--- a/includes/search.php Wed Nov 21 20:45:50 2007 -0500
+++ b/includes/search.php Wed Nov 21 21:03:48 2007 -0500
@@ -31,15 +31,13 @@
}
/**
- * Algorithm to actually do the searching. This system usually works pretty fast (tested and developed on a site with 22 pages) but one
- * caveat of this algorithm is that it has to load the entire index into memory. It also requires manual parsing of the search query
- * which can be quite CPU-intensive. On the flip side this algorithm is extremely flexible and can be adapted for other uses very easily.
+ * In Enano versions prior to 1.0.2, this class provided a search function that was keyword-based and allowed boolean searches. It was
+ * cut from Coblynau and replaced with perform_search(), later in this file, because of speed issues. Now mostly deprecated. The only
+ * thing remaining is the buildIndex function, which is still used by the path manager and the new search framework.
*
- * Most of the time, this system is disabled. It is only used when MySQL can't or won't allow FULLTEXT indices.
- *
* @package Enano
* @subpackage Page management frontend
- * @license GNU General Public License http://enanocms.org/Special:GNU_General_Public_License
+ * @license GNU General Public License <http://enanocms.org/Special:GNU_General_Public_License>
*/
class Searcher
@@ -50,30 +48,11 @@
var $warnings;
var $match_case = false;
- function __construct()
- {
- $this->warnings = Array();
- }
-
- function Searcher()
- {
- $this->__construct();
- }
-
- function warn($t)
- {
- if(!in_array($t, $this->warnings)) $this->warnings[] = $t;
- }
-
- function convertCase($text)
- {
- return ( $this->match_case ) ? $text : strtolower($text);
- }
-
function buildIndex($texts)
{
$this->index = Array();
-
+ $stopwords = get_stopwords();
+
foreach($texts as $i => $l)
{
$seed = md5(microtime(true) . mt_rand());
@@ -96,7 +75,7 @@
$words = explode(' ', $letters);
foreach($words as $c => $w)
{
- if(strlen($w) < 4)
+ if(strlen($w) < 2 || in_array($w, $stopwords))
unset($words[$c]);
else
$words[$c] = $w;
@@ -121,516 +100,808 @@
$this->index[$k] = implode(',', $this->index[$k]);
}
}
+}
+
+/**
+ * Searches the site for the specified string and returns an array with each value being an array filled with the following:
+ * page_id: string, self-explanatory
+ * namespace: string, self-explanatory
+ * page_length: integer, the length of the full page in bytes
+ * page_text: string, the contents of the page (trimmed to ~150 bytes if necessary)
+ * score: numerical relevance score, 1-100, rounded to 2 digits and calculated based on which terms were present and which were not
+ * @param string Search query
+ * @param string Will be filled with any warnings encountered whilst parsing the query
+ * @param bool Case sensitivity - defaults to false
+ * @return array
+ */
+
+function perform_search($query, &$warnings, $case_sensitive = false)
+{
+ global $db, $session, $paths, $template, $plugins; // Common objects
+ $warnings = array();
- function search($query, $texts)
+ $query = parse_search_query($query, $warnings);
+
+ // Segregate search terms containing spaces
+ $query_phrase = array(
+ 'any' => array(),
+ 'req' => array()
+ );
+
+ foreach ( $query['any'] as $i => $_ )
{
-
- // OK, let's establish some basics here. Here is the procedure for performing the search:
- // * search for items that matches all the terms in the correct order.
- // * search for items that match in any order
- // * eliminate one term and do the loop all over
-
- $this->results = Array();
- $query = $this->parseQuery($query);
- $querybak = $query;
- for($i = sizeof($query['any'])-1; $i >= 0; $i--)
+ $term =& $query['any'][$i];
+ $term = trim($term);
+ // the indexer only indexes words a-z with apostrophes
+ if ( preg_match('/[^A-Za-z\']/', $term) )
+ {
+ $query_phrase['any'][] = $term;
+ unset($term, $query['any'][$i]);
+ }
+ }
+ unset($term);
+ $query['any'] = array_values($query['any']);
+
+ foreach ( $query['req'] as $i => $_ )
+ {
+ $term =& $query['req'][$i];
+ $term = trim($term);
+ if ( preg_match('/[^A-Za-z\']/', $term) )
{
- $res = $this->performCoreSearch($query, $texts, true);
- $this->results = enano_safe_array_merge($this->results, $res);
- $res = $this->performCoreSearch($query, $texts, false);
- $this->results = enano_safe_array_merge($this->results, $res);
- unset($query['any'][$i]);
+ $query_phrase['req'][] = $term;
+ unset($term, $query['req'][$i]);
}
-
- // Last resort - search for any of the terms instead of all of 'em
- $res = $this->performCoreSearch($querybak, $texts, false, true);
- $this->results = enano_safe_array_merge($this->results, $res);
-
- $this->highlightResults($querybak);
+ }
+ unset($term);
+ $query['req'] = array_values($query['req']);
+
+ $results = array();
+ $scores = array();
+
+ // FIXME: Update to use FULLTEXT algo when available.
+
+ // Build an SQL query to load from the index table
+ if ( count($query['any']) < 1 && count($query['req']) < 1 && count($query_phrase['any']) < 1 && count($query_phrase['req']) < 1 )
+ {
+ // This is both because of technical restrictions and devastation that would occur on shared servers/large sites.
+ $warnings[] = 'You need to have at least one keyword in your search query. Searching only for pages not containing a term is not allowed.';
+ return array();
}
- // $texts should be a textual MySQL query!
- // @todo document
- function searchMySQL($query, $texts)
+ //
+ // STAGE 1
+ // Get all possible result pages from the search index. Tally which pages have the most words, and later sort them by boolean relevance
+ //
+
+ // Skip this if no indexable words are included
+
+ if ( count($query['any']) > 0 || count($query['req']) > 0 )
{
- global $db;
- // OK, let's establish some basics here. Here is the procedure for performing the search:
- // * search for items that matches all the terms in the correct order.
- // * search for items that match in any order
- // * eliminate one term and do the loop all over
-
- $this->results = Array();
- $query = $this->parseQuery($query);
- $querytmp = $query;
- $querybak = $query;
- for($i = sizeof($querytmp['any'])-1; $i >= 0; $i--)
+ $where_any = array();
+ foreach ( $query['any'] as $term )
{
- $res = $this->performCoreSearchMySQL($querytmp, $texts, true);
- $this->results = enano_safe_array_merge($this->results, $res);
- $res = $this->performCoreSearchMySQL($querytmp, $texts, false);
- $this->results = enano_safe_array_merge($this->results, $res);
- unset($querytmp['any'][$i]);
+ $term = escape_string_like($term);
+ if ( !$case_sensitive )
+ $term = strtolower($term);
+ $where_any[] = $term;
+ }
+ foreach ( $query['req'] as $term )
+ {
+ $term = escape_string_like($term);
+ if ( !$case_sensitive )
+ $term = strtolower($term);
+ $where_any[] = $term;
}
- // Last resort - search for any of the terms instead of all of 'em
- $res = $this->performCoreSearchMySQL($querybak, $texts, false, true);
- $this->results = enano_safe_array_merge($this->results, $res);
+ $col_word = ( $case_sensitive ) ? 'word' : 'lcase(word)';
+ $where_any = ( count($where_any) > 0 ) ? '( ' . $col_word . ' = \'' . implode('\' OR ' . $col_word . ' = \'', $where_any) . '\' )' : '';
+
+ // generate query
+ // using a GROUP BY here ensures that the same word with a different case isn't counted as 2 words - it's all melted back
+ // into one later in the processing stages
+ $group_by = ( $case_sensitive ) ? '' : ' GROUP BY lcase(word);';
+ $sql = "SELECT word, page_names FROM " . table_prefix . "search_index WHERE {$where_any}{$group_by}";
+ if ( !($q = $db->sql_unbuffered_query($sql)) )
+ $db->_die('Error is in perform_search(), includes/search.php, query 1');
- $this->highlightResults($querybak);
- }
-
- /**
- * This method assumes that $query is already parsed and $texts is an (associative) array of possible results
- * @param array $query A search query parsed with Searcher::parseQuery()
- * @param array $texts The list of possible results
- * @param bool $exact_order If true, only matches results with the terms in the same order as the terms in the query
- * @return array An associative array of results
- * @access private
- */
- function performCoreSearch($query, $texts, $exact_order = false, $any = false)
- {
- $textkeys = array_keys($texts);
- $results = Array();
- if($exact_order)
+ $word_tracking = array();
+ if ( $row = $db->fetchrow() )
{
- $query = $this->concatQueryTerms($query);
- }
- $query['trm'] = array_merge($query['any'], $query['req']);
- # Find all remotely possible results first
- // Single-word terms
- foreach($this->index as $term => $keys)
- {
- foreach($query['trm'] as $userterm)
+ do
{
- if($this->convertCase($userterm) == $this->convertCase($term))
+ // get page list
+ $pages =& $row['page_names'];
+ $ns_list = '(' . implode('|', array_keys($paths->nslist)) . ')';
+ if ( strpos($pages, ',') )
{
- $k = explode(',', $keys);
- foreach($k as $idxkey)
+ // the term occurs in more than one page
+
+ // Find page IDs that contain commas
+ // This should never happen because commas are escaped by sanitize_page_id(). Nevertheless for compatibility with older
+ // databases, and to alleviate the concerns of hackers, we'll accommodate for page IDs with commas here by checking for
+ // IDs that don't match the pattern for stringified page ID + namespace. If it doesn't match, that means it's a continuation
+ // of the previous ID and should be concatenated to the previous entry.
