php mysql 分词 模糊查询 并根据分词匹配度排序

Posted 2021-01-04 masterccc

中文分词用 SCWS 的api

http://www.xunsearch.com/scws/api.php

1.php中用 curl获取分词结果

protected function http_curl($url,$type="get",$res="json",$arr=array()){


        $ch =curl_init();
        curl_setopt($ch,CURLOPT_URL,$url);
        curl_setopt($ch,CURLOPT_RETURNTRANSFER,1);

        if($type=="post"){
            curl_setopt($ch,CURLOPT_POST,1);
            curl_setopt($ch,CURLOPT_POSTFIELDS,$arr);
        }
        curl_setopt($ch,CURLOPT_HTTP_VERSION,CURL_HTTP_VERSION_1_0);
        $output = curl_exec($ch);

        curl_close($ch);
        if($res=="json"){
          return json_decode($output,true);
        }else{
          return $output;
        }
    }

　　

protected function get_fenci($str){
        $url="http://www.xunsearch.com/scws/api.php";
        $type="post";
        $arr=array(
            ‘data‘=>$str,
            ‘respond‘=>"json",
            );

       $res= $this->http_curl($url,$type,"json",$arr);
       return $res;
    }

　　

2.根据分词循环sql where语句

foreach ($fenci as $k => $v) {
                    if(empty($where)){
                        $where.="name LIKE ‘%{$v}%‘";
                    }else{
                        $where.=" OR name LIKE ‘%{$v}%‘";
                    }
                }

3.根据分词循环sql order语句

用CASE WHEN THEN ELSE END 语句；

排序也是通过模糊查询，分别匹配词组，如果匹配到就定义一个数值，这些数值相加，词匹配到的越多，相加值的结果越大，越靠前值也越大，结果就相当于匹配度了， 然后进行降序排序；

 

protected function fenci_order($fenci){
        $order_str;
        foreach ($fenci as $k => $v) {

            if(empty($order_str)){
                $order_str.="(CASE WHEN name LIKE ‘%".$v."%‘ THEN ".(1000-10*$k)." ELSE 0 END)";
            }else{
                $order_str.=" + (CASE WHEN name LIKE ‘%".$v."%‘  THEN ".(1000-10*$k)." ELSE 0 END)";
            }
        }
        $order_str=$order_str." DESC";
        // echo count($fenci);

        return $order_str;
    }