DELETE THIS PREFIX BEFORE RUNNING!   #! /usr/bin/perl -w

use strict;

# Hack to concatenate all of the ITS RDF Subgroup minutes into one big page,
# to assist in finding things.   Runs on Linux.  

# NOTE: The last attempted minutes download will result in 'ERROR 404: 
# Not Found' message.  This is normal.
# 
# Run as:
#    ./hl7-its-fhir-rdf-minutes.perl > /tmp/all-minutes.html
#
# Author: David Booth
# Date: 9/14/2015
# License: Apache 2.0

my $startYear = 2014;
my $yearListUrl = "http://wiki.hl7.org/index.php?title=Category:ITS_RDF_Minutes_2015";
my $reverse = 1;	# Reverse chronological order?

my $tmp = "/tmp/its-rdf-minutes-tmp.html";
my $isFirstPage = 1;
my $year = $startYear;
my @bodies = ();
my $beforeBodies = "";
my $afterBodies = "";
while (1) {
	die if $year > 2035;
	my $yurl = $yearListUrl;
	$yurl =~ s/2015/$year/;
	my $yresponse = `wget -k -O $tmp '$yurl' ; cat $tmp`;
	last if $yresponse =~ m|There is currently no text in this page|i;
	last if !$yresponse;
	$yresponse =~ m|Sub category of| || die "Unknown response from $yurl\n ";
	my @urls = split(/\n/, &Pipe($yresponse, '/home/dbooth/bin/grepurls'));
	@urls = grep {m/^http(s?)\:/} @urls;
	@urls = grep {m/ITS_RDF_Concall_Minutes/} @urls;
	# print "URLS: @urls\n";
	# exit 0;
	foreach my $url (@urls) {
		my $body = "";
		warn "Getting $url ...\n";
		my $page = `wget -k -O $tmp '$url' ; cat $tmp`;
		$page =~ m|ITS RDF .* Minutes|i || die "Unknown response from $url : \n$page\n ";

		# Most wiki minutes pages are stubs that point to
		# W3C minutes pages.
		# Meeting log: <a rel="nofollow" class="external free" href="http://www.w3.org/2014/11/04-hcls-minutes.html">http://www.w3.org/2014/11/04-hcls-minutes.html</a>
		if ($page =~ m/Meeting log:\s*<a (.*?)href="(http(s?)\:([^\"\<\>]+))"/ims) {
			$url = $2;
			$page = `wget -k -O $tmp '$url' ; cat $tmp`;
			$page =~ m|Attendees|i || die "Unknown response from $url : \n$page\n ";
			}
		$page =~ m/\A.*\<body\b(.*?)\>/ims || die;
		my $beforeBody = $&;
		my $afterBody = $';
		if ($isFirstPage) {
			# print "$beforeBody\n";
			$beforeBodies = "$beforeBody\n";
			}
		$page = $afterBody;
		$page =~ s/\<\/body\>.*\Z//ims || die;
		# print "<!-- =============================================  -->\n";
		$body .= "<!-- =============================================  -->\n";
		# print "<br>\n<hr>\nMinutes downloaded from: <a href=\"$url\">$url</a><br>\n";
		$body .= "<br>\n<hr>\nMinutes downloaded from: <a href=\"$url\">$url</a><br>\n";
		# print "$page\n";
		$body .= "$page\n";
		push(@bodies, $body);
		$isFirstPage = 0;
		}
	$year++;
	}
if (!$isFirstPage) {
	# print "\n</body>\n</html>\n";
	$afterBodies = "\n</body>\n</html>\n";
	}

if ($reverse) {
	@bodies = reverse @bodies;
	}
my $allBodies = join("", @bodies);
print $beforeBodies;
print $allBodies;
print $afterBodies;

# unlink($tmp);
exit 0;


################ Pipe ###############
sub Pipe
# Perl code snippet to pipe a string (as stdin) through a command,
# returning the stdout of the command either as a list or a string.
{
@_ == 2 || die;
my ($s, $cmd) = @_;
my $tmp = "/tmp/pipe.$$.tmp";
open(my $fh,">$tmp") || die "ERROR: Cannot open for write: $tmp
";
print $fh $s;
close($fh);
my $fullcmd = "cat $tmp | $cmd";
if (wantarray)
        {
        my @result = `$fullcmd`;
        unlink($tmp);
        return(@result);
        }
else
        {
        my $result = `$fullcmd`;
        unlink($tmp);
        return($result);
        }
}