/[drupal]/contributions/modules/search404/search404.module
ViewVC logotype

Diff of /contributions/modules/search404/search404.module

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

revision 1.16, Mon Jul 28 08:05:14 2008 UTC revision 1.16.2.1, Tue Jul 14 12:24:03 2009 UTC
# Line 1  Line 1 
1  <?php  <?php
2  /* $Id: search404.module,v 1.15 2008/03/04 12:14:30 zyxware Exp $ */  // $Id$
3    
4    /* @file
5     * The search404 module files, does all the searching
6     * when a 404 occurs
7     */
8    
9  /**  /**
10   * Implementation of hook_menu().   * Implementation of hook_menu().
11   */   */
12  function search404_menu() {  function search404_menu() {
13    $items = array();    $items = array();
14    
15    $items['search404'] = array(    $items['search404'] = array(
16      'title' => t('Page not found'),      'title' => 'Page not found',
17                  'access callback' => TRUE,      'access callback' => TRUE,
18      'page callback' => 'search404_page',      'page callback' => 'search404_page',
19                  'type' => MENU_LOCAL_TASK      'type' => MENU_LOCAL_TASK
20          );    );
21    
22    $items['admin/settings/search404'] = array(    $items['admin/settings/search404'] = array(
23      'title' => t('Search 404 settings'),      'title' => 'Search 404 settings',
24      'description' => t('Administer search 404.'),      'description' => 'Administer search 404.',
25      'page callback' => 'drupal_get_form',      'page callback' => 'drupal_get_form',
26      'page arguments' => array('search404_settings'),      'page arguments' => array('search404_settings'),
27                  'access callback' => 'user_access',      'access callback' => 'user_access',
28      'access arguments' => array('administer site configuration'),      'access arguments' => array('administer site configuration'),
29      'type' => MENU_NORMAL_ITEM,      'type' => MENU_NORMAL_ITEM,
30    );    );
# Line 33  function search404_get_keys() { Line 41  function search404_get_keys() {
41    // Abort query on certain extensions, e.g: gif jpg jpeg png    // Abort query on certain extensions, e.g: gif jpg jpeg png
42    $extensions = preg_split('/\s+/', variable_get('search404_ignore_query', 'gif jpg jpeg bmp png'));    $extensions = preg_split('/\s+/', variable_get('search404_ignore_query', 'gif jpg jpeg bmp png'));
43    $extensions = trim(implode('|', $extensions));    $extensions = trim(implode('|', $extensions));
44    if (!empty($extensions) && preg_match("/\.($extensions)$/", $_REQUEST['destination'])) {    $keys = $_REQUEST['q'];
45      return false;    if (!empty($extensions) && preg_match("/\.($extensions)$/", $keys)) {
46        return FALSE;
47      }
48      $regex_filter = variable_get('search404_regex', '');
49      $keys_array[] = $keys;
50      if (!empty($regex_filter)) {
51        $keys = preg_replace("/" . $regex_filter . "/", '', $keys);
52    }    }
   $keys = $_REQUEST['destination'];  
   $misc_var = variable_get('search404_regex', '');  
   if (!empty($misc_var)) {  
         $keys = preg_grep($misc_var, $keys);  
         $keys = $keys[0];  
  }  
53    // Ingore certain extensions from query    // Ingore certain extensions from query
54    $extensions = preg_split('/\s+/', variable_get('search404_ignore_extensions', 'htm html php'));    $extensions = preg_split('/\s+/', variable_get('search404_ignore_extensions', 'htm html php'));
55    $extensions = trim(implode('|', $extensions));    $extensions = trim(implode('|', $extensions));
56    if (!empty($extensions)) {    if (!empty($extensions)) {
57       $keys = preg_replace("/\.($extensions)$/", '', $keys);      $keys = preg_replace("/\.($extensions)$/", '', $keys);
58    }    }
59    
60    $keys = preg_split('/['. PREG_CLASS_SEARCH_EXCLUDE .']+/u', $keys);    $keys = preg_split('/[' . PREG_CLASS_SEARCH_EXCLUDE . ']+/u', $keys);
61    
62    // Ignore certain words    // Ignore certain words
63    $keys = array_diff($keys, explode(' ', variable_get('search404_ignore', 'and or the')));    $keys = array_diff($keys, explode(' ', variable_get('search404_ignore', 'and or the')));
# Line 57  function search404_get_keys() { Line 65  function search404_get_keys() {
65    foreach ($keys as $a => $b) {    foreach ($keys as $a => $b) {
66      $keys[$a] = check_plain($b);      $keys[$a] = check_plain($b);
67    }    }
68    $modifier = variable_get('search404_use_or', false) ? ' OR ' : ' ';    $modifier = variable_get('search404_use_or', FALSE) ? ' OR ' : ' ';
69    $keys = trim(implode($modifier, $keys));    $keys = trim(implode($modifier, $keys));
70    return $keys;    return $keys;
71  }  }
72    
73  /**  /**
74   * Detect search from search engine (BETA)   * Detect search from search engine
75   */   */
76  function search404_search_engine_query() {  function search404_search_engine_query() {
77    $engines = array(    $engines = array(
78      'altavista' => 'q',      'altavista' => 'q',
