#! /bin/bash

########################################################################
##  Grabs newest movie trailers off http://www.apple.com/trailers and ##
##  saves them on the disk for viewing.                               ##
##                                                                    ##
##  by Kevin Welch, 2005                                              ##
########################################################################

# move to proper directory
PWD=`pwd`
cd /home/kevin/sbin/trailers

if [ -z "$1" ]; then
	# grab fresh copy of the trailer list for screenscraping
	echo "Grabbing fresh page..."
	wget -q http://www.apple.com/trailers/ -O trailers.html

	# scrape trailers.html for the "Newest Trailers" and "Exclusive Trailers"
	# columns and strip HTML:
	# resulting list is in format "DATE \t URL \t TITLE"
	echo "Filtering links..."
	sed -nr -e "
	/Trailers Exclusive/,/class=\"clearer\"/ {
		s/^\t*//
		s/<\/?div[^<]*>//g
		s/<\/?h[1-4]>//g
		s/<a href=\"/\thttp:\/\/www.apple.com\/trailers\//
		s/\">/\t/
		s/<\/a>.*//
		p
		}
	" -e "
	/Newest Trailers/,/class=\"clearer\"/ {
		s/^\t*//
		s/<\/?div[^<]*>//g
		s/<\/?h[1-4]>//g
		s/<a href=\"/\thttp:\/\/www.apple.com\/trailers\//
		s/\">/\t/
		s/<\/a>.*//
		p
		}
	" trailers.html | egrep ^[0-9]{2}[.][0-9]{2} | awk '{print $2}' > trailerlist.tmp
elif [ -e $PWD/$1 ]; then 
	# argument is filelist
	cp "$PWD/$1" trailerlist.tmp
elif [[ "$1" == http* ]]; then
	# argument is url
	echo "$1" > trailerlist.tmp
else
	# Dunno what it is
	echo "Parameter not recognized."
	exit 1
fi

# loop through trailer list
echo "Grabbing trailers..."
numtrailers=`wc -l trailerlist.tmp | awk '{print $1}'`
numhistory=0
for url in `cat trailerlist.tmp`
do
	# set temp short name for file
	name=`echo $url | sed -e 's/[.]html$//' | awk -F / '{print $6}'`

	# check if url is in history
	if [ -z "$1" ]; then
		if grep -q $url trailerhistory; then
			inhistory=TRUE
		else
			inhistory=FALSE
		fi
	else
		# force no history check for command-line parameters
		inhistory=FALSE
	fi
	
	if [[ $inhistory == TRUE ]]; then
		# it is; move on
		echo "History: $url"
		let "numhistory += 1"

		# w00! new trailer! grab it.
		else
			wget -q $url -O ./$name.html
			if grep -iq embed $name.html; then

					# we can look directly at the file
					# scour html file for <EMBED> tag, and nab reference quicktime mov
					refurl=`cat ./$name.html | sed -rne '/<(EMBED|embed|Embed)/,/<\/(EMBED|embed|Embed)>/p' | egrep -io '(HREF|SRC)=\"[^\"]*ref[^\"]*\"' | awk -F \" '{print $2}'`
	
				# we need to dig further into the directory to find an html file
				else
					wget -q $url/large.html -O large.html
					# check to see if large.html has an <EMBED> tag -- boy, does that make things easy
					if grep -iq embed large.html; then
						# yes, it does ... extract as before.
						echo -n ""
	
						# nope, large.html doesn't exist. We'll have to fish for it.
						else
							wget -q $url -O $name.html
	
							# look for images that are links:
							# search for "large" and "high"
							temp=`cat $name.html | sed -rne 's/^[ \t]*//;s/[ \t]*$//;/^$/d;s/parentalguide.org//;/<(a|A)/,/<\/(a|A)>/p' | egrep -i '<(a|A)[^<>]*(large|high|-lrg|_lrg|-lg|_lg|lg)' | egrep -io '<a[^<>]*' | egrep -io 'href="[^"]*(large|high|-lrg|_lrg|-lg|_lg|lg)[^"]*"' | awk -F \" '{print $2}'`
							if [[ `echo "$temp" | wc -l` -gt 1 ]]; then
								echo "More than one trailer present for $name... must be downloaded by hand."
								echo "$temp"
								largehtml=""
							else
								if [[ $url == *html ]]; then
									largehtml=`echo $url | sed -e "s/\/[^\/]*$/\/$temp/"`
								elif [[ $temp == http* ]]; then
									largehtml=$temp
								else
									largehtml=$url$temp
								fi
							fi
							[ -n "$largehtml" ] && wget -q "$largehtml" -O large.html > /dev/null
					fi
					# scour html file for <EMBED> tag, and nab reference quicktime mov
					refurl=`cat ./large.html | sed -rne '/<(EMBED|embed|Embed)/,/<\/(EMBED|embed|Embed)>/p' | egrep -io '(HREF|SRC)=\"[^\"]*mov\"' | grep -v 'apple.*images' | awk -F \" '{print $2}'`
			fi
		[ -n "$refurl" ] && wget -q $refurl -O $name-ref.mov

		# search reference quicktime mov for filenames
		[ -e $name-ref.mov ] && filename=`strings $name-ref.mov | grep '[.]mov' | sed -e 's/^[^a-z]//' | sed -ne '$p'`

		# create downloadable url and grab it, or if the file we downloaded is not a reference, move it
		if [ -n "$filename" ]; then
				downloadfile=`echo $refurl | sed "s/\/[^\/]*$/\/$filename/"`
				wget --spider $downloadfile 2> filesize.tmp
				filesizecommas=`cat filesize.tmp | grep Length | egrep -o '[0-9]+,[0-9]+,{0,1}[0-9]*'`
				filesize=`echo $filesizecommas | sed -e 's/,//g'`
				currentsize=0
				echo -n "Downloading $filename [$filesizecommas bytes] ... 00%"
				wget -qb $downloadfile -O new/$filename > /dev/null
				sleep 2
				while [ "$currentsize" -lt "$filesize" ]
				do
					currentsize=`du -b new/$filename | awk '{print $1}'`
					percent=$(($currentsize * 100 / $filesize ))
					[ $percent -lt 10 ] && percent=0$percent
					echo -ne "\b\b\b$percent%"
					sleep 1
				done
				echo -ne "\b\b\b\bdone!\n"
			else
				[ -e $name-ref.mov ] && mv -v $name-ref.mov new/$name_m480.mov
		fi

		# add url to history file so we don't grab the trailer twice
		echo $url >> trailerhistory

		# clean up temp files
		[ -e $name.html ] && rm ./$name.html
		[ -e $name-ref.mov ] && rm ./$name-ref.mov
		[ -e large.html ] && rm ./large.html
	fi
done

# final output
numdownloaded=$(( $numtrailers - $numhistory ))
if [[ "$numdownloaded" == "0" ]]; then
	echo "No new trailers."
else
	echo "Downloaded $numdownloaded new trailers."
fi

# clean up
[ -e filesize.tmp ] && rm filesize.tmp
[ -e trailerlist.tmp ] && rm trailerlist.tmp
[ -e trailers.html ] && rm trailers.html
rm wget-log*

# move back to original directory
cd "$PWD"