#!/usr/bin/perl


######################### VARIABLES ######################
#

$number_cites_wanted=300;
$urls_per_altavista_page = 10;
$orig_number_cites_wanted = $number_cites_wanted;
$altavista_hits=0;
$altavista_wordcount=0;
$goodcontext=1;
$contextsize=40;

$number_of_other_results_pages_found = 0;
$number_of_other_results_pages_needed = 0;
$highest_number_URLs_allowed = 300;
$URL_to_search = "";

$number_urls_searched = 0;
$total_number_urls = 0;

$inaccessible_server = 0;
$accessible_server = 0;

$fourofour = 0;
$http_two_hundred = 0;
$other_http_error_code = 0;

$term_not_there = 0;
$term_in_page = 0;

$good_html_page = 0;
$crap_html = 0;

$title_hits = 0;
$body_hits = 0;
$neither_title_nor_body = 0;

$non_meta_hits = 0;
$metadata_hits = 0;

$table_list_citation = 0;
$potential_goodhits = 0;
$goodhits = 0;

$success = 0;
$total_errors = 0;


$separator = "******************************\n";
$http = "http\:\/\/";

#
#
#
######## ALTA VISTA RESULT PAGE REFERENCES  ##############
#
#
# URL example
# b>URL:</b> <font color="#808080">www.newadvent.org/cathen/p.htm</font>
#
# next pages example
# <a href\=\"\/cgi\-bin\/query\?pg\=q\&stype\=stext\&enc\=iso88591\&q\=nosocomial&stq=180">19</a>

$alta_vista = "www.altavista.com";

$null_search = "AltaVista found no document matching your query\. \n";

$single_page = "1 page found";

#
##########################################################


$output_autoflush = 1;

open(TESTFILE, ">test1");

$argument = $ARGV[0];
if (-e $argument)
{
    open(FILE, $argument);
    @search_terms = <FILE>;
}
else
{
    $search_terms[0] = $argument;
}

print TESTFILE @search_terms;

###########################################################
#                     MAIN PROGRAM                        #
###########################################################

foreach $term (@search_terms)
{
    if ($term =~ /\n/)
    {
	$term =~ s/\n//;
    }
    $term =~ s/[ ]+$//;
    $term =~ s/[ ]+/\+/g;
    $term =~ s/[\+]+/\+/g;

    $URL = "http://www.altavista.com/cgi-bin/query?mss=simple&pg=q&what=web&enc=iso88591&kl=XX&locale=xx&q=$term&search=Search";

#    &start_output_page;

# GETTING 1st ALTA-VISTA RESULTS PAGE

    @first_results_page = &get_page_by_URL($URL);
 
# CHECK FOR SOCKET ERRORS
# 1st POSSIBLE MISS

    $socket_error = $first_results_page[0];
    if ($socket_error < 0)
    {
	print "\tAlta-Vista is currently down - too bad\n\n";
    }
    else # if AltaVista is up
    {
	@altavista_stats = &get_altavista_stats(@first_results_page);
	$altavista_hits = @altavista_stats[0];
	$altavista_wordcount = @altavista_stats[1];
	$number_of_other_results_pages_found = @altavista_stats[2];

	if ($altavista_hits == 0)
	{
	    print "AltaVista contained no tokens of \"$term\"\n";
	}
	else # there is a term or there are terms
	{
 	    if ($number_cites_wanted > $highest_number_URLs_allowed)
	    {
		$number_cites_wanted = $highest_number_URLs_allowed;
	    }
	    if ($number_cites_wanted > $urls_per_altavista_page)
	    {
		$number_of_other_results_pages_needed = int(($number_cites_wanted + ($urls_per_altavista_page - 1)) / $urls_per_altavista_page);
		if ($number_of_other_results_pages_needed > $number_of_other_results_pages_found)
		{
		    $number_of_other_results_pages_needed = $number_of_other_results_pages_found;
		}
	    }
	    @other_results_pages = &get_other_resultpages($number_of_other_results_pages_needed, @first_results_page);
	    
