mnoGoSearch是一个功能完整的Web搜索引擎，你可以用它来在HTTP，HTTPS，FTP，NTTP服务器，本地文件和数据库文本字段之上构建搜索引擎。它内置支持XML，HTML，eml和文本文件。外部转换器，可用于支持其它类型的文档。支持自动 70多种语言/字符集猜测。还支持群集，从而能够实现通过多台计算机分发数据，能够索引高达几百亿文件巨大集合。Mnogosearch是由dpsearch改良而来的，与php整合比较好用。
此扩展已被移至 PECL 资源库且不再与 PHP 捆绑。下载 http://www.mnogosearch.org/
PHP包含内置的MySQL访问库，可以用来访问MySQL。众所周知，mnoGoSearch与此内置库不兼容，只能与通用MySQL库一起使用。因此，如果将mnoGoSearch与MySQL一起使用，则在PHP配置期间，必须指明在mnoGoSearch配置期间使用的MySQL安装目录，例如：--with-mnogosearch --with-mysql=/usr。
Well since the function udm_get_res_param used with UDM_PARAM_NUM_ROWS does infact not get the number of rows on the current page, it can actualy find more rows than specified by UDM_PARAM_PAGE_SIZE, ??? well i had to rewrite the some stuff I also added highlightning since the parameter UDM_PARAM_HLBEG nor UDM_PARAM_HLEND doesnt seem to work, alot of things here are fairly undocumented, I also found out after reading the Search_Mnogosearch pear class that replacing "\2" and "\3" in the text with HLBEG and HLEND works. well here is the class rewritten (without fetch, just find):
<?php
class MnogoSearch
{
public $matches = NULL;
public $total = NULL;
public $hilites = array('title', 'text');
public $hlbeg = '<span class="hilite">';
public $hlend = '</span>';
protected $agent = NULL;
protected $res = NULL;
public function __construct() {
$this->agent = udm_alloc_agent('mysql://user:pass@localhost/mnogosearch/');
$this->set(UDM_PARAM_SEARCH_MODE, UDM_MODE_ALL);
$this->set(UDM_PARAM_CHARSET, 'iso-8859-1');
$this->set(UDM_PARAM_BROWSER_CHARSET, 'iso-8859-1');
$this->set(UDM_PARAM_WORD_MATCH, UDM_MATCH_SUBSTR);
}
private function hiLite($t) {
if ($t == '')
return '';
$t = str_replace("\2", $this->hlbeg, $t);
$t = str_replace("\3", $this->hlend, $t);
return $t;
}
public function set($k, $v) {
udm_set_agent_param($this->agent, $k, $v);
}
public function find($q, $page=0, $rows=10) {
if ($page < 0 || $rows <= 0) {
$page = 0;
$rows = 10;
}
$this->set(UDM_PARAM_PAGE_SIZE, $rows);
$this->set(UDM_PARAM_PAGE_NUM, $page);
$this->res = udm_find($this->agent, $q);
$this->total = udm_get_res_param($this->res, UDM_PARAM_FOUND);
$found = udm_get_res_param($this->res, UDM_PARAM_NUM_ROWS);
if($found) {
$b = udm_get_res_param($this->res, UDM_PARAM_FIRST_DOC);
$e = udm_get_res_param($this->res, UDM_PARAM_LAST_DOC);
$rows = $e - $b + 1;
$fields = array('urlid', 'url', 'content', 'title', 'keywords', 'desc',
'text', 'size', 'rating', 'modified', 'order', 'crc', 'category',
'lang', 'charset', 'siteid', 'pop_rank', 'originid');
for($i=0; $i<$rows; $i++){
for($j=0; $j<count($fields); $j++) {
$this->matches[$i][$fields[$j]] = udm_get_res_field($this->res, $i, $j+1);
}
foreach($this->hilites as $hilite) {
$this->matches[$i][$hilite] = $this->hiLite($this->matches[$i][$hilite]);
}
}
return $rows;
} else {
return FALSE;
}
}
public function __destruct() {
udm_free_agent($this->agent);
if (isset($this->res)) {
udm_free_res($this->res);
}
}
}
?>
Im thinking, maybe swish-e is better ? :)for win32 users, mnogosearch since version 3.2.x
support COM interface
for Reflection API of COM use Visual Studio .NET
-> Tools ..
