convert fruux tasks ICS into plain text

fruux-tasks-ics-to-plaintext.sh

#!/bin/sh -x

###########
# Problem this solves:
#
#   I use Fruux (fruux.com) as the backend to my iOS Reminders app.  Rather than for actual
#   reminders, I used it for general notes, roughly one Reminders entry to a paragraph of notes.
#   Later I decided to switch to using the actual iOS Notes app, with my email provider
#   (fastmail.fm) as the backend.
#
#   But how to transfer my existing data?  Fruux lets me download each set of Reminders
#   (they call it Tasks) as an .ics file.  This little shell script takes this .ics file
#   as its input and outputs my content as plain text.  The only remaining problem is that
#   the content is out of order.  I suspect it uses timestamps in the file to know how to
#   order it on the device.  So I convert to plaintext using this script, then reorder
#   manually in a text editor by looking at the original content on Fruux web interface,
#   then finally paste the plaintext into my new notes backend.
#
# How to use:
#
# 1. Download the .ics file from Fruux
# 2. cat {ics-file} | ./this-script.sh > plaintext.txt
#
###########

###########
# Each command in the pipeline documented, since I cannot put inline comments:
#
# 1. `awk`: Text content we care about is only in the SUMMARY field, so take only lines
#    between SUMMARY: and the next field, UID:.
# 2. `grep -v`: But we don't actually want the UID field, so exclude those lines.
# 3. `sed`: Replace SUMMARY prefix with a paragraph linebreak. Now each entry of the
#    original file is its own paragraph, separated by an empty line, with the first line of
#    each paragraph starting at the first column and all subsequent lines starting with a
#    single space.
# 4. `sed`: Join all lines within each paragraph.  Now each paragraph is on a single line.
# 5. `sed`: Commas and semicolons are backslash-escaped, so remove the escaping.
# 
###########

awk '/SUMMARY:/, /UID:/ { print $0 }' \
  | grep -v UID \
  | sed "s/SUMMARY:/\\"$'\r'"\\"$'\n'"/" \
  | sed -E -e ':a' -e '$!N' -e "s/"$'\r'$'\\n'" //" -e 't a' -e 'P' -e 'D' \
  | sed 's/\\\([;,]\)/\1/g'