#!/usr/bin/perl
use CGI;
use LWP;
use HTML::TreeBuilder::XPath qw();
use WWW::Wikipedia;
use Text::MeCab;
use HTML::Entities;
my $cgi = new CGI;
print $cgi->header(-charset=>"utf-8");
my $title = "Association Search";
#&mecab('perl', 'q', 'r');
#exit;
if(defined $cgi->param('action')){
&mecab($cgi->param('q'), $cgi->param('qid'), $cgi->param('rid'));
} else {
&print_index;
}
sub mecab($$$){
my $q = shift;
my $qid = shift;
my $rid = shift;
my %querys = ();
return if $q eq "";
my $text = "";
my $wiki = WWW::Wikipedia->new();
foreach my $query(split(/ /,$q)){
$querys{$query} = 1;
my $result = $wiki->search( $query );
if ( $result ) {
#$result->language( 'ja' );
$text .= $result->text();
}
# list any related items we can look up
# print join( "\n", $result->related() );
}
my $ua = LWP::UserAgent->new;
$ua->agent('Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0;');
my $request = HTTP::Request->new('GET' , 'http://www.google.co.jp/search?num=100&q=' . $q);
my $response = $ua->request($request);
my $tree = HTML::TreeBuilder::XPath->new;
$tree->parse($response->content);
for my $node ($tree->findnodes('//h3[@class = "r"]')) {
$text .= $_->as_text() for @{ $node->content_array_ref };
}
for my $node ($tree->findnodes('//span[@class = "st"]')) {
$text .= $node->as_text;
}
my $m = Text::MeCab->new();
my $n = $m->parse($text);
my %keywords = {};
while ($n) {
my @features = split(/,/, $n->feature);
# 国 名詞,普通名詞,*,*,国,こく,* 43138243
# reference /home/dot/work/kurobox/work/cbot/module/relation.pl
if(@features[0] eq "名詞"){
if(exists $keywords{$n->surface}){
$keywords{$n->surface}++;
} else {
$keywords{$n->surface} = 1;
}
}
$n = $n->next;
}
my $i = 0;
print '';
}
sub escape($){
my $str = shift;
return encode_entities($str, q{&<>"'});
#return encode_entities($str, "<>&");
}
sub print_index(){
print <<"_HTML_";
Developers Convention - $title
_HTML_
}