+ $matches = explode(',', $pages);
+ $prev = false;
+ foreach ( $matches as $i => $_ )
{
- if(isset($texts[$idxkey]))
+ $match =& $matches[$i];
+ if ( !preg_match("/^ns=$ns_list;pid=(.+)$/", $match) && $prev )
{
- $results[$idxkey] = $texts[$idxkey];
+ $matches[$prev] .= ',' . $match;
+ unset($match, $matches[$i]);
+ continue;
+ }
+ $prev = $i;
+ }
+ unset($match);
+
+ // Iterate through each of the results, assigning scores based on how many times the page has shown up.
+ // This works because this phase of the search is strongly word-based not page-based. If a page shows up
+ // multiple times while fetching the result rows from the search_index table, it simply means that page
+ // contains more than one of the terms the user searched for.
+
+ foreach ( $matches as $match )
+ {
+ if ( isset($scores[$match]) )
+ {
+ $scores[$match]++;
}
else
{
- if(preg_match('#^([0-9]+)$#', $idxkey))
- {
- $idxkey = intval($idxkey);
- if(isset($texts[$idxkey])) $results[$idxkey] = $texts[$idxkey];
- }
+ $scores[$match] = 1;
}
- }
- }
- }
- }
- // Quoted terms
- foreach($query['trm'] as $userterm)
- {
- if(!preg_match('/[\s"\'~`!@#\$%\^&\*\(\)\{\}:;<>,.\/\?_-]/', $userterm)) continue;
- foreach($texts as $k => $t)
- {
- if(strstr($this->convertCase($t), $this->convertCase($userterm)))
- {
- // We have a match!
- if(!isset($results[$k])) $results[$k] = $t;
- }
- }
- }
- // Remove excluded terms
- foreach($results as $k => $r)
- {
- foreach($query['not'] as $not)
- {
- if(strstr($this->convertCase($r), $this->convertCase($not))) unset($results[$k]);
- }
- }
- if(!$any)
- {
- // Remove results not containing all terms
- foreach($results as $k => $r)
- {
- foreach($query['any'] as $term)
- {
- if(!strstr($this->convertCase($r), $this->convertCase($term))) unset($results[$k]);
- }
- }
- }
- // Remove results not containing all required terms
- foreach($results as $k => $r)
- {
- foreach($query['req'] as $term)
- {
- if(!strstr($this->convertCase($r), $this->convertCase($term))) unset($results[$k]);
- }
- }
- return $results;
- }
-
- /**
- * This is the same as performCoreSearch, but $texts should be a MySQL result resource. This can save tremendous amounts of memory on large sites.
- * @param array $query A search query parsed with Searcher::parseQuery()
- * @param string $texts A text MySQL query that selects the text as the first column and the index key as the second column
- * @param bool $exact_order If true, only matches results with the terms in the same order as the terms in the query
- * @return array An associative array of results
- * @access private
- */
- function performCoreSearchMySQL($query, $texts, $exact_order = false, $any = false)
- {
- global $db;
- $results = Array();
- if($exact_order)
- {
- $query = $this->concatQueryTerms($query);
- }
- $query['trm'] = array_merge($query['any'], $query['req']);
- # Find all remotely possible results first
- $texts = $db->sql_query($texts);
- if ( !$texts )
- $db->_die('The error is in the search engine.');
- if ( $r = $db->fetchrow_num($texts) )
- {
- do
- {
- foreach($this->index as $term => $keys)
- {
- foreach($query['trm'] as $userterm)
- {
- if($this->convertCase($userterm) == $this->convertCase($term))
+ if ( isset($word_tracking[$match]) )
{
- $k = explode(',', $keys);
- foreach($k as $idxkey)
- {
- $row[0] = $r[0];
- $row[1] = $r[1];
- if(!isset($row[1]))
- {
- echo('PHP PARSER BUG: $row[1] is set but not set... includes/search.php:'.__LINE__);
- $GLOBALS['template']->footer();
- exit;
- }
- if($row[1] == $idxkey)
- $results[$idxkey] = $row[0];
- else
- {
- if(preg_match('#^([0-9]+)$#', $idxkey))
- {
- $idxkey = intval($idxkey);
- if($row[1] == $idxkey) $results[$idxkey] = $row[0];
- }
- }
- }
+ $word_tracking[$match][] = $row['word'];
+ }
+ else
+ {
+ $word_tracking[$match] = array($row['word']);
}
}
}
- // Quoted terms
- foreach($query['trm'] as $userterm)
+ else
{
- if(!preg_match('/[\s"\'~`!@#\$%\^&\*\(\)\{\}:;<>,.\/\?_-]/', $userterm)) continue;
- if(strstr($this->convertCase($r[0]), $this->convertCase($userterm)))
+ // the term only occurs in one page
+ if ( isset($scores[$pages]) )
{
- // We have a match!
- if(!isset($results[$r[1]])) $results[$r[1]] = $r[0];
+ $scores[$pages]++;
}
- }
- } while( $r = $db->fetchrow_num($texts) );
- }
- // Remove excluded terms
- foreach($results as $k => $r)
- {
- foreach($query['not'] as $not)
- {
- if(strstr($this->convertCase($r), $this->convertCase($not))) unset($results[$k]);
- }
- }
- if(!$any)
- {
- // Remove results not containing all terms
- foreach($results as $k => $r)
- {
- foreach($query['any'] as $term)
- {
- if(!strstr($this->convertCase($r), $this->convertCase($term))) unset($results[$k]);
+ else
+ {
+ $scores[$pages] = 1;
+ }
+ if ( isset($word_tracking[$pages]) )
+ {
+ $word_tracking[$pages][] = $row['word'];
+ }
+ else
+ {
+ $word_tracking[$pages] = array($row['word']);
+ }
}
}
- }
- // Remove results not containing all terms
- foreach($results as $k => $r)
- {
- foreach($query['req'] as $term)
- {
- if(!strstr($this->convertCase($r), $this->convertCase($term))) unset($results[$k]);
- }
+ while ( $row = $db->fetchrow() );
}
- return $results;
- }
-
- function concatQueryTerms($query)
- {
- $tmp = implode(' ', $query['any']);
- unset($query['any']);
- $query['any'] = Array(0 => $tmp);
- return $query;
- }
+ $db->free_result();
- /**
- * Builds a basic assoc array with a more organized version of the query
- */
-
- function parseQuery($query)
- {
- $ret = array(
- 'any' => array(),
- 'req' => array(),
- 'not' => array()
- );
- $terms = array();
- $in_quote = false;
- $start_term = 0;
- $just_finished = false;
- for ( $i = 0; $i < strlen($query); $i++ )
+ //
+ // STAGE 2: FIRST ELIMINATION ROUND
+ // Iterate through the list of required terms. If a given page is not found to have the required term, eliminate it
+ //
+
+ foreach ( $query['req'] as $term )
{
- $chr = $query{$i};
- $prev = ( $i > 0 ) ? $query{ $i - 1 } : '';
- $next = ( ( $i + 1 ) < strlen($query) ) ? $query{ $i + 1 } : '';
-
- if ( ( $chr == ' ' && !$in_quote ) || ( $i + 1 == strlen ( $query ) ) )
+ foreach ( $word_tracking as $i => $page )
{
- $len = ( $next == '' ) ? $i + 1 : $i - $start_term;
- $word = substr ( $query, $start_term, $len );
- $terms[] = $word;
- $start_term = $i + 1;
- }
-
- elseif ( $chr == '"' && $in_quote && $prev != '\\' )
- {
- $word = substr ( $query, $start_term, $i - $start_term + 1 );
- $start_pos = ( $next == ' ' ) ? $i + 2 : $i + 1;
- $in_quote = false;
- }
-
- elseif ( $chr == '"' && !$in_quote )
- {
- $in_quote = true;
- $start_pos = $i;
- }
-
- }
-
- $ticker = 0;
-
- foreach ( $terms as $element => $__unused )
- {
- $atom =& $terms[$element];
-
- $ticker++;
-
- if ( $ticker == 20 )
- {
- $this->warn('Some of your search terms were excluded because searches are limited to 20 terms to prevent excessive server load.');
- break;
- }
-
- if ( substr ( $atom, 0, 2 ) == '+"' && substr ( $atom, ( strlen ( $atom ) - 1 ), 1 ) == '"' )
- {
- $word = substr ( $atom, 2, ( strlen( $atom ) - 3 ) );
- if ( strlen ( $word ) < 4 )
- {
- $this->warn('One or more of your search terms was excluded because terms must be at least 4 characters in length.');
- $ticker--;
- continue;
- }
- if(in_array($word, $ret['req']))
- {
- $this->warn('One or more of your search terms was excluded because duplicate terms were encountered.');
- $ticker--;
- continue;
- }
- $ret['req'][] = $word;
- }
- elseif ( substr ( $atom, 0, 2 ) == '-"' && substr ( $atom, ( strlen ( $atom ) - 1 ), 1 ) == '"' )
- {
- $word = substr ( $atom, 2, ( strlen( $atom ) - 3 ) );
