/[drupal]/contributions/modules/xapian/xapian.module
ViewVC logotype

Diff of /contributions/modules/xapian/xapian.module

Parent Directory Parent Directory | Revision Log Revision Log | View Revision Graph Revision Graph | View Patch Patch

revision 1.10.2.5, Fri Jan 30 14:24:13 2009 UTC revision 1.10.2.6, Thu Feb 5 23:10:50 2009 UTC
# Line 1  Line 1 
1  <?php  <?php
2  // $Id: xapian.module,v 1.10.2.4 2009/01/16 00:18:39 simon Exp $  // $Id: xapian.module,v 1.10.2.5 2009/01/30 14:24:13 simon Exp $
3    
4  /**  /**
5   * @file   * @file
# Line 24  define('XAPIAN_FLUSH', TRUE); Line 24  define('XAPIAN_FLUSH', TRUE);
24   * @param int $errline   * @param int $errline
25   * @param unknown_type $errcontext   * @param unknown_type $errcontext
26   */   */
27  function _xapian_requirements_error_handler($errno, $errstr, $errfile = NULL, $errline = NULL, $errcontext = NULL) {  function _xapian_requirements_error_handler($errno, $errstr, $errfile = NULL, $errline = NULL, $errcontext = NULL) {
28    $GLOBALS['xapian_include_error'] = $errstr;    $GLOBALS['xapian_include_error'] = $errstr;
29  }  }
30    
# Line 47  function xapian_available() { Line 47  function xapian_available() {
47        $available = TRUE;        $available = TRUE;
48      }      }
49    }    }
50    
51    return $available;    return $available;
52  }  }
53    
# Line 121  function xapian_menu() { Line 121  function xapian_menu() {
121  function xapian_admin() {  function xapian_admin() {
122    $form = array();    $form = array();
123    $form['#cache'] = FALSE;    $form['#cache'] = FALSE;
124    
125    $database_type = variable_get('xapian_database_type', 0);    $database_type = variable_get('xapian_database_type', 0);
126    
127    // Create a database fieldset    // Create a database fieldset
128    $form['database'] = array(    $form['database'] = array(
129      '#type' => 'fieldset',      '#type' => 'fieldset',
130      '#title' => t('Xapian database')      '#title' => t('Xapian database')
131    );    );
132    
133    // Database type    // Database type
134    $form['database']['xapian_database_type'] = array(    $form['database']['xapian_database_type'] = array(
135      '#type' => 'radios',      '#type' => 'radios',
# Line 243  function xapian_admin() { Line 243  function xapian_admin() {
243      '#default_value' => variable_get('xapian_log_queries', FALSE),      '#default_value' => variable_get('xapian_log_queries', FALSE),
244    );    );
245    
246          // Node Type Settings    // Node Type Settings
247          $form['node_types'] = array(    $form['node_types'] = array(
248      '#type' => 'fieldset',      '#type' => 'fieldset',
249      '#title' => t('Node Types')      '#title' => t('Node Types')
250    );    );
# Line 333  function _xapian_init_database($writable Line 333  function _xapian_init_database($writable
333    if (!$writable && is_object($database)) {    if (!$writable && is_object($database)) {
334      return $database;      return $database;
335    }    }
336    
337    if ($writable && is_object($writable_database)) {    if ($writable && is_object($writable_database)) {
338      return $writable_database;      return $writable_database;
339    }    }
# Line 384  function _xapian_init_database($writable Line 384  function _xapian_init_database($writable
384   * The xapian_query function queries the database using both a query string   * The xapian_query function queries the database using both a query string
385   * and application-defined terms.   * and application-defined terms.
386   *   *
387   * @param $query_string A string (perhaps supplied by the user) containing   * @param $query_string A string (perhaps supplied by the user) containing
388   *                      terms to search for.  This string will be parsed and   *                      terms to search for.  This string will be parsed and
389   *                      stemmed automatically.   *                      stemmed automatically.
