/**
* Construct a Zend_Search_Lucene_Document object out of a document db row.
*
* @global string $urlServer
* @param object $docu
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($docu) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'doc_' . $docu->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $docu->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'doc', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $docu->course_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($docu->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('content', Indexer::phonetics($docu->description), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('filename', Indexer::phonetics($docu->filename), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('comment', Indexer::phonetics($docu->comment), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('creator', Indexer::phonetics($docu->creator), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('subject', Indexer::phonetics($docu->subject), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('author', Indexer::phonetics($docu->author), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('visible', $docu->visible, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('public', $docu->public, $encoding));
$urlAction = ($docu->format == '.dir') ? 'openDir' : 'download';
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer
. 'modules/document/index.php?course=' . course_id_to_code($docu->course_id)
. '&' . $urlAction . '=' . $docu->path, $encoding));
return $doc;
}
/**
* This is the default 'index' action that is invoked
* when an action is not explicitly requested by users.
*/
public function actionIndexing()
{
ini_set('max_execution_time', 0);
ob_start();
$index = new Zend_Search_Lucene(Yii::getPathOfAlias($this->_indexFilesPath), true);
$criteria = new CDbCriteria();
$criteria->compare('t.publish', 1);
$criteria->order = 'album_id DESC';
//$criteria->limit = 10;
$model = Albums::model()->findAll($criteria);
foreach ($model as $key => $item) {
if ($item->media_id != 0) {
$images = Yii::app()->request->baseUrl . '/public/album/' . $item->album_id . '/' . $item->cover->media;
} else {
$images = '';
}
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('id', CHtml::encode($item->album_id), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Text('media', CHtml::encode($images), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Text('title', CHtml::encode($item->title), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Text('body', CHtml::encode(Utility::hardDecode(Utility::softDecode($item->body))), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Text('url', CHtml::encode(Utility::getProtocol() . '://' . Yii::app()->request->serverName . Yii::app()->createUrl('album/site/view', array('id' => $item->album_id, 't' => Utility::getUrlTitle($item->title)))), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('date', CHtml::encode(Utility::dateFormat($item->creation_date, true) . ' WIB'), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('creation', CHtml::encode($item->user->displayname), 'utf-8'));
$index->addDocument($doc);
}
echo 'Album Lucene index created';
$index->commit();
$this->redirect(Yii::app()->createUrl('article/search/indexing'));
ob_end_flush();
}
protected function _indexate($url)
{
if (!stristr($url, 'http://')) {
$url = HTTP_HOST . $url;
}
$url = substr($url, -1) == '/' ? substr($url, 0, -1) : $url;
if (!in_array($url, $this->_indexedUrl)) {
if (stristr($url, HTTP_HOST)) {
array_push($this->_indexedUrl, $url);
$html = file_get_contents($url);
libxml_use_internal_errors(true);
$doc = Zend_Search_Lucene_Document_Html::loadHTML($html);
libxml_use_internal_errors(false);
if (preg_match('/<\\!--index-->(.*)<\\!--\\/index-->/isu', $html, $matches)) {
$html = $matches[1];
}
$html = preg_replace('#<script(.*?)>(.*?)</script>#is', '', $html);
$html = strip_tags($html);
$doc->addField(Zend_Search_Lucene_Field::Text('content', $html, 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('body', '', 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Text('url', $url, 'utf-8'));
$this->_indexHandle->addDocument($doc);
Zend_Registry::get('Logger')->info('Search index is created: ' . $url, Zend_Log::INFO);
foreach ($doc->getLinks() as $link) {
$temp = explode('.', $link);
$ext = end($temp);
if ($link == $ext || in_array($ext, array('php', 'html', 'txt', 'htm'))) {
$this->_indexate($link);
}
}
}
}
}
/**
* Adds attributes to be indexed
*
* @return string
**/
protected function addAttributes()
{
$this->addField(Zend_Search_Lucene_Field::Text('name', $this->getSourceModel()->getName(), self::ENCODING));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('short_content', $this->getSourceModel()->getShortDescription(), self::ENCODING));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('url', $this->getSourceModel()->getProductUrl(), self::ENCODING));