-> OLE/COM object viewer
- > C:\Program Files\mnoGoSearch\searchcom.dll
simple code
<pre>
<?
define('MNOGOSEARCH_WIN32',
(strtoupper(substr(PHP_OS,0,3))=='WIN') && extension_loaded('COM'));
if (!MNOGOSEARCH_WIN32)
die('MnoGoSearch COM not loaded');
/** Create COM object */
$agent = new COM('MnoGoSearch.GoSearch')
or die('Can\'t create COM object MnoGoSearch.GoSearch');
com_load_typelib('MnoGoSearch.GoSearch');
//mysql://user:passwd@host:port/database/?mode=multi
$agent->DBAddr = $params['DBAddr'];
//buggy, not work
//$agent->DBMode = $params['DBMode'];
$agent->SetCharset($params['Charset']);
$agent->SetCacheMode(true);
$q = 'q='. $query .'&np='. $_REQUEST['page'] .'&ps='. $_REQUEST['pg_size'] .'&m='. $_REQUEST['match'];
$q .= '&wm='. $_REQUEST['word_match'] .'&ul='. $_REQUEST['url_match']. '&wf='. $_REQUEST['where_find'];
$q .= '&typ='. $_REQUEST['content_type'] .'&s='. $_REQUEST['sortby'].'';
//!!! EXECUTE SEARCH QUERY
$result = $agent->Find($q);
if ($agent->ErrorCode>0 || !$result->Valid)
die('MnogoSearch win32 COM Error #'.$agent->ErrorCode.' - '.$agent->ErrorDescription);
$last=$result->LastDoc;
$first=$result->FirstDoc;
$hl_begin = '<span class="mnogosearch_hilight">';
$hl_end= '</span>';
$fetched_result=array();
$i=0;
for($row=$first;$row<=$last;$row++) {
$i=$row-$first;
$line = $result->Line($row);
$fetched_result[$i] = array(
'order'=>$line->Section('Order', $hl_begin, $hl_end),
'url'=> $line->Section('URL', $hl_begin, $hl_end),
'relevance'=>$line->Section('Score', $hl_begin, $hl_end),
'pagerank'=> $line->Section('Pop_Rank', $hl_begin, $hl_end),
'content_type'=> $line->Section('Content-Type', $hl_begin, $hl_end),
'content_length'=> $line->Section('Content-Length', $hl_begin, $hl_end),
'last_time'=> $line->Section('Last-Modified', $hl_begin, $hl_end),
'title'=> $line->Section('title', $hl_begin, $hl_end),
'body'=> $line->Section('body', $hl_begin, $hl_end),
'metadesc'=> $line->Section('meta.description', $hl_begin, $hl_end),
'metakeywords'=> $line->Section('meta.keywords', $hl_begin, $hl_end)
);
}
var_dump($fetched_result);
echo '<br/>search time : '.$result->SearchSeconds;
echo '<br/>words stats :'.$result->WordInfo;
echo '<br/>pg_count: '.$result->NPages;
echo '<br/>total_rows: '. $result->Count;
echo '<br/>first_doc: '.$result->FirstDoc;
echo '<br/>last_doc: '.$result->LastDoc;
echo '<br/>rows_in_page: '.$result->RowsOnPage;
?>
</pre>I have found that if you compile php with "--with-ftp and --with-mnogosearch=/dir" you must compile and build mnogosearch with "--disable-ftp" first. There are some naming conflicts between php's ftp libs and mno's ftp libs. Hope this helps