390   *   *
391   * @param $start        An integer defining the first document to return   * @param $start        An integer defining the first document to return
# Line 453  function xapian_query($query_string, $st Line 453  function xapian_query($query_string, $st
453        case 1: // Lower bound        case 1: // Lower bound
454          $count = $matches->get_matches_lower_bound();          $count = $matches->get_matches_lower_bound();
455          break;          break;
456    
457        case 2: // Upper bound        case 2: // Upper bound
458          $count = $matches->get_matches_upper_bound();          $count = $matches->get_matches_upper_bound();
459          break;          break;
# Line 511  if (!function_exists('do_search')) { Line 511  if (!function_exists('do_search')) {
511        if (is_numeric($condition)) {        if (is_numeric($condition)) {
512          $term = taxonomy_get_term($condition);          $term = taxonomy_get_term($condition);
513          if ($term) {          if ($term) {
514            $terms[] = 'XTAXONOMY'. strtolower($term->name);            $terms[] = 'XTAXONOMY'. strtolower($term->name);
515          }          }
516        }        }
517        else {        else {
# Line 525  if (!function_exists('do_search')) { Line 525  if (!function_exists('do_search')) {
525    
526      $pager_total[0] = (int)($count / variable_get('xapian_search_results_per_page', 10)) + 1;      $pager_total[0] = (int)($count / variable_get('xapian_search_results_per_page', 10)) + 1;
527      $pager_page_array[0] = $page;      $pager_page_array[0] = $page;
528    
529      return $results;      return $results;
530    }    }
531  }  }
# Line 544  function _xapian_index_node($node, $dele Line 544  function _xapian_index_node($node, $dele
544      watchdog('xapian', 'Could not get writable database.');      watchdog('xapian', 'Could not get writable database.');
545      return FALSE;      return FALSE;
546    }    }
547    
548    // Track which content is being indexed until the Xapian cache is flushed.    // Track which content is being indexed until the Xapian cache is flushed.
549    // If we  fail to flush the cache to disk for any reason, we will re-attempt    // If we  fail to flush the cache to disk for any reason, we will re-attempt
550    // to index this content.  See bug #272140 for full details.    // to index this content.  See bug #272140 for full details.
551    db_query('UPDATE {xapian_index_queue} SET status = status + 1 WHERE nid = %d', $node->nid);    db_query('UPDATE {xapian_index_queue} SET status = status + 1 WHERE nid = %d', $node->nid);
552    
553    // Provide mechanism to override default indexing behaviour    // Provide mechanism to override default indexing behaviour
554    $function = $node->type .'_xapian_index';    $function = $node->type .'_xapian_index';
555    
556    // But fallback to standard node indexing    // But fallback to standard node indexing
557    if (!function_exists($function)) {    if (!function_exists($function)) {
558      $function = 'node_xapian_index';      $function = 'node_xapian_index';
559    }    }
560    
561    $terms = $function($node);    $terms = $function($node);
562    if (is_array($terms)) {    if (is_array($terms)) {
563    
564      try {      try {
   
565        if ($delete) {        if ($delete) {
566          watchdog('xapian', 'Delete attempt Node with id %node', array('%node' => $node->nid));          watchdog('xapian', 'Delete attempt Node with id %node', array('%node' => $node->nid));
567          $db->delete_document('Q'. $node->nid);          $db->delete_document('Q'. $node->nid);
568        }        }
569        else{        else{
570          if (!is_object($indexer)) {          if (!is_object($indexer)) {
571            $indexer = new XapianTermGenerator();            $indexer = new XapianTermGenerator();
572          }          }
573    
574          if (!is_object($stemmer)) {          if (!is_object($stemmer)) {
575            $stemmer = new XapianStem("english");            $stemmer = new XapianStem("english");
576          }          }
