/[swish]/trunk/spider/filter.pm
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Diff of /trunk/spider/filter.pm

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 56 by dpavlin, Fri Jan 23 13:10:40 2004 UTC revision 74 by dpavlin, Wed Apr 7 12:54:21 2004 UTC
# Line 1  Line 1 
1  sub filter {  sub filter {
2          my $contents = shift || return;          my $contents = shift || return;
3    
4            my $verbose = 0;
5            
6          # if you don't want content to be indexed, include it in          # if you don't want content to be indexed, include it in
7          # <noindex> foobar </noindex> tags or surround it with comments          # <noindex> foobar </noindex> tags or surround it with comments
8          # <!-- noindex --> foobar <!-- /noindex -->          # <!-- noindex --> foobar <!-- /noindex -->
# Line 19  sub filter { Line 22  sub filter {
22          # remove comments between <html> and <head> texi2html inserts them          # remove comments between <html> and <head> texi2html inserts them
23          # there and swish can't find document title then (libxml or swish bug?)          # there and swish can't find document title then (libxml or swish bug?)
24          while ($contents =~ s/(<html>.*)<!--.*?-->(.*<head>)/$1$2/msi) { };          while ($contents =~ s/(<html>.*)<!--.*?-->(.*<head>)/$1$2/msi) { };
25            
26            # remove empty lines before/after <html>
27            $contents =~ s/^\s+(<html>)/$1/is;
28            $contents =~ s/(<\/html>)\s+$/$1/is;
29            # remove cr
30            $contents =~ s/\r//gs;
31    
32            # remove SQL Magazine header and footer
33            $contents =~ s/<!-- begin topnav area -->.+?<!-- end topnav area -->/<\/table>/is;
34            $contents =~ s/<!--Begin Footer-->.+?<\/table>/<\/table>/is;
35    
36          # remote TPJ left column          # remote TPJ left column
37          if ($contents =~ s,<!-- BEGIN LEFT SIDE BAR CELL -->.+?<!-- END LEFT SIDE BAR CELL -->,,isg) {          if ($contents =~ s,<!-- BEGIN LEFT SIDE BAR CELL -->.+?<!-- END LEFT SIDE BAR CELL -->,,isg) {
# Line 45  sub filter { Line 58  sub filter {
58    
59          if ($contents =~ m,<!--SafTocEntry="([^"]+)"-->,is) {          if ($contents =~ m,<!--SafTocEntry="([^"]+)"-->,is) {
60                  $new_title = $1;                  $new_title = $1;
61          } elsif ($contents =~ m,<h\d\sclass="docPartTitle"[^>]*>([^<]+)</h\d>,is) {                  print STDERR "using title '$new_title' from <!--SafTocEntry-->\n" if ($verbose);
62                  $new_title = $1;          } elsif ($contents =~ m,<(h\d)\s+class="docPartTitle"[^>]*>(.+?)</\1>,is) {
63          } elsif ($contents =~ m,<h\d\sclass="docChapterTitle"[^>]*>([^<]+)</h\d>,is) {                  $new_title = $2;
64                  $new_title = $1;                  print STDERR "using title '$new_title' from docPartTitle\n" if ($verbose);
65          } elsif ($contents =~ m,<h\d\sclass="docSection1Title"[^>]*>([^<]+)</h\d>,is) {          } elsif ($contents =~ m,<(h\d)\s+class="docChapterTitle"[^>]*>(.+?)</\1>,is) {
66                  $new_title = $1;                  $new_title = $2;
67                    print STDERR "using title '$new_title' from docChapterTitle\n" if ($verbose);
68            } elsif ($contents =~ m,<(h\d)\s+class="docSection1Title"[^>]*>(.+?)</\1>,is) {
69                    $new_title = $2;
70                    print STDERR "using title '$new_title' from docSection1Title\n" if ($verbose);
71            } elsif ($contents =~ m,<(h\d)\s+class="doc[^"]*Title"[^>]*>(.+?)</\1>,is) {
72                    $new_title = $2;
73                    print STDERR "using title '$new_title' from doc.+Title\n" if ($verbose);
74            } elsif ($contents =~ m,<(h\d)\s+class="chapter"[^>]*>(.+?)</\1>,is) {
75                    $new_title = $2;
