#!/bin/bash
# Original version by Linc Fessenden 12/1/2004
# Find the latest script at:
# http://lincgeek.org/bashpodder
#
#
#
# Changelog for revisions by Chess Griffin
#		<chess at chessgriffin dot com>
#
# v0.1	2005-12-07:	Initial set of revisions
# v0.2	2006-12-01:	Second set of revisions incorporating 
#			command line flags (based on some of
#			Brian Hefferan's changes)
# v0.3  2007-06-02:	Add podshow.com fix, add m3u switch, and
#			tweak the echo statements
# v0.4  2007-06-19:	Add ability to remove spaces from filenames
# v0.5  2008-01-06:	Add in MSNBC feed fix from Brian Reichert


# Default values can be set here. Command-line flags override these.

first_only=	# default: nothing is all files; 1 is first only
m3u=		# default: no m3u playlist is created; 1 will create m3u
update_all= 	# default: nothing is normal download; 1 is update all 
verbose=1	# default: nothing is quiet; 1 is verbose
fetchlist='bp.conf' # default is bp.conf; can change name to other file
wget_quiet='' # default: -q is quiet, can remove the -q to get output

# No changes should be necessary below this line

function usage
{
echo "
Chess Griffin's modified BashPodder v0.5 2008-01-06
Usage: $0 [OPTIONS]
Options are:

-c, --config		Use a different config file other than bp.conf

-f, --first_only        Grabs only the first new enclosed file found in 
			each feed.  The --update_all flag will not work with
		       	this option. If you want to download the first file
		       	and also permanently ignore the other files, run
			bashpodder with this option, and then run it again
		       	with --update_all.

-h, --help             	Display this help message

-m, --m3u		Create m3u playlists

-u, --update_all	Write all episode names to the log file without
			downloading the actual files. This is useful
			if you want to subscribe to some podcasts but 
			do not want to download all the back episodes. 
			Once this switch has been run and the episodes
			names have been updated in the podcast.log file,
			you can edit the podcast.log file afterwards to
			delete any episode you still wish to download 
			the next time bashpodder is run.

-v, --verbose		Display verbose messages. 

bp.conf is the standard configuration file.  It should contain a list of
podcast feeds and local directories where each podcast's episodes will be
saved.  There should be a space between the feed and the local directory.
For example, one line in the bp.conf file may look like this:

http://feeds.feedburner.com/linuxreality /home/user/podcasts/LinuxReality

The first part of the line contains the URL of the podcast feed and the
second part of the line is the local directory where the episodes of that
podcast will be saved.  Note that you can use the '-c' switch to pass a
different configuration file when executing the script.  One can also
rename the default configuration file in the default settings located near
the top of the script.

Please note that one also needs the parse_enclosure.xsl file found at
the BashPodder page: http://lincgeek.org/bashpodder/.
"
}

# Make script crontab friendly:
cd $(dirname $0)

# Here are the command line switches 
while [ "$1" != "" ];do
   case $1 in
	-c|--config )		config=1
				fetchlist=$2
		;;
	-f|--first_only )	first_only=1
		;;
	-h|--help )		usage
				exit
		;;
	-m|--m3u )		m3u=1
		;;
	-u|--update_all )	update_all=1
    		;;
	-v|--verbose )		verbose=1
	       	;;
   esac
   shift
done

# Make sure a config file is passed when -c switch is used
if [ -z "$fetchlist" ]; 
then 
echo "You used the -c switch but did not include a configuration file.
Run $0 -h for usage.  Exiting.";
exit 1;
fi

# Make sure the bp.conf file or the file passed with -c switch exists
if test ! -f "$fetchlist";
then echo "The file $fetchlist does not exist.  Run $0 -h for usage.  
Exiting.";
exit 1;
fi

# Print the date
if [ -n "$verbose" ]; 
then echo "Starting BashPodder on";
date; 
fi

# Use an incoming folder to hold feeds prior to editing the file name
# from James Stone
if [ -n "$verbose" ]; then echo "Creating temp folders...";fi
incoming="incoming"
rm -rf $incoming
mkdir $incoming

# Delete any temp file:
rm -f temp.log

# If this is the first time running, create the podcast.log 
# from Scott Maxwell
if ! [ -e podcast.log ]
	then
	    if [ -n "$verbose" ]; 
	    then echo "Creating podcast.log file...";fi
	touch podcast.log
fi