79      'aol' => 'query',      'aol' => 'query',
80      'google' => 'q',      'google' => 'q',
81      'live' => 'q',      'bing' => 'q',
82      'lycos' => 'query',      'lycos' => 'query',
     'msn' => 'q',  
83      'yahoo' => 'p',      'yahoo' => 'p',
84    );    );
85    $parsed_url = parse_url($_SERVER['HTTP_REFERER']);    $parsed_url = parse_url($_SERVER['HTTP_REFERER']);
# Line 80  function search404_search_engine_query() Line 87  function search404_search_engine_query()
87    $query_string = $parsed_url['query'];    $query_string = $parsed_url['query'];
88    parse_str($query_string, $query);    parse_str($query_string, $query);
89    
90    if (!$parsed_url === false && !empty($remote_host) && !empty($query_string) && count($query)) {    if (!$parsed_url === FALSE && !empty($remote_host) && !empty($query_string) && count($query)) {
91      foreach ($engines as $host => $key) {      foreach ($engines as $host => $key) {
92        if (strpos($remote_host, $host) !== false && array_key_exists($key, $query)) {        if (strpos($remote_host, $host) !== FALSE && array_key_exists($key, $query)) {
93          return trim($query[$key]);          return trim($query[$key]);
94        }        }
95      }      }
96    }    }
97    
98    return false;    return FALSE;
99  }  }
100    
101  /**  /**
# Line 98  function search404_search_engine_query() Line 105  function search404_search_engine_query()
105   * Beware of messy code   * Beware of messy code
106   */   */
107  function search404_page() {  function search404_page() {
108    $output = '<p>'. t('The page you requested was not found.') .'</p>';    drupal_set_title(t('Page not found'));
109    if (module_exists('search') && user_access('search content')) {    if (module_exists('search') && user_access('search content')) {
110      $keys = "";      $keys = "";
111        // if apachesolr_search is installed use apachesolr_search instead default node search
112        $type_search = (module_exists('apachesolr_search')) ? 'apachesolr_search' : 'node';
113      if (variable_get('search404_use_search_engine', false)) {      if (variable_get('search404_use_search_engine', false)) {
114        $keys = search404_search_engine_query();        $keys = search404_search_engine_query();
115      }      }
# Line 109  function search404_page() { Line 118  function search404_page() {
118      }      }
119      if ($keys) {      if ($keys) {
120        // TODO: watchdog?        // TODO: watchdog?
121        $results = module_invoke('node', 'search', 'search', $keys);        if (module_exists('google') && user_access('search Google CSE') && variable_get('search404_do_google_cse', FALSE)) {
122        if (isset($results) && is_array($results) && count($results) == 1 && variable_get('search404_jump', false)) {          drupal_set_message(t('The page you requested does not exist. For your convenience, a google search was performed using the query %keys.', array('%keys' => check_plain($keys))), 'error');
123          // First, check to see if there is exactly 1 result          drupal_goto('search/google/' . $keys);
         drupal_set_message(t('The page you requested does not exist. A search for %keys resulted in this page.', array('%keys' => check_plain($keys))), 'status');  
         // overwrite $_REQUEST['destination'] because it is set by drupal_not_found()  
         $_REQUEST['destination'] = 'node/'. $results[0]['node']->nid;  
         drupal_goto();  
       }  
       elseif (isset($results) && is_array($results) && count($results) > 1 && variable_get('search404_first', false)) {  
         drupal_set_message(t('The page you requested does not exist. A search for %keys resulted in this page.', array('%keys' => check_plain($keys))), 'status');  
         // overwrite $_REQUEST['destination'] because it is set by drupal_not_found()  
         $_REQUEST['destination'] = 'node/'. $results[0]['node']->nid;  
         drupal_goto();  
124        }        }
125        else {        else {
126          drupal_set_message(t('The page you requested does not exist. For your convenience, a search was performed using the query %keys.', array('%keys' => check_plain($keys))), 'error');          $results = module_invoke($type_search, 'search', 'search', $keys);
127          if (isset($results) && is_array($results) && count($results) > 0) {drupal_set_message('Inside Else1');          if (isset($results) && is_array($results) && count($results) == 1 && variable_get('search404_jump', FALSE)) {
128            drupal_add_css(drupal_get_path('module', 'search') .'/search.css', 'module', 'all', FALSE);            // First, check to see if there is exactly 1 result
129            // EVIL HAXX!            drupal_set_message(t('The page you requested does not exist. A search for %keys resulted in this page.', array('%keys' => check_plain($keys))), 'status');
130              $oldgetq = $_GET['q'];            // overwrite $_REQUEST['destination'] because it is set by drupal_not_found()