- if ( strlen ( $word ) < 4 )
+ if ( !in_array($term, $page) )
{
- $this->warn('One or more of your search terms was excluded because terms must be at least 4 characters in length.');
- $ticker--;
- continue;
- }
- if(in_array($word, $ret['not']))
- {
- $this->warn('One or more of your search terms was excluded because duplicate terms were encountered.');
- $ticker--;
- continue;
- }
- $ret['not'][] = $word;
- }
- elseif ( substr ( $atom, 0, 1 ) == '+' )
- {
- $word = substr ( $atom, 1 );
- if ( strlen ( $word ) < 4 )
- {
- $this->warn('One or more of your search terms was excluded because terms must be at least 4 characters in length.');
- $ticker--;
- continue;
- }
- if(in_array($word, $ret['req']))
- {
- $this->warn('One or more of your search terms was excluded because duplicate terms were encountered.');
- $ticker--;
- continue;
- }
- $ret['req'][] = $word;
- }
- elseif ( substr ( $atom, 0, 1 ) == '-' )
- {
- $word = substr ( $atom, 1 );
- if ( strlen ( $word ) < 4 )
- {
- $this->warn('One or more of your search terms was excluded because terms must be at least 4 characters in length.');
- $ticker--;
- continue;
- }
- if(in_array($word, $ret['not']))
- {
- $this->warn('One or more of your search terms was excluded because duplicate terms were encountered.');
- $ticker--;
- continue;
- }
- $ret['not'][] = $word;
- }
- elseif ( substr ( $atom, 0, 1 ) == '"' && substr ( $atom, ( strlen($atom) - 1 ), 1 ) == '"' )
- {
- $word = substr ( $atom, 1, ( strlen ( $atom ) - 2 ) );
- if ( strlen ( $word ) < 4 )
- {
- $this->warn('One or more of your search terms was excluded because terms must be at least 4 characters in length.');
- $ticker--;
- continue;
- }
- if(in_array($word, $ret['any']))
- {
- $this->warn('One or more of your search terms was excluded because duplicate terms were encountered.');
- $ticker--;
- continue;
- }
- $ret['any'][] = $word;
- }
- else
- {
- $word = $atom;
- if ( strlen ( $word ) < 4 )
- {
- $this->warn('One or more of your search terms was excluded because terms must be at least 4 characters in length.');
- $ticker--;
- continue;
- }
- if(in_array($word, $ret['any']))
- {
- $this->warn('One or more of your search terms was excluded because duplicate terms were encountered.');
- $ticker--;
- continue;
- }
- $ret['any'][] = $word;
- }
- }
- return $ret;
- }
-
- function highlightResults($query, $starttag = '<b>', $endtag = '</b>')
- {
- $query['trm'] = array_merge($query['any'], $query['req']);
- //die('<pre>'.print_r($query, true).'</pre>');
- foreach($query['trm'] as $q)
- {
- foreach($this->results as $k => $r)
- {
- $startplace = 0;
- //$this->results[$k] = htmlspecialchars($this->results[$k]);
- for($i = 0; $i < strlen($r); $i++)
- {
- $word = substr($r, $i, strlen($q));
- if($this->convertCase($word) == $this->convertCase($q))
- {
- $word = $starttag . $word . $endtag;
- $this->results[$k] = substr($r, 0, $i) . $word . substr($r, $i + strlen($q), strlen($r)+999999);
- $startplace = $i - 75;
- if($startplace < 0) $startplace = 0;
- $this->results[$k] = '...'.trim(substr($this->results[$k], $startplace, strlen($word) + 150)).'...';
- continue 2;
- }
+ unset($word_tracking[$i], $scores[$i]);
}
}
}
}
+ //
+ // STAGE 3: PHRASE SEARCHING
+ // Use LIKE to find pages with specified phrases. We can do a super-picky single query without another elimination round because
+ // at this stage we can search the full page_text column instead of relying on a word list.
+ //
+
+ // We can skip this stage if none of these special terms apply
+
+ $text_col = ( $case_sensitive ) ? 'page_text' : 'lcase(page_text)';
+
+ if ( count($query_phrase['any']) > 0 || count($query_phrase['req']) > 0 )
+ {
+
+ $where_any = array();
+ foreach ( $query_phrase['any'] as $term )
+ {
+ $term = escape_string_like($term);
+ if ( !$case_sensitive )
+ $term = strtolower($term);
+ $where_any[] = $term;
+ }
+
+ $where_any = ( count($where_any) > 0 ) ? "( $text_col LIKE '%" . implode("%' OR $text_col LIKE '%", $where_any) . "%' )" : '';
+
+ // Also do required columns, but use AND to ensure that all required terms are included
+ $where_req = array();
+ foreach ( $query_phrase['req'] as $term )
+ {
+ $term = escape_string_like($term);
+ if ( !$case_sensitive )
+ $term = strtolower($term);
+ $where_req[] = $term;
+ }
+ $and_clause = ( $where_any != '' ) ? 'AND ' : '';
+ $where_req = ( count($where_req) > 0 ) ? "{$and_clause}$text_col LIKE '%" . implode("%' AND $text_col LIKE '%", $where_req) . "%'" : '';
+
+ $sql = 'SELECT CONCAT("ns=",namespace,";pid=",page_id) AS id FROM ' . table_prefix . "page_text WHERE $where_any $where_req;";
+ if ( !($q = $db->sql_unbuffered_query($sql)) )
+ $db->_die('Error is in perform_search(), includes/search.php, query 2. Parsed query dump follows:<pre>(indexable) ' . htmlspecialchars(print_r($query, true)) . '(non-indexable) ' . htmlspecialchars(print_r($query_phrase, true)) . '</pre>');
+
+ if ( $row = $db->fetchrow() )
+ {
+ do
+ {
+ $id =& $row['id'];
+ if ( isset($scores[$id]) )
+ {
+ $scores[$id]++;
+ }
+ else
+ {
+ $scores[$id] = 1;
+ }
+ }
+ while ( $row = $db->fetchrow() );
+ }
+ $db->free_result();
+ }
+
+ //
+ // STAGE 4 - SELECT PAGE TEXT AND ELIMINATE NOTS
+ // At this point, we have a complete list of all the possible pages. Now we want to obtain the page text, and within the same query
+ // eliminate any terms that shouldn't be in there.
+ //
+
+ // Generate master word list for the highlighter
+ $word_list = array_values(array_merge($query['any'], $query['req'], $query_phrase['any'], $query_phrase['req']));
+
+ $text_where = array();
+ foreach ( $scores as $page_id => $_ )
+ {
+ $text_where[] = $db->escape($page_id);
+ }
+ $text_where = '( CONCAT("ns=",t.namespace,";pid=",t.page_id) = \'' . implode('\' OR CONCAT("ns=",t.namespace,";pid=",t.page_id) = \'', $text_where) . '\' )';
+
+ if ( count($query['not']) > 0 )
+ $text_where .= ' AND';
+
+ $where_not = array();
+ foreach ( $query['not'] as $term )
+ {
+ $term = escape_string_like($term);
+ if ( !$case_sensitive )
+ $term = strtolower($term);
+ $where_not[] = $term;
+ }
+ $where_not = ( count($where_not) > 0 ) ? "$text_col NOT LIKE '%" . implode("%' AND $text_col NOT LIKE '%", $where_not) . "%'" : '';
+
+ $sql = 'SELECT CONCAT("ns=",t.namespace,";pid=",t.page_id) AS id, t.page_id, t.namespace, CHAR_LENGTH(t.page_text) AS page_length, t.page_text, p.name AS page_name FROM ' . table_prefix . "page_text AS t
+ LEFT JOIN " . table_prefix . "pages AS p
+ ON ( p.urlname = t.page_id AND p.namespace = t.namespace )
+ WHERE $text_where $where_not;";
+ if ( !($q = $db->sql_unbuffered_query($sql)) )
+ $db->_die('Error is in perform_search(), includes/search.php, query 3');
+
+ $page_data = array();
+ if ( $row = $db->fetchrow() )
+ {
+ do
+ {
+ $row['page_text'] = htmlspecialchars($row['page_text']);
+ $row['page_name'] = htmlspecialchars($row['page_name']);
+
+ // Highlight results (this is wonderfully automated)
+ $row['page_text'] = highlight_and_clip_search_result($row['page_text'], $word_list, $case_sensitive);
+ if ( strlen($row['page_text']) > 250 && !preg_match('/^\.\.\.(.+)\.\.\.$/', $row['page_text']) )
+ {
+ $row['page_text'] = substr($row['page_text'], 0, 150) . '...';
+ }
+ $row['page_name'] = highlight_search_result($row['page_name'], $word_list, $case_sensitive);
+
+ $page_data[$row['id']] = $row;
+ }
+ while ( $row = $db->fetchrow() );
+ }
+ $db->free_result();
+
+ //
+ // STAGE 5 - SPECIAL PAGE TITLE SEARCH
+ // Iterate through $paths->pages and check the titles for search terms. Score accordingly.
