extract_words.sh

#!/bin/bash
# This script prints each word in $dict_file, one per line.
# $dict_file is a file from http://www.csse.monash.edu.au/~jwb/edict_doc.html.


# See http://www.csse.monash.edu.au/~jwb/jmdict_dtd_h.html to understand why
# <keb> and </keb> are used.
start_marker="<keb>"
end_marker="</keb>"
dict_file=JMdict_e


# This grep-based method will break if any of the "keb" elements in $dict_file
# spans multiple lines.
for line in $(grep $start_marker $dict_file); do
  word=${line#$start_marker}
  word=${word%$end_marker}
  echo $word
done