$this->addFilterableAttributes();
$this->addSearchableAttributes();
}
/**
* Constructor. Creates our indexable document and adds all
* necessary fields to it using the passed in document
*/
public function __construct($document)
{
$this->addField(Zend_Search_Lucene_Field::UnIndexed('id_entry', $document->getId()));
$this->addField(Zend_Search_Lucene_Field::Keyword('url', $document->getUrl()));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('creation_date', $document->getCreationDate()));
$this->addField(Zend_Search_Lucene_Field::Text('name', $document->getName()), 'utf-8');
$this->addField(Zend_Search_Lucene_Field::Text('content', $document->getDetails()));
$this->addField(Zend_Search_Lucene_Field::Text('tag', $document->getImplodedTags()));
}
protected function _indexDocument($doc, $fields)
{
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('checkSum', md5($doc->body)));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('lastIndexed', time()));
foreach ($fields as $name => $value) {
$doc->addField(Zend_Search_Lucene_Field::Text($name, $value));
}
$this->_index->addDocument($doc);
Bbx_Log::write('Added ' . urldecode($doc->url) . ' to index', null, self::LOG);
}
private static function _insert($index, $item, $tags)
{
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Text('title', $item->name));
$doc->addField(Zend_Search_Lucene_Field::Text('item_id', strval($item->id)));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('image_id', strval($item->image_id)));
$doc->addField(Zend_Search_Lucene_Field::Text('description', $item->description));
$doc->addField(Zend_Search_Lucene_Field::Text('tag', $tags));
$index->addDocument($doc);
$index->commit();
}
public function testUnIndexed()
{
$field = Zend_Search_Lucene_Field::UnIndexed('field', 'value');
$this->assertEquals($field->boost, 1);
$this->assertEquals($field->encoding, '');
$this->assertEquals($field->isBinary, false);
$this->assertEquals($field->isIndexed, false);
$this->assertEquals($field->isStored, true);
$this->assertEquals($field->isTokenized, false);
$this->assertEquals($field->name, 'field');
$this->assertEquals($field->value, 'value');
}
public function __construct(&$doc, &$data, $course_id, $group_id, $user_id, $path, $additional_keyset = null)
{
$encoding = 'UTF-8';
//document identification and indexing
$this->addField(Zend_Search_Lucene_Field::Keyword('docid', $doc->docid, $encoding));
//document type : the name of the Moodle element that manages it
$this->addField(Zend_Search_Lucene_Field::Keyword('doctype', $doc->documenttype, $encoding));
//allows subclassing information from complex modules.
$this->addField(Zend_Search_Lucene_Field::Keyword('itemtype', $doc->itemtype, $encoding));
//caches the course context.
$this->addField(Zend_Search_Lucene_Field::Keyword('course_id', $course_id, $encoding));
//caches the originator's group.
$this->addField(Zend_Search_Lucene_Field::Keyword('group_id', $group_id, $encoding));
//caches the originator if any
$this->addField(Zend_Search_Lucene_Field::Keyword('user_id', $user_id, $encoding));
// caches the context of this information. i-e, the context in which this information
// is being produced/attached. Speeds up the "check for access" process as context in
// which the information resides (a course, a module, a block, the site) is stable.
$this->addField(Zend_Search_Lucene_Field::UnIndexed('context_id', $doc->contextid, $encoding));
//data for document
$this->addField(Zend_Search_Lucene_Field::Text('title', $doc->title, $encoding));
$this->addField(Zend_Search_Lucene_Field::Text('author', $doc->author, $encoding));
$this->addField(Zend_Search_Lucene_Field::UnStored('contents', $doc->contents, $encoding));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('url', $doc->url, $encoding));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('date', $doc->date, $encoding));
//additional data added on a per-module basis
$this->addField(Zend_Search_Lucene_Field::Binary('data', serialize($data)));
// adding a path allows the document to know where to find specific library calls
// for checking access to a module or block content. The Lucene records should only
// be responsible to bring back to that call sufficient and consistent information
// in order to perform the check.
$this->addField(Zend_Search_Lucene_Field::UnIndexed('path', $path, $encoding));
/*
// adding a capability set required for viewing. -1 if no capability required.
// the capability required for viewing is depending on the local situation
// of the document. each module should provide this information when pushing
// out search document structure. Although capability model should be kept flat
// there is no exclusion some module or block developpers use logical combinations
// of multiple capabilities in their code. This possibility should be left open here.