+ //
+
+ foreach ( $paths->pages as $page )
+ {
+ if ( $page['namespace'] != 'Special' )
+ continue;
+ $idstring = 'ns=' . $page['namespace'] . ';pid=' . $page['urlname_nons'];
+ $any = array_merge($query['any'], $query_phrase['any']);
+ foreach ( $any as $term )
+ {
+ if ( $case_sensitive )
+ {
+ if ( strstr($page['name'], $term) || strstr($page['urlname_nons'], $term) )
+ {
+ ( isset($scores[$idstring]) ) ? $scores[$idstring]++ : $scores[$idstring] = 1;
+ }
+ }
+ else
+ {
+ if ( strstr(strtolower($page['name']), strtolower($term)) || strstr(strtolower($page['urlname_nons']), strtolower($term)) )
+ {
+ ( isset($scores[$idstring]) ) ? $scores[$idstring]++ : $scores[$idstring] = 1;
+ }
+ }
+ }
+ if ( isset($scores[$idstring]) )
+ {
+ $page_data[$idstring] = array(
+ 'page_name' => $page['name'],
+ 'page_text' => '',
+ 'page_id' => $page['urlname_nons'],
+ 'namespace' => $page['namespace'],
+ 'score' => $scores[$idstring],
+ 'page_length' => 1,
+ 'page_note' => '[Special page]'
+ );
+ }
+ }
+
+ //
+ // STAGE 6 - SECOND ELIMINATION ROUND
+ // Iterate through the list of required terms. If a given page is not found to have the required term, eliminate it
+ //
+
+ $required = array_merge($query['req'], $query_phrase['req']);
+ foreach ( $required as $term )
+ {
+ foreach ( $page_data as $id => $page )
+ {
+ if ( ( $page['namespace'] == 'Special' || ( $page['namespace'] != 'Special' && !strstr($page['page_text'], $term) ) ) && !strstr($page['page_id'], $term) && !strstr($page['page_name'], $term) )
+ {
+ unset($page_data[$id]);
+ }
+ }
+ }
+
+ // At this point, all of our normal results are in. However, we can also allow plugins to hook into the system and score their own
+ // pages and add text, etc. as necessary.
+ // Plugins are COMPLETELY responsible for using the search terms and handling Boolean logic properly
+
+ $code = $plugins->setHook('search_global_inner');
+ foreach ( $code as $cmd )
+ {
+ eval($cmd);
+ }
+
+ // a marvelous debugging aid :-)
+ // die('<pre>' . htmlspecialchars(print_r($page_data, true)) . '</pre>');
+
+ //
+ // STAGE 7 - HIGHLIGHT, TRIM, AND SCORE RESULTS
+ // We now have the complete results of the search. We need to trim text down to show only portions of the page containing search
+ // terms, highlight any search terms within the page, and sort the final results array in descending order of score.
+ //
+
+ // Sort scores array
+ arsort($scores);
+
+ // Divisor for calculating relevance scores
+ $divisor = count($query['any']) + count($query_phrase['any']) + count($query['req']) + count($query_phrase['not']);
+
+ foreach ( $scores as $page_id => $score )
+ {
+ if ( !isset($page_data[$page_id]) )
+ // It's possible that $scores contains a score for a page that was later eliminated because it contained a disallowed term
+ continue;
+
+ // Make a copy of the datum, then delete the original (it frees up a LOT of RAM)
+ $datum = $page_data[$page_id];
+ unset($page_data[$page_id]);
+
+ // This is an internal value used for sorting - it's no longer needed.
+ unset($datum['id']);
+
+ // Calculate score
+ if ( $score > $divisor )
+ $score = $divisor;
+ $datum['score'] = round($score / $divisor, 2) * 100;
+
+ // Store it in our until-now-unused results array
+ $results[] = $datum;
+ }
+
+ // Our work here is done. :-D
+ return $results;
}
/**
- * Developer-friendly way to do searches. :-) Uses the MySQL FULLTEXT index type.
- * @package Enano
- * @subpackage Search
+ * Parses a search query into an associative array. The resultant array will be filled with the following values, each an array:
+ * any: Search terms that can optionally be present
+ * req: Search terms that must be present
+ * not: Search terms that should not be present
+ * @param string Search query
+ * @param array Will be filled with parser warnings, such as query too short, words too short, etc.
+ * @return array
*/
-class MySQL_Fulltext_Search {
-
- /**
- * Performs a search.
- * @param string The search query
- * @return resource MySQL result resource - this is an UNBUFFERED query.
- */
-
- function search($query)
+function parse_search_query($query, &$warnings)
+{
+ $stopwords = get_stopwords();
+ $ret = array(
+ 'any' => array(),
+ 'req' => array(),
+ 'not' => array()
+ );
+ $warnings = array();
+ $terms = array();
+ $in_quote = false;
+ $start_term = 0;
+ $just_finished = false;
+ for ( $i = 0; $i < strlen($query); $i++ )
{
- global $db, $session, $paths, $template, $plugins; // Common objects
+ $chr = $query{$i};
+ $prev = ( $i > 0 ) ? $query{ $i - 1 } : '';
+ $next = ( ( $i + 1 ) < strlen($query) ) ? $query{ $i + 1 } : '';
- $fulltext_col = 'MATCH(t.page_id,t.namespace,p.name,t.page_text) AGAINST (\'' . $db->escape($query) . '\' IN BOOLEAN MODE)';
- $sql = "SELECT t.page_text,CONCAT('ns=',t.namespace,';pid=',t.page_id) AS page_identifier, $fulltext_col AS score, CHAR_LENGTH(t.page_text) AS length FROM ".table_prefix."page_text AS t
- LEFT JOIN ".table_prefix."pages AS p
- ON ( p.urlname=t.page_id AND p.namespace=t.namespace)
- WHERE $fulltext_col > 0
- AND p.visible=1
- ORDER BY score DESC;";
- $q = $db->sql_unbuffered_query($sql);
- if ( !$q )
- $db->_die();
+ if ( ( $chr == ' ' && !$in_quote ) || ( $i + 1 == strlen ( $query ) ) )
+ {
+ $len = ( $next == '' ) ? $i + 1 : $i - $start_term;
+ $word = substr ( $query, $start_term, $len );
+ $terms[] = $word;
+ $start_term = $i + 1;
+ }
- return $q;
+ elseif ( $chr == '"' && $in_quote && $prev != '\\' )
+ {
+ $word = substr ( $query, $start_term, $i - $start_term + 1 );
+ $start_pos = ( $next == ' ' ) ? $i + 2 : $i + 1;
+ $in_quote = false;
+ }
+
+ elseif ( $chr == '"' && !$in_quote )
+ {
+ $in_quote = true;
+ $start_pos = $i;
+ }
+
}
- function highlight_result($query, $result)
+ $ticker = 0;
+
+ foreach ( $terms as $element => $__unused )
{
- global $db, $session, $paths, $template, $plugins; // Common objects
- $search = new Searcher();
- $parsed_query = $search->parseQuery($query);
- return $this->highlight_result_inner($query, $result);
+ $atom =& $terms[$element];
+
+ $ticker++;
+
+ if ( $ticker == 20 )
+ {
+ $warnings[] = 'Some of your search terms were excluded because searches are limited to 20 terms to prevent excessive server load.';
+ break;
+ }
+
+ if ( substr ( $atom, 0, 2 ) == '+"' && substr ( $atom, ( strlen ( $atom ) - 1 ), 1 ) == '"' )
+ {
+ $word = substr ( $atom, 2, ( strlen( $atom ) - 3 ) );
+ if ( strlen ( $word ) < 2 || in_array($word, $stopwords) )
+ {
+ $warnings[] = 'One or more of your search terms was excluded because either it was less than 2 characters in length or is a common word (a stopword) that is typically found on a large number of pages. Examples of stopwords include "the", "this", "which", "with", etc.';
+ $ticker--;
+ continue;
+ }
+ if(in_array($word, $ret['req']))
+ {
+ $warnings[] = 'One or more of your search terms was excluded because duplicate terms were encountered.';
+ $ticker--;
+ continue;
+ }
+ $ret['req'][] = $word;
+ }
+ elseif ( substr ( $atom, 0, 2 ) == '-"' && substr ( $atom, ( strlen ( $atom ) - 1 ), 1 ) == '"' )
+ {
+ $word = substr ( $atom, 2, ( strlen( $atom ) - 3 ) );
+ if ( strlen ( $word ) < 4 )
+ {
+ $warnings[] = 'One or more of your search terms was excluded because terms must be at least 4 characters in length.';
+ $ticker--;
+ continue;
+ }
+ if(in_array($word, $ret['not']))
+ {
+ $warnings[] = 'One or more of your search terms was excluded because duplicate terms were encountered.';
+ $ticker--;
+ continue;
+ }
+ $ret['not'][] = $word;
+ }
+ elseif ( substr ( $atom, 0, 1 ) == '+' )
+ {
+ $word = substr ( $atom, 1 );
+ if ( strlen ( $word ) < 2 || in_array($word, $stopwords) )
+ {
+ $warnings[] = 'One or more of your search terms was excluded because either it was less than 2 characters in length or is a common word (a stopword) that is typically found on a large number of pages. Examples of stopwords include "the", "this", "which", "with", etc.';
+ $ticker--;
+ continue;
+ }
+ if(in_array($word, $ret['req']))
+ {
+ $warnings[] = 'One or more of your search terms was excluded because duplicate terms were encountered.';
+ $ticker--;
+ continue;
+ }
+ $ret['req'][] = $word;
+ }
+ elseif ( substr ( $atom, 0, 1 ) == '-' )
+ {
+ $word = substr ( $atom, 1 );
+ if ( strlen ( $word ) < 2 || in_array($word, $stopwords) )
+ {
+ $warnings[] = 'One or more of your search terms was excluded because either it was less than 2 characters in length or is a common word (a stopword) that is typically found on a large number of pages. Examples of stopwords include "the", "this", "which", "with", etc.';
+ $ticker--;
+ continue;
+ }
+ if(in_array($word, $ret['not']))
+ {
+ $warnings[] = 'One or more of your search terms was excluded because duplicate terms were encountered.';
+ $ticker--;
+ continue;
+ }
+ $ret['not'][] = $word;
+ }
+ elseif ( substr ( $atom, 0, 1 ) == '"' && substr ( $atom, ( strlen($atom) - 1 ), 1 ) == '"' )
+ {
+ $word = substr ( $atom, 1, ( strlen ( $atom ) - 2 ) );
+ if ( strlen ( $word ) < 2 || in_array($word, $stopwords) )
+ {
+ $warnings[] = 'One or more of your search terms was excluded because either it was less than 2 characters in length or is a common word (a stopword) that is typically found on a large number of pages. Examples of stopwords include "the", "this", "which", "with", etc.';
+ $ticker--;
+ continue;
+ }
+ if(in_array($word, $ret['any']))
+ {
+ $warnings[] = 'One or more of your search terms was excluded because duplicate terms were encountered.';
+ $ticker--;
+ continue;
+ }
+ $ret['any'][] = $word;
+ }
+ else
+ {
+ $word = $atom;
+ if ( strlen ( $word ) < 2 || in_array($word, $stopwords) )
+ {
+ $warnings[] = 'One or more of your search terms was excluded because either it was less than 2 characters in length or is a common word (a stopword) that is typically found on a large number of pages. Examples of stopwords include "the", "this", "which", "with", etc.';
+ $ticker--;
+ continue;
+ }
+ if(in_array($word, $ret['any']))
+ {
+ $warnings[] = 'One or more of your search terms was excluded because duplicate terms were encountered.';
+ $ticker--;
+ continue;
+ }
+ $ret['any'][] = $word;
+ }
+ }
+ return $ret;
+}
+
+/**
+ * Escapes a string for use in a LIKE clause.
