docx-change-comments

#!/bin/bash
set -euo pipefail #bash strict mode
if [ "$#" -ne 2 ] || [ ! -f "$1" ] || [ -f "$2" ] ; then
  echo USAGE: "$0 file_name sed_script" && echo && echo "IMPLEMENTATION:" && cat "$0" && echo && exit 22 #usage message for invalid number of arguments #also, in this case, for wrong arguments / mixed up order of arguments
fi
# $1: the file to operate on. A new file, "changed.$1", will be created by this process.
# $2: the sed script. Something like "s/word1/word2/g; s/replacemealso/secondreplacementvalue/g" See https://www.gnu.org/software/sed/manual/html_node/sed-script-overview.html https://www.gnu.org/software/sed/manual/html_node/Multiple-commands-syntax.html and https://www.gnu.org/software/sed/manual/html_node/The-_0022s_0022-Command.html

tmpdir=tmp-docx-unzipped-"$1"
unzip "$1" -d "$tmpdir"
cd "$tmpdir"/
sed --in-place -e "$2" word/comments.xml # oddly, rm word/comments.xml cripples the document so it won't display all of its pages. For some reason. So... don't do that.
zip -r ../changed."$1" *
cd ../
rm -r "$tmpdir"

: <<'HEREDOC'
  OK, the shell script is over now. Here is a guide on how to do it manually, in case that knowledge is useful:

  This guide is me explaining how to convert comments in a docx to have proper author names attached instead of system user names (which are not very useful). It does, however, assume each system user has a unique user name. The process is probably similar for .odt. Also scripting this way means you can change any stuff about the comments.

  This whole process is not very useful, because libreoffice is also slow at editing large documents like this :/, so there's no point having a local copy instead of just a google doc . Open Office is way better, though. Though also looks weird 🤔 not sure if it renders them right.

  1. rename file to end with .zip
  2. unzip it (you can't just open it with file explorer, because you're gonna have to edit a file)
  3. Navigate to word/comments.xml
  4. You can edit that file. Run a script on it, whatever. Or just delete it I guess.
  5. Zip up again. (Do NOT commit the common mistake of making a zip file that contains a single folder with all the contents inside that. The contents must be directly under root, like in the original docx.)

  And that's it!
HEREDOC