首页  :: 新随笔  :: 联系 :: 订阅 订阅  :: 管理

实现coreseek based on sphinx 对drupal和mediawiki的查询，总算基本搞定。

Posted on 2011-01-19 23:26  昕旸灿烂  阅读(447)  评论(0)    收藏  举报

#crontab -e

@hourly /usr/local/coreseek/bin/indexer --quiet --rotate --all

 

#vi /usr/local/coreseek/etc/csoft.conf

csoft.conf 
#
# Minimal Sphinx configuration sample (clean, simple, functional)
#

source src_technet_node
{
    type                = mysql
    sql_host            = drupal_dbhost
    sql_user            = drupal_dbusr
    sql_pass            = drupal_dbpwd
    sql_db              = drupal_dbname
    sql_port            = 3306
    sql_query_pre        = SET NAMES utf8
    sql_query            = SELECT nid, timestamp, title, body FROM node_revisions
    sql_attr_timestamp    = timestamp
    sql_query_info        = SELECT * FROM node_revisions WHERE nid=$id
}

source src_technet_comment
{
        type                        = mysql
        sql_host                    = drupal_dbhost
        sql_user                    = drupal_dbusr
        sql_pass                    = drupal_dbpwd
        sql_db                        = drupal_dbname
        sql_port                    = 3306
        sql_query_pre               = SET NAMES utf8
        sql_query                   = SELECT cid, nid, timestamp, subject, comment FROM comments
        sql_attr_uint               = nid
        sql_attr_timestamp          = timestamp
        sql_query_info              = SELECT * FROM comments WHERE cid=$id
}

source src_technet_wiki
{
        type                        = mysql
        sql_host                    = mediawiki_dbhost
        sql_user                    = mediawiki_dbusr
        sql_pass                    = mediawiki_dbpwd
        sql_db                        = mediawiki_dbname
        sql_port                    = 3306
        sql_query_pre               = SET NAMES utf8
        sql_query                    = SELECT rev_id, old_text FROM page, revision, text WHERE rev_id=page_latest AND old_id=rev_text_id
        sql_query_info               = SELECT page_title FROM page WHERE page_id=$id
}

index technet_drupal_node
{
    source                = src_technet_node
    path                = /var/data/technet_drupal_node
    docinfo                = extern
    mlock                = 0
    morphology            = none
    min_word_len            = 1
    html_strip            = 0
    charset_dictpath         = /usr/local/mmseg3/etc/
    charset_type            = zh_cn.utf-8
    ngram_len            = 0
}

index technet_drupal_comment
{
    source                          = src_technet_comment
    path                            = /var/data/technet_drupal_comment
    docinfo                         = extern
    mlock                           = 0
    morphology                      = none
    min_word_len                    = 1
    html_strip                      = 0
    charset_dictpath                = /usr/local/mmseg3/etc/
    charset_type                    = zh_cn.utf-8
    ngram_len                       = 0
}

index technet_wiki
{
    source                          = src_technet_wiki
    path                            = /var/data/technet_wiki
    docinfo                         = extern
    mlock                           = 0
    morphology                      = none
    min_word_len                    = 1
    html_strip                      = 0
    charset_dictpath                = /usr/local/mmseg3/etc/
    charset_type                    = zh_cn.utf-8
    ngram_len                       = 0
}

indexer
{
    mem_limit            = 256M
}

searchd
{
    port                = 3312
    log                = /var/log/technet-search.log
    query_log            = /var/log/technet-query.log
    read_timeout            = 5
    max_children            = 30
    pid_file            = /var/log/technet-searchd.pid
    max_matches            = 1000
    seamless_rotate            = 1
    preopen_indexes            = 0
    unlink_old            = 1
}

# vi search.php

功能实现，还需要再加工咯！

search.php 
<html><title>search for r&d</title><head><meta http-equiv = " content-type " content = " text/html; charset=UTF-8 " /></head><body>

<form action ="search.php" method = "post">
<input type="text" name="q" size="40"/> 
<input type="submit">
</form>

<?php
if (isset($_POST["q"])) {    
    $q = $_POST["q"];
    unset($_POST["s"]);
    search($q);
}

function search($q)    {
    require ( "sphinxapi.php" );
    $cl = new SphinxClient ();
    $cl->SetServer ( '168.2.4.246', 3312);
    $cl->SetMatchMode(SPH_MATCH_PHRASE); 
    $res_node = $cl->Query ( $q, 'technet_drupal_node' );
    $res_comment = $cl->Query ( $q, 'technet_drupal_comment' );
    $res_wiki = $cl->Query ( $q, 'technet_wiki' );

    printf("查询：[ %s ]，共找到 %d(%d/%d/%d) 条结果。 <hr/>", $q, count($res_node['matches']) + count($res_comment['matches']) + count($res_wiki['matches']), count($res_node['matches']), count($res_comment['matches']), count($res_wiki['matches']));    
    echo "<ul>";

    if (count($res_node['matches']) > 0) {
        foreach ($res_node['matches'] as $id => $value) 
            printf("<li>node: <a href=\"http://technet.baosight.com/drupal/node/%d\">http://technet.baosight.com/drupal/node/%d</a></li>", $id, $id);
    }
    
    if (count($res_comment['matches']) > 0) {
        foreach ($res_comment['matches'] as $id => $value)
            printf("<li>comment: <a href=\"http://technet.baosight.com/drupal/node/%d/#comment-%d\">http://technet.baosight.com/drupal/node/%d/#comment-%d</a></li>", $value[attrs][nid], $id, $value[attrs][nid], $id);
    }

    if (count($res_wiki['matches']) > 0) {
        foreach ($res_wiki['matches'] as $id => $value)
            printf("<li>wiki: <a href=\"http://technet.baosight.com/mediawiki/index.php?oldid=%d\">http://technet.baosight.com/mediawiki/index.php?oldid=%d</a></li>", $id, $id);
    }

    echo "</ul>";

    echo '<hr /><pre>';
    print_r($cl->GetLastError());
    print_r($cl->GetLastWarning());
    echo '</pre>';
}

?>

</body></html>