76                    print STDERR "using title '$new_title' from chapter\n" if ($verbose);
77            } elsif ($contents =~ m,<(h\d)\s+class="sect1"[^>]*>(.+?)</\1>,is) {
78                    $new_title = $2;
79                    print STDERR "using title '$new_title' from sect1\n" if ($verbose);
80          } else {          } else {
81                  if ($contents =~ m,<title>([^<]+)</title>,is) {                  if ($contents =~ m,<title>([^<]+)</title>,is) {
82                          $new_title = $1;                          $new_title = $1;
83                            print STDERR "using title '$new_title' from <title>\n" if ($verbose);
84                  } elsif ($contents =~ m,<h\d[^>]*>([^<]+)</h\d>,is) {                  } elsif ($contents =~ m,<h\d[^>]*>([^<]+)</h\d>,is) {
85                          $new_title = $1;                          $new_title = $1;
86                            print STDERR "using title '$new_title' from <h_>\n" if ($verbose);
87                  }                  }
88          }          }
89    
90          if ($new_title) {          if ($new_title) {
91                    # nuke html in title
92                    $new_title =~ s/<br>\s+/: /gs;
93                    $new_title =~ s/<\/*[^>]+>//gs;
94            
95                  # check if new title is same as collection name                  # check if new title is same as collection name
96                  my ($a,$b) = ($new_title,$collection);                  my ($a,$b) = ($new_title,$collection);
97                  $a =~ s/([^a-zA-Z])+/ /gs;                  $a =~ s/([^a-zA-Z])+/ /gs;
98                  $b =~ s/([^a-zA-Z])+/ /gs;                  $b =~ s/([^a-zA-Z])+/ /gs;
99                  if ($a =~ m/$b/i) {                  if ($a =~ m/$b/i) {
100                          $new_title = $collection;                          $new_title = $collection;
101                            print STDERR "new_title and collection are same! [$new_title]\n" if ($verbose);
102                  } else {                  } else {
103                          $new_title = "$collection :: $new_title";                          $new_title = "$collection :: $new_title";
104                  }                  }
# Line 76  sub filter { Line 109  sub filter {
109    
110          $new_title =~ s/\s\s+/ /g;          $new_title =~ s/\s\s+/ /g;
111    
112          if ($contents =~ s,<title>(.*)</title>,$new_title,is) {          if ($contents =~ s,<title>(.*)</title>,<title>$new_title</title>,is) {
113                  print STDERR "replace title '$1' with '$new_title'\n";                  print STDERR "replace title '$1' with '$new_title'\n" if ($verbose);
114            } elsif ($contents =~ s,(<head>),$1<title>$new_title</title>,is) {
115                    print STDERR "adding title '$new_title' after <head>\n" if ($verbose);
116            
117            } elsif ($contents =~ s,(<html>),$1<head><title>$new_title</title></head>,is) {
118                    print STDERR "adding title '$new_title' after <html>\n" if ($verbose);
119    
120            } elsif ($contents =~ s,^,<title>$new_title</title>,)  {
121                    print STDERR "adding new title '$new_title'\n" if ($verbose);
122          } else {          } else {
123                  print STDERR "adding new title '$new_title'\n";                  print STDERR "WARNING: filter couldn't add new title '$new_title' anywhere!";
                 # try to insert after <head>, <html> or at top  
                 $contents =~ s,(<head>),$1<title>$new_title</title>,is ||  
                 $contents =~ s,(<html>),$1<title>$new_title</title>,is ||  
                 $contents =~ s,^,<title>$new_title</title>,;  
124          }          }
125    
126          return $contents;          return $contents;
Legend:
Removed from v.56  
changed lines
  Added in v.74
  ViewVC Help
Powered by ViewVC 1.1.26