#!/usr/bin/perl -w
#
# file:         build/htmlpostprocess
# purpose:      postprocess html for my homepage
# created:      pasha oct 29 2000
# modified:     pasha sep 24 2007
# modification: border="0" in <wikipedia />
# synopsis:     htmlpostprocess --webroot=<webserver_root> 
#                               --pagedir=<location_of_this_page> 
#               < input > output
# pending:      strip html comments
# notes:
# 
# it does 12 things:
#  1. inserts width and height attributes of <img> tag
#  2. adds type="square" to <ul> tag
#  3. converts translit to cyrillic inside <koi8> and <win> tags (my own invention)
#  4. substitute <down date="" /> with the tumbdown icon
#  5. converts <amazon asin="" /> tag to amazon associates link
#  6. converts <ozon id="" cover="" /> tag to ozon associates link
#  7. converts <exclam /> tag to a red exclamation mark sign
#  8. converts <wikipedia /> tag to a small wikipedia image
#  9. converts <rss url=""> tag to link to rss feed with xml icon
# 10. converts <mr mr="" /> tag to MR link
# 11. converts <zbl an=""></zbl> tag to Zbl link
# 12. converts <gprint id="" /> tag to a google print link
#

use strict;
use Getopt::Long;

use pasha::common;


######### define subclass of HTML::Filter ###########

package homepagefilter;
require HTML::Filter;
use vars qw(@ISA);
@ISA = qw(HTML::Filter);

use GD;
use translit2koi8;

use constant GIF_HEADER_STRUCT => 'C6S2';   # 6 characters (gif version), and
                                            # 2 unsigned shorts (width & length)
use constant GIF_HEADER_LENGTH => length (pack (GIF_HEADER_STRUCT, ()));


# constructor
sub new
{
    my $class = shift;
    my $this = $class->SUPER::new (@_);
	for (qw(inside_koi8 inside_win))
	{
		$this->{$_} = 0;
	}
    return ($this);
}


sub start
{
    my ($this, $tag, $attr, undef, $orig) = @_;
    for ($tag)
    {
		/^img$/ && do
		{
		  ######### process <img> tag ###########
		  # don't process remote images
		  if ((exists ($attr->{'src'})) && ($attr->{'src'} !~ /^http:\/\//) &&
			  ((! exists ($attr->{'width'})) || (! exists ($attr->{height}))))
			{
			  my ($width, $height) = get_size (web2path ($attr->{src}));
			  my $add = '';
			  if (! exists ($attr->{width}))
				{
				  $add .= " width=\"$width\"";
				}
			  if (! exists ($attr->{height}))
				{
				  $add .= " height=\"$height\"";
				}
			  $orig =~ s/(\/|)>$/${add} $1>/;  
	        }
		  last;
        };

	    /^ul$/ && do
		{
			######### process <ul> tag ############
			if (! exists $attr->{type})
			{
				$orig =~ s/>$/ type="square">/o;
			}
			last;
		};

	    /^koi8$/ && do
		{
			$this->{'inside_koi8'}++;
			$orig = '';
			last;
		};

	    /^win$/ && do
		{
			$this->{'inside_win'}++;
			$orig = '';
			last;
		};

	    /^down$/ && do
		  {
			$orig = 
			  '<small>[<img src="/image/md09_thumbdown.gif" alt="down" align="middle" /> as of ' . 
				$attr->{'date'} . ']</small>';
		  };

	    /^amazon$/ && do
		  {
			#$orig = '<table><tr><td>a<br />m<br />a<br />z<br />o<br />n<br /></td><td><a href="http://www.amazon.com/gp/associates/link-types/marketplace.html?t=pashazusmansh-20&asin=' . $attr->{'asin'} . 
			$orig = '<table><tr><td>a<br />m<br />a<br />z<br />o<br />n<br /></td><td><a href="http://www.amazon.com/exec/obidos/ASIN/' . 
			  $attr->{'asin'} . '/ref=nosim/pashazusmansh-20"><img src="http://images.amazon.com/images/P/' .
				$attr->{'asin'} . 
				  '.01.MZZZZZZZ.jpg" border="0" alt="amazon cover" align="middle" /></a></td></tr></table>';
			last;
		};

	    /^amazontxt$/ && do
		{
			$orig = '<a href="http://www.amazon.com/exec/obidos/ASIN/' . $attr->{'asin'} . 
'/ref=nosim/pashazusmansh-20">';
			last;
		};

	    /^ozon$/ && do
		{
			$orig = '<table><tr><td>o<br />z<br />o<br />n<br /></td><td><a href="http://www.ozon.ru/context/detail/id/' . $attr->{'id'} . '/?partner=justpasha"><img src="http://www.ozon.ru/multimedia/books_covers/small/' . $attr->{'cover'} . '" border="0" alt="ozon cover" /></a></td></tr></table>';
			last;
		};