$this->addField(Zend_Search_Lucene_Field::UnIndexed('capabilities', $caps));
*/
/*
// Additional key set allows a module to ask for extensible criteria based search
// depending on the module internal needs.
*/
if (!empty($additional_keyset)) {
foreach ($additional_keyset as $keyname => $keyvalue) {
$this->addField(Zend_Search_Lucene_Field::Keyword($keyname, $keyvalue, $encoding));
}
}
}
/**
* @param Zend_Feed_Atom $atom
*/
private function processAtom($atom)
{
$this->addField(Zend_Search_Lucene_Field::UnIndexed('id', $this->id));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('url', $this->url));
$this->addField(Zend_Search_Lucene_Field::Text('title', $atom->title()));
// Loop over each channel item and store relevant data
$text = '';
foreach ($atom as $item) {
$text .= $item->title();
$text .= ' ';
$text .= $item->content();
}
$this->addField(Zend_Search_Lucene_Field::Text('body', $text));
}
/**
* Construct a Zend_Search_Lucene_Document object out of a link db row.
*
* @global string $urlServer
* @param object $link
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($link) {
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'link_' . $link->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $link->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'link', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $link->course_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($link->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('content', Indexer::phonetics(strip_tags($link->description)), $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $link->url, $encoding));
return $doc;
}
/**
* Construct a Zend_Search_Lucene_Document object out of a video db row.
*
* @global string $urlServer
* @param object $video
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($video) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'video_' . $video->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $video->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'video', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $video->course_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($video->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('content', Indexer::phonetics($video->description), $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer . 'modules/video/file.php?course=' . course_id_to_code($video->course_id) . '&id=' . $video->id, $encoding));
return $doc;
}
public function index(Zfplanet_Model_Entry $entry)
{
if (is_null($this->_index)) {
return;
}
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('id', $entry->id, 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('publishedDate', $entry->publishedDate, 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Keyword('uri', $entry->uri, 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Text('title', $entry->title, 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::UnStored('content', $entry->content, 'utf-8'));
$this->_index->addDocument($doc);
$this->_index->commit();
$this->_index->optimize();
}
static function AddToIndex(SearchableObject $object, $commitOnEnd = true)
{
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('combinedid', $object->getRelObjectManager() . $object->getRelObjectId()));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('objectid', $object->getRelObjectId()));
$doc->addField(Zend_Search_Lucene_Field::Keyword('manager', $object->getRelObjectManager()));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('column', $object->getColumnName()));
$doc->addField(Zend_Search_Lucene_Field::UnStored('text', $object->getContent()));
$doc->addField(Zend_Search_Lucene_Field::Text('workspaces', "ws" . $object->getProjectId() . " "));
$doc->addField(Zend_Search_Lucene_Field::Text('isprivate', ($object->getIsPrivate() ? '1' : '0') . " "));
self::GetIndex()->addDocument($doc);
if ($commitOnEnd) {
self::GetIndex()->commit();
}
return true;
}
/**
* Construct a Zend_Search_Lucene_Document object out of an exercise db row.
*
* @global string $urlServer
* @param object $exercise
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($exercise) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'exercise_' . $exercise->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $exercise->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'exercise', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $exercise->course_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($exercise->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('content', Indexer::phonetics(strip_tags($exercise->description)), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('visible', $exercise->active, $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer . 'modules/exercise/exercise_submit.php?course=' . course_id_to_code($exercise->course_id) . '&exerciseId=' . $exercise->id, $encoding));
return $doc;
}
/**
* Construct a Zend_Search_Lucene_Document object out of an announcement db row.
*
* @global string $urlServer
* @param object $announce
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($announce) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'announce_' . $announce->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $announce->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'announce', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $announce->course_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($announce->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('content', Indexer::phonetics(strip_tags($announce->content)), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('visible', $announce->visible, $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer . 'modules/announcements/index.php?course=' . course_id_to_code($announce->course_id) . '&an_id=' . $announce->id, $encoding));
return $doc;
}
/**
* Construct a Zend_Search_Lucene_Document object out of a forum topic db row.