+ * @param string
+ * @return string
+ */
+
+function escape_string_like($string)
+{
+ global $db, $session, $paths, $template, $plugins; // Common objects
+ $string = $db->escape($string);
+ $string = str_replace(array('%', '_'), array('\%', '\_'), $string);
+ return $string;
+}
+
+/**
+ * Wraps <highlight></highlight> tags around all words in both the specified array. Does not perform any clipping.
+ * @param string Text to process
+ * @param array Word list
+ * @param bool If true, searches case-sensitively when highlighting words
+ * @return string
+ */
+
+function highlight_search_result($pt, $words, $case_sensitive = false)
+{
+ $words2 = array();
+ for ( $i = 0; $i < sizeof($words); $i++)
+ {
+ if(!empty($words[$i]))
+ $words2[] = preg_quote($words[$i]);
}
- function highlight_result_inner($query, $fulltext, $starttag = '<b>', $endtag = '</b>')
+ $flag = ( $case_sensitive ) ? '' : 'i';
+ $regex = '/(' . implode('|', $words2) . ')/' . $flag;
+ $pt = preg_replace($regex, '<highlight>\\1</highlight>', $pt);
+
+ return $pt;
+}
+
+/**
+ * Wraps <highlight></highlight> tags around all words in both the specified array and the specified text and clips the text to
+ * an appropriate length.
+ * @param string Text to process
+ * @param array Word list
+ * @param bool If true, searches case-sensitively when highlighting words
+ * @return string
+ */
+
+function highlight_and_clip_search_result($pt, $words, $case_sensitive = false)
+{
+ $cut_off = false;
+
+ $space_chars = Array("\t", "\n", "\r", " ");
+
+ $pt = highlight_search_result($pt, $words, $case_sensitive);
+
+ foreach ( $words as $word )
{
- $result = false;
- $query['trm'] = array_merge($query['any'], $query['req']);
- //die('<pre>'.print_r($query, true).'</pre>');
- foreach($query['trm'] as $q)
+ // Boldface searched words
+ $ptlen = strlen($pt);
+ for ( $i = 0; $i < $ptlen; $i++ )
{
- $startplace = 0;
- //$this->results[$k] = htmlspecialchars($this->results[$k]);
- for($i = 0; $i < strlen($r); $i++)
+ $len = strlen($word);
+ if ( strtolower(substr($pt, $i, $len)) == strtolower($word) )
{
- $word = substr($r, $i, strlen($q));
- if($this->convertCase($word) == $this->convertCase($q))
+ $chunk1 = substr($pt, 0, $i);
+ $chunk2 = substr($pt, $i, $len);
+ $chunk3 = substr($pt, ( $i + $len ));
+ $pt = $chunk1 . $chunk2 . $chunk3;
+ $ptlen = strlen($pt);
+ // Cut off text to 150 chars or so
+ if ( !$cut_off )
{
- $word = $starttag . $word . $endtag;
- $result = substr($fulltext, 0, $i) . $word . substr($r, $i + strlen($q), strlen($r)+99999999);
- $startplace = $i - 75;
- if($startplace < 0) $startplace = 0;
- $result = '...'.trim(substr($result, $startplace, strlen($word) + 150)).'...';
- continue 2;
+ $cut_off = true;
+ if ( $i - 75 > 0 )
+ {
+ // Navigate backwards until a space character is found
+ $chunk = substr($pt, 0, ( $i - 75 ));
+ $final_chunk = $chunk;
+ for ( $j = strlen($chunk); $j > 0; $j = $j - 1 )
+ {
+ if ( in_array($chunk{$j}, $space_chars) )
+ {
+ $final_chunk = substr($chunk, $j + 1);
+ break;
+ }
+ }
+ $mid_chunk = substr($pt, ( $i - 75 ), 75);
+
+ $clipped = '...' . $final_chunk . $mid_chunk . $chunk2;
+
+ $chunk = substr($pt, ( $i + strlen($chunk2) + 75 ));
+ $final_chunk = $chunk;
+ for ( $j = 0; $j < strlen($chunk); $j++ )
+ {
+ if ( in_array($chunk{$j}, $space_chars) )
+ {
+ $final_chunk = substr($chunk, 0, $j);
+ break;
+ }
+ }
+
+ $end_chunk = substr($pt, ( $i + strlen($chunk2) ), 75 );
+
+ $clipped .= $end_chunk . $final_chunk . '...';
+
+ $pt = $clipped;
+ }
+ else if ( strlen($pt) > 200 )
+ {
+ $mid_chunk = substr($pt, ( $i - 75 ), 75);
+
+ $clipped = $chunk1 . $chunk2;
+
+ $chunk = substr($pt, ( $i + strlen($chunk2) + 75 ));
+ $final_chunk = $chunk;
+ for ( $j = 0; $j < strlen($chunk); $j++ )
+ {
+ if ( in_array($chunk{$j}, $space_chars) )
+ {
+ $final_chunk = substr($chunk, 0, $j);
+ break;
+ }
+ }
+
+ $end_chunk = substr($pt, ( $i + strlen($chunk2) ), 75 );
+
+ $clipped .= $end_chunk . $final_chunk . '...';
+
+ $pt = $clipped;
+
+ }
+ break 2;
}
}
}
- return $result;
+ $cut_off = false;
}
+ return $pt;
+}
+
+/**
+ * Returns a list of words that shouldn't under most circumstances be indexed for searching. Kudos to MySQL.
