#!/bin/bash
# By Linc 10/1/2004
# Find the latest script at http://linc.homeunix.org:8080/scripts/bashpodder
# Last revision 07/01/2005 - Many Contributers!
# If you use this and have made improvements or have comments
# drop me an email at linc dot fessenden at gmail dot com
# I'd appreciate it!
# Amended by Nik Butler 2005 10 12
# Dropped creation of date in favour of downloading podcasts by title.
# using wgets -N timestamp and -c Continue feature to grab the files.
# wget knows what the filename is and if we have seen it or not
# of course the above screws up if the file names change !
# also we accept 1 argument ... the count of the shows to download from the feed.
# this stops the files going MAD trying to grab everything from the feed.
# ive split the xsl files up to spot the title in the feed.
#
# 2005 11 20 - Split bp.conf into everyday and weekly downloads
#
# 2005 12 10 - test the result of file to see if its long enough to be a http:// line.
#              it does get it more than neccesay ... must fix this !
#              with lincs help ive added a method for reading feeburner/itunes style css functions


# Make script crontab friendly:
cd $(dirname $0)
rm current_shows temp_shows
# Make a List of Shows we currently Have available.
du -a | grep -i mp3 | awk '{FS=" "; print $2}' >temp_shows

TODAY=`date +%A`.bp.conf
if [ -e $TODAY ]
then
       FILES="$TODAY bp.conf"
       else
       FILES="bp.conf "
fi

# This will read bp.conf every day and a file named WEEKDAY.bp.conf
cat $FILES |
while read podcast
       do
       echo "$podcast"
       # So First we grab the title of the shows we are grabbing.
       title=$(wget -q $podcast -O - | xsltproc parse_enclosure_title.xsl - 2>/dev/null)
       # drop spaces and annoying characters from the name.
       # i could do this all on one line but im lazy !
       title=${title//\ /_}
       title=${title//\'/}
       title=${title//\\/}
       title=${title//\//}
       # then create the directory.
       if ! test  -d "$title"
               then
               mkdir $title
       fi
       # grab the links to the files download from the Rss Feed.
       ## test the return length i guess it needs to be longer than http://
       file=$(wget -q $podcast -O - | xsltproc parse_enclosure_url.xsl - 2>/dev/null )

       if [ ${#file} -lt 8 ]
       then
               file=$(wget -q $podcast -O - | tr '\r' '\n' | tr \' \" | sed -n 's/.*url="\([^"]*\)".*/\1/p')
       fi

       if [ ${#file} -lt 8 ]
       then
               file=$(wget -q $podcast -O - | xsltproc parse_retarded_feed.xsl - 2>/dev/null )
       fi
       # count the files we are downloading.
       fileCount=0
       for url in $file
       do

               echo $url
               # We need to make sure that the Output file looks like filename.mp3 since
               # this makes life easier for other applications looking for them.
               # note a lot of feeds are getting wierd redirect strings and renames .. so watch out for wierd
               # urls turning up from time to time .
               outfile=$(echo $url | awk 'BEGIN { FS = "/" } ; { print $(NF) }')
               outfile=${outfile//\?[a-z]*/}
               if [ -n "$1" ]
                       then
                       if [ $fileCount -eq $1  ]
                       then
                               break
                       fi
               fi
               wget -c -N -P $title -O ./${title}/${outfile} "$url"
               fileCount=$(($fileCount+1))
       done
done
du -a | grep -i mp3 | awk '{FS=" "; print $2}' >>temp_shows
datenow=$(date +%Y-%m-%d)
sort temp_shows | uniq -u >> PodCasts-$datenow
# Now Ive added this bit to set the Podcast Albums to the word 'Podcast' and the Genre to Speech.
# making this easier for me to locate them on my ipod

while read mp3file
do
       id3v2 -A Podcast -g 101 $mp3file

done < PodCasts-$datenow