Tool to fix epub encoding for use with devices like the Amazon Kindle which do not handle high UTF characters and require HTML Entities

fixepubencoding.sh

#!/bin/sh
#

IFS='|'
base=${1%%.*}
ext=${1#*.}

#if [[ $# -eq 0 ]] ; then
if [ "$ext" != "epub" ]; then
    echo "Usage: $0 [EPUB FILE]"
    echo "\nNo source epub file specified. Please specify a valid epub document."
    echo "\nThis tool will fix HTML entities in badly formatted epub documents so they can be"
    echo "used on devices that do not allow non-HTML entities for high ASCII characters."
    exit 255;
fi

if type he >/dev/null 2>&1; then

  # Be safe
  tmp=$(mktemp -d epubfix.XXXXXX)
  pwd=$(pwd) # Store original working directory to avoid ".." nonsense

  echo "=== Processing $base ==="
  mkdir -p $tmp/orig $tmp/work || true

  echo "=== Creating Working Copies ==="
  unzip -aoq "$base.$ext" -d "$tmp/orig/$base/"
  unzip -aoq "$base.$ext" -d "$tmp/work/$base/"
  echo "=== Processing HTML Encoding ==="
  printf "\tFixing HTML encoding..."
  for b in $(find $tmp/orig/ -type f -regextype posix-egrep -regex ".*?html" -printf "%P|")
  do
   cat "$tmp/orig/$b" | he --encode --use-named-refs --allow-unsafe > "$tmp/work/$b"
   #printf "$b\t"
   printf "."
   # \diff -U 0 "orig/$b" "work/$b" | grep ^@ | wc -l  # Show number of changed lines, 
  done
  printf " Done!\n"
  echo "=== Creating new epub book ==="
  cd "$tmp/work/$base"
  epub="$pwd/$base-html5.epub"
  zip -X0q "$epub" mimetype
  zip -rDX9q "$epub" * -x "*.DS_Store" -x mimetype
  echo "\t$base-html5.epub written."
  echo "=== Cleaning up workspace ==="
  #echo "Removing $tmp"
  cd $pwd
  rm -rf $tmp
  echo "=== Done ==="

else

  echo "Missing he -- (https://github.com/mathiasbynens/he)"
  echo "Please install he by running"
  echo "\n\tnpm install -g he"

fi