+ * @return array
+ * @see http://dev.mysql.com/doc/refman/5.0/en/fulltext-stopwords.html
+ */
+
+function get_stopwords()
+{
+ static $stopwords;
+ if ( is_array($stopwords) )
+ return $stopwords;
+ $stopwords = array('a\'s', 'able', 'after', 'afterwards', 'again',
+ 'against', 'ain\'t', 'all', 'almost', 'alone', 'along', 'already', 'also', 'although', 'always',
+ 'am', 'among', 'amongst', 'an', 'and', 'another', 'any', 'anybody', 'anyhow', 'anyone', 'anything', 'anyway',
+ 'anyways', 'anywhere', 'apart', 'appear', 'appreciate', 'appropriate', 'are', 'aren\'t', 'around', 'as', 'aside',
+ 'ask', 'asking', 'associated', 'at', 'available', 'away', 'awfully', 'be', 'became', 'because', 'become', 'becomes',
+ 'becoming', 'been', 'before', 'beforehand', 'behind', 'being', 'believe', 'below', 'beside', 'besides', 'best',
+ 'better', 'between', 'beyond', 'both', 'brief', 'but', 'by', 'c\'mon', 'c\'s', 'came', 'can', 'can\'t', 'cannot',
+ 'cant', 'cause', 'causes', 'certain', 'certainly', 'changes', 'clearly', 'co', 'com', 'come', 'comes', 'concerning',
+ 'consequently', 'consider', 'considering', 'contain', 'containing', 'contains', 'corresponding', 'could',
+ 'couldn\'t', 'course', 'despite', 'did', 'didn\'t', 'different', 'do',
+ 'does', 'doesn\'t', 'doing', 'don\'t', 'done', 'down', 'downwards', 'during', 'each', 'edu', 'eg', 'eight',
+ 'either', 'else', 'elsewhere', 'enough', 'entirely', 'especially', 'et', 'etc', 'even', 'ever', 'every',
+ 'everybody', 'everyone', 'everything', 'everywhere', 'ex', 'exactly', 'example', 'except', 'far', 'few', 'fifth',
+ 'first', 'five', 'followed', 'following', 'follows', 'for', 'former', 'formerly', 'forth', 'four', 'from',
+ 'further', 'get', 'gets', 'getting', 'given', 'gives', 'go', 'goes', 'going', 'gone', 'got',
+ 'gotten', 'had', 'hadn\'t', 'happens', 'hardly', 'has', 'hasn\'t', 'have', 'haven\'t', 'having',
+ 'he', 'he\'s', 'hello', 'help', 'hence', 'her', 'here', 'here\'s', 'hereafter', 'hereby', 'herein', 'hereupon',
+ 'hers', 'herself', 'hi', 'him', 'himself', 'his', 'hither', 'hopefully', 'how', 'howbeit', 'however', 'i\'d',
+ 'i\'ll', 'i\'m', 'i\'ve', 'ie', 'if', 'ignored', 'immediate', 'in', 'inasmuch', 'inc', 'indeed', 'indicate',
+ 'indicated', 'indicates', 'inner', 'insofar', 'instead', 'into', 'inward', 'is', 'isn\'t', 'it', 'it\'d', 'it\'ll',
+ 'it\'s', 'its', 'itself', 'just', 'keep', 'keeps', 'kept', 'know', 'knows', 'known', 'last', 'lately', 'later',
+ 'latter', 'latterly', 'least', 'less', 'lest', 'let', 'let\'s', 'like', 'liked', 'likely', 'little', 'look',
+ 'looking', 'looks', 'ltd', 'mainly', 'many', 'may', 'maybe', 'me', 'mean', 'meanwhile', 'merely', 'might', 'more',
+ 'moreover', 'most', 'mostly', 'much', 'must', 'my', 'myself', 'name', 'namely', 'nd', 'near', 'nearly', 'necessary',
+ 'need', 'needs', 'neither', 'never', 'nevertheless', 'new', 'next', 'nine', 'no', 'nobody', 'non', 'none', 'noone',
+ 'nor', 'normally', 'not', 'nothing', 'novel', 'now', 'nowhere', 'obviously', 'of', 'off', 'often', 'oh', 'ok',
+ 'okay', 'old', 'on', 'once', 'one', 'ones', 'only', 'onto', 'or', 'other', 'others', 'otherwise', 'ought', 'our',
+ 'ours', 'ourselves', 'out', 'outside', 'over', 'overall', 'own', 'particular', 'particularly', 'per', 'perhaps',
+ 'placed', 'please', 'plus', 'possible', 'presumably', 'probably', 'provides', 'que', 'quite', 'qv', 'rather', 'rd',
+ 're', 'really', 'reasonably', 'regarding', 'regardless', 'regards', 'relatively', 'respectively', 'right', 'said',
+ 'same', 'saw', 'say', 'saying', 'says', 'second', 'secondly', 'see', 'seeing', 'seem', 'seemed', 'seeming', 'seems',
+ 'seen', 'self', 'selves', 'sensible', 'sent', 'serious', 'seriously', 'seven', 'several', 'shall', 'she', 'should',
+ 'shouldn\'t', 'since', 'six', 'so', 'some', 'somebody', 'somehow', 'someone', 'something', 'sometime', 'sometimes',
+ 'somewhat', 'somewhere', 'soon', 'sorry', 'specified', 'specify', 'specifying', 'still', 'sub', 'such', 'sup',
+ 'sure', 't\'s', 'take', 'taken', 'tell', 'tends', 'th', 'than', 'thank', 'thanks', 'thanx', 'that', 'that\'s',
+ 'thats', 'the', 'their', 'theirs', 'them', 'themselves', 'then', 'thence', 'there', 'there\'s', 'thereafter',
+ 'thereby', 'therefore', 'therein', 'theres', 'thereupon', 'these', 'they', 'they\'d', 'they\'ll', 'they\'re',
+ 'they\'ve', 'think', 'third', 'this', 'thorough', 'thoroughly', 'those', 'though', 'three', 'through', 'throughout',
+ 'thru', 'thus', 'to', 'together', 'too', 'took', 'toward', 'towards', 'tried', 'tries', 'truly', 'try', 'trying',
+ 'twice', 'two', 'un', 'under', 'unfortunately', 'unless', 'unlikely', 'until', 'unto', 'up', 'upon', 'us', 'use',
+ 'used', 'useful', 'uses', 'using', 'usually', 'value', 'various', 'very', 'via', 'viz', 'vs', 'want', 'wants',
+ 'was', 'wasn\'t', 'way', 'we', 'we\'d', 'we\'ll', 'we\'re', 'we\'ve', 'welcome', 'well', 'went', 'were', 'weren\'t',
+ 'what', 'what\'s', 'whatever', 'when', 'whence', 'whenever', 'where', 'where\'s', 'whereafter', 'whereas',
+ 'whereby', 'wherein', 'whereupon', 'wherever', 'whether', 'which', 'while', 'whither', 'who', 'who\'s', 'whoever',
+ 'whole', 'whom', 'whose', 'why', 'will', 'willing', 'wish', 'with', 'within', 'without', 'won\'t', 'wonder',
+ 'would', 'would', 'wouldn\'t', 'yes', 'yet', 'you', 'you\'d', 'you\'ll', 'you\'re', 'you\'ve', 'your', 'yours',
+ 'yourself', 'yourselves', 'zero');
+ return $stopwords;
}
?>
--- a/plugins/SpecialSearch.php Wed Nov 21 20:45:50 2007 -0500
+++ b/plugins/SpecialSearch.php Wed Nov 21 21:03:48 2007 -0500
@@ -52,7 +52,12 @@
function page_Special_Search()
{
global $db, $session, $paths, $template, $plugins; // Common objects
- if(!$q = $paths->getParam(0)) $q = ( isset($_GET['q']) ) ? $_GET['q'] : false;
+ global $aggressive_optimize_html;
+ $aggressive_optimize_html = false;
+
+ if ( !$q = $paths->getParam(0) )
+ $q = ( isset($_GET['q']) ) ? $_GET['q'] : '';
+
if(isset($_GET['words_any']))
{
$q = '';
@@ -85,184 +90,167 @@
}
$q = trim($q);
- if ( !empty($q) && !isset($_GET['search']) )
+ $template->header();
+
+ $qin = ( isset($q) ) ? str_replace('"', '\"', htmlspecialchars($q)) : '';
+ $search_form = '<form action="' . makeUrlNS('Special', 'Search') . '">
+ <input type="text" tabindex="1" name="q" size="50" value="' . $qin . '" /> <input tabindex="2" type="submit" value="Search" />
+ ' . ( $session->auth_level > USER_LEVEL_MEMBER ? '<input type="hidden" name="auth" value="' . $session->sid_super . '" />' : '' ) . '
+ </form>';
+
+ if ( !empty($q) )
{
- list($pid, $ns) = RenderMan::strToPageID($q);
- $pid = sanitize_page_id($pid);
- $key = $paths->nslist[$ns] . $pid;
- if ( isPage($key) )
- {
- redirect(makeUrl($key), 'Results', 'found page', 0);
- }
- }
-
- $template->header();
- if(!empty($q))
- {
- // See if any pages directly match the title
+ $search_start = microtime_float();
+
+ $results = perform_search($q, $warn, ( isset($_GET['match_case']) ));
+ $warn = array_unique($warn);
- if ( strlen($q) >= 4 )
+ if ( file_exists( ENANO_ROOT . '/themes/' . $template->theme . '/search-result.tpl' ) )
+ {
+ $parser = $template->makeParser('search-result.tpl');
+ }
+ else
{
- for ( $i = 0; $i < count ( $paths->pages ) / 2; $i++ )
- {
- $pg =& $paths->pages[$i];
- $q_lc = strtolower( str_replace(' ', '_', $q) );
- $q_tl = strtolower( str_replace('_', ' ', $q) );
- $p_lc = strtolower($pg['urlname']);
- $p_tl = strtolower($pg['name']);
- if ( strstr($p_tl, $q_tl) || strstr($p_lc, $q_lc) && $pg['visible'] == 1 )
- {
- echo '<div class="usermessage">Perhaps you were looking for <b><a href="' . makeUrl($pg['urlname'], false, true) . '">' . htmlspecialchars($pg['name']) . '</a></b>?</div>';
- break;
- }
- }
+ $tpl_code = <<<LONGSTRING
+
+ <!-- Start search result -->
+
+ <div class="search-result">
+ <p>
+ <h3><a href="{RESULT_URL}"><span class="search-result-annotation">{PAGE_NOTE}</span>{PAGE_TITLE}</a></h3>
+ {PAGE_TEXT}
+ <span class="search-result-url">{PAGE_URL}</span> -