	    @all_results_pages = (@first_results_page, @other_results_pages);
	
	    @total_urls = &extract_urls_from_results_pages(@all_results_pages);
	
	
	    if ($number_cites_wanted > $total_number_urls)
	    {
		$orig_number_cites_wanted = $number_cites_wanted;
		$number_cites_wanted = $total_number_urls;
	    }
	    
	    for ($i = $number_cites_wanted; $i >= 1; $i--)
	    {
		$number_urls_searched++;
		$URL_to_search = shift(@total_urls);
		&find_target_page($URL_to_search, $term);
	    }
	    &print_output;
	}
    }
}

######################################################
#                  SUB ROUTINES                      #
######################################################
#
# sub get_altavista_stats
# sub print_output
# sub find_target_page
# sub extract_urls_from_results_pages
# sub get_other_resultpages
# sub search_for_original_search_term
# sub get_page_by_URL 
# sub start_output_page
# sub end_output_page
# sub remove_font_crap
# sub delete_metadata_hits
# sub delete_list_and_table_crap
# sub delete_table_crap
# sub delete_list_crap
# sub tidy_sentence_breaks
# sub delete_all_other_tags
#
######################################################



sub get_altavista_stats
{
    local(@first_results_page) = @_;
    local($line) = "";

    foreach $line (@first_results_page)
    {
	if  ($line eq $null_search)
	{
	    $altavista_hits = 0;
	}
	if  ($line =~ m/([0-9\,]+) pages found/i)
	{
	    $altavista_hits = $1;
	    $altavista_hits =~ s/\,//g;
	}
	if  ($line eq $single_page)
	{
	    $altavista_hits = 1;
	}
	if  ($line =~ /word count.*\: ([0-9\,]+)/)
	{
	    $altavista_wordcount = $1;
	}
	if ($line =~ m/<a href\=\"\/cgi\-bin\/query\?pg\=q\&stype\=stext\&enc\=iso88591\&q\=/i)
	{
	    $number_of_other_results_pages_found++;
	}
    }
    push(@altavista_stats, $altavista_hits, $altavista_wordcount, $number_of_other_results_pages_found);
    return @altavista_stats;
}



sub print_output
{
######################################
#            OUTPUTTING
######################################

	    

    $success = $title_hits + $goodhits;
    
    $total_errors = $inaccessible_server + $fourofour + $other_http_error_code + $term_not_there + $crap_html + $neither_title_nor_body + $metadata_hits + $table_list_citation;
    
    print $separator;
    print $separator;
    
    print "Citations requested - $orig_number_cites_wanted\n";
    print "Alta-vista hits (ie pages) - $altavista_hits\n";
    print "Alta-vista wordcount - $altavista_wordcount\n";
    print $separator;
    
    print "Found $total_number_urls URLs\n";
    print "Searching $number_urls_searched URLs\n";

    
    print $separator;

    print "Of $number_urls_searched URLs searched there were\n";
    print "   $inaccessible_server inaccessible servers and\n";
    print "   $accessible_server accessible servers\..\n";
    print "\n";
    print "Of $accessible_server servers searched there were\n";
    print "   $fourofour 404 or similar HTML errors\n";
    print "   $other_http_error_code other HTML errors and\n";
    print "   $http_two_hundred HTTP 200 pages.\n";
    print "\n";
    print "Of $http_two_hundred HTTP pages searched there were\n";
    print "   $term_not_there not containing the \"$term\" and\n";
    print "   $term_in_page containing \"$term\"\.\n";
    print "\n";
    print "Of $term_in_page matching pages searched there were\n";
    print "   $crap_html with bad html and\n";
    print "   $good_html_page with good html\.\n";
    print "\n";
    print "Of $good_html_page good html pages found there were\n";
    print "   $title_hits with the term only in the title\n";
    print "   $neither_title_nor_body with the term neither in the title or the body and\n";
    print "   $body_hits with \"$term\" in the body\.\n";
    print "\n";
    print "Of $body_hits matching bodies searched there were\n";
    print "   $metadata_hits in metadata and\n";
    print "   $non_meta_hits in text data\.\n";
    print "\n";
    print "Of $non_meta_hits pages with better hits searched there were\n";
    print "   $table_list_citation with table/list citations and\n";
    print "   $potential_goodhits prose texts.\n";
    print "\n";
    print "   Of $potential_goodhits prose texts searched there were\n";
    print "   $goodhits sentences containing \"$term\".\n";
    print $separator;
    print "Total URLs searched - $number_urls_searched\n";
    print "Total citations found - $success\n";
    print "Total errors - $total_errors\n";
	
