--- lib/Grep/Source.pm 2007/02/23 11:48:39 73 +++ lib/Grep/Source.pm 2007/03/14 20:02:19 110 @@ -7,8 +7,14 @@ use Carp qw/verbose/; use Module::Pluggable search_path => 'Grep::Source', sub_name => 'sources', require => 1; -use base qw(Class::Accessor); -Grep::Source->mk_accessors( qw(feed uri q new_items collection) ); +use base qw(Class::Accessor Jifty::Object); +Grep::Source->mk_accessors( qw(feed uri q new_items collection search_obj) ); + +use HTML::TreeBuilder; +use WWW::Mechanize; +use XML::Feed; +use URI; +use HTML::ResolveLink; use Data::Dump qw/dump/; @@ -95,16 +101,19 @@ $self->uri( $uri ); - Jifty->log->info( $message ); + $self->log->info( $message ); $self->collection( Grep::Model::ItemCollection->new() ); my $class = $self->feed->source || 'Grep::Source::Feed'; - Jifty->log->debug("using $class"); + $self->log->debug("using $class"); + + $self->search_obj( Grep::Search->new() ); + $self->log->debug("created " . $self->search_obj); $class->fetch( $self ); - Grep::Search->finish if $self->new_items; + $self->search_obj->finish; return $self->collection; } @@ -123,20 +132,28 @@ sub add_record { my $self = shift; + $self->log->confess("no search_obj") unless ($self->search_obj); + my $i = Grep::Model::Item->new(); - my ($ok,$msg) = $i->load_or_create( @_ ); + my $rec = {@_}; + + $self->log->debug("resolving links using base ", $rec->{link}); + my $resolver = HTML::ResolveLink->new( base => $rec->{link} ); + $rec->{content} = $resolver->resolve( $rec->{content} ); + + my ($ok,$msg) = $i->load_or_create( %$rec ); $msg ||= ''; if ( $ok ) { - Jifty->log->debug("item ", $i->id, ": $msg"); + $self->log->debug("item ", $i->id, ": $msg"); $self->collection->add_record( $i ); # is new record? if ( $msg !~ m/^Found/ ) { - Grep::Search->add( $i ); - $self->new_items( $self->new_items + 1 ); + $self->search_obj->add( $i ); + $self->new_items( ( $self->new_items || 0 ) + 1 ); } } else { warn "can't add entry ", dump( @_ ), "\n"; @@ -157,11 +174,151 @@ my $content = shift or die "no content?"; foreach my $s ( $self->sources ) { - Jifty->log->debug("testing source class $s"); - if ($s->can('content_have') && $s->content_have( $content ) ) { - Jifty->log->debug("${s}->content_have succesful"); - return "$s"; + $self->log->debug("testing source class $s"); + if ( $s->can('content_have') ) { + my $regex = $s->content_have( $content ) or + die "${s}->content_have didn't return anything"; + die "${s}->content_have didn't return regex but ", dump( $regex ), " ref ", ref( $regex ) + unless ( ref($regex) eq 'Regexp' ); + if ( $content =~ $regex ) { + $self->log->debug("${s}->content_have succesful"); + return $s; + } + } + } +} + +=head2 scrape + +Create semi-complex L rules to scrape page + + +=cut + +sub scrape { + my $self = shift; + + my $args = {@_}; + + $self->log->debug("scrape with args ",dump($args)); + + my ($feed,$uri,$q) = ($self->feed, $self->uri,$self->q); + die "no uri" unless ($uri); + die "feed is not a Grep::Model::Feed but ", ref $feed unless $feed->isa('Grep::Model::Feed'); + + sub mech_warn { + my $m = shift || return; + warn $m; + } + + my $mech = WWW::Mechanize->new( + cookie_jar => {}, + onwarn => \&mech_warn, + onerror => \&mech_warn, + ); + + $mech->get( $uri ); + + $self->save( 'get.html', $mech->content ); + + if ( my $form = $args->{submit_form} ) { + $self->log->debug("submit form on $uri with ", dump( $form )); + $mech->submit_form( %$form ) or die "can't submit form ", dump( $form ); + $self->save( 'submit.html', $mech->content ); + } + + $self->log->debug("parse result page"); + + my $tree = HTML::TreeBuilder->new or die "can't create html tree"; + $tree->parse( $mech->content ) or die "can't parse fetched content"; + + die "wrapper doesn't have 3 elements but ", $#{ $args->{wrapper} } unless ( $#{ $args->{wrapper} } == 2 ); + my ( $el,$attr,$value ) = @{ $args->{wrapper} }; + + $self->log->debug("looking for <$el $attr=\"$value\">"); + + my $div = $tree->look_down( '_tag', $el, sub { + ( $_[0]->attr( $attr ) || '' ) eq $value; + }); + + if ( ! $div ) { + warn "can't find results wrapper <$el $attr=\"$value\">"; + return; + } + + my $max = 15; + my $nr = 1; + + my $base_uri = $uri; + $base_uri =~ s!\?.*$!!; + + foreach my $dt ( $div->look_down( '_tag', $args->{results} ) ) { + my $a = $dt->look_down( '_tag', 'a', sub { $_[0]->attr('href') } ); + if ( $a ) { + + my $href = $a->attr('href') or die "can't find href inside <", $args->{results}, ">"; + my $page_uri = URI->new_abs( $a->attr('href'), $base_uri ); + $page_uri->query( undef ); + $page_uri = $page_uri->canonical; + + $self->log->debug("fetching page: ",$a->as_text," from $page_uri"); + if ( $mech->follow_link( url => $a->attr('href') ) ) { + + $self->save( "page-${nr}.html", $mech->content ); + + my $page_tree = HTML::TreeBuilder->new or die "can't create page tree"; + $page_tree->parse( $mech->content ) or die "can't parse page at $page_uri"; + + ( $el,$attr,$value ) = @{ $args->{scrape} }; + $div = $page_tree->look_down( '_tag', $el, sub { ( $_[0]->attr( $attr ) || '' ) eq $value } ); + + die "can't find <$el $attr=\"$value\">" unless ($div); + + $self->add_record( + in_feed => $feed, + title => $mech->title, + link => $page_uri, + content => $div->as_HTML, +# summary => +# category => +# author => +# issued => +# modified => + ); + + $mech->back; + $page_tree->delete; + + } else { + warn "can't follow uri $page_uri: $!\n"; + } } + + last if ($nr == $max); + $nr++; + } + + $tree->delete; # clear memory! + +} + +=head2 save + + save( 'name', $content ); + +Save dumps into C if writable + +=cut + +sub save { + my $self = shift; + my ( $file, $content ) = @_; + return unless ( defined($file) && defined($content) ); + if ( -w '/tmp/grep' ) { + open(my $f, '>', "/tmp/grep/$file") or die "can't open $file: $!"; + print $f $content or die "can't write to $file: $!"; + close $f or die "can't close $file: $!"; + $self->log->debug("saved $file ",length($content)," bytes"); } }