+ <!-- BEGINNOT special_page --><span class="search-result-info">{PAGE_LENGTH} {PAGE_LENGTH_UNIT}</span> -<!-- END special_page -->
+ <span class="search-result-info">Relevance: {RELEVANCE_SCORE}%</span>
+ </p>
+ </div>
+
+ <!-- Finish search result -->
+
+LONGSTRING;
+ $parser = $template->makeParserText($tpl_code);
}
-
- switch(SEARCH_MODE)
+ foreach ( $results as $i => $_ )
{
+ $result =& $results[$i];
+ $result['page_text'] = str_replace(array('<highlight>', '</highlight>'), array('<span class="highlight">', '</span>'), $result['page_text']);
+ if ( !empty($result['page_text']) )
+ $result['page_text'] .= '<br />';
+ $result['page_name'] = str_replace(array('<highlight>', '</highlight>'), array('<span class="highlight">', '</span>'), $result['page_name']);
+ if ( $result['page_length'] >= 1048576 )
+ {
+ $result['page_length'] = round($result['page_length'] / 1048576, 1);
+ $length_unit = 'MB';
+ }
+ else if ( $result['page_length'] >= 1024 )
+ {
+ $result['page_length'] = round($result['page_length'] / 1024, 1);
+ $length_unit = 'KB';
+ }
+ else
+ {
+ $length_unit = 'bytes';
+ }
+ $url = makeUrlComplete($result['namespace'], $result['page_id']);
+ $url = preg_replace('/\?.+$/', '', $url);
+ $parser->assign_vars(array(
+ 'PAGE_TITLE' => $result['page_name'],
+ 'PAGE_TEXT' => $result['page_text'],
+ 'PAGE_LENGTH' => $result['page_length'],
+ 'RELEVANCE_SCORE' => $result['score'],
+ 'RESULT_URL' => makeUrlNS($result['namespace'], $result['page_id'], false, true),
+ 'PAGE_LENGTH_UNIT' => $length_unit,
+ 'PAGE_URL' => $url,
+ 'PAGE_NOTE' => ( isset($result['page_note']) ? $result['page_note'] . ' ' : '' )
+ ));
+ $has_content = ( $result['namespace'] == 'Special' );
- case "FULLTEXT":
- if ( isset($_GET['offset']) )
- {
- $offset = intval($_GET['offset']);
- }
- else
- {
- $offset = 0;
- }
- $sql = $db->sql_query('SELECT search_id FROM '.table_prefix.'search_cache WHERE query=\''.$db->escape($q).'\';');
- if(!$sql)
- {
- $db->_die('Error scanning search query cache');
- }
- if($db->numrows() > 0)
- {
- $row = $db->fetchrow();
- $db->free_result();
- search_fetch_fulltext_results(intval($row['search_id']), $offset);
- }
- else
- {
- // Perform search
-
- $search = new MySQL_Fulltext_Search();
-
- // Parse the query
- $parse = new Searcher();
- $query = $parse->parseQuery($q);
- unset($parse);
+ $code = $plugins->setHook('search_global_results');
+ foreach ( $code as $cmd )
+ {
+ eval($cmd);
+ }
+
+ $parser->assign_bool(array(
+ 'special_page' => $has_content
+ ));
+ $result = $parser->run();
+ }
+ unset($result);
+
+ $per_page = 10;
+ $start = ( isset($_GET['start']) ? intval($_GET['start']) : 0 );
+ $start_string = $start + 1;
+ $per_string = $start_string + $per_page - 1;
+ $num_results = count($results);
+ if ( $per_string > $num_results )
+ $per_string = $num_results;
+
+ $search_time = microtime_float() - $search_start;
+ $search_time = round($search_time, 3);
+
+ $q_trim = ( strlen($q) > 30 ) ? substr($q, 0, 27) . '...' : $q;
+ $q_trim = htmlspecialchars($q_trim);
+
+ $result_string = ( count($results) > 0 ) ? "Results <b>$start_string</b> - <b>$per_string</b> of about <b>$num_results</b> for <b>" . $q_trim . "</b> in {$search_time}s." : 'No results.';
+
+ echo '<div class="search-hibar">
+ <div style="float: right;">
+ ' . $result_string . '
+ </div>
+ <b>Site search</b>
+ </div>
+ <div class="search-lobar">
+ ' . $search_form . '
+ </div>';
- // Send query to MySQL
- $sql = $search->search($q);
- $results = Array();
- if ( $row = $db->fetchrow($sql) )
- {
- do {
- $results[] = $row;
- } while ( $row = $db->fetchrow($sql) );
- }
- else
- {
- // echo '<div class="warning-box">No pages that matched your search criteria could be found.</div>';
- }
- $texts = Array();
- foreach ( $results as $result )
- {
- $texts[] = render_fulltext_result($result, $query);
- }
-
- // Store the result in the search cache...if someone makes the same query later we can skip searching and rendering
- // This cache is cleared when an affected page is saved.
-
- $results = serialize($texts);
-
- $sql = $db->sql_query('INSERT INTO '.table_prefix.'search_cache(search_time,query,results) VALUES('.time().', \''.$db->escape($q).'\', \''.$db->escape($results).'\');');
- if($sql)
- {
- search_render_fulltext_results(unserialize($results), $offset, $q);
- }
- else
- {
- $db->_die('Error inserting search into cache');
- }
-
- }
- break;
-
- case "BUILTIN":
- $titles = $paths->makeTitleSearcher(isset($_GET['match_case']));
- if ( isset($_GET['offset']) )
- {
- $offset = intval($_GET['offset']);
- }
- else
- {
- $offset = 0;
- }
- $sql = $db->sql_query('SELECT search_id FROM '.table_prefix.'search_cache WHERE query=\''.$db->escape($q).'\';');
- if(!$sql)
- {
- $db->_die('Error scanning search query cache');
- }
- if($db->numrows() > 0)
- {
- $row = $db->fetchrow();
- $db->free_result();
- search_show_results(intval($row['search_id']), $offset);
- }
- else
- {
- $titles->search($q, $paths->get_page_titles());
- $search = $paths->makeSearcher(isset($_GET['match_case']));
- $texts = $paths->fetch_page_search_resource();
- $search->searchMySQL($q, $texts);
-
- $results = Array();
- $results['text'] = $search->results;
- $results['page'] = $titles->results;
- $results['warn'] = $search->warnings;
-
- $results = serialize($results);
-
- $sql = $db->sql_query('INSERT INTO '.table_prefix.'search_cache(search_time,query,results) VALUES('.time().', \''.$db->escape($q).'\', \''.$db->escape($results).'\');');
- if($sql)
- {
- search_render_results(unserialize($results), $offset, $q);
- }
- else
- {
- $db->_die('Error inserting search into cache');
- }
- }
- break;
+ if ( count($warn) > 0 )
+ {
+ echo '<div class="warning-box" style="margin: 10px 0 0 0;">';
+ echo '<b>Some problems were encountered during your search.</b><br />
+ There was a problem with your search query, and as a result there may be a reduced number of search results.';
+ echo '<ul><li>' . implode('</li><li>', $warn) . '</li></ul>';
+ echo '</div>';
}
- $code = $plugins->setHook('search_results'); // , Array('query'=>$q));
+
+ if ( count($results) > 0 )
+ {
+ $html = paginate_array(
+ $results,
+ count($results),
+ makeUrlNS('Special', 'Search', 'q=' . str_replace('%', '%%', htmlspecialchars(urlencode($q))) . '&start=%s'),
+ $start,
+ $per_page
+ );
+ echo $html;
+ }
+ else
+ {
+ // No results for the search
+ echo '<h3 style="font-weight: normal;">Your search for <b>"' . htmlspecialchars($q) . '"</b> didn\'t turn up any results.</h3>';
+ echo '<p>There are a few things you can try:</p>';
+ echo '<ul>
+ <li>Were you looking for a specific Special page? Special pages are not searchable. You may want to see a <a href="' . makeUrlNS('Special', 'SpecialPages') . '">list of special pages</a>.</li>
+ <li>If you have the appropriate permissions, you can <a href="' . makeUrl($q) . '#do:edit">start the ' . htmlspecialchars($q) . ' page</a>.</li>
+ <li>Try using fewer keywords. You can get broader results if you remove quotes from your search query.</li>
+ <li>Did your search trigger any warnings? Sometimes a search can be cancelled if there aren\'t any terms in a search query that are 4 characters or greater in length.</li>
+ </ul>';
+ }
+ $code = $plugins->setHook('search_results');
foreach ( $code as $cmd )
{
eval($cmd);
}
- ?>
- <form action="<?php echo makeUrl($paths->page); ?>" method="get">
- <p>
- <?php if ( $session->sid_super ): ?>
- <input type="hidden" name="auth" value="<?php echo $session->sid_super; ?>" />
- <?php endif; ?>
- <?php if ( urlSeparator == '&' ): ?>
- <input type="hidden" name="title" value="<?php echo $paths->nslist['Special'] . 'Search'; ?>" />
- <?php endif; ?>
- <input type="text" name="q" size="40" value="<?php echo htmlspecialchars( $q ); ?>" /> <input type="submit" value="Go" style="font-weight: bold;" /> <input name="search" type="submit" value="Search" /> <small><a href="<?php echo makeUrlNS('Special', 'Search'); ?>">Advanced Search</a></small>
- </p>
- </form>