		/^exclam$/ && do
		  {
			$orig = '<b><font color="red">[!]</font></b> ';
			last;
		  };

		/^wikipedia$/ && do
		  {
			$orig = '<img src="/image/wikipedia.png" align="middle" border="0" />';
			last;
		  };

		/^rss$/ && do
		  {
			die ('no href attribute for rss tag') if (! defined ($attr->{'href'}));
			$orig = '&nbsp; <a href="' . $attr->{'href'} . 
			  '"><img src="/math/img/rss.gif" align="middle" border="0" /></a>';
			last;
		  };

		/^mr$/ && do
		  {
			(my $hash_to_23 = $attr->{'mr'}) =~ s/#/%23/;;
			$orig = 
			  "<a href=\"http://www.ams.org/mathscinet-getitem?mr=${hash_to_23}\">MR " . 
				$attr->{'mr'} . '</a>';
			last;
		  };

		/^zbl/ && do
		  {
			# old:
			#$orig = '<a href="http://www.emis.de/cgi-bin/zmen/ZMATH/en/full.html?type=html&an=' .
			#  $attr->{'an'} . '&format=complete">';
			# new:
			$orig = '<a href="http://www.zentralblatt-math.org/zmath/en/search/?q=an:%22' .
			  $attr->{'an'} . '%22&format=complete">';
			last;
		  };

		/^gprint/ && do
		  {
			# it could be in two forms: <gprint vid="" /> and <gprint id="" />
			my $set = 0;
			for (qw(vid id))
			  {
				if (defined ($attr->{$_}))
				  {
					$orig = "<a href=\"http://books.google.com/books?${_}=" . $attr->{$_} . 
					  '">google books</a>';
					$set = 1;
					last;
				  }
			  }
			if ($set == 0)
			  {
				::me_die ('<gprint> tag without either vid of id attribute');
			  }
			last;
		  };
		
    } # end of switch thru the custom tags

    $this->output ($orig);
} # end of start()
 


sub end
{
    my $this = shift;
	for ($_[0])
	{
		/^koi8$/ && do
		{
			$this->{'inside_koi8'}--;
			last;
		};
		/^win$/ && do
		{
			$this->{'inside_win'}--;
			last;
		};
		/^amazontxt$/ && do
		{
			$this->output ('</a>');
			last;
		};
		/^zbl/ && do
		  {
			$this->output ('</a>');
			last;
		  };
		$this->SUPER::end (@_);
	}
}



sub text
  {
    my ($this, $text) = @_;
    if ($this->{'inside_koi8'} > 0)
	  {
		$text = translit2koi8 ($text);
	  }
	elsif ($this->{'inside_win'} > 0)
	  {
		$text = translit2win ($text);
	  }
    $this->SUPER::text ($text);
  }



# note that the following functions are not blessed
# into this class, hence first argument isn't class reference

# get image size
# input  - image filename
# output - ($width, $height) list
sub get_size ($)
{
    my ($width, $height);
    if (::get_ext ($_[0]) eq 'gif')
    {
	# GD like any other good modern software does not support gifs,
	# but we, as the majority of other web suckers, still use them,
	# so do it ourselves

	my $gif;
	open (F, "<$_[0]") || ::me_die ("unable open $_[0]");
	read (F, $gif, GIF_HEADER_LENGTH) || 
	    me_die ('unable read ' . GIF_HEADER_LENGTH . " bytes from $_[0]"); 
	close (F);
	if (! $gif =~ /^GIF/o)
	{
	    ::me_die ("$_[0] is not a gif file");
	}
	($width, $height) = (unpack (GIF_HEADER_STRUCT, $gif))[6,7];
    }
    else
    {
	my $gd = new GD::Image->new ($_[0]) || ::me_die ("error opening $_[0] with GD");
	($width, $height) = $gd->getBounds();
    }

    if ((! defined ($width)) || (! defined ($height)))
    {
	::me_die ("unable get size of $_[0]");
    }

    return ($width, $height);
} # end of get_size()



# transforms webserver path to real path
sub web2path ($)
{
    return (($_[0] =~ /^\//o ? $::WEBROOT : $::PAGEDIR . '/') . $_[0]);
}

######### end of subclass of HTML::Filter ###########



package main;

use vars qw($WEBROOT $PAGEDIR);
local $WEBROOT;   # we need them to be visible in homepagefilter
local $PAGEDIR;

GetOptions (
	    'webroot=s' => \$WEBROOT,
	    'pagedir=s' => \$PAGEDIR
	    );
if ((! defined ($WEBROOT)) || (! defined ($PAGEDIR)))
{
    print (STDERR 'Usage: ' . ME . " --webroot=<webserver_root> --pagedir=<location_of_this_page> < input > output\n");
    exit (1);
}

homepagefilter->new->parse_file (*STDIN) || 
    me_die ('error from HTML::Filter parsing STDIN');



__END__