xyzshop/helper/sensitive/SensitiveWordInit.php

<?php

/**
 * 初始化敏感词库<br>
 * 将敏感词加入到HashMap中<br>
 * 构建DFA算法模型
 *
 * @author dxm
 *
 */

namespace sensitive_word;

require_once(BASE_ROOT_PATH . "/helper/sensitive/DFAItem.php");

class SensitiveWordInit
{
    // 字符编码
    const ENCODING = "UTF-8";

    /**
     * 初始化敏感字库
     *
     * @return
     */
    public function initKeyWord()
    {
        $word_array = $this->readSensitiveWordFile();
        return $this->addSensitiveWordToHashMap($word_array);
    }

    /**
     * 读取敏感词库，将敏感词放入HashSet中，构建一个DFA算法模型：<br>
     * 中 = { isEnd = 0 国 = {<br>
     * isEnd = 1 人 = {isEnd = 0 民 = {isEnd = 1} } 男 = { isEnd = 0 人 = { isEnd =
     * 1 } } } } 五 = { isEnd = 0 星 = { isEnd = 0 红 = { isEnd = 0 旗 = { isEnd = 1
     * } } } }
     *
     */
    public function addSensitiveWordToHashMap($words)
    {
        $dfa = new DFAItem();
        foreach ($words as $word) {
            $dfa->addwords(explode(" ",$word));
        }
        return $dfa;
    }

    /**
     * 读取敏感词库中的内容，将内容添加到array中
     *
     * @return
     * @throws Exception
     */
    private function readSensitiveWordFile()
    {
        $word_array = array();
        array_push($word_array, '中 国');
        array_push($word_array, '中 央');
        array_push($word_array, '国 家');
        array_push($word_array, '他 妈 的');
        return $word_array;
    }
}