# Read the bp.conf file and wget any url not already in the 
# podcast.log file:
while read podcast datadir
	do
	seenfirst=
# Skip lines beginning with '#' as comment lines - from Rick Slater
# Also skip blank lines
	if echo $podcast | grep -E '^#|^$' > /dev/null
		then
		continue
	fi
	# Check to make sure datadir is listed on each line; if not, quit
	if test ! $datadir
		then 
echo "No feed directory specified in $fetchlist for the
$podcast feed.
Please edit $fetchlist and name a directory for this feed.  Exiting." 
		exit 1;
	fi
	# Check for and create datadir if necessary:
	# (Uses Tony Whitmore's addition to bp.conf where 
	# name of directory comes after feed, e.g.,
	# http://www.lugradio.org/episodes.rss LUGRadio)
	if test ! -d $datadir && test "$update_all" != "1";
	        then
		    if [ -n "$verbose" ]; 
		    then
		    echo "Creating $datadir directory to store feeds...";
		    fi
	        mkdir -p $datadir
	fi
	if [ -n "$verbose" ]; then echo "Checking "$podcast; fi
	file=$(wget -q $podcast -O - | \
	xsltproc parse_enclosure.xsl - 2> /dev/null) || \
	file=$(wget -q $podcast -O - | tr '\r' '\n' | tr \' \" | \
	sed -n 's/.*url="\([^"]*\)".*/\1/p')
	for url in $file
	do
	    if [ -n "$first_only" ] && [ -n "$seenfirst" ]; then break;fi
	# The realurl fix for dynamically generated feeds (from Huw
	# Lynes and James Stone)
	realurl=`curl -s -I -L -w %{url_effective} --url $url | tail -n 1`
	# Figure out the short url (i.e. filename!) from James Stone
	urlss=`echo $realurl|awk -F / '{print $NF}'`
	urls=`echo $urlss|awk -F ? '{print $1}'`
	# Replace URL encoded spaces in filenames with regular space
	nospace=`echo $urls|sed -e 's/%20/ /g'`
	# Remove all spaces from filenames
	nospace2=`echo $nospace|sed -e 's/ //g'`
	# Fix Podshow.com numbers that keep changing
	urlsfix=`echo $nospace2|sed s/_pshow_[0-9]*//`
	# Fix MSNBC podcast names for audio feeds from Brian Reichart
	if echo $realurl | grep -q 'msnbc.*pd_.*mp3$'; then
		urlsfix=`echo $urlss | sed -e 's/.*\(pd_.*mp3$\)/\1/'`
	fi
	if echo $realurl | grep -q 'msnbc.*pdm_.*mp3$'; then
		urlsfix=`echo $urlss | sed -e 's/.*\(pdm_.*mp3$\)/\1/'`
	fi
	if echo $realurl | grep -q 'msnbc.*vh-.*mp3$'; then
		urlsfix=`echo $urlss | sed -e 's/.*\(vh-.*mp3$\)/\1/'`
	fi
	# Fix MSNBC podcast names for video feeds
	if echo $realurl | grep -q 'msnbc.*pdv_.*m4v$'; then
		urlsfix=`echo $urlss | sed -e 's/.*\(pdv_.*m4v$\)/\1/'`
	fi
	echo $urlsfix >> temp.log
	if [ -n "$update_all" ];
	then
	    if [ -n "$verbose" ]; then echo "Catching up $realurl...";fi
	elif ! grep "$urlsfix" podcast.log > /dev/null
    		then
		# Added Christian Daven's wget switches
		if [ -n "$verbose" ]; then echo "Fetching $realurl...";fi
		# File names longer than 255 characters break wget, this should
		# fix that.
		if echo $realurl | grep -q 'msnbc'; then
			wget $wget_quiet -c -T 2 -O $incoming/$urlsfix "$realurl"
			mv $incoming/$urlsfix $datadir
		else
			wget $wget_quiet -c -T 2 -P $incoming "$realurl"
	      		# Fix for ?podcast end to recent lugradio feeds
			# rom James Stone
			mv $incoming/"$nospace"* $datadir/$urlsfix
		fi
	fi
	seenfirst=1
	done
	# Create an m3u playlist:
	if [ -z "$update_all" ];
	then
		if [ -n "$m3u" ]; 
		then
			if [ -n "$verbose" ]; 
			then 
			    echo "Creating $datadir m3u playlist...";
			fi
			ls $datadir | grep -v m3u > $datadir/podcast.m3u
		fi
	fi
	done < $fetchlist

if test ! -f temp.log && [ -n "$verbose" ];
then 
    echo "Nothing to download.";
fi

# Move dynamically created log file to permanent log file:
if [ -n "$verbose" ]; then echo "Cleaning up...";fi
cat podcast.log >> temp.log
sort temp.log | uniq > podcast.log
rm temp.log

if [ -n "$verbose" ]; then echo "Done."; fi;