    print $separator;
#    &end_output_page;
}


sub find_target_page
{
    local($url) = $_[0];
    if ($url =~ /\n/)
    {
	$url =~ s/\n//g;
    }
    local($term) = $_[1];

    if ($url !~ /^http/)
    {
	$url = "$http$url";
    }

    local(@web_page) = &get_page_by_URL($url);

# 2nd POSSIBLE ERROR

    $socket_error = $web_page[0];

    if ($socket_error < 0)
    {
	$inaccessible_server++;
#	print "Server for $url inaccessible\n";
    }
    else
    {
	$accessible_server++;
	&search_for_original_search_term($term,$url,@web_page);
    }
}

sub extract_urls_from_results_pages
{
    local(@web_page) = @_;
    local($URL) = "";
    local(@URLS) = ();

    foreach $line (@web_page)
    {
	if ($line =~ /<b>URL\:<\/b> <font color\=\"\#808080\">/)
	{
	    $total_number_urls++;
	    $line =~ /<b>URL\:<\/b> <font color\=\"\#808080\">(.*?)<\/font>/;
	    $URL = "$http$1\n";
	    
	    push (@URLS, $URL);
	}
    }
    return @URLS;
}

sub get_other_resultpages
{
    local($number_of_other_results_pages_needed) = shift(@_);
    local(@first_results_page) = @_;
    local(@URL) = [];
    local($URL) = "";
    local($needed_URL) = "";
    local($i) = 0;

    foreach $line (@first_results_page)
    {
	if ($line =~ m/<a href\=\"\/cgi\-bin\/query\?pg\=q\&stype\=stext\&enc\=iso88591\&q\=.*?\&stq\=[0-9]+\">[0-9]/i)
	{
	    $line =~ /<a href\=\"(\/cgi\-bin\/query\?pg\=q\&stype\=stext\&enc\=iso88591\&q\=.*?\&stq\=[0-9]+)\">[0-9]+<\/a>/i;
	    $URL = "$http$alta_vista$1";

	    push (@URLS, $URL);
	}
    }
    for ($i = $number_of_other_results_pages_needed; $i > 0; $i--)
    {
	$needed_URL = shift(@URLS);
	@next_results_page = &get_page_by_URL($needed_URL);
	push (@extra_results_pages, @next_results_page);
    }
    return @extra_results_pages;
}

sub search_for_original_search_term
{
    local($length) = 0;

    local($term) = shift(@_);
    if ($term =~ /\+/)
    {
	$term =~ s/\+/ /g;
    }
    local($url) = shift(@_);

    local($web_page) = join " ", @_, "[RECORDBREAKER]";

    $/ = "[RECORDBREAKER]";

    $length = length($web_page);

#    print STDOUT "$length \n $term \n $ url \n <DAMN> $web_page <damn>\n";

	
    $web_page =~ s/\&nbsp\;/ /gi;
    $web_page =~ s/\cM/ /g;
    $web_page =~ s/\n/ /g;
    $web_page =~ s/\t/ /g;
    $web_page =~ s/[ ]+/ /g;
    $web_page =~ s/[ ]+(<\/P>)/$1/gi;
    $web_page =~ s/[ ]+(<BR>)/$1/gi;
    $web_page =~ s/[ ]+(<\/HR>)/$1/gi;
    $web_page =~ s/<BR.*?>/<BR>/gi;
    $web_page =~ s/<HR.*?>/<HR>/gi;


    if ($web_page =~ /HTTP\/[0-9]\.[0-9] 200/)
    {
	$http_two_hundred++;
# 4th POSSIBLE ERROR
	if ($web_page !~ m/$term/i)
	{
	    $term_not_there++;
#	    print "$separator NO CITATION for $term IN $url\n";
	}
	else
	{
	    $term_in_page++;


	    if ($web_page =~ m/<TITLE>/i)
	    {
		$good_html_page++;

