discuzX使用sphinx实现全文检索教程

发布:smiling　来源: PHP粉丝网　添加日期：2015-04-04 10:21:00　浏览: 评论：0　

Sphinx是一个基于SQL的全文检索引擎,可以结合MySQL,PostgreSQL做全文搜索,它可以提供比数据库本身更专业的搜索功能,使得应用程序更容易实现专业化的全文检索,这儿为大家介绍两点,第一个是基于discuz的索引配置文件,这个配置文件比较灵活,可以根据不同的需求来配置,代码如下:

#

# LinuxTone full index search configure file

#

source lt_posts

{

    type = mysql

    sql_host = 127.0.0.1

    sql_user = root

    sql_pass =

    sql_db = lt_bbs

    sql_port = 3306

    sql_query_pre = SET NAMES utf8

    sql_query = SELECT pid,tid,fid,dateline,subject,message,author FROM cdb_posts where first=1 #此处是基于posts表来做索引的，这样的目的是可以同时检索到subject,message,author 三个字段的值

    sql_attr_uint = fid

    sql_attr_timestamp = dateline

    sql_query_info = SELECT * FROM cdb_posts WHERE pid=$id

}

index lt_posts

{

    source = lt_posts

    path = /data/sphinx/data/lt_posts

    docinfo = extern

    mlock = 0

    morphology = none

    min_word_len = 2

    html_strip = 1

    charset_dictpath = /usr/local/mmseg-3.2.13/etc/

    charset_type = zh_cn.utf-8

    ngram_len = 0

}

########## 增量索引 ##################

source delta

{

    type = mysql

    sql_host = 127.0.0.1

    sql_user = root

    sql_pass =

    sql_db = lt_bbs

    sql_port = 3306 # optional, default is 3306

    sql_query_pre = SET NAMES utf8

    sql_query = SELECT pid,tid,fid,dateline,subject,message,author FROM cdb_posts where first=1 and dateline > unix_timestamp()-3600*10 #增量索引采用当前时间戳减去一个需要间隔的时间来新建新增的数据索引

    sql_attr_uint = fid

    sql_attr_timestamp = dateline

    sql_query_info = SELECT * FROM cdb_posts WHERE pid=$id

}

index delta

{

    source = delta

    path = /data/sphinx/data/lt_delta

    docinfo = extern

    mlock = 0

    morphology = none

    min_word_len = 2

    html_strip = 1

    charset_dictpath = /usr/local/mmseg-3.2.13/etc/

    charset_type = zh_cn.utf-8 //开源软件:phpfensi.com

    ngram_len = 0

}

indexer

{

    mem_limit = 32M

}

searchd

{

    port = 9312

    log = /data/sphinx/var/log/searchd.log

    query_log = /data/sphinx/var/log/query.log

    read_timeout = 5

    max_children = 30

    pid_file = /data/sphinx/var/log/searchd.pid

    max_matches = 10000

    seamless_rotate = 1

    preopen_indexes = 0

    unlink_old = 1

}

sphinx最主要的就是这个配置文件,当然在增量索引部分可以写一个脚本放到crontab里面来定时跑.

下面介绍下sphinx的PHP调用部分,sphinx的接口采用PHP的扩展,可以通过pecl或者http://pecl.php.net/package/sphinx来安装,代码如下:

<?php

/**

*全文搜索服务

*/

define('IN_DISCUZ', true);

require_once './include/common.inc.php';

$q = isset($_GET['q']) && !emptyempty($_GET['q']) ? $_GET['q'] : '';

$q = str_replace(array('<', '>', ' ', ''', ','), array('', '', ' ', '', ''), strip_tags($q));

$page = isset($_GET['page']) && intval($_GET['page']) > 0 ? intval($_GET['page']) : 1;

$perNum = 20;

$offset = ($page - 1) * $perNum;

$search = new SphinxClient();

$search -> setServer('127.0.0.1', 9312);

$search -> setConnectTimeout(2);

$search -> setArrayResult(true);

$search -> setMatchMode(SPH_MATCH_ANY);

$search -> setRankingMode(SPH_RANK_PROXIMITY_BM25);

$search -> setSortMode(SPH_SORT_EXTENDED, '@relevance desc,@weight desc');

$search -> setLimits($offset, $perNum);

$search -> setFieldWeights(array('subject' => 2000, 'message' => 0));

$rs = array();

$query_totals = $query_time = 0;

if (!emptyempty($q)) {

$rs = $search -> Query($q, "*");

$pages = ceil($rs['total'] / $perNum);

$query_totals = $rs['total_found'];

$query_time = $rs['time'];

}

$data = $title = $content = array();

if (!emptyempty($rs) && $page <= $pages) {

$pids = array();

foreach($rs['matches'] as $v) {

  $pids[] = $v['id'];

}

$pid = implode(',', $pids);

$sql = "select pid,tid,author,authorid,subject,message,dateline from cdb_posts where pid IN($pid) and status ='0' and invisible='0'";

$query = $db -> query($sql);

while ($row = $db -> fetch_array($query)) {

  $data[] = $row;

  $title[] = $row['subject'];

  $content[] = preg_replace('/[[/]?(b|img|url|color|s|hr|p|list|i|align|email|u|font|code|hide|table|tr|td|th|attach|list|indent|float).*]/', '', strip_tags($row['message']));

}

// 搜索词高亮

$opts = array();

$opts['before_match'] = '<em>';

$opts['after_match'] = '</em>';

$title = $search -> BuildExcerpts($title, 'lt_posts', $q, $opts);

$content = $search -> BuildExcerpts($content, 'lt_posts', $q, $opts);

foreach($data as $k => $v) {

  $data[$k]['subject'] = $title[$k];

  $data[$k]['message'] = $content[$k];

}

$url = "s.php?q=" . urlencode($q);

$multipage = multi($rs['total'], $perNum, $page, $url);

}

include template("lt_search");

?>

跑主索引的shell脚本search-index.sh,代码如下:

#!/bin/bash

#

# The BBS search exec full index

#

/usr/local/csft-3.2.13/bin/indexer -c /usr/local/csft-3.2.13/etc/lt_posts.conf --rotate lt_posts >> /data/sphinx/var/`date "+%Y-%m-%d-%H"`.log

跑增量索引的shell脚本search-delta.sh,代码如下:

#!/bin/bash

#

# The BBS search exec delta index

#

＃跑增量索引

/usr/local/csft-3.2.13/bin/indexer -c /usr/local/csft-3.2.13/etc/lt_posts.conf --rotate delta

#合并主索引和增量索引

#/usr/local/csft-3.2.13/bin/indexer --config /usr/local/csft-3.2.13/etc/lt_posts.conf --rotate --merge lt_posts delta