- <?php
}
else
{
- ?>
- <br />
+ ?>
<form action="<?php echo makeUrl($paths->page); ?>" method="get">
<?php if ( urlSeparator == '&' ): ?>
<input type="hidden" name="title" value="<?php echo $paths->nslist['Special'] . 'Search'; ?>" />
- <?php endif; ?>
+ <?php
+ echo ( $session->auth_level > USER_LEVEL_MEMBER ? '<input type="hidden" name="auth" value="' . $session->sid_super . '" />' : '' );
+ endif; ?>
<div class="tblholder">
<table border="0" style="width: 100%;" cellspacing="1" cellpadding="4">
<tr><th colspan="2">Advanced Search</th></tr>
@@ -298,255 +286,10 @@
</table>
</div>
</form>
- <?php
+ <?php
}
+
$template->footer();
}
-function search_show_results($search_id, $start = 0)
-{
- global $db, $session, $paths, $template, $plugins; // Common objects
- $q = $db->sql_query('SELECT query,results,search_time FROM '.table_prefix.'search_cache WHERE search_id='.intval($search_id).';');
- if(!$q)
- return $db->get_error('Error selecting cached search results');
- $row = $db->fetchrow();
- $db->free_result();
- $results = unserialize($row['results']);
- search_render_results($results, $start, $row['query']);
-}
-
-function search_render_results($results, $start = 0, $q = '')
-{
- global $db, $session, $paths, $template, $plugins; // Common objects
- $nr1 = sizeof($results['page']);
- $nr2 = sizeof($results['text']);
- $nr = ( $nr1 > $nr2 ) ? $nr1 : $nr2;
- $results['page'] = array_slice($results['page'], $start, SEARCH_RESULTS_PER_PAGE);
- $results['text'] = array_slice($results['text'], $start, SEARCH_RESULTS_PER_PAGE);
-
- // Pagination
- $pagination = '';
- if ( $nr1 > SEARCH_RESULTS_PER_PAGE || $nr2 > SEARCH_RESULTS_PER_PAGE )
- {
- $pagination .= '<div class="tblholder" style="padding: 0; display: table; margin: 0 0 0 auto; float: right;">
- <table border="0" style="width: 100%;" cellspacing="1" cellpadding="4">
- <tr>
- <th>Page:</th>';
- $num_pages = ceil($nr / SEARCH_RESULTS_PER_PAGE);
- $j = 0;
- for ( $i = 1; $i <= $num_pages; $i++ )
- {
- if ($j == $start)
- $pagination .= '<td class="row1"><b>' . $i . '</b></td>';
- else
- $pagination .= '<td class="row1"><a href="' . makeUrlNS('Special', 'Search', 'q=' . urlencode($q) . '&offset=' . $j, true) . '">' . $i . '</a></td>';
- $j = $j + SEARCH_RESULTS_PER_PAGE;
- }
- $pagination .= '</tr></table></div>';
- }
-
- echo $pagination;
-
- if ( $nr1 >= $start )
- {
- echo '<h3>Page title matches</h3>';
- if(count($results['page']) < 1)
- {
- echo '<div class="error-box">No pages with a title that matched your search criteria could be found.</div>';
- }
- else
- {
- echo '<p>';
- foreach($results['page'] as $page => $text)
- {
- echo '<a href="'.makeUrl($page).'">'.$paths->pages[$page]['name'].'</a><br />';
- }
- echo '</p>';
- }
- }
- if ( $nr2 >= $start )
- {
- echo '<h3>Page text matches</h3>';
- if(count($results['text']) < 1)
- {
- echo '<div class="error-box">No page text that matched your search criteria could be found.</div>';
- }
- else
- {
- foreach($results['text'] as $kpage => $text)
- {
- preg_match('#^ns=('.implode('|', array_keys($paths->nslist)).');pid=(.*?)$#i', $kpage, $matches);
- $page = $paths->nslist[$matches[1]] . $matches[2];
- echo '<p><span style="font-size: larger;"><a href="'.makeUrl($page).'">'.$paths->pages[$page]['name'].'</a></span><br />'.$text.'</p>';
- }
- }
- }
- if(count($results['warn']) > 0)
- echo '<div class="warning-box"><b>Your search may not include all results.</b><br />The following errors were encountered during the search:<br /><ul><li>'.implode('</li><li>', $results['warn']).'</li></ul></div>';
- echo $pagination;
-}
-
-function render_fulltext_result($result, $query)
-{
- global $db, $session, $paths, $template, $plugins; // Common objects
- preg_match('#^ns=('.implode('|', array_keys($paths->nslist)).');pid=(.*?)$#i', $result['page_identifier'], $matches);
- $page = $paths->nslist[$matches[1]] . $matches[2];
- //$score = round($result['score'] * 100, 1);
- $score = number_format($result['score'], 2);
- $char_length = $result['length'];
- $result_template = <<<TPLCODE
- <div class="search-result">
- <h3><a href="{HREF}">{TITLE}</a></h3>
- <p>{TEXT}</p>
- <p>
- <span class="search-result-info">{NAMESPACE} - Relevance score: {SCORE} ({LENGTH} bytes)</span>
- </p>
- </div>
-TPLCODE;
- $parser = $template->makeParserText($result_template);
-
- $pt =& $result['page_text'];
- $space_chars = Array("\t", "\n", "\r", " ");
-
- $words = array_merge($query['any'], $query['req']);
- $pt = htmlspecialchars($pt);
- $words2 = array();
-
- for ( $i = 0; $i < sizeof($words); $i++)
- {
- if(!empty($words[$i]))
- $words2[] = preg_quote($words[$i]);
- }
-
- $regex = '/(' . implode('|', $words2) . ')/i';
- $pt = preg_replace($regex, '<span class="search-term">\\1</span>', $pt);
-
- $title = preg_replace($regex, '<span class="title-search-term">\\1</span>', htmlspecialchars($paths->pages[$page]['name']));
-
- $cut_off = false;
-
- foreach ( $words as $word )
- {
- // Boldface searched words
- $ptlen = strlen($pt);
- for ( $i = 0; $i < $ptlen; $i++ )
- {
- $len = strlen($word);
- if ( strtolower(substr($pt, $i, $len)) == strtolower($word) )
- {
- $chunk1 = substr($pt, 0, $i);
- $chunk2 = substr($pt, $i, $len);
- $chunk3 = substr($pt, ( $i + $len ));
- $pt = $chunk1 . $chunk2 . $chunk3;
- $ptlen = strlen($pt);
- // Cut off text to 150 chars or so
- if ( !$cut_off )
- {
- $cut_off = true;
- if ( $i - 75 > 0 )
- {
- // Navigate backwards until a space character is found
- $chunk = substr($pt, 0, ( $i - 75 ));
- $final_chunk = $chunk;
- for ( $j = strlen($chunk); $j > 0; $j = $j - 1 )
- {
- if ( in_array($chunk{$j}, $space_chars) )
- {
- $final_chunk = substr($chunk, $j + 1);
- break;
- }
- }
- $mid_chunk = substr($pt, ( $i - 75 ), 75);
-
- $clipped = '...' . $final_chunk . $mid_chunk . $chunk2;
-
- $chunk = substr($pt, ( $i + strlen($chunk2) + 75 ));
- $final_chunk = $chunk;
- for ( $j = 0; $j < strlen($chunk); $j++ )
- {
- if ( in_array($chunk{$j}, $space_chars) )
- {
- $final_chunk = substr($chunk, 0, $j);
- break;
- }
- }
-
- $end_chunk = substr($pt, ( $i + strlen($chunk2) ), 75 );
-
- $clipped .= $end_chunk . $final_chunk . '...';
-
- $pt = $clipped;
- }
- else if ( strlen($pt) > 200 )
- {
- $mid_chunk = substr($pt, ( $i - 75 ), 75);
-
- $clipped = $chunk1 . $chunk2;
-
- $chunk = substr($pt, ( $i + strlen($chunk2) + 75 ));
- $final_chunk = $chunk;
- for ( $j = 0; $j < strlen($chunk); $j++ )
- {
- if ( in_array($chunk{$j}, $space_chars) )
- {
- $final_chunk = substr($chunk, 0, $j);
- break;
- }
- }
-
- $end_chunk = substr($pt, ( $i + strlen($chunk2) ), 75 );
-
- $clipped .= $end_chunk . $final_chunk . '...';
-
- $pt = $clipped;
-
- }
- break 2;
- }
- }
- }
- $cut_off = false;
- }
-
- $parser->assign_vars(Array(
- 'TITLE' => $title,
- 'TEXT' => $pt,
- 'NAMESPACE' => $matches[1],
- 'SCORE' => $score,
- 'LENGTH' => $char_length,
- 'HREF' => makeUrl($page)
- ));
-
- return $parser->run();
-
-}
-
-function search_fetch_fulltext_results($search_id, $offset = 0)
-{
- global $db, $session, $paths, $template, $plugins; // Common objects
- $q = $db->sql_query('SELECT query,results,search_time FROM '.table_prefix.'search_cache WHERE search_id='.intval($search_id).';');
- if(!$q)
- return $db->get_error('Error selecting cached search results');
- $row = $db->fetchrow();
- $db->free_result();
- $results = unserialize($row['results']);
- search_render_fulltext_results($results, $offset, $row['query']);
-}
-
-function search_render_fulltext_results($results, $offset = 0, $query)
-{
- $num_results = sizeof($results);
- $slice = array_slice($results, $offset, SEARCH_RESULTS_PER_PAGE);
-
- if ( $num_results < 1 )
- {
- echo '<div class="warning-box" style="margin-left: 0;">No page text that matched your search criteria could be found.</div>';
- return null;
- }
-
- $html = paginate_array($results, sizeof($results), makeUrlNS('Special', 'Search', 'q=' . urlencode($query) . '&offset=%s'), $offset, 10);
- echo $html . '<br />';
-
-}
-
?>