		$web_page =~ /(.*?)<TITLE>(.*?)<\/TITLE>(.*)/i;
		$prelim = $1;
		$title = $2;
		$body = $3;

		if ($body =~ m/$term/i)
		{
		    $body_hits++;
		    $removed_fonts = &remove_font_crap($body);
		    $removed_meta = &delete_metadata_hits($removed_fonts);
		    
		    if ($removed_meta =~ m/$term/i)
		    {
			$non_meta_hits++;
			$non_list_table = &delete_list_and_table_crap($removed_meta);

			if ($non_list_table =~ m/$term/i)
			{
			    $potential_goodhits++;

			    $neat_and_tidy = &tidy_sentence_breaks($no_html);
			    
			    while ($neat_and_tidy =~ s/(.*?[\.\?\!\n])//)
			    {
				$sentence = $1;
				
				if ($sentence =~ m/$term/i)
				{
				    $goodhits++;
				    $sentence =~ s/\[stop\]/\./g;
				    $sentence =~ s/\[quot\]/\"/g;
				    $sentence =~ s/\[ellipsis\]/\.\.\./g;
				    
				    print $separator;	
				    print "\nCIT = $sentence\n";
				    print "LENGTH = $length\n";
				    print "URL = $url\n";
				}
			    }
			}
			else   # 7th POSSIBLE ERROR - Cit was in a list or table
			{
			    $table_list_citation++;
#			    print "$separator CITATION IN TABLE/LIST AT $url\n";
			}
		    }
		    else     # 6th POSSIBLE ERROR - metadata hit
		    {
			$metadata_hits++;
#			print "$separator CITATION IN METADATA, HREF, ETC., AT $url\n";
		    }
		}
		elsif ($title =~ m/$term/i)
		{
		    # only return the <title> as citation if the body doesn't
		    # contain the search term
		
		    $title_hits++;
		    $citation = $title;
		    print $separator;
		    print "TITLE CITITATION = $citation\n";
		    print "LENGTH = $length\n";
		    print "URL = $url\n";
		}
		else
		{
		    $neither_title_nor_body++;
		}
	    }
# 5th POSSIBLE ERROR - crap html
	    else
	    {
		$crap_html++;
#		print "$separator CRAP HTML at $url\n";
	    }
	}
    }
# 3rd POSSIBLE ERROR, 404, etc.
    elsif ($web_page =~ /HTTP\/[0-9]\.[0-9] [345][0-9][0-9]/)
    {
	$fourofour++;
#	print "$separator ERROR FINDING THE PAGE AT $url\n";
    }
    else
    {
	$other_http_error_code++;
    }
}

sub get_page_by_URL 
{
    local($URL) = @_;

    if ($URL !~ m/^http:\/\/.*/i)
    { 
	return;
    }
    else
    {
	if ($URL =~ m/^http:\/\/([\w-\.]+):?(\d*)($|\/(.*))/)
	{
	    $host = $1;
	    $port = $2;
	    $path = $3;
	    if ($path eq "")
	    {
		$path = '/';
	    }
	    if ($port eq "")
	    {
		$port = 80;
	    }
	    $path =~ s/#.*//;
	    