*
* @global string $urlServer
* @param object $ftopic
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($ftopic) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'ftopic_' . $ftopic->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $ftopic->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'ftopic', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $ftopic->course_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('forumid', $ftopic->forum_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($ftopic->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer . 'modules/forum/viewforum.php?course=' . course_id_to_code($ftopic->course_id)
. '&forum=' . intval($ftopic->forum_id), $encoding));
return $doc;
}
/**
* Construct a Zend_Search_Lucene_Document object out of a forum post row.
*
* @global string $urlServer
* @param object $fpost
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($fpost) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'fpost_' . $fpost->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $fpost->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'fpost', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $fpost->course_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('topicid', $fpost->topic_id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('content', Indexer::phonetics(strip_tags($fpost->post_text)), $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer . 'modules/forum/viewtopic.php?course=' . course_id_to_code($fpost->course_id)
. '&topic=' . intval($fpost->topic_id)
. '&forum=' . intval($fpost->forum_id), $encoding));
return $doc;
}
/**
* Adds all values of related category to the search document.
*
* @return void
**/
protected function addAttributes()
{
$content = strip_tags($this->getStaticBlock());
$this->addField(Zend_Search_Lucene_Field::UnStored('content', $content, self::ENCODING));
$this->addField(Zend_Search_Lucene_Field::Text('name', $this->getSourceModel()->getName(), self::ENCODING));
$this->addField(Zend_Search_Lucene_Field::Keyword('category', $this->getSourceModel()->getParentCategory()->getName(), self::ENCODING));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('short_content', substr($content, 0, self::SHORT_CONTENT_CHAR_COUNT), self::ENCODING));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('url', $this->getSourceModel()->getUrl(), self::ENCODING));
if ($this->getSourceModel()->getImage()) {
try {
$image = Mage::getModel('catalog/product_image')->setBaseFile('../category/' . $this->getSourceModel()->getImage())->setHeight(100)->setWidth(100)->resize()->saveFile()->getUrl();
$this->addField(Zend_Search_Lucene_Field::UnIndexed('image', $image, self::ENCODING));
} catch (Exception $e) {
/* no image for category, so none will be added to index */
}
}
}
/**
* Construct a Zend_Search_Lucene_Document object out of a note db row.
*
* @global string $urlServer
* @param object $note
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($note) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'note_' . $note->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $note->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'note', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('userid', $note->user_id, $encoding));
if(isset($note->course_id)){
$doc->addField(Zend_Search_Lucene_Field::Keyword('courseid', $note->course_id, $encoding));
}
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($note->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('content', Indexer::phonetics(strip_tags($note->content)), $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer . 'modules/notes/index.php?an_id=' . $note->id, $encoding));
return $doc;
}
public function add($content, $section, $mtime)
{
foreach ($this->split_headings($content) as $headers) {
$doc = new Zend_Search_Lucene_Document();
$link = "index.php?page=" . preg_replace('/\\/|\\\\/', '.', $section);
$link = str_replace('.page', '', $link) . '#' . $headers['section'];
//unsearchable text
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('link', $link));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('mtime', $mtime));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('title', $headers['title']));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('text', $headers['content']));
//searchable text
$doc->addField(Zend_Search_Lucene_Field::Keyword('page', strtolower($headers['title'])));
$body = strtolower($this->sanitize($headers['content'])) . ' ' . strtolower($headers['title']);
$doc->addField(Zend_Search_Lucene_Field::Unstored('contents', $body));
$this->_index->addDocument($doc);
}
}
public static function indexationAdd($indexationData)