577          $indexer->set_stemmer($stemmer);          $indexer->set_stemmer($stemmer);
578    
579          $document = new XapianDocument();          $document = new XapianDocument();
580          $indexer->set_document($document);          $indexer->set_document($document);
581    
582          foreach ($terms as $term) {          foreach ($terms as $term) {
583            if (!$term['weight']) {            if (!$term['weight']) {
584              $term['weight'] = 1;              $term['weight'] = 1;
# Line 596  function _xapian_index_node($node, $dele Line 595  function _xapian_index_node($node, $dele
595                break;                break;
596            }            }
597          }          }
598    
599          $document->set_data($node->nid);          $document->set_data($node->nid);
600    
601          $document->add_term('Q'. $node->nid);          $document->add_term('Q'. $node->nid);
602          $document->add_term('XNODETYPE'. strtolower($node->type));          $document->add_term('XNODETYPE'. strtolower($node->type));
603    
604          $db->replace_document('Q'. $node->nid, $document);          $db->replace_document('Q'. $node->nid, $document);
605        }        }
606      }      }
607      catch (Exception $e) {      catch (Exception $e) {
608        watchdog('xapian', $e->getMessage());        watchdog('xapian', $e->getMessage());
609      }      }
610    }    }
611    
612    return TRUE;    return TRUE;
613  }  }
614    
# Line 633  function xapian_nodeapi(&$node, $op) { Line 632  function xapian_nodeapi(&$node, $op) {
632  }  }
633    
634  /**  /**
635   * Implementation of hook_comment to catch inserts, updates and deletes and   * Implementation of hook_comment to catch inserts, updates and deletes and
636   * update the index.   * update the index.
637   *   *
638   * @param unknown type $a1   * @param unknown type $a1
# Line 660  function xapian_comment($a1, $op) { Line 659  function xapian_comment($a1, $op) {
659  }  }
660    
661  /**  /**
662   * Queue a node for indexing   * Queue a node for indexing
663   *   *
664   * @param object $node   * @param object $node
665   */   */
# Line 681  function _xapian_queue_index($node) { Line 680  function _xapian_queue_index($node) {
680  function xapian_index_node(&$node) {  function xapian_index_node(&$node) {
681    $exclusions = array_values(variable_get('xapian_excluded_nodes', array()));    $exclusions = array_values(variable_get('xapian_excluded_nodes', array()));
682    if (in_array($node->type, $exclusions)) {    if (in_array($node->type, $exclusions)) {
683          return;      return;
684    }    }
685    if (variable_get('xapian_index_immediately', TRUE)) {    if (variable_get('xapian_index_immediately', TRUE)) {
686      if (_xapian_index_node($node)) {      if (_xapian_index_node($node)) {
# Line 727  function xapian_index_queued_nodes() { Line 726  function xapian_index_queued_nodes() {
726    $failed = 0;    $failed = 0;
727    $delete = FALSE;    $delete = FALSE;
728    $start_time = microtime(TRUE);    $start_time = microtime(TRUE);
729    
730    // Check to see if some indexed nodes failed to be flushed to disk last time    // Check to see if some indexed nodes failed to be flushed to disk last time
731    // we tried indexing nodes.    // we tried indexing nodes.
732    $not_flushed = db_result(db_query('SELECT COUNT(xid) as count FROM {xapian_index_queue} WHERE status > 0'));    $not_flushed = db_result(db_query('SELECT COUNT(xid) as count FROM {xapian_index_queue} WHERE status > 0'));
# Line 739  function xapian_index_queued_nodes() { Line 738  function xapian_index_queued_nodes() {
738    }    }
739    else if ($not_flushed == 1) {    else if ($not_flushed == 1) {
740      // Re-index a node that previously failed to index, removing it from the      // Re-index a node that previously failed to index, removing it from the
741      // queue now as it may not be indexable, such as a php node calling      // queue now as it may not be indexable, such as a php node calling
742      // drupal_goto()...      // drupal_goto()...