	    $AF_INET = 2;
	    $SOCK_STREAM = 1;
	    $sockaddr = 'S n a4 x8';
	    chop($hostname = `hostname`);
	    ($name,$aliases,$proto) = getprotobyname('tcp');
	    ($name,$aliases,$port) = getservbyname($port,'tcp') unless $port =~ /^\d+$/;
	    ($name,$aliases,$type,$len,$thisaddr) = gethostbyname($hostname);
	    
	    if (!(($name,$aliases,$type,$len,$thataddr) = gethostbyname($host)))
	    {
		return -1;
	    }
	    $this = pack($sockaddr, $AF_INET, 0, $thisaddr);
	    $that = pack($sockaddr, $AF_INET, $port, $thataddr);
	    
	    if (!(socket(S, $AF_INET, $SOCK_STREAM, $proto)))
	    {
		$SOCK_STREAM = 2;
		if (!(socket(S, $AF_INET, $SOCK_STREAM, $proto)))
		{
		    return -2;
		}
	    }
	    if (!(bind(S, $this)))
	    {
		return -3;
	    }
	    if (!(connect(S,$that)))
	    {
		return -4;
	    }
	    select(S); $| = 1;
	    select(STDOUT);
	    print S "GET $path HTTP/1.0\n\n";
	    @response = <S>;
	    ($protocol, $status) = split(/ /, $response);
	    close(S);

#	    print STDOUT "<HI>  @response <hi>\n";
	    return @response;

	}
    }
}

sub start_output_page
{
    print "<HTML><HEAD><TITLE> ${term}: webcorpus search result</TITLE></HEAD>";
    print "<BODY><H1> ${term}: webcorpus search result</H1>\n<TABLE>\n";
#    print "${term}: webcorpus search result\n";
}
sub end_output_page
{
    print "</TABLE>\n<TABLE><TR><TD><B>HITS $hits </B>";
    print "<TD><B>WC $wordcount </B>";
    print "<TD><B>GOODIES $goodies </B>";
    print "<TD><B>DUDS $duds </B>";
    print "</TABLE></BODY></HTML>\n";
}

#    foreach $line (split /<(body|p|td|br|li)( [^>]*)?>/i, $web_page)
#    {
#	$line =~ s/\cM$/\n/g;
#	$line =~ s/\cM/ /g;
#	$line =~ s/[ \t]+/ /g;
#	$line =~ s/<[^>]+>//gi;
#
#	if ($line =~ m/(^.*)($term)(.*)/i)
#	{
#
# break a "hit" up into 3 parts so I can show it prettily in html table
#	    $this_term=$2; 
#	    $pre=substr($1, length($pre)-$contextsize, $contextsize);
#	    $post=substr($3, 0, $contextsize); 
#delete trailing part-words
#	    $pre=~s/^[a-z\.,;:?\!\)]*//;
#	    $post=~s/[^ \.,;:?\!\)]*$//;
#	    $termcount++;
#
#	    if (scalar(split(/\s/, $pre)) + scalar(split(/\s/, $post)) > $goodcontext)
#	    {
#		$goodies++;
# if we're going to do duplicate-checking, we shoudl either do it here 
# (or store everything in an arrary and check at the end)
#		
#		print "<TR><TD><a href=http:\"$url\">all</a></TD>";
#		print "<TD align=\"right\">$pre</TD><TD>${this_term}</TD><TD>${post}</TD></TR>\n";
#	    }
#	}
#    }
#    print $web_page;

sub remove_font_crap
{
    local($rec) = $_[0];
    
    $rec =~ s/<B>//gi;
    $rec =~ s/<I>//gi;
    $rec =~ s/<U>//gi;
    $rec =~ s/<\/B>//gi;
    $rec =~ s/<\/I>//gi;
    $rec =~ s/<\/U>//gi;
    $rec =~ s/<FONT.*?>//gi;
    $rec =~ s/<BASEFONT.*?>//gi;
    $rec =~ s/<\/FONT>//gi;
    
    return $rec;
}

sub delete_metadata_hits
{
    local($rec) = $_[0];

    $rec =~ s/<BODY.*?>//gi;
    $rec =~ s/<IMG.*?>/\[image\]/gi;
    $rec =~ s/<A HREF.*?>(.*?)<\/A>/$1/gi;
    $rec =~ s/<LINK.*?>.*?<\/LINK>/ /gi;
    $rec =~ s/<META.*?>/ /gi;
    
    return $rec;
}


sub delete_list_and_table_crap
{
    local($rec) = $_[0];

