あるユーザのツイートを抽出してマルコフ連鎖によって文章を生成

markov_tweet.pl

!/usr/bin/perl

use strict;
use warnings;
use Text::MeCab;
use Algorithm::MarkovChain;
use List::Util qw(shuffle);
use Net::Twitter::Lite;
use Encode;
use utf8;
#binmode(STDOUT,":utf8");

my $t = Net::Twitter::Lite->new(
        consumer_key=>'こんしゅーまーきー',
        consumer_secret=>'こんしゅーまーしーくれっと',
        access_token=>'あくせすとーくん',
        access_token_secret=>'あくせすとーくんしーくれっと',
        legacy_lists_api=>1
);

my $status = $t->user_timeline({screen_name => '元にしたいユーザーのID',count => 200});
my @tweets ;
my $tmp;
foreach (@$status){
        $tmp = $_->{'text'};
        #print $tmp."\n";
        push(@tweets,$tmp);
}

my $tweet_connected = "";
@tweets = shuffle @tweets;
foreach(@tweets){
        $tweet_connected .= "$_ ";
}
$tweet_connected =~ s/@\S*//gm;
$tweet_connected =~ s/http:\S*//gm;
#print $tweet_connected;

my $mecab = Text::MeCab->new({});
my @symbols;
for(my $node = $mecab->parse($tweet_connected); $node; $node = $node->next){
        #print $node->surface;
        #print "\n";
        push @symbols , $node->surface;
}

my $chain = Algorithm::MarkovChain->new;
$chain->seed(
        symbols=>\@symbols,
        longest=>7,
);

my @newness = $chain->spew(
        #length=>int(rand 30)+1,
        length=>10,
        strict_start=>1,
);

my $post;
for my $chunk(@newness){
        print $chunk;
        $post = $post.$chunk;
        #print "\n\n";
}

$post = decode('utf-8',$post);
$t->update({status=>$post});