searchfuncs.php

开源的蜘蛛程序

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

error_reporting(E_ALL ^ E_NOTICE);
$include_dir = "include";
include "$include_dir/commonfuncs.php";

	function cmp($a, $b) {
		if ($a['weight'] == $b['weight'])
			return 0;

		return ($a['weight'] > $b['weight']) ? -1 : 1;
	}

	function addmarks($a) {
		$a = eregi_replace("[ ]+", " ", $a);
		$a = str_replace(" +", "+", $a);
		$a = str_replace(" ", "+", $a);
		return $a;
	}

	function makeboollist($a) {
		global $entities;
		while ($char = each($entities)) {
			$a = eregi_replace($char[0], $char[1], $a);
		}
		$a = trim($a);
		$a = eregi_replace("&quot;", "\"", $a);
		$returnWords = array();
		//get all phrases
		$regs = Array();
		while (eregi("([-]?)\"([^\"]+)\"", $a, $regs)) {
			if ($regs[1] == '') {
				$returnWords['+s'][] = $regs[2];
				$returnWords['hilight'][] = $regs[2];
			} else {
				$returnWords['-s'][] = $regs[2];
			}
			$a = str_replace($regs[0], "", $a);
		}
		$a = strtolower(eregi_replace("[ ]+", " ", $a));
//		$a = remove_accents($a);
		$a = trim($a);
		$words = explode(' ', $a);
		if ($a=="") {
			$limit = 0;
		} else {
		$limit = count($words);
		}


		$k = 0;
		//get all words (both include and exlude)
		$includeWords = array();
		while ($k < $limit) {
			if (substr($words[$k], 0, 1) == '+') {
				$includeWords[] = substr($words[$k], 1);
			} else if (substr($words[$k], 0, 1) == '-') {
				$returnWords['-'][] = substr($words[$k], 1);
			} else {
				$includeWords[] = $words[$k];
				if (!ignoreWord($words[$k])) {
					$returnWords['hilight'][] =  $words[$k];
				}
			}
			$k++;
		}
		//add words from phrases to includes
		if (isset($returnWords['+s'])) {
			foreach ($returnWords['+s'] as $phrase) {
				$phrase = strtolower(eregi_replace("[ ]+", " ", $phrase));
				$phrase = trim($phrase);
				$temparr = explode(' ', $phrase);
				foreach ($temparr as $w)
					$includeWords[] = $w;
			}
		}

		foreach ($includeWords as $word) {
			if (!($word =='')) {
				if (ignoreWord($word)) {

					$returnWords['ignore'][] = $word;
				} else {
					$returnWords['+'][] = $word;
				}	
			}

		}
		

		return $returnWords;
	}

	function ignoreword($word) {
		global $common;
		global $min_word_length;
		global $index_numbers;
		if ($index_numbers == 1) {
			$pattern = "[a-z0-9]+";
		} else {
			$pattern = "[a-z]+";
		}
		if (strlen($word) < $min_word_length || (!eregi($pattern, remove_accents($word))) || ($common[$word] == 1)) {
			return 1;
		} else {
			return 0;
		}
	}

	function search($searchstr, $category, $start, $per_page, $type) {
		global $length_of_link_desc,$mysql_table_prefix, $show_meta_description;

		//find all sites that should not be included in the result
		if (count($searchstr['+']) == 0) {
			return null;
		}
		$wordarray = $searchstr['-'];
		$notlist = array();
		$not_words = 0;
		while ($not_words < count($wordarray)) {
			$searchword = addslashes($wordarray[$not_words]);
			$query1 = "SELECT link_id from ".$mysql_table_prefix."link_keyword, ".$mysql_table_prefix."keywords where ".$mysql_table_prefix."link_keyword.keyword_id= ".$mysql_table_prefix."keywords.keyword_id and keyword='$searchword'";
			echo mysql_error();
			$result = mysql_query($query1);

			while ($row = mysql_fetch_row($result)) {	
				$notlist[$not_words]['id'][$row[0]] = 1;
			}
			$not_words++;
		}
		

		//find all sites containing the search phrase
		$wordarray = $searchstr['+s'];
		$phrase_words = 0;
		while ($phrase_words < count($wordarray)) {
			$searchword = addslashes($wordarray[$phrase_words]);
			$query1 = "SELECT link_id from ".$mysql_table_prefix."links where fulltxt like '% $searchword%'";
			echo mysql_error();
			$result = mysql_query($query1);
			$num_rows = mysql_num_rows($result);
			if ($num_rows == 0) {
				return null;
			}
			while ($row = mysql_fetch_row($result)) {	
				$phraselist[$phrase_words]['id'][$row[0]] = 1;
			}
			$phrase_words++;
		}
		

		if ($category> 0) {
			$allcats = get_cats($category);
			$catlist = implode(",", $allcats);
			$query1 = "select link_id from ".$mysql_table_prefix."links, ".$mysql_table_prefix."sites, ".$mysql_table_prefix."categories, ".$mysql_table_prefix."site_category where ".$mysql_table_prefix."links.site_id = ".$mysql_table_prefix."sites.site_id and ".$mysql_table_prefix."sites.site_id = ".$mysql_table_prefix."site_category.site_id and ".$mysql_table_prefix."site_category.category_id in ($catlist)";
			$result = mysql_query($query1);
			echo mysql_error();
			$num_rows = mysql_num_rows($result);
			if ($num_rows == 0) {
				return null;
			}
			while ($row = mysql_fetch_row($result)) {	
				$category_list[$row[0]] = 1;
			}
		}

