<?php
|
|
//require 'sphinxapi.php';
|
class sphinx{
|
/**
|
* @var SphinxClient
|
**/
|
protected $client;
|
|
/**
|
* @var string
|
**/
|
protected $keywords;
|
|
//同义词搜索
|
private $tyWordSearch;
|
|
//判断sphinx的查询服务searchd进程是否运行中（未运行则强制不使用sphinx查询，仍然用原来的mysql查询）
|
public static function isRun($host = 'localhost', $port = 9312)
|
{
|
$c = new SphinxClient();
|
$c->setServer($host, $port);
|
|
$c->SetConnectTimeout(0.01);
|
if($c->Status() === false){
|
return false;
|
}
|
return true;
|
}
|
|
/**
|
* Constructor
|
**/
|
public function __construct($options = array())
|
{
|
$defaults = array(
|
'query_mode' => SPH_MATCH_EXTENDED2,
|
// 'query_mode' => SPH_MATCH_ALL,
|
'ranking_mode' => SPH_RANK_PROXIMITY_BM25,//keyword匹配相关度根据：文本相似度 + 关键词出现次数
|
'sort_mode' => SPH_SORT_EXTENDED, //排序模式：混合排序模式，需要结合排序子句
|
|
'field_weights' => array(),
|
'snippet_enabled' => false
|
);
|
$this->options = array_merge($defaults, $options);
|
$this->client = new SphinxClient();
|
|
global $config;
|
|
$host = $config['sphinxhost'] ? $config['sphinxhost'] : 'localhost';
|
$host = isset($config['host']) ? $options['host'] : $host;
|
$port = $config['sphinxport'] ? $config['sphinxport'] : 9312;
|
$port = isset($options['port']) ? $options['port'] : $port;
|
|
$this->client->setServer( $host, $port);
|
|
//sy_indexpage : 前台页面列表最大展示数
|
$this->options['max_matches'] = $config['sy_indexpage'] ? $config['sy_indexpage'] : 0;
|
|
$this->client->setMatchMode($this->options['query_mode']);
|
$this->client->setRankingMode($this->options['ranking_mode']);
|
$this->client->setSortMode(SPH_SORT_EXTENDED, '@weight desc, @id desc');
|
if ($this->options['field_weights'] !== array()) {
|
$this->client->setFieldWeights($this->options['field_weights']);
|
}
|
|
$this->tyWordSearch = new ChineseTyWordSearch(APP_PATH . 'data/sphinx/dic_ty.txt', 60);
|
}
|
|
private function _getQueryString($where)
|
{
|
//setFilter
|
//setFilterString
|
//setFilterRange
|
//setSelect
|
//keywords
|
//queryStr
|
|
foreach($where as $v){
|
if(is_array($v) && count($v) > 1){
|
$k = $v[0];
|
if($k == 'setFilter' || $k == 'setFilterString'){
|
$exclude = isset($v[3]) ? $v[3] : false;
|
$this->client->$k($v[1], $v[2], $exclude);
|
}
|
else if($k == 'setFilterRange'){
|
$exclude = isset($v[4]) ? $v[4] : false;
|
$this->client->setFilterRange($v[1], $v[2], $v[3], $exclude);
|
}
|
else if($k == 'setSelect'){
|
$this->client->setSelect($v[1]);
|
}
|
}
|
}
|
|
$keywords = '';
|
if(array_key_exists('keywords', $where)){
|
$keywords = $this->wordSplit($where['keywords']);
|
// var_dump($keywords);
|
}
|
|
$queryStr = '';
|
if(array_key_exists('queryStr', $where)){
|
$queryStr = implode(' & ', $where['queryStr'] );
|
}
|
|
if($keywords == '' && $queryStr == ''){
|
return '';
|
}
|
else if($keywords == ''){
|
return $queryStr;
|
}
|
else if($queryStr == ''){
|
return $keywords;
|
}
|
else{
|
return "($keywords) & " . $queryStr;
|
}
|
}
|
|
private function _executeQuery($where, $offset = 0, $limit = 10, $index = '*', $type = 'getCount')
|
{
|
//分站
|
global $SiteTable,$config;
|
if(is_numeric($config['did']) && $config['did'] > 0 && in_array($index, $SiteTable) ){
|
$where [] = array('setFilter', 'did', array($config['did']) );
|
}
|
$this->client->ResetFilters();//查询之前，清空setFilter等查询条件
|
$queryStr = $this->_getQueryString($where);
|
|
|
if($this->options['max_matches'] == 0){
|
