conf.php

开源的蜘蛛程序

<?
/*******************************************
* Sphider Version 1.2
* This program is licensed under the GNU GPL.
* By Ando Saabas          ando(a t)cs.ioc.ee
********************************************/

$includes = array('./languages', 'languages', '../languages');
if( !in_array($language_dir, $includes) && $language_dir != "")  {
	die("Illegal include.");
	
} else if ($language_dir != "" ){
	include "$language_dir/cn-language.php";  //The language file
} 


// Number of words a webpage must have to be indexed
$min_words_per_page = 10;

// Words shorter than this will not be indexed
$min_word_length = 3;

// Keyword weight depending on the number of times it appears in a page is capped at $word_upper_bound
$word_upper_bound = 100; 

$bound_search_result = 0; //can speed up searches on large database (should be 0)

// Number of query results shown per page
$results_per_page = 10;

// The length of the description string queried when displaying search results. 
// If set to 0 (default), makes a query for the whole page text,
// otherwise queries "substring(fulltext, 1, $length_of_link_desc))"
// Can significantly speed up searching on slower machines, if set to a lower value (eg 250 or 1000).
$length_of_link_desc = 0;

// Number of links shown to "next" pages
$links_to_next = 9;



// Show meta description in results page if it exists, otherwise show
// an extract from the page text.
$show_meta_description = 1;

// Query scores are not shown if set to 0
$show_query_scores = 1;

// if this value is set to 1, word in domain name and url path are also indexed,
// so that for example the index of www.php.net returns a positive answer to query "php" even 
// if the word is not included in the page itself. 
$index_host = 0;


// Use advanced search form: radio buttons for AND, OR and Phrase search and
// "results per page" selector
$advanced_search = 0;

// if this value is set to 1, keywords in a meta tag are also indexed
$index_meta_keywords = 1;

//Length of page description given in results page
$desc_length = 250;

//The number of colums the category list should have
$cat_columns = 2;

//If this is set to 1, numbers are also indexed and searchable
$index_numbers = 1;

//Set the user agent string
$user_agent = "Sphider";


//function to calculate the weight of pages
function calc_weight ($words_in_page, $word_in_title, $word_in_domain, $word_in_path, $path_depth, $meta_keyword) {

//Relative weight of a word in the title of a webpage
$title_weight = 20;

//Relative weight of a word in the domain name
$domain_weight = 60; 

//Relative weight of a word in the domain name
$path_weight = 10;

//Relative weight of a word in meta_keywords
$meta_weight = 5;

//weight function	
	$weight = ($words_in_page + $word_in_title * $title_weight +
			  $word_in_domain * $domain_weight +
			  $word_in_path * $path_weight + $meta_keyword * $meta_weight) *10 / (0.8 +0.2*$path_depth);

	return $weight;
}


//index pdf files
$index_pdf = 0;

//index pdf files
$index_doc = 0;

//temporary directory, this should be readable and writable
$tmp_dir='tmp/';

//path to pdf converter (including the file name itself)
$pdftotext_path = 'pdftotext.exe';


//path to doc converter (including the file name itself)
$catdoc_path = 'catdoc.exe';

?>