public function run()
{
require_once 'inc/SweteSite.class.php';
require_once 'inc/SweteWebpage.class.php';
require_once 'inc/PageProcessor.php';
require_once 'modules/tm/lib/XFTranslationMemory.php';
$wpids = array();
$numPages = 0;
foreach ($this->changedTranslationMemories as $tmid => $strids) {
if (!$strids) {
$strids = array(0);
}
$strids = implode(',', $strids);
$sql = "select distinct wp.webpage_id, s.source_language\n\t\t\t\tfrom \n\t\t\t\t\twebpage_properties wp \n\t\t\t\t\tinner join webpage_strings ws on wp.webpage_id=ws.webpage_id\n\t\t\t\t\tinner join webpages w on wp.webpage_id=w.webpage_id\n\t\t\t\t\tinner join websites s on w.website_id=s.website_id\n\t\t\t\twhere\n\t\t\t\t\twp.effective_translation_memory_id='" . addslashes($tmid) . "' and\n\t\t\t\t\tws.string_id in (" . $strids . ")";
$res = df_q($sql);
$numPages += mysql_num_rows($res);
$wplangs = array();
while ($row = mysql_fetch_row($res)) {
list($webpageId, $sourceLanguage) = $row;
$wpids[] = $webpageId;
$wplangs[] = $sourceLanguage;
}
@mysql_free_result($res);
}
df_q("update background_processes set \n\t\t\t\tstatus_message='" . addslashes('Updating webpages with new translations') . "',\n\t\t\t\tstatus_current_position=0,\n\t\t\t\tstatus_total='" . addslashes($numPages) . "'\n\t\t\t\twhere process_id='" . addslashes($this->getProcessId()) . "'");
$count = 1;
foreach ($wpids as $k => $webpageId) {
$pageWrapper = SweteWebpage::loadById($webpageId, $wplangs[$k]);
$pageWrapper->setSite($this->getSite($pageWrapper->getRecord()->val('website_id')));
$tmid = $pageWrapper->getTranslationMemoryId(true);
if ($tmid) {
$tm = $this->getTranslationMemory($tmid);
if ($tm) {
$processor = new PageProcessor();
$processor->site = $pageWrapper->getSite();
$processor->translationMemory = $tm;
$processor->page = $pageWrapper;
$processor->translateMinStatus = 3;
$processor->translateMaxStatus = 5;
$processor->logTranslationMisses = true;
$processor->savePage = true;
$processor->saveTranslationLogRecord = true;
$processor->process();
}
}
df_q("update background_processes set \n\t\t\t\tstatus_current_position='" . addslashes($count) . "'\n\t\t\t\twhere process_id='" . addslashes($this->getProcessId()) . "'");
$count++;
}
df_q("update background_processes set \n\t\t\t\tstatus_message='" . addslashes('New translations successfully applied to ' . $numPages . ' pages.') . "',\n\t\t\t\tstatus_current_position=0,\n\t\t\t\tstatus_total='" . addslashes($numPages) . "'\n\t\t\t\twhere process_id='" . addslashes($this->getProcessId()) . "'");
}
/**
* Set-ups the any indexing plugins associated with this page
* processor
*
* @param array $plugins an array of indexing plugins which might
* do further processing on the data handles by this page
* processor
* @param int $max_description_len maximal length of a page summary
* @param int $summarizer_option CRAWL_CONSTANT specifying what kind
* of summarizer to use self::BASIC_SUMMARIZER or
* self::CENTROID_SUMMARIZER
*/
function __construct($plugins = array(), $max_description_len = NULL, $summarizer_option = self::BASIC_SUMMARIZER)
{
$this->plugin_instances = $plugins;
$this->summarizer_option = $summarizer_option;
if ($max_description_len != NULL) {
self::$max_description_len = $max_description_len;
} else {
self::$max_description_len = MAX_DESCRIPTION_LEN;
}
}
/**
* @brief Imports a node and all of its children into the database.
*
* @param stdClass $node The root node to import.
* @param Dataface_Record $parentPage A record of the @e webpages table
* that represents the parent page of the current page.
*
* @see SiteCrawler for more information about nodes and the properties
* they can have.
*/
public function importNode(stdClass $node, $parentPage = null)
{
$page = null;
if (isset($node->path) and isset($node->httpStatus)) {
$page = df_get_record('webpages', array('webpage_url' => '=' . $node->path, 'website_id' => '=' . $this->site->getRecord()->val('website_id')));
if (!$page) {
$page = new Dataface_Record('webpages', array());
$page->setValues(array('website_id' => $this->site->getRecord()->val('website_id'), 'webpage_url' => $node->path, 'active' => -1, 'posted_by' => $this->username, 'parent_id' => $parentPage ? $parentPage->val('webpage_id') : null, 'is_loaded' => 0));
$res = $page->save();
if (PEAR::isError($res)) {
throw new Exception($res->getMessage());
}
$this->pagesAdded[] = $page;
} else {
$this->pagesUpdated[] = $page;
}
$page->setValues(array('last_checked' => date('Y-m-d H:i:s'), 'last_checked_response_code' => $node->httpStatus, 'last_checked_content_type' => $node->contentType, 'last_checked_by' => $this->username));
$updateRefreshLog = false;
$translationStats = array();
//if ( $node->content and $this->loadContent and (!$page->val('locked') or $this->overrideLocks) ){
if (@$node->content and @$this->loadContent) {
if ($page->val('locked') and !$this->overrideLocks) {
error_log("Skipping refresh of " . @$node->path . " because the page is locked.");
} else {
$page->setValues(array('last_refresh' => date('Y-m-d H:i:s'), 'last_refresh_response_code' => $node->httpStatus, 'is_loaded' => 1, 'webpage_content' => $node->content));
// Now log the check
$logEntry = new Dataface_Record('webpage_refresh_log', array());
$logEntry->setValues(array('webpage_id' => $page->val('webpage_id'), 'date_checked' => date('Y-m-d H:i:s'), 'response_code' => $node->httpStatus, 'content_type' => $node->contentType, 'content' => $node->content, 'checked_by' => $this->username));
$res = $logEntry->save();
if (PEAR::isError($res)) {
throw new Exception($res->getMessage());
}
if ($this->translate) {
$pageWrapper = new SweteWebpage($page);
$tmid = $pageWrapper->getTranslationMemoryId(true);
if ($tmid) {
$tm = $this->getTranslationMemory($tmid);
if ($tm) {
import('inc/PageProcessor.php');
$processor = new PageProcessor();
$processor->webpageRefreshLogId = $logEntry->val('refresh_log_id');
$processor->site = $this->site;
$processor->translationMemory = $tm;
$processor->page = $pageWrapper;
$processor->translateMinStatus = $this->translateMinStatus;
$processor->translateMaxStatus = $this->translateMaxStatus;
$processor->logTranslationMisses = $this->logTranslationMisses;
$processor->savePage = false;
$processor->saveTranslationLogRecord = true;
$processor->process();
}
}
}
}
}
$res = $page->save();
// Now log the check
$logEntry = new Dataface_Record('webpage_check_log', array());
$logEntry->setValues(array('webpage_id' => $page->val('webpage_id'), 'date_checked' => date('Y-m-d H:i:s'), 'response_code' => $node->httpStatus, 'content_type' => $node->contentType, 'checked_by' => $this->username));
$res = $logEntry->save();
if (PEAR::isError($res)) {
throw new Exception($res->getMessage());
}
}
if (isset($node->children) and is_array($node->children)) {
foreach ($node->children as $child) {
$this->importNode($child, $page);
}
}
}