$this->options['max_matches'] = $config["{$index}_count"] + $config['sphinx_day_add_num'];
|
}
|
$max_matches = $limit > $this->options['max_matches'] ? $limit : $this->options['max_matches'];
|
$this->client->setLimits($offset, $limit, $max_matches);
|
|
//查询主索引 + 增量索引
|
$index = "{$index},{$index}_add";
|
$query_results = $this->client->query($queryStr, $index);
|
|
if ($query_results === false) {
|
$this->log('sphinx error:' . $this->client->getLastError());
|
}
|
if($type == 'getCount'){//只查询满足条件的数据总数
|
return empty($query_results['matches']) ? 0 : $query_results['total'];
|
}
|
else if($type == 'getFields'){//查询满足条件的数据的指定字段
|
return empty($query_results['matches']) ? array() : $query_results['matches'];
|
}
|
else{//get id array 查询满足条件的数据的id
|
return empty($query_results['matches']) ? array() : array_keys($query_results['matches']);
|
}
|
}
|
|
public function getCount($where, $offset = 0, $limit = 10, $index = '*')
|
{
|
return $this->_executeQuery($where, $offset, $limit , $index , 'getCount');
|
}
|
|
public function getIds($where, $offset = 0, $limit = 10, $index = '*')
|
{
|
return $this->_executeQuery($where, $offset, $limit , $index , 'getIds');
|
}
|
|
public function getFields($where, $offset = 0, $limit = 10, $index = '*')
|
{
|
return $this->_executeQuery($where, $offset, $limit , $index , 'getFields');
|
}
|
|
/**
|
* Chinese words segmentation
|
*
|
**/
|
public function wordSplit($keywords) {
|
global $config;
|
|
$fpath = ini_get('scws.default.fpath');
|
$so = scws_new();
|
$so->set_charset('utf-8');
|
$so->add_dict($fpath . '/dict.utf8.xdb');
|
$so->add_dict($fpath .'/custom_dict.txt', SCWS_XDICT_TXT);
|
|
$so->set_rule($fpath . '/rules.utf8.ini');
|
$so->set_ignore(true);//设置分词结果是否忽略标点符号。
|
|
/**
|
说明：设定分词返回结果时是否复合分割，如“中国人”返回“中国＋人＋中国人”三个词。
|
返回：无。
|
参数：mode 设定值，1 ~ 15。
|
按位异或的 1 | 2 | 4 | 8 分别表示: 短词 | 二元 | 主要单字 | 所有单字
|
*/
|
// $so->set_multi(1);
|
$so->set_multi(1);
|
|
/**
|
说明：设定是否将闲散文字自动以二字分词法聚合。
|
返回：无。
|
参数：yes 设定值，如果为 true 则结果中多个单字会自动按二分法聚分，如果为 false 则不处理，缺省为 false。
|
*/
|
$so->set_duality(false);
|
|
$so->send_text($keywords);
|
|
|
//说明：根据 send_text 设定的文本内容，返回系统计算出来的最关键词汇列表。
|
//返回：成功返回切好的词汇组成的数组， 若无更多词汇，返回 false。
|
$tops = $so->get_tops(5);
|
|
$words = array();
|
$words[] = $keywords;
|
foreach ($tops as $res) {
|
if($keywords != $res['word']){
|
$words[] = $res['word'];
|
}
|
}
|
foreach($words as &$w){
|
$w = '(' . $w . ')';
|
}
|
return join('|', $words);
|
}
|
|
// //判断字符串是否全部中文
|
// private function _isAllChineseWord($str)
|
// {
|
// return preg_match("/^[\x7f-\xff]+$/", $str);
|
// }
|
|
// private function _getChineseWord($str)
|
// {
|
// preg_match_all("/[\x7f-\xff]+/", $str, $matches);
|
// if(isset($matches[0]) && count($matches[0]) > 1){
|
// return $matches[0];
|
// }
|
// return array();
|
// }
|
|
/**
|
* log error
|
**/
|
private function log($msg) {
|
// log errors here
|
trigger_error($msg);
|
}
|
|
/**
|
* magic methods
|
**/
|
public function __call($method, $args)
|
{
|
$rc = new ReflectionClass('SphinxClient');
|
if ( !$rc->hasMethod($method) ) {
|
throw new Exception('invalid method :' . $method);
|
}
|
return call_user_func_array(array($this->client, $method), $args);
|
}
|
}
|
|
/**
|
* 中文同义词查询
|
* 设计思路：
|
* 把文件重新格式化为二进制文件，进行二分查找
|
* 调取方法：
|
* $m = new FileSearch("test.log", 200);
|
$resultArr = $m->search('司机'); //在文件中搜索
|
* 说明：第一次使用会很慢，因为要生成缓存文件
|
*/
|
class ChineseTyWordSearch
|
{
|
