#!/bin/csh -f
#
#	Get a show's episode list from tvrage.com
#	and generate an editor script for numbering the episodes.
#	(Based on my get.episodes script.)
#
#	Gets one show at a time.  The argument can be the URL for
#	the show's episode list, or it can be the program name;
#	multiple words are allowed, capitalization is NOT significant.
#	Given a title, the script will guess at the TVrage URL.
#	It is often the program name with the spaces changed
#	to underscores, and /episode_list/all added.
#
#	Brent Chivers				2010 Jan 5
#
#	Solaris-10 seems to need "nawk" for that 2nd awk. 8-(	2005/Jun/27
#
#	Show title is no longer on the "Change Layout" line.	2010/Jun/09
#	Find it 2 lines later; force it to sort first into final awk.
#	Pick up "Unaired" episodes
#
#	lynx and sed path changes from Darwin to Linux.		2011/Jan/08
#	No preceding number on the episode-date lines
#	Episode number is not on the line with the date;
#	  it's on the line before the date.
#
#	TVRage.com page layout has changed;			2011/Apr/17
#	  actually simpler to use now.
#	Not handling "unaired" episodes 'cause I don't know how they look now.
#	Get rid of "View Trailer" crap				2011/Apr/29
#	TVRage.com page layout changed again; show title	2011/Jul/18
#	  is now the line after the 2nd empty line.
#	TVRage.com page layout changed again; show title	2012/Oct/23
#	  is now the line after the 5th empty line.
#	TVRage.com page layout changed again; show title is	2013/Apr/06
#	  now the line preceeding the line "^Classification: ".
#	Adding "tee" commands to catch the scraped web page was clever.
#	  (It's normally a waste of CPU and disk, but it really speeds
#	  up diagnosis when the format changes, and I can't count on
#	  being as clever next time.)
#	Pick up specials (and sort them back into sequence).	2013/Apr/12

# set verbose		# see commands before variable substitution
# set echo		# see commands after variable substitution
unset histchars		# get other users' history characters out of the way
unalias *		# get other users' .cshrc stuff out of the way
 
# $0:t construction doesn't work!
set cmd_name = $0
set cmd_name = $cmd_name:t	# the name of this command

# set webpage = /tmp/$cmd_name.$$	# temp file for data from TVRage

if ($#argv == 0) then		# no arguments provided
	echo "usage: $cmd_name show name [or tvrage.com URL]"
	exit
    endif

if ($#argv == 1) then		# one argument provided
	if ( "$1" =~ *tvrage.com* ) then
		set url = "$1"	# looks like a TVRage URL
	    else		# guess a probable TVRage URL
		set url = "http://www.tvrage.com/$1/episode_list/all"
	    endif
    else			# guess at the TVRage URL
	set show = `echo "$*" | tr ' ' '_'`
	set url = "http://www.tvrage.com/$show/episode_list/all"
    endif			# this works surprisingly often....
				# but not always. 8-(
	# get data from TVRage
	# strip off "Buy" and "Rating"
	# convert any html "&" to "&" character.
	# strip out icons
	# join episode lines to preceeding episode-number/air-date lines (reformated for sorting)
	# convert month to digits for sorting
	# keep show title
	# sort episode lines by air date
	# strip out text found with episode number
	# format program title, episode number, and episode title into editor commands

## echo "Using URL $url"

/usr/bin/lynx -dump -nolist -pseudo_inlines -width=300 $url | tee /tmp/tvrage1 | /bin/sed -e 's/^  *//'	\
-e 's=  *[NY]  *[0-9][0-9]*  *Amazon  *[0-9\.]*$=='	\
-e 's=  *[NY]  *[0-9][0-9]*  *Amazon  *N/A$=='	\
-e 's=  *[0-9][0-9]*  *Amazon  *[0-9\.]*$=='	\
-e 's=  *[0-9][0-9]*  *Amazon  *N/A$=='	\
-e 's=  *[0-9][0-9]*  *N/A$=='	\
-e 's=  *[NY]  *Amazon  *[0-9\.]*$=='	\
-e 's=  *[NY]  *Amazon  *N/A$=='	\
-e 's=  *Amazon  *[0-9\.]*$=='	\
-e 's=  *Amazon  *N/A$=='	\
-e 's=  *N/A$=='	\
-e 's=  *[NY]  *[0-9]*  *[0-9\.]*$=='	\
-e 's=  *[0-9]*  *[0-9\.]*$=='	\
-e 's=  *[NY] *[0-9\.]*$=='	\
-e 's=  *View Trailer=='	\
-e 's/&/\&/' -e 's/.favicon.ico. //' -e 's/.film.gif. //'	\
-e 's=^[1-9][0-9]* *\([1-9][0-9]*\)x\([0-9][0-9]*\) *\([0-3][0-9]\)/\([JFMASOND][aepuco][nbrylgptvc]\)/\([12][0-9][0-9][0-9]\) *\(.*\)$=\1.\2	\5/\4/\3	\6='	\
-e 's=^\(S\)[0-9]* - \(#*[1-9][0-9]*\) *\([0-3][0-9]\)/\([JFMASOND][aepuco][nbrylgptvc]\)/\([12][0-9][0-9][0-9]\) *\(.*\)$=\1p\2	\5/\4/\3	\6='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Jan/\([0-3][0-9]	\)=\1/01/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Feb/\([0-3][0-9]	\)=\1/02/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Mar/\([0-3][0-9]	\)=\1/03/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Apr/\([0-3][0-9]	\)=\1/04/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/May/\([0-3][0-9]	\)=\1/05/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Jun/\([0-3][0-9]	\)=\1/06/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Jul/\([0-3][0-9]	\)=\1/07/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Aug/\([0-3][0-9]	\)=\1/08/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Sep/\([0-3][0-9]	\)=\1/09/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Oct/\([0-3][0-9]	\)=\1/10/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Nov/\([0-3][0-9]	\)=\1/11/\2='	\
-e 's=\(	[12][0-9][0-9][0-9]\)/Dec/\([0-3][0-9]	\)=\1/12/\2=' | /usr/bin/tee /tmp/tvrage2 | /usr/bin/awk '\
/^$/	{ blank +=1 }	\
/^Classification: /	{ getline ; printf ("++++\t0000/00/00\t%s\n", prev) ; blank = 100; next }	\
$2 ~ /[12][0-9][0-9][0-9]\/[01][0-9]\/[0-3][0-9]/	{ print }	\
	{ prev = $0 }' | /usr/bin/sort -k2 | /usr/bin/tee /tmp/tvrage3 | /usr/bin/awk '\
BEGIN {	FS="	" }	\
/^++++	/	{ show = $3 ; next }	\
NF==3	{ printf ("g_^%s\t%s\t_s_\t%s\t_\t%s %s\t_p\n", show, $3, $3, $1, $3) }' | /bin/sed -e 's/ (unaired)//' -e 's/ (unaired)//'

# /bin/rm $webpage