#!/bin/bash

# grickr

# Todd Slater <dontodd@gmail.com>
# Revised 30 May 2007

# Finds flickr images in a google reader feed and downloads the original image

# Requires: grep, curl, wget

# set up variables

# where to store all pics and working files
baseDir="$HOME/flickr"

# cookie file - the 7dayzq8m below is random, yours will differ
cookies="$HOME/.mozilla/firefox/7dayzq8m.default/cookies.txt"

# feed file
feedfile=${baseDir}/starred

# feed url (you could probably substitute any feed, though not tested)
feedurl="http://www.google.com/reader/atom/user/your_user_id/state/com.google/starred"

# no need to edit beyond this point

# store all pic url's
masterlist=${baseDir}/master # all pictures that exist, all sessions
goodlist=${baseDir}/goodlist # pictures that exist, current session
badlist=${baseDir}/badlist # pictures that are unavailable, all sessions

cd $baseDir

# fetch feed

curl -b "$cookies" -o "$feedfile" "$feedurl"

# find all flickr pics and rewrite url to get original size
# then, test if pic is available and write to appropriate list

for file in `grep -o -E "http://farm1.static.flickr.com/[0-9]{2,}/[[:alnum:]]{9}_[[:alnum:]]{10}_[mo].jpg" "$feedfile"|sed s/_m/_o/g`
do
	curl -I "$file" | grep "HTTP/1.0 200 OK" && echo "$file" >> $goodlist || echo "$file" >> $badlist.tmp
done

# for unavailable pics, let's try to find a big version
# and write to appropriate list, but not if we've already
# downloaded it due to multiple links/enclosures in the feed

while read line
do
	bigVersion=`echo "$line" | sed s/_o.jpg/_b.jpg/`
	dupepart=`echo "$line" | cut -b 36-44`
	if [ ! `grep "$dupepart" $goodlist` ] ; then
		curl -I "$bigVersion" | grep "HTTP/1.0 200 OK" && echo "$bigVersion" >> $goodlist
	fi
done<$badlist.tmp

# ok, got list of available pics, let's fetch them
# using wget for its noclobber goodness

wget -nc -i $goodlist

# combine recent list and master list, check for dupes

cat $goodlist >> $masterlist
sort $masterlist | uniq > $masterlist.$$
mv $masterlist.$$ $masterlist

# combine recent bad with master bad list, check for dupes

cat $badlist.tmp >> $badlist
sort $badlist | uniq > $badlist.$$
mv $badlist.$$ $badlist

# clean up temp files

rm -f $feedfile
rm -f $goodlist
rm -f $badlist.tmp