    $no_tables = &delete_table_crap($rec);
    $no_lists = &delete_list_crap($no_tables);

    if ($no_lists =~ m/</)
    {
	$no_html = &delete_all_other_tags($no_lists);
	return $no_html;    
    }
    else
    {
	return $no_lists;    
    }
}

sub delete_table_crap
{
    local($rec) = $_[0];

    $rec =~ s/<TABLE.*?>//gi;
    $rec =~ s/<\/TABLE>//gi;
    $rec =~ s/<TR.*?>//gi;
    $rec =~ s/<\/TR>//gi;
    $rec =~ s/\.<\/TD>/\./gi;
    $rec =~ s/<\/TD>/\./gi;
    $rec =~ s/<TD.*?>/ /gi;
    $rec =~ s/\.<\/TH>/\./gi;
    $rec =~ s/<\/TH>/\./gi;
    $rec =~ s/<TH.*?>/\[heading\] /gi;
    $rec =~ s/<COLGROUP.*?>.*?<\/COLGROUP>//gi;
    $rec =~ s/<COL.*?>.*?<\/COL>//gi;
    $rec =~ s/[ ]+/ /gi;

    return $rec;
}

sub delete_list_crap
{
    local($rec) = $_[0];

    $rec =~ s/<UL.*?>//gi;
    $rec =~ s/<\/UL>/\./gi;
    $rec =~ s/<OL.*?>//gi;
    $rec =~ s/<\/OL>/\./gi;
    $rec =~ s/<DL.*?>//gi;
    $rec =~ s/<\/DL>/\./gi;
    $rec =~ s/<DT>/ /gi;
    $rec =~ s/<DD>/ \- /gi;
    $rec =~ s/<\/DD>//gi;
    $rec =~ s/<DIR.*?>.*?<\/DIR>//gi;
    $rec =~ s/<MENU.*?>.*?<\/MENU>//gi;
    $rec =~ s/[ ]+(<LI>)/$1/gi;
    $rec =~ s/[ ]+(<\/LI>)/$1/gi;
    $rec =~ s/\.(<\/LI>)/\./gi;
    $rec =~ s/(<\/LI>)/\./gi;
    $rec =~ s/\.(<LI>)/\./gi;
    $rec =~ s/(<LI>)/\./gi;

    return $rec;
}

sub tidy_sentence_breaks
{
    local($rec) = $_[0];

# converting horizontal rules, line breaks and paragraph ends to fullstops
    $rec =~ s/<HR>/\. /gi;
    $rec =~ s/<BR>/ | /gi;
    $rec =~ s/<P .*?>/<P>/gi;
    $rec =~ s/<P><\/P>//gi;
    $rec =~ s/<P>[ ]+<\/P>//gi;
    $rec =~ s/\.<\/P>/\. /gi;
    $rec =~ s/<\/P>/\. /gi;
    $rec =~ s/[ ]+(<P>)/$1/gi;
    $rec =~ s/\.<P>/\. /gi;
    $rec =~ s/<P>/\. /gi;
    $rec =~ s/[ ]+/ /gi;
    $rec =~ s/\. \./\. /gi;
    $rec =~ s/<BLOCKQUOTE.*?>//gi;
    $rec =~ s/<\/BLOCKQUOTE>/\./gi;
    $rec =~ s/<Q.*?>//gi;
    $rec =~ s/<\/Q>/\./gi;
    $rec =~ s/<H[0-9].*?>/\[heading\]/gi;
    $rec =~ s/<\/H[0-9]>/\./gi;
    $rec =~ s/<CAPTION.*?>/\[caption\]/gi;
    $rec =~ s/<\/CAPTION>/\./gi;