{
$directory = Zend_Registry::get('lucene_index');
Zend_Search_Lucene_Analysis_Analyzer::setDefault(new Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8());
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pageID', $indexationData['pageID']));
$doc->addField(Zend_Search_Lucene_Field::Keyword('moduleID', $indexationData['moduleID']));
$doc->addField(Zend_Search_Lucene_Field::Keyword('contentID', $indexationData['contentID']));
$doc->addField(Zend_Search_Lucene_Field::Keyword('languageID', $indexationData['languageID']));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Cible_FunctionsGeneral::html2text($indexationData['title'])));
$doc->addField(Zend_Search_Lucene_Field::Text('text', Cible_FunctionsGeneral::html2text($indexationData['text'])));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('link', $indexationData['link']));
$doc->addField(Zend_Search_Lucene_Field::UnStored('contents', strtolower(Cible_FunctionsGeneral::removeAccents(Cible_FunctionsGeneral::html2text($indexationData['contents'])))));
$newIndex = !is_dir($directory);
$index = new Zend_Search_Lucene($directory, $newIndex);
$index->addDocument($doc);
$index->commit();
}
public function updateLuceneIndex()
{
$index = GcMailboxPeer::getLuceneIndex();
//borrar las entradas existentes
if ($hit = $index->find('pk:' . $this->getId())) {
$index->delete($hit->id);
}
//no indexar buzones expirados
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('pk', $this->getId()));
$doc->addField(Zend_Search_Lucene_Field::unStored('username', $this->getUsername(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::unStored('name', $this->getName(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::unStored('email', $this->getEmail(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::unStored('domain', $this->getDomainName(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::unStored('area', $this->getGcArea(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::unStored('group', $this->getGcGroup(), 'utf-8'));
$index->addDocument($doc);
$index->commit();
}
public function __construct($class, $key, $title, $contents, $summary, $createdBy, $dateCreated, $keywords = array())
{
$this->addField(Zend_Search_Lucene_Field::Keyword('docRef', "{$class}:{$key}"));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('class', $class));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('key', $key));
$this->addField(Zend_Search_Lucene_Field::Text('title', $title));
$this->addField(Zend_Search_Lucene_Field::UnStored('contents', $contents));
$this->addField(Zend_Search_Lucene_Field::UnIndexed('summary', $summary));
$this->addField(Zend_Search_Lucene_Field::Keyword('createdBy', $createdBy));
$this->addField(Zend_Search_Lucene_Field::Keyword('dateCreated', $dateCreated));
if (!is_array($keywords)) {
$keywords = explode('', $keywords);
}
foreach ($keywords as $name => $value) {
if (!empty($name) && !empty($value)) {
$this->addField(Zend_Search_Lucene_Field::keyword($name, $value));
}
}
}
/**
* Construct a Zend_Search_Lucene_Document object out of a course db row.
*
* @global string $urlServer
* @param object $course
* @return Zend_Search_Lucene_Document
*/
protected function makeDoc($course) {
global $urlServer;
$encoding = 'utf-8';
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', 'course_' . $course->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('pkid', $course->id, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Keyword('doctype', 'course', $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('code', Indexer::phonetics($course->code), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('title', Indexer::phonetics($course->title), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('keywords', Indexer::phonetics($course->keywords), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('visible', $course->visible, $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('prof_names', Indexer::phonetics($course->prof_names), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('public_code', Indexer::phonetics($course->public_code), $encoding));
$doc->addField(Zend_Search_Lucene_Field::Text('units', Indexer::phonetics(strip_tags($course->units)), $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('created', $course->created, $encoding));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('url', $urlServer . 'courses/' . $course->code, $encoding));
return $doc;
}
public function updateLuceneIndex()
{
$index = trackTable::getLuceneIndex();
// remove existing entries
foreach ($index->find('pk:' . $this->getId()) as $hit) {
$index->delete($hit->id);
}
$doc = new Zend_Search_Lucene_Document();