private $filename; //源文件名
|
private $maxLength; //源文件中单行的最大长度(按字节算)
|
private $sorted; //源文件是否已经排好序
|
private $formatFile; //重新格式化存储的文件名
|
private $fd;
|
|
/**
|
* 初始化
|
* @param string $filename 需要检索的文件
|
* @param int $maxLength 单行的最大长度
|
* @param int $sorted 源文件是否已经排好序
|
* @param int $forceReForm 是否强制重新生成索引文件
|
* @return [type] [description]
|
*/
|
public function __construct($filename, $maxLength, $sorted = 1, $forceReForm = 0)
|
{
|
$this->filename = $filename;
|
$this->maxLength = $maxLength;
|
$this->sorted = $sorted;
|
$this->formatFile = rtrim(dirname($this->filename), '/') . '/' . md5($this->filename);
|
if ($forceReForm || !file_exists($this->formatFile)) {
|
$this->formatFile();
|
}
|
}
|
|
//源文件排序
|
public function sort($path, $outPath = '')
|
{
|
$file = fopen($path, "r");
|
$content = array();
|
$i=0;
|
//输出文本中所有的行，直到文件结束为止。
|
while(! feof($file))
|
{
|
$content[$i]= trim(fgets($file));//fgets()函数从文件指针中读取一行
|
$i++;
|
}
|
fclose($file);
|
|
$content = array_unique($content);
|
|
|
usort($content,$this->mb_sort($prev, $next));
|
|
$content = implode(PHP_EOL, $content);
|
$outPath = $outPath == '' ? $path : $outPath;
|
file_put_contents($outPath, $content);
|
}
|
function mb_sort($prev, $next){
|
$p = explode('=', $prev);
|
$n = explode('=', $next);
|
if($p[0] == $n[0]) return 0;
|
return $p[0] > $n[0] ? 1 : -1;//升序排序
|
}
|
/**
|
* 格式化源文本文件为二进制文件
|
*/
|
public function formatFile()
|
{
|
if ($this->sorted == 0) {
|
//对源文件排序
|
}
|
//读源文件，写入到新的索引文件
|
$readfd = fopen($this->filename, 'rb');
|
$writefd = fopen($this->formatFile.'_tmp', 'wb+');
|
if ($readfd === false || $writefd === false) {
|
return false;
|
}
|
$i = 0;
|
while (!feof($readfd)) {
|
$line = fgets($readfd, 8192);
|
fseek($writefd, $i * $this->maxLength);
|
fwrite($writefd, pack("a".$this->maxLength, $line));
|
$i ++;
|
}
|
// echo "\n reformat ok\n";
|
fclose($readfd);
|
fclose($writefd);
|
rename($this->formatFile.'_tmp', $this->formatFile);
|
}
|
/**
|
* 在索引文件中进行二分查找
|
* @param int $id 进行二分查找的id
|
* @return [type] [description]
|
*/
|
public function search($key)
|
{
|
$filesize = filesize($this->formatFile);
|
$fd = fopen($this->formatFile, "rb");
|
$left = 0; //行号
|
$right = $total = ($filesize / $this->maxLength) - 1; //行号
|
while ($left <= $right) {
|
$middle = intval(($right + $left)/2);
|
fseek($fd, ($middle) * $this->maxLength);
|
$info = unpack("a*", fread($fd, $this->maxLength));
|
$info = explode('=', current($info));
|
|
if ($info[0] > $key) {
|
$right = $middle - 1;
|
} elseif ($info[0] < $key) {
|
$left = $middle + 1;
|
} else {
|
$retval = $this->getResult($fd, $middle, $key, $info, $total);
|
fclose($fd);
|
return $retval;
|
}
|
}
|
fclose($fd);
|
return false;
|
}
|
|
private function getResult($fd, $index, $key, $info, $total)
|
{
|
$retval = array();
|
$retval[] = trim($info[1]);
|
|
$i = -1; $j = 1;
|
|
while($index + $i >= 0){
|
fseek($fd, ($index + $i) * $this->maxLength);
|
$info = unpack("a*", fread($fd, $this->maxLength));
|
$info = explode('=', current($info));
|
if($info[0] == $key){
|
$retval[] = trim($info[1]);
|
}
|
else{
|
break;
|
}
|
$i --;
|
}
|
|
while($index + $j < $total){
|
fseek($fd, ($index + $j) * $this->maxLength);
|
$info = unpack("a*", fread($fd, $this->maxLength));
|
$info = explode('=', current($info));
|
if($info[0] == $key){
|
$retval[] = $info[1];
|
}
|
else{
|
break;
|
}
|
$j ++;
|
}
|
return $retval;
|
}
|
}
|