743      $delete = db_result(db_query_range('SELECT nid FROM {xapian_index_queue} WHERE status > 0', 0, 1));      $delete = db_result(db_query_range('SELECT nid FROM {xapian_index_queue} WHERE status > 0', 0, 1));
744      $sql = 'SELECT nid FROM {xapian_index_queue} WHERE nid = '. $delete;      $sql = 'SELECT nid FROM {xapian_index_queue} WHERE nid = '. $delete;
# Line 749  function xapian_index_queued_nodes() { Line 748  function xapian_index_queued_nodes() {
748      $sql = 'SELECT nid FROM {xapian_index_queue} ORDER BY COALESCE(priority, 0) DESC, added ASC';      $sql = 'SELECT nid FROM {xapian_index_queue} ORDER BY COALESCE(priority, 0) DESC, added ASC';
749      $limit = variable_get('xapian_indexing_throttle', 100);      $limit = variable_get('xapian_indexing_throttle', 100);
750    }    }
751    
752    $result = db_query_range($sql, 0, $limit);    $result = db_query_range($sql, 0, $limit);
753    
754    if (!$result) {    if (!$result) {
# Line 762  function xapian_index_queued_nodes() { Line 761  function xapian_index_queued_nodes() {
761    
762    $exclusions = array_values(variable_get('xapian_excluded_nodes', array()));    $exclusions = array_values(variable_get('xapian_excluded_nodes', array()));
763    while ($node = db_fetch_object($result)) {    while ($node = db_fetch_object($result)) {
764      $node = node_load($node->nid);      $node = node_load(array('nid' => $node->nid));
765      if ($node) {      if ($node) {
766        if (in_array($node->type, $exclusions)) {        if (in_array($node->type, $exclusions)) {
767          // None indexable a node, log and clear it from the queue          // None indexable a node, log and clear it from the queue
# Line 794  function xapian_index_queued_nodes() { Line 793  function xapian_index_queued_nodes() {
793    if ($status) {    if ($status) {
794      db_query('DELETE FROM {xapian_index_queue} WHERE status = %d', $status);      db_query('DELETE FROM {xapian_index_queue} WHERE status = %d', $status);
795    }    }
796    
797    if (variable_get('xapian_log_queries', FALSE)) {    if (variable_get('xapian_log_queries', FALSE)) {
798      $time_taken = (microtime(TRUE) - $start_time) * 1000;      $time_taken = (microtime(TRUE) - $start_time) * 1000;
799      watchdog('xapian', '<p>Indexed %indexed nodes successfully, %failed failed in: %time ms</p>',      watchdog('xapian', '<p>Indexed %indexed nodes successfully, %failed failed in: %time ms</p>',
# Line 835  function xapian_reindex_all_confirm_subm Line 834  function xapian_reindex_all_confirm_subm
834    
835  /**  /**
836   * Default node index function   * Default node index function
837   * Indexes title, body, terms in the node (as text), and the taxonomy terms   * Indexes title, body, terms in the node (as text), and the taxonomy terms
838   * as well.   * as well.
839   *   *
840   * @param object $node   * @param object $node
# Line 853  function node_xapian_index($node) { Line 852  function node_xapian_index($node) {
852    
853    $terms[] = array('type' => 'text', 'data' => strip_tags($node->title), 'weight' => 5);    $terms[] = array('type' => 'text', 'data' => strip_tags($node->title), 'weight' => 5);
854    $terms[] = array('type' => 'text', 'data' => strip_tags($node->body), 'weight' => 2);    $terms[] = array('type' => 'text', 'data' => strip_tags($node->body), 'weight' => 2);
855    
856    // If the node type doesn't have taxonomy, Drupal 6 doesn't create the taxonomy array,    // If the node type doesn't have taxonomy, Drupal 6 doesn't create the taxonomy array,
857    // and this triggers an error in taxonomy_node_update_index() called next.    // and this triggers an error in taxonomy_node_update_index() called next.
858    // Make sure the array exists    // Make sure the array exists
Legend:
Removed from v.1.10.2.5  
changed lines
  Added in v.1.10.2.6
  ViewVC Help
Powered by ViewVC 1.1.2