Hyperlink ODS text (LibreOffice Calc)

ods-hyperlink.sh

#!/usr/bin/env bash
# Add hyperlinks to a LibreOffice Calc file
# We appreciate https://ask.libreoffice.org/en/question/81492/convert-clickable-hyperlink-to-viewable-url/
# but don't want to handwrite the action every time nor trust a macro (though I guess a bash script isn't much better)
# Intended to be used on `.ods` with no hyperlinks at all, would need XML parser otherwise
# More recent version might exist in https://github.com/twolfson/dotfiles but no promises

# Exit on first error, unset variable, or pipe failure
set -euo pipefail

# Retrieve our parameters
set +u
ods_filepath="$1"
set -u
if test "$ods_filepath" = ""; then
  echo "Usage: $0 <ods-filepath>" 1>&2
  exit 1
fi

# Create a temporary folder to work in and set up cleanup hooks
tmp_folder="$(mktemp --directory)"
cleanup () {
  rm -r "$tmp_folder"
}
trap cleanup EXIT

# Extract content from our `.ods` as a `.zip`
unzip "$ods_filepath" "content.xml" -d "$tmp_folder" 1> /dev/null
# <text:p>Non-link text <text:a xlink:href="https://google.com/" xlink:type="simple">https://google.com/</text:a></text:p>
# DEV: Regexp to match URLs without leading/trailing quotes or angel brackets (as likely in HTML attribute/content)
# DEV: Use list to know when to terminate URL
#   https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/encodeURIComponent
#   Except removing ",;()" since they could be in sentences but `.` is critical obviously for `.com`
sed --in-place -E 's/([^"])(https?:\/\/[-A-Za-z0-9_!~*;.\/?:@&=+$#]+)([^"])/\1<text:a xlink:href="\2" xlink:type="simple">\2<\/text:a>\3/g' "$tmp_folder/content.xml"

# Add our `.xml` back into our `.ods` (as a zip)
# DEV: `-u` is `update`
old_pwd="$PWD"
cd "$tmp_folder"
zip --verbose "$old_pwd/$ods_filepath" "content.xml" 1> /dev/null
cd - 1> /dev/null