131              $olddestination = $_REQUEST['destination'];            $_REQUEST['destination'] = 'node/' . $results[0]['node']->nid;
132              unset($_REQUEST['destination']);            drupal_goto();
133              $_GET['q'] = "search/node/$keys";          }
134              $results = theme('search_results', $results, 'node');          elseif (isset($results) && is_array($results) && count($results) > 1 && variable_get('search404_first', FALSE)) {
135              $_GET['q'] = $oldgetq;            drupal_set_message(t('The page you requested does not exist. A search for %keys resulted in this page.', array('%keys' => check_plain($keys))), 'status');
136              $_REQUEST['destination'] = $olddestination;            // overwrite $_REQUEST['destination'] because it is set by drupal_not_found()
137            // END OF EVIL HAXX!            $_REQUEST['q'] = 'node/' . $results[0]['node']->nid;
138              drupal_goto();
139          }          }
140          else {          else {
141            $results = search_help('search#noresults', drupal_help_arg());            drupal_set_message(t('The page you requested does not exist. For your convenience, a search was performed using the query %keys.', array('%keys' => check_plain($keys))), 'error');
142              if (isset($results) && is_array($results) && count($results) > 0) {
143                drupal_add_css(drupal_get_path('module', 'search') . '/search.css', 'module', 'all', FALSE);
144                // EVIL HAXX!
145                  $oldgetq = $_GET['q'];
146                  $olddestination = $_REQUEST['q'];
147                  unset($_REQUEST['q']);
148                  $_GET['q'] = "search/$type_search/$keys";
149                  $results = theme('search_results', $results, 'node');
150                  $_GET['q'] = $oldgetq;
151                  $_REQUEST['destination'] = $olddestination;
152                // END OF EVIL HAXX!
153              }
154              else {
155                $results = search_help('search#noresults', drupal_help_arg());
156              }
157              $results = theme('search_results_listing', t('Search results'), $results);
158          }          }
         $results = theme('box', t('Search results'), $results);  
159        }        }
160      }      }
161      // Construct the search form.      // Construct the search form.
162      $output .= drupal_get_form('search_form', NULL, $keys, 'node');      $output = drupal_render(drupal_get_form('search_form', NULL, $keys, $type_search)) . $results;
     $output .= $results;  
163    }    }
164    
165    return $output;    return $output;
166  }  }
167    
# Line 157  function search404_settings() { Line 172  function search404_settings() {
172    $form['search404_jump'] = array(    $form['search404_jump'] = array(
173      '#type' => 'checkbox',      '#type' => 'checkbox',
174      '#title' => t('Jump directly to the search result when there is only one result.'),      '#title' => t('Jump directly to the search result when there is only one result.'),
175      '#default_value' => variable_get('search404_jump', false),      '#default_value' => variable_get('search404_jump', FALSE),
176    );    );
177    $form['search404_first'] = array(    $form['search404_first'] = array(
178      '#type' => 'checkbox',      '#type' => 'checkbox',
179      '#title' => t('Jump directly to the first search result when there are multiple results.'),      '#title' => t('Jump directly to the first search result even when there are multiple results.'),
180      '#default_value' => variable_get('search404_first', false),      '#default_value' => variable_get('search404_first', FALSE),
181    );    );
182    
183      $form['search404_do_google_cse'] = array(
184        '#type' => 'checkbox',
185        '#title' => t('Do a Google CSE Search instead of a Drupal Search when a 404 Occurs'),
186        '#description' => t('Requires Google CSE and Google CSE Search Modules to be Enabled'),
187        '#attributes' => module_exists('google') ? array()  : array('disabled' => 'disabled'),
188        '#default_value' => variable_get('search404_do_google_cse', FALSE),
189      );
190    
191    $form['advanced'] = array(    $form['advanced'] = array(
192      '#type' => 'fieldset',      '#type' => 'fieldset',
193      '#title' => t('Advanced settings'),      '#title' => t('Advanced settings'),
194      '#description' => t("WARNING. Some of these settings can mess up stuff, don't touch unless you know what you are      //'#description' => t("WARNING. Some of these settings can mess up stuff, don't touch unless you know what you are doing."),
                 doing."),  
195      '#collapsible' => TRUE,      '#collapsible' => TRUE,
196      '#collapsed' => TRUE,      '#collapsed' => TRUE,
197    );    );
198    $form['advanced']['search404_use_or'] = array(    $form['advanced']['search404_use_or'] = array(