		//find all sites that include the search word		
		$wordarray = $searchstr['+'];
		$words = 0;
		$starttime = getmicrotime();
		while ($words < count($wordarray)) {
			$searchword = addslashes($wordarray[$words]);
			$query1 = "SELECT distinct link_id, weight from ".$mysql_table_prefix."link_keyword, ".$mysql_table_prefix."keywords where ".$mysql_table_prefix."link_keyword.keyword_id= ".$mysql_table_prefix."keywords.keyword_id and keyword='$searchword' order by weight desc";
			$result = mysql_query($query1);
			$num_rows = mysql_num_rows($result);
			if ($num_rows == 0) {
				return null;
			}
			if ($type == "or") {
				$indx = 0;
			} else {
				$indx = $words;
			}
			
			while ($row = mysql_fetch_row($result)) {	
				$linklist[$indx]['id'][] = $row[0];
				$linklist[$indx]['weight'][$row[0]] = $row[1]; 			}
			$words++;
		}

		if ($type == "or") {
			$words = 1;
		}
		$result_array = Array();

		if ($words == 1 && $not_words == 0 && $category < 1) { //if there is only one search word, we already have the result
			$result_array = $linklist[0]['weight'];
		} else { //otherwise build an intersection of all the results
			$j= 1;
			$min = 0;
			while ($j < $words) {
				if (count($linklist[$min]['id']) > count($linklist[$j]['id'])) {
					$min = $j;
				}
				$j++;
			}

			$j = 0;


			$temp_array = $linklist[$min]['id'];
			$count = 0;
			while ($j < count($temp_array)) {
				$k = 0; //and word counter
				$n = 0; //not word counter
				$o = 0; //phrase word counter
				$weight = 1;
				$break = 0;
				while ($k < $words && $break== 0) {
					if ($linklist[$k]['weight'][$temp_array[$j]] > 0) {
						$weight = $weight + $linklist[$k]['weight'][$temp_array[$j]];
					} else {
						$break = 1;
					}
					$k++;
				}
				while ($n < $not_words && $break== 0) {
					if ($notlist[$n]['id'][$temp_array[$j]] > 0) {
						$break = 1;
					}
					$n++;
				}				

				while ($o < $phrase_words && $break== 0) {
					if ($phraselist[$n]['id'][$temp_array[$j]] != 1) {
						$break = 1;
					}
					$o++;
				}
				if ($break== 0 && $category > 0 && $category_list[$temp_array[$j]] != 1) {
					$break = 1;
				}

				if ($break == 0) {
					$result_array[$temp_array[$j]] = $weight;
					$count ++;
				}
				$j++;
			}
		}//word == 1

		$end = getmicrotime()- $starttime;

		$results = count($result_array);
		if ($results == 0) {
			return null;
		}
		arsort ($result_array);


		$keys = array_keys($result_array);
		$maxweight = $result_array[$keys[0]];

		for ($i = ($start -1)*$per_page; $i <min($results, ($start -1)*$per_page + $per_page) ; $i++) {
			$in[] = $keys[$i];

		}
		if (!is_array($in)) {
			$res['results'] = $results;
			return $res;
		}

		$inlist = implode(",", $in);


		if ($length_of_link_desc == 0) {
			$fulltxt = "fulltxt";
		} else {
			$fulltxt = "substring(fulltxt, 1, $length_of_link_desc)";
		}

		$query1 = "SELECT distinct link_id, url, title, description,  $fulltxt, size FROM ".$mysql_table_prefix."links WHERE link_id in ($inlist)";
		$result = mysql_query($query1);
		echo mysql_error();

		$i = 0;
		while ($row = mysql_fetch_row($result)) {
			$res[$i]['title'] = $row[2];
			$res[$i]['url'] = $row[1];
			if ($row[3] != null && $show_meta_description == 1)
				$res[$i]['fulltxt'] = $row[3];
			else 
				$res[$i]['fulltxt'] = $row[4];
			$res[$i]['size'] = $row[5];
			$res[$i]['weight'] = $result_array[$row[0]];
			$i++;
		}


		usort($res, "cmp"); 
		echo mysql_error();
		$res['maxweight'] = $maxweight;
		$res['results'] = $results;
		return $res;
	/**/
	}
?>