#!/bin/bash
#
# Hanlinize EPUB: make Hanlin and other e-readers display epub files correctly
# 
# Problem:    the font for displaying EPUB files doesn't have unicode characters and
#             even though the reader supports changing fonts, not for the epub format;
#             this is fixed in newer firmware, but Czech Hanlin reseller didn't update
#             the localized firmware to either tackle this problem (despite knowing
#             about it for years) or to at least include the option to select custom 
#             fonts for epub files; shame on you, guys!
#
# Solution:   embed unicode-aware fonts into the epub - Hanlin supports embedded fonts;
#             beware that this significantly increases the epub size - there is also 
#             the possibility to load external fonts from within the epub file, but not
#             using this tool
#
# Usage:      ./hanlinize file1.epub file2.epub file3.epub
#
# Fonts:      set fonts and their locations by changing variables FONTS and FONTLOC
#             note that for the FONTS variable, you need to include 4 fonts (filenames)
#             in the right order - regular, bold, italic and bold-italic
#
# Warning:    ugly code warning - the tool is very, very far from being perfect, it
#             uses bad approach and isn't guaranteed to work on all epubs - chances
#             are it may do weird stuff - use it at your own risk!
#             
# License:    DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE 
#             Version 2, December 2004 
#
#             Copyright (C) 2011 Michal Docekal <docekal@gmail.com> 
#
#             Everyone is permitted to copy and distribute verbatim or modified 
#             copies of this license document, and changing it is allowed as long 
#             as the name is changed. 
#
#             DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE 
#             TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION 
#
#             0. You just DO WHAT THE FUCK YOU WANT TO.
#
#

# SET THESE VARIABLES TO WHATEVER SUITS YOU

FONTS="DejaVuSerifCondensed.ttf DejaVuSerifCondensed-Bold.ttf DejaVuSerifCondensed-Italic.ttf DejaVuSerifCondensed-BoldItalic.ttf"
FONTLOC="/usr/share/fonts/TTF"

# THE ACTUAL CODE IS BELOW, DON'T MESS WITH IT IF YOU DON'T KNOW WHAT YOU'RE DOING

if [ "$#" == "0" ]; then
   echo 'No parameter entered. Expecting one or more epub files.'
   exit 0
fi

function error {
   echo
   echo "Fatal error: $@"
   echo "Terminating."
   exit 1
}

function gen-css {
   echo """@font-face {
font-family : HanlinEmbedFix;
font-weight : normal;
font-style: normal;
src : url(../Fonts/"""$(echo $FONTS | awk '{print $1}')""");
}

@font-face {
font-family : HanlinEmbedFix;
font-weight : bold;
font-style: normal;
src : url(../Fonts/"""$(echo $FONTS | awk '{print $2}')""");
}
 
@font-face {
font-family : HanlinEmbedFix;
font-weight : normal;
font-style: italic;
src : url(../Fonts/"""$(echo $FONTS | awk '{print $3}')""");
}

@font-face {
font-family : HanlinEmbedFix;
font-weight : bold;
font-style: italic;
src : url(../Fonts/"""$(echo $FONTS | awk '{print $4}')""");
}

body {
font-family: "HanlinEmbedFix", serif !important;
}
 
strong, b {
font-family: "HanlinEmbedFix", serif !important;
}
 
em, i {
font-family: "HanlinEmbedFix", serif !important;
}"""
}

ORIGIN=$(pwd)

while [ "$#" != "0" ]; do
   TMPDIR=$(mktemp -d)
   OUTPUT="$(basename $1 .epub)-hanlinized.epub"
   BACKUP="$(basename $1)~"

   echo -n "Converting file $1 to $OUTPUT: "

   unzip "$1" -d "$TMPDIR/" 2>/dev/null >/dev/null          || error "unzipping epub failed"
   cd "$TMPDIR/"                                            || error "changing directory to \"$TMPDIR\""
   mkdir -p "$TMPDIR/OEBPS/Styles"                          || error "creating directory \"$TMPDIR/OEBPS/Styles\""
   mkdir -p "$TMPDIR/OEBPS/Fonts"                           || error "creating directory \"$TMPDIR/OEBPS/Fonts\""
   for font in $FONTS; do
      cp "$FONTLOC/$font" "$TMPDIR/OEBPS/Fonts/"            || error "copying /usr/share/fonts/TTF/$font to $TMPDIR/OEBPS/Fonts"
   done
   gen-css > "$TMPDIR/OEBPS/Styles/hanlin-embed-fonts-fix.css"

   # add font files to manifest
   # sorry for the messy regexps, sed probably has a way to do case insensitive search, but I'm too lazy to keep trying finding it
   sed -i '/<\/[mM][aA][nN][iI][fF][eE][sS][tT]\>/i '"<item href=\"Styles/hanlin-embed-fonts-fix.css\" id=\"hanlin-embed-fonts-fix.css\" media-type=\"text/css\" />" "$TMPDIR/OEBPS/content.opf"
   for font in $FONTS; do
      sed -i '/<\/[mM][aA][nN][iI][fF][eE][sS][tT]\>/i '"<item href=\"Fonts/$font\" id=\"$font\" media-type=\"application/x-font-ttf\" />" "$TMPDIR/OEBPS/content.opf"
   done

   # apply styles to content
   ls -1 "$TMPDIR/OEBPS/Text/" | while read line; do
      sed -i '/<[hH][eE][aA][dD]>/i '"<link rel=\"stylesheet\" type=\"text/css\" href=\"../Styles/hanlin-embed-fonts-fix.css\"/>" "$TMPDIR/OEBPS/Text/$line"
   done

   zip -Xr "$ORIGIN/$OUTPUT" mimetype META-INF OEBPS 2>/dev/null >/dev/null || error "packing output file \"$ORIGIN/$OUTPUT\""
   rm -rf "$TMPDIR"                                       || error "erasing \"$TMPDIR\""
   cd "$ORIGIN"                                           || error "changing directory to \"$ORIGIN\""
   # switcheroo
   mv -n "$1" "$BACKUP"                                   || error "error moving \"$1\" to \"$BACKUP\""
   mv -n "$OUTPUT" "$1"                                   || error "error moving \"$OUTPUT\" to \"$1\""
   echo "Done."
   shift
done