199      '#type' => 'checkbox',      '#type' => 'checkbox',
200      '#title' => t('Use OR between keywords when searching.'),      '#title' => t('Use OR between keywords when searching.'),
201      '#default_value' => variable_get('search404_use_or', false),      '#default_value' => variable_get('search404_use_or', FALSE),
202    );    );
203    $form['advanced']['search404_use_search_engine'] = array(    $form['advanced']['search404_use_search_engine'] = array(
204      '#type' => 'checkbox',      '#type' => 'checkbox',
205      '#title' => t('Use auto-detection of keywords from search engine referer. BETA! Not for production sites, use at your own risk.'),      '#title' => t('Use auto-detection of keywords from search engine referer.'),
206      '#default_value' => variable_get('search404_use_search_engine', false),      '#description' => t('This feature will conduct a search based on the query string got from a search engine if the URL of the search result points to a 404 page in the current website. Currently supported search engines: Google, Yahoo, Altavista, Lycos, Bing and AOL'),
207        '#default_value' => variable_get('search404_use_search_engine', FALSE),
208    );    );
209    $form['advanced']['search404_ignore'] = array(    $form['advanced']['search404_ignore'] = array(
210      '#type' => 'textfield',      '#type' => 'textfield',
211      '#title' => t('Words to ignore'),      '#title' => t('Words to ignore'),
212      '#description' => t('These words will be ignored from query. Separate words with a space, e.g.: "and or the".'),      '#description' => t('These words will be ignored from the search query. Separate words with a space, e.g.: "and or the".'),
213      '#default_value' => variable_get('search404_ignore', 'and or the'),      '#default_value' => variable_get('search404_ignore', 'and or the'),
214    );    );
215    $form['advanced']['search404_ignore_extensions'] = array(    $form['advanced']['search404_ignore_extensions'] = array(
216      '#type' => 'textfield',      '#type' => 'textfield',
217      '#title' => t('Extensions to ignore'),      '#title' => t('Extensions to ignore'),
218      '#description' => t('These extensions will be ignored from query. Separate extensions with a space, e.g.: "htm html php". Do not include leading dot.'),      '#description' => t('These extensions will be ignored from the search query, e.g.: http://www.example.com/invalid/page.php will only search for "invalid page". Separate extensions with a space, e.g.: "htm html php". Do not include leading dot.'),
219      '#default_value' => variable_get('search404_ignore_extensions', 'htm html php'),      '#default_value' => variable_get('search404_ignore_extensions', 'htm html php'),
220    );    );
221    $form['advanced']['search404_ignore_query'] = array(    $form['advanced']['search404_ignore_query'] = array(
222      '#type' => 'textfield',      '#type' => 'textfield',
223      '#title' => t('Extensions to abort search'),      '#title' => t('Extensions to abort search'),
224      '#description' => t('A search will not be performed for a query ending in the following extensions. Separate extensions with a space, e.g.: "gif jpg jpeg bmp png". Do not include leading dot.'),      '#description' => t('A search will not be performed for a query ending in these extensions. Separate extensions with a space, e.g.: "gif jpg jpeg bmp png". Do not include leading dot.'),
225      '#default_value' => variable_get('search404_ignore_query', 'gif jpg jpeg bmp png'),      '#default_value' => variable_get('search404_ignore_query', 'gif jpg jpeg bmp png'),
226    );    );
227    $form['advanced']['search404_regex'] = array(    $form['advanced']['search404_regex'] = array(
228      '#type' => 'textfield',      '#type' => 'textfield',
229      '#title' => t('PCRE REGEX'),      '#title' => t('PCRE filter'),
230      '#description' => t('This regex will applied to all queries. It uses the code:<p>%code</p>Look directly at the source code to understand underlying syntax. See also <a href="http://php.net/pcre">PCRE pages in the PHP Manual</a>.', array('%code' => "\$keys = preg_grep(variable_get('search404_regex'), \$keys);\n\$keys = \$keys[0];", '%function' => 'search404_get_keys()')),      '#description' => t('This regular expression will be applied to filter all queries. The parts of the path that match the expression will be EXCLUDED from the search. You do NOT have to enclose the regex in forward slashes when defining the PCRE. e.g.: use "[foo]bar" instead of "/[foo]bar/". On how to use a PCRE Regex please refer <a href="http://php.net/pcre">PCRE pages in the PHP Manual</a>.'),
231      '#default_value' => variable_get('search404_regex',''),      '#default_value' => variable_get('search404_regex', ''),
232    );    );
233    return system_settings_form($form);    return system_settings_form($form);
234  }  }
Legend:
Removed from v.1.16  
changed lines
  Added in v.1.16.2.1
  ViewVC Help
Powered by ViewVC 1.1.2