$doc->addField(Zend_Search_Lucene_Field::Keyword('pk', $this->getId()));
$doc->addField(Zend_Search_Lucene_Field::text('track_name', $this->getName(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('track_url', $this->getUrl(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Keyword('play_it_user_id', $this->getSfGuardUser()->getId()));
$doc->addField(Zend_Search_Lucene_Field::text('track_name', $this->getName(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::text('track_type', $this->getPlayList()->getObjectType(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::text('playlist_name', $this->getPlayList()->getTitle(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::Keyword('play_owner_id', $this->getPlayList()->getPlayOwner()->getId()));
$doc->addField(Zend_Search_Lucene_Field::text('play_owner_name', $this->getPlayList()->getPlayOwner()->getName(), 'utf-8'));
$doc->addField(Zend_Search_Lucene_Field::UnStored('play_owner_name_fr', $this->getPlayList()->getPlayOwner()->getNameFr(), 'utf-8'));
$index->addDocument($doc);
$index->commit();
}
function create_index()
{
echo "Building search index...\n";
$files = $this->get_files($this->_api);
$count = 0;
foreach ($files as $file) {
$content = $this->get_details($file, $this->_api);
$doc = new Zend_Search_Lucene_Document();
$title = $content['class'];
echo " Adding " . $title . "\n";
//unsearchable text
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('link', $content['link']));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('title', $title));
//$doc->addField(Zend_Search_Lucene_Field::UnIndexed('text', $content['content']));
//searchable
$body = strtolower($this->sanitize($content['content'])) . ' ' . strtolower($title);
$doc->addField(Zend_Search_Lucene_Field::Keyword('page', strtolower(str_replace('.', ' ', $title))));
$doc->addField(Zend_Search_Lucene_Field::Unstored('contents', $body));
$this->_index->addDocument($doc);
$count++;
}
$this->_index->commit();
echo "\n {$count} files indexed.\n";
}
/**
* Method to add a record to the search index
*
* @param string $docId Unique Document Id - Should be in the format of {module}_{type}_{id}
* @param string $docDate Date and Time document has been created or should become available
* @param string $url URL to the page
* @param string $title Title of the Document
* @param string $contents Contents of the Document
* @param string $teaser Short Overview of Contents - Appears in Search Results
* @param string $module Module Item is coming from
* @param string $userId User adding the item
* @param array $tags Tags for the story
* @param string $license License applicable to the content
* @param string $context Context content is being created in, duplicate if used in various contexts
* @param string $workgroup Workgroup content is being created in, duplicate if used in various workgroups
* @param string $permissions Permissions to be checked before being displayed in Search Results. module|action, which could then be used as $this->isValid('action', 'module')
* @param date/time $dateAvailable Date story becomes available
* @param date/time $dateUnavailable Date story expires
* @param array $extra Any extra fields that might be peculiar to a module, in the format of $item=>$value,
* @param boolean $doOptimize Should Index be automatically optimized
*
* The last one is useful for categories, eg. forum. $extra = array('forum'=>'id') allows one to do a search in a particular forum
*/
public function luceneIndex($docId, $docDate, $url, $title, $contents, $teaser, $module, $userId, $tags = NULL, $license = NULL, $context = 'nocontext', $workgroup = 'noworkgroup', $permissions = NULL, $dateAvailable = NULL, $dateUnavailable = NULL, $extra = NULL, $doOptimize = TRUE)
{
if (!$this->enableLucene) {
return;
}
// Remove Index if it exists
$this->removeIndex($docId);
// Get Indexer Object
$this->index = $this->checkIndexPath();
// Setup Lucene Document
$document = new Zend_Search_Lucene_Document();
// Set the properties that we want to use in our index
// Document
$document->addField(Zend_Search_Lucene_Field::Keyword('docid', $docId));
// Date
$document->addField(Zend_Search_Lucene_Field::Keyword('date', $docDate));
// URL
$document->addField(Zend_Search_Lucene_Field::UnIndexed('url', $url));
// Title
$document->addField(Zend_Search_Lucene_Field::Text('title', iconv('ISO-8859-1', 'ASCII//TRANSLIT', $title)));
// Contents
$document->addField(Zend_Search_Lucene_Field::UnStored('contents', iconv('ISO-8859-1', 'ASCII//TRANSLIT', mb_strtolower(stripslashes(strip_tags($title . ' ' . $contents))))));
// Teaser
$document->addField(Zend_Search_Lucene_Field::Text('teaser', $teaser));
// Module