# converting fullstops in name abbreviations, URLs, etc.
    $rec =~ s/([A-Z])\.( [A-Z])\.( [A-Z][a-z]+)/$1\[stop\]$2\[stop\]$3/g;
    $rec =~ s/([A-Z])\.([A-Z])\.( [A-Z][a-z]+)/$1\[stop\]$2\[stop\]$3/g;
    $rec =~ s/( [A-Z])\.( [A-Z][a-z]+)/$1\[stop\]$2/g;
    $rec =~ s/([a-zA-Z]+)\.([cC][oO][mM])\.([a-zA-Z]+ )/$1\[stop\]$2\[stop\]$3/g;
    $rec =~ s/([a-zA-Z]+)\.([cC][oO][mM] )/$1\[stop\]$2/g;
    $rec =~ s/ \"([a-zA-Z])+\" / \[quot\]$1\[quot\] /g;
    $rec =~ s/\.[\.]+/\[ellipsis\]/g;

    $rec =~ s/([0-9])\.([0-9])/$1\[stop\]$2/g;

    return $rec;
}

sub delete_all_other_tags
{
    local($rec) = $_[0];

    $rec =~ s/<\!\-\-.*?\-\->/ /g;
    $rec =~ s/<A NAME.*?>(.*?)<\/A>/$1/gi;
    $rec =~ s/<SCRIPT.*?<\/SCRIPT>/ /gi;
    $rec =~ s/<NOSCRIPT.*?<\/NOSCRIPT>/ /gi;
    $rec =~ s/<APPLET.*?<\/APPLET>/ /gi;
    $rec =~ s/<SERVER>.*?<\/SERVER>/ /gi;
    $rec =~ s/<BGSOUND.*?>/ /gi;
    $rec =~ s/<MAP.*?>.*<\/MAP>/ /gi;
    $rec =~ s/<AREA.*?>/ /gi;
    $rec =~ s/<EMBED.*?>/ /gi;
    $rec =~ s/<OBJECT>.*<\/OBJECT>/ /gi;
    $rec =~ s/<PARAM>/ /gi;
    $rec =~ s/<FORM.*?>.*?<\/FORM>/ /gi;
    $rec =~ s/<INPUT.*?>/ /gi;
    $rec =~ s/<BUTTON.*?>.*?<\/BUTTON>/ /gi;
    $rec =~ s/<LABEL.*?>.*?<\/LABEL>/ /gi;
    $rec =~ s/<SELECT.*?>.*?<\/SELECT>/ /gi;
    $rec =~ s/<OPT.*?>/ /gi;
    $rec =~ s/<\/OPTGROUP>/ /gi;
    $rec =~ s/<TEXTAREA.*?>.*?<\/TEXTAREA>/ /gi;
    $rec =~ s/<FIELDSET>.*?<\/FIELDSET>/ /gi;
    $rec =~ s/<LEGEND.*?>.*?<\/LEGEND>/ /gi;
    $rec =~ s/<IFRAME.*?>//gi;
    $rec =~ s/<FRAME.*?>//gi;
    $rec =~ s/<\/FRAMESET>//gi;
    $rec =~ s/<\/IFRAME>//gi;
    $rec =~ s/<EM>//gi;
    $rec =~ s/<\/EM>//gi;
    $rec =~ s/<BLINK>//gi;
    $rec =~ s/<\/BLINK>//gi;
    $rec =~ s/<SMALL>//gi;
    $rec =~ s/<\/SMALL>//gi;
    $rec =~ s/<BIG>//gi;
    $rec =~ s/<\/BIG>//gi;
    $rec =~ s/<STRONG>//gi;
    $rec =~ s/<\/STRONG>//gi;
    $rec =~ s/<STRIKE>//gi;
    $rec =~ s/<\/STRIKE>//gi;
    $rec =~ s/<S>//gi;
    $rec =~ s/<\/S>//gi;
    $rec =~ s/<TT>//gi;
    $rec =~ s/<\/TT>//gi;
    $rec =~ s/<PRE.*?>//gi;
    $rec =~ s/<\/PRE>//gi;
    $rec =~ s/<CENTER>//gi;
    $rec =~ s/<\/CENTER>//gi;
    $rec =~ s/<MARQUEE>//gi;
    $rec =~ s/<\/MARQUEE>//gi;
    $rec =~ s/<CITE>//gi;
    $rec =~ s/<\/CITE>//gi;
    $rec =~ s/<CODE>//gi;
    $rec =~ s/<\/CODE>//gi;
    $rec =~ s/<SAMP>//gi;
    $rec =~ s/<\/SAMP>//gi;
    $rec =~ s/<KBD>//gi;
    $rec =~ s/<\/KBD>//gi;
    $rec =~ s/<VAR>//gi;
    $rec =~ s/<\/VAR>//gi;
    $rec =~ s/<DFN>//gi;
    $rec =~ s/<\/DFN>//gi;
    $rec =~ s/<\/BODY>//gi;
    $rec =~ s/<\/HEAD>//gi;
    $rec =~ s/<\/HTML>//gi;
    $rec =~ s/<DIV.*?>//gi;
    $rec =~ s/<\/DIV>//gi;
    $rec =~ s/<SPAN.*?>//gi;
    $rec =~ s/<\/SPAN>//gi;
    $rec =~ s/<INS.*?>//gi;
    $rec =~ s/<\/INS>//gi;
    $rec =~ s/<DEL.*?>//gi;
    $rec =~ s/<\/DEL>//gi;
    $rec =~ s/<ACRONYM>//gi;
    $rec =~ s/<\/ACRONYM>//gi;
    $rec =~ s/<ABBR>//gi;
    $rec =~ s/<\/ABBR>//gi;
    $rec =~ s/<ADDRESS>//gi;
    $rec =~ s/<\/ADDRESS>//gi;
    $rec =~ s/<MULTICOL.*?>//gi;
    $rec =~ s/<\/MULTICOL>//gi;
    $rec =~ s/<SPACER.*?>//gi;
    $rec =~ s/<\/SPACER>//gi;
    $rec =~ s/<LAYER.*?>//gi;
    $rec =~ s/<\/LAYER>//gi;
    $rec =~ s/<ILAYER.*?>//gi;
    $rec =~ s/<\/ILAYER>//gi;
    $rec =~ s/<NOLAYER.*?>//gi;
    $rec =~ s/<\/NOLAYER>//gi;
    $rec =~ s/\&quot\;/\`/gi;

