/[Grep]/lib/Grep/Source.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 /lib/Grep/Source.pm

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

revision 88 by dpavlin, Fri Feb 23 21:52:29 2007 UTC revision 96 by dpavlin, Sat Feb 24 11:56:18 2007 UTC
# Line 14  use HTML::TreeBuilder; Line 14  use HTML::TreeBuilder;
14  use WWW::Mechanize;  use WWW::Mechanize;
15  use XML::Feed;  use XML::Feed;
16  use URI;  use URI;
17    use HTML::ResolveLink;
18    
19  use Data::Dump qw/dump/;  use Data::Dump qw/dump/;
20    
# Line 107  sub search { Line 108  sub search {
108          my $class = $self->feed->source || 'Grep::Source::Feed';          my $class = $self->feed->source || 'Grep::Source::Feed';
109          Jifty->log->debug("using $class");          Jifty->log->debug("using $class");
110    
111          $class->fetch( $self );          my $parent = $self;
112            $class->fetch( $parent );
113            undef $parent;
114    
115          Grep::Search->finish if $self->new_items;          Grep::Search->finish if $self->new_items;
116    
# Line 130  sub add_record { Line 133  sub add_record {
133    
134          my $i = Grep::Model::Item->new();          my $i = Grep::Model::Item->new();
135    
136          my ($ok,$msg) = $i->load_or_create( @_ );          my $rec = {@_};
137    
138            warn "resolving links";
139            my $resolver = HTML::ResolveLink->new( base => $rec->{link} );
140            $rec->{content} = $resolver->resolve( $rec->{content} );
141    
142            my ($ok,$msg) = $i->load_or_create( %$rec );
143    
144          $msg ||= '';          $msg ||= '';
145    
# Line 194  sub scrape { Line 203  sub scrape {
203          die "no uri" unless ($uri);          die "no uri" unless ($uri);
204          die "feed is not a Grep::Model::Feed but ", ref $feed unless $feed->isa('Grep::Model::Feed');          die "feed is not a Grep::Model::Feed but ", ref $feed unless $feed->isa('Grep::Model::Feed');
205    
206          my $mech = WWW::Mechanize->new();          sub mech_warn {
207                    my $m = shift || return;
208                    warn $m;
209            }
210    
211            my $mech = WWW::Mechanize->new(
212                    cookie_jar => {},
213                    onwarn => \&mech_warn,
214                    onerror => \&mech_warn,
215            );
216    
217          $mech->get( $uri );          $mech->get( $uri );
218    
219          $self->save( 'get.html', $mech->content );          $self->save( 'get.html', $mech->content );
220    
221          if ( $args->{submit_form} ) {          if ( my $form = $args->{submit_form} ) {
222                  warn "submit form on $uri\n";                  warn "submit form on $uri with ", dump( $form ),"\n";
223                  $mech->submit_form( %{ $args->{submit_form} } ) or die "can't submit form";                  $mech->submit_form( %$form ) or die "can't submit form ", dump( $form );
224                  $self->save( 'submit.html', $mech->content );                  $self->save( 'submit.html', $mech->content );
225          }          }
226    
# Line 249  sub scrape { Line 267  sub scrape {
267                                  my $page_tree = HTML::TreeBuilder->new or die "can't create page tree";                                  my $page_tree = HTML::TreeBuilder->new or die "can't create page tree";
268                                  $page_tree->parse( $mech->content ) or die "can't parse page at $page_uri";                                  $page_tree->parse( $mech->content ) or die "can't parse page at $page_uri";
269    
270                                  my ( $el,$attr,$value ) = @{ $args->{scrape} };                                  ( $el,$attr,$value ) = @{ $args->{scrape} };
271                                  my $div = $page_tree->look_down( '_tag', $el, sub { ( $_[0]->attr( $attr ) || '' ) eq $value } );                                  $div = $page_tree->look_down( '_tag', $el, sub { ( $_[0]->attr( $attr ) || '' ) eq $value } );
272    
273                                  die "can't find <$el $attr=\"$value\">" unless ($div);                                  die "can't find <$el $attr=\"$value\">" unless ($div);
274    

Legend:
Removed from v.88  
changed lines
  Added in v.96

  ViewVC Help
Powered by ViewVC 1.1.26