$document->addField(Zend_Search_Lucene_Field::Keyword('module', $module));
// UserId
$document->addField(Zend_Search_Lucene_Field::Keyword('userId', $userId));
// Tags // Check if Array - if yes, convert to string
if (is_array($tags) && count($tags) > 0) {
$newTags = '';
$divider = '';
foreach ($tags as $tag) {
$newTags .= $divider . ' ' . $tag;
$divider = ',';
}
$tags = $newTags;
}
// Add Tags
$document->addField(Zend_Search_Lucene_Field::UnStored('tags', iconv('ISO-8859-1', 'ASCII//TRANSLIT', mb_strtolower(stripslashes(strip_tags($tags))))));
// License
$document->addField(Zend_Search_Lucene_Field::Keyword('license', $license));
// Context
$document->addField(Zend_Search_Lucene_Field::Keyword('context', $context));
// Workgroup
$document->addField(Zend_Search_Lucene_Field::Keyword('workgroup', $workgroup));
// Permissions
$document->addField(Zend_Search_Lucene_Field::UnIndexed('permissions', $permissions));
// Date Available
$document->addField(Zend_Search_Lucene_Field::UnIndexed('dateavailable', $dateAvailable));
// Date Unavailable
$document->addField(Zend_Search_Lucene_Field::UnIndexed('dateunavailable', $dateUnavailable));
// Add Extra items into arrray
if (is_array($extra)) {
// Todo: check that fields dont clash with existing
foreach ($extra as $item => $value) {
$document->addField(Zend_Search_Lucene_Field::Keyword($item, $value));
}
}
// Add the document to the index
$this->index->addDocument($document);
//optimize the thing
if ($doOptimize) {
$this->index->optimize();
}
}
* @package Zend_Search_Lucene
* @subpackage Demos
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/**
* @see Zend_Feed
*/
require_once 'Zend/Feed.php';
/**
* @see Zend_Search_Lucene
*/
require_once 'Zend/Search/Lucene.php';
//create the index
$index = new Zend_Search_Lucene('/tmp/feeds_index', true);
// index each item
$rss = Zend_Feed::import('http://feeds.feedburner.com/ZendDeveloperZone');
foreach ($rss->items as $item) {
$doc = new Zend_Search_Lucene_Document();
if ($item->link && $item->title && $item->description) {
$link = htmlentities(strip_tags($item->link()));
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('link', $link));
$title = htmlentities(strip_tags($item->title()));
$doc->addField(Zend_Search_Lucene_Field::Text('title', $title));
$contents = htmlentities(strip_tags($item->description()));
$doc->addField(Zend_Search_Lucene_Field::Text('contents', $contents));
echo "Adding {$item->title()}...\n";
$index->addDocument($doc);
}
}
$index->commit();
/**
* Add an entry
*
* @param Searchable $item Model implememting Searchable interface
* @param bool $create_new whether or not to create new index when adding item - only used when index is rebuilt
* @return Search return this instance for method chaining
*/
public function add($item, $create_new = FALSE)
{
// ensure item implements Searchable interface
if (!is_a($item, "Searchable")) {
throw new Kohana_User_Exception('Invalid Object', 'Object must implement Searchable Interface');
}
if (!$create_new) {
$this->open_index();
}
$doc = new Zend_Search_Lucene_Document();
// get indexable fields;
$fields = $item->get_indexable_fields();
// index the object type - this allows search results to be grouped/searched by type
$doc->addField(Zend_Search_Lucene_Field::Keyword('type', $item->get_type()));
// index the object's id - to avoid any confusion, we call it 'identifier' as Lucene uses 'id' attribute internally.
$doc->addField(Zend_Search_Lucene_Field::UnIndexed('identifier', $item->get_identifier()));
// store, but don't index or tokenize
// index the object type plus identifier - this gives us a unique identifier for later retrieval - e.g. to delete
$doc->addField(Zend_Search_Lucene_Field::Keyword('uid', $item->get_unique_identifier()));
// index all fields that have been identified by Interface
foreach ($fields as $field) {
// get attribute value from model
$value = $item->__get($field->name);
// html decode value if required
$value = $field->html_decode ? htmlspecialchars_decode($value) : $value;
// add field value based on type
switch ($field->type) {
case Searchable::KEYWORD:
$doc->addField(Zend_Search_Lucene_Field::Keyword($field->name, $value));
break;
case Searchable::UNINDEXED:
$doc->addField(Zend_Search_Lucene_Field::UnIndexed($field->name, $value));
break;
case Searchable::BINARY:
$doc->addField(Zend_Search_Lucene_Field::Binary($field->name, $value));
break;
case Searchable::TEXT:
$doc->addField(Zend_Search_Lucene_Field::Text($field->name, $value));
break;
case Searchable::UNSTORED:
$doc->addField(Zend_Search_Lucene_Field::UnStored($field->name, $value));
break;
}
}
$this->index->addDocument($doc);
// return this so we can have chainable methods
return $this;
}
