WAIT/Parse/HTML.pm

#!/usr/bin/perl
#                              -*- Mode: Perl -*- 
# $Basename: HTML.pm $
# $Revision: 1.2 $
# Author          : Ulrich Pfeifer with Andreas König
# Created On      : Sat Nov 1 1997
# Last Modified By: Ulrich Pfeifer
# Last Modified On: Wed Nov  5 16:48:17 1997
# Language        : CPerl
# Update Count    : 1
# Status          : Unknown, Use with caution!
# 
# (C) Copyright 1997, Ulrich Pfeifer, all rights reserved.
# 
# 

package WAIT::Parse::HTML;
use vars qw(@ISA);
require HTML::Parse;
require HTML::FormatText;
use HTML::Entities qw(decode_entities);
@ISA = qw(WAIT::Parse::Base);


sub split {
  my ($self, $html_source) = @_;

  my (undef,$title) = $html_source =~ /<(title|h1|h2|h3|h4)[^>]*>(.*?)<\/\1\s*>/si;

  my $html = HTML::Parse::parse_html($html_source);
  my $formatter = HTML::FormatText->new;

  {
   'text',  $formatter->format($html),
   'title', $title || 'no title',
  };
}

sub tag {
  my ($self, $html_source) = @_;

  $html_source =~ tr/\r/\n/;

  my ($pre,$title,$body) 
      = $html_source =~ /^(.*?<title\s*>)(.*?)(<\/title\s*>.+)/si;

  $pre .= '';
  $title .= '';
  $body .= '';

  (
   {'text'  => 1},  decode_entities($pre),
   {'title' => 2},  decode_entities($title),
   {'text'  => 1},  decode_entities($body),
  );
}

1;
1	#!/usr/bin/perl
2	# -- Mode: Perl --
3	# $Basename: HTML.pm $
4	# $Revision: 1.2 $
5	# Author : Ulrich Pfeifer with Andreas König
6	# Created On : Sat Nov 1 1997
7	# Last Modified By: Ulrich Pfeifer
8	# Last Modified On: Wed Nov 5 16:48:17 1997
9	# Language : CPerl
10	# Update Count : 1
11	# Status : Unknown, Use with caution!
12	#
13	# (C) Copyright 1997, Ulrich Pfeifer, all rights reserved.
14	#
15	#
16
17	package WAIT::Parse::HTML;
18	use vars qw(@ISA);
19	require HTML::Parse;
20	require HTML::FormatText;
21	use HTML::Entities qw(decode_entities);
22	@ISA = qw(WAIT::Parse::Base);
23
24
25	sub split {
26	my ($self, $html_source) = @_;
27
28	my (undef,$title) = $html_source =~ /<(title\|h1\|h2\|h3\|h4)[^>]>(.?)<\/\1\s*>/si;
29
30	my $html = HTML::Parse::parse_html($html_source);
31	my $formatter = HTML::FormatText->new;
32
33	{
34	'text', $formatter->format($html),
35	'title', $title \|\| 'no title',
36	};
37	}
38
39	sub tag {
40	my ($self, $html_source) = @_;
41
42	$html_source =~ tr/\r/\n/;
43
44	my ($pre,$title,$body)
45	= $html_source =~ /^(.?<title\s>)(.?)(<\/title\s>.+)/si;
46
47	$pre .= '';
48	$title .= '';
49	$body .= '';
50
51	(
52	{'text' => 1}, decode_entities($pre),
53	{'title' => 2}, decode_entities($title),
54	{'text' => 1}, decode_entities($body),
55	);
56	}
57
58	1;