/**
* Static method __cutDAG
*
* @param string $sentence # input sentence
* @param array $options # other options
*
* @return array $words
*/
public static function __cutDAG($sentence, $options = array())
{
$defaults = array('mode' => 'default');
$options = array_merge($defaults, $options);
$words = array();
$N = mb_strlen($sentence, 'UTF-8');
$i = 0;
$j = 0;
$p = self::$trie;
$DAG = array();
$word_c = array();
while ($i < $N) {
$c = mb_substr($sentence, $j, 1, 'UTF-8');
if (count($word_c) == 0) {
$next_word_key = $c;
} else {
$next_word_key = implode('.', $word_c) . '.' . $c;
}
if (self::$trie->exists($next_word_key)) {
array_push($word_c, $c);
$next_word_key_value = self::$trie->get($next_word_key);
if ($next_word_key_value == array("end" => "") || isset($next_word_key_value["end"]) || isset($next_word_key_value[0]["end"])) {
if (!isset($DAG[$i])) {
$DAG[$i] = array();
}
array_push($DAG[$i], $j);
}
$j += 1;
if ($j >= $N) {
$word_c = array();
$i += 1;
$j = $i;
}
} else {
$word_c = array();
$i += 1;
$j = $i;
}
}
for ($i = 0; $i < $N; $i++) {
if (!isset($DAG[$i])) {
$DAG[$i] = array($i);
}
}
self::calc($sentence, $DAG);
$x = 0;
$buf = '';
while ($x < $N) {
$current_route_keys = array_keys(self::$route[$x]);
$y = $current_route_keys[0] + 1;
$l_word = mb_substr($sentence, $x, $y - $x, 'UTF-8');
if ($y - $x == 1) {
$buf = $buf . $l_word;
} else {
if (mb_strlen($buf, 'UTF-8') > 0) {
if (mb_strlen($buf, 'UTF-8') == 1) {
array_push($words, $buf);
$buf = '';
} else {
$regognized = Finalseg::__cut($buf);
foreach ($regognized as $key => $word) {
array_push($words, $word);
}
$buf = '';
}
}
array_push($words, $l_word);
}
$x = $y;
}
if (mb_strlen($buf, 'UTF-8') > 0) {
if (mb_strlen($buf, 'UTF-8') == 1) {
array_push($words, $buf);
} else {
$regognized = Finalseg::__cut($buf);
foreach ($regognized as $key => $word) {
array_push($words, $word);
}
}
}
return $words;
}
#!/usr/bin/php
<?php
/**
* demo_extract_tags.php
*
* PHP version 5
*
* @category PHP
* @package /src/cmd/
* @author Fukuball Lin <*****@*****.**>
* @license MIT Licence
* @version GIT: <fukuball/jieba-php>
* @link https://github.com/fukuball/jieba-php
*/
ini_set('memory_limit', '600M');
require_once dirname(dirname(__FILE__)) . "/vendor/multi-array/MultiArray.php";
require_once dirname(dirname(__FILE__)) . "/vendor/multi-array/Factory/MultiArrayFactory.php";
require_once dirname(dirname(__FILE__)) . "/class/Jieba.php";
require_once dirname(dirname(__FILE__)) . "/class/Finalseg.php";
require_once dirname(dirname(__FILE__)) . "/class/JiebaAnalyse.php";
use Fukuball\Jieba;
use Fukuball\Finalseg;
use Fukuball\JiebaAnalyse;
Jieba::init(array('mode' => 'test', 'dict' => 'samll'));
Finalseg::init();
JiebaAnalyse::init();
$top_k = 10;
$content = file_get_contents(dirname(dirname(__FILE__)) . "/dict/lyric.txt", "r");
$tags = JiebaAnalyse::extractTags($content, $top_k);
var_dump($tags);
public function testFinalsegCut()
{
$case_array = array("怜香惜", "玉", "也", "得", "要", "看", "对象", "啊");
$seg_list = Finalseg::cut("怜香惜玉也得要看对象啊！");
$this->assertEquals($case_array, $seg_list);
}