    return $rec;    
}


# This program searches the first page of hits of an AltaVista search
#
# Written by
#         James Lambert - 15/9/99
#
# additions by Adam Kilgarriff 26/10/99
# main changes: outputs an html page, tries to be clever about the object
# that is " a linguistic context"

# do we want ALL instances from a doc, or one per doc?
# do we want only instances with decent linguistic context, or all?
# what counts to keep -- maybe all of the following:
#    docs containing at least one good hit
#    docs with atleast one hit (with or without decent context)
#    timed-out items
#    no-such-page items
#    DUPLICATES --
#       what do we check for duplicates -- probably, just the context
#         we shall be showing the lexicographer
#       how do we count duplicates -- eg if 2 docs are duplicates of
#         each other but each have 2 good instances... do we show
#         one of the good instances from each... how many duplicates
#         do we then record?
#  "documnets found" and "word count" from altavista response

# we should also log "time taken" and put it in the output page

# do we carry on paging through, looking at *all* hits, or do we set 
# some limit, eg first 200?  (NEXT JOB is to get this to page thorugh 
# the "next" web pages - I've started but it's not working yet

#output for multiple search terms needs multiple html files - haven't done that  

# good context is the number of words of context required to make a hit 
# lexicographically useful - should probably be 3 or more
# contextsize - eg how many characters fit in half (or less) of 
# a netscape screen

#sub find_next_lists 
#{
#    foreach $line (@_)
#    {
#	if (/^\&nbsp\;<a href\=\"cgi\-bin\/query\?/)
#	{
#	    for $nextref (split /<a href=\"/)
#	    {
#		$nextref =~ s/\">([0-9]+)<\/a> //;
#		$nextref = "http://altavista.yellowpages.com.au/" . $nextref . "\"";
#		if ($1 > 1)
#		{
#		    print "\n\n<!-- now I look for $1 th page -->\n\n";
#		    push @next_list, $nextref;
#		}
#	    }
#	    return @next_list;
#	}
#    }
#}

close(TESTFILE);