEFSOURCE"
    echo "  refs and labels: $REFS"
    echo "  table model: $TABLEMODEL"
    echo "  draw table grid lines: $TABLEGRID"
    echo "  custom evolve-hub: $INPUTCUSTOMDRIVER"
    echo "  custom xslt: $CUSTOMXSL"
    echo "  Java heap: $HEAP"
    echo ""
fi

# docx2tex xproc pipeline
HEAP=$HEAP $CALABASH \
    $INPUTCUSTOMDRIVER \
    -o result="$OUT_DIR_PSEUDOURI/$BASENAME.tex" \
    -o hub="$OUT_DIR_PSEUDOURI/$BASENAME.xml" \
    "$DIR/xpl/docx2tex.xpl" \
    docx="$FILE_CP" \
    conf="$CONF" \
    custom-xsl="$CUSTOMXSL" \
    conf-template="$OUT_DIR/$BASENAME.csv" \
    mtef-source="$MTEFSOURCE" \
    table-model=$TABLEMODEL \
    table-grid=$TABLEGRID \
    list-mode=$LISTMODE \
    refs=$REFS \
    debug=$DEBUG \
    debug-dir-uri=$DEBUG_DIR_URI \
    status-dir-uri=$DEBUG_DIR_URI/status \
    2>&1 2>>"$LOG" || exitonerror $? docx2tex
echo "writing texfile => $OUT_DIR/$BASENAME.tex"

# run pdflatex with "-p"
if [ "$PDF" = "yes" ]; then
    cd "$OUT_DIR" && "$PDFLATEX" \
    -interaction=nonstopmode \
    -output-directory="$OUT_DIR" \
    "$PDFLATEX_OUT/$BASENAME.tex" \
    1>>"$LOG" || exitonerror $? pdflatex
    echo "writing pdf => $OUT_DIR/$BASENAME.pdf"
fi
echo ""
echo "docx2tex finished."