290 lines
12 KiB
Plaintext
290 lines
12 KiB
Plaintext
<?php
|
|
|
|
/**
|
|
* @file
|
|
* Used to aggregate syndicated content (RSS, RDF, and Atom).
|
|
*/
|
|
|
|
use Drupal\aggregator\FeedInterface;
|
|
use Drupal\Component\Plugin\Exception\PluginException;
|
|
|
|
/**
|
|
* Denotes that a feed's items should never expire.
|
|
*/
|
|
const AGGREGATOR_CLEAR_NEVER = 0;
|
|
|
|
/**
|
|
* Implements hook_help().
|
|
*/
|
|
function aggregator_help($path, $arg) {
|
|
switch ($path) {
|
|
case 'admin/help#aggregator':
|
|
$output = '';
|
|
$output .= '<h3>' . t('About') . '</h3>';
|
|
$output .= '<p>' . t('The Aggregator module is an on-site syndicator and news reader that gathers and displays fresh content from RSS-, RDF-, and Atom-based feeds made available across the web. Thousands of sites (particularly news sites and blogs) publish their latest headlines in feeds, using a number of standardized XML-based formats. For more information, see the <a href="!aggregator-module">online documentation for the Aggregator module</a>.', array('!aggregator-module' => 'https://drupal.org/documentation/modules/aggregator')) . '</p>';
|
|
$output .= '<h3>' . t('Uses') . '</h3>';
|
|
$output .= '<dl>';
|
|
$output .= '<dt>' . t('Viewing feeds') . '</dt>';
|
|
$output .= '<dd>' . t('Users view feed content in the <a href="!aggregator">main aggregator display</a>, or by <a href="!aggregator-sources">their source</a> (usually via an RSS feed reader). The most recent content in a feed can be displayed as a block through the <a href="!admin-block">Blocks administration page</a>.', array('!aggregator' => \Drupal::url('aggregator.page_last'), '!aggregator-sources' => \Drupal::url('aggregator.sources'), '!admin-block' => \Drupal::url('block.admin_display'))) . '</dd>';
|
|
$output .= '<dt>' . t('Adding, editing, and deleting feeds') . '</dt>';
|
|
$output .= '<dd>' . t('Administrators can add, edit, and delete feeds, and choose how often to check each feed for newly updated items on the <a href="!feededit">Feed aggregator administration page</a>.', array('!feededit' => \Drupal::url('aggregator.admin_overview'))) . '</dd>';
|
|
$output .= '<dt>' . t('<abbr title="Outline Processor Markup Language">OPML</abbr> integration') . '</dt>';
|
|
$output .= '<dd>' . t('A <a href="!aggregator-opml">machine-readable OPML file</a> of all feeds is available. OPML is an XML-based file format used to share outline-structured information such as a list of RSS feeds. Feeds can also be <a href="!import-opml">imported via an OPML file</a>.', array('!aggregator-opml' => \Drupal::url('aggregator.opml_page'), '!import-opml' => \Drupal::url('aggregator.opml_add'))) . '</dd>';
|
|
$output .= '<dt>' . t('Configuring cron') . '</dt>';
|
|
$output .= '<dd>' . t('A correctly configured <a href="!cron">cron maintenance task</a> is required to update feeds automatically.', array('!cron' => \Drupal::url('system.cron_settings'))) . '</dd>';
|
|
$output .= '</dl>';
|
|
return $output;
|
|
case 'admin/config/services/aggregator':
|
|
// Don't use placeholders for possibility to change URLs for translators.
|
|
$output = '<p>' . t('Thousands of sites (particularly news sites and blogs) publish their latest headlines and posts in feeds, using a number of standardized XML-based formats. Formats supported by the aggregator include <a href="http://cyber.law.harvard.edu/rss/">RSS</a>, <a href="http://www.w3.org/RDF/">RDF</a>, and <a href="http://www.atomenabled.org">Atom</a>.') . '</p>';
|
|
$output .= '<p>' . t('Current feeds are listed below, and <a href="!addfeed">new feeds may be added</a>. At the <a href="!block">blocks administration page</a> you can enable for each feed the block <em>Aggregator feed</a> that contains the <em>latest items</em> .', array('!addfeed' => \Drupal::url('aggregator.feed_add'), '!block' => \Drupal::url('block.admin_display'))) . '</p>';
|
|
return $output;
|
|
case 'admin/config/services/aggregator/add/feed':
|
|
return '<p>' . t('Add a feed in RSS, RDF or Atom format. A feed may only have one entry.') . '</p>';
|
|
case 'admin/config/services/aggregator/add/opml':
|
|
return '<p>' . t('<abbr title="Outline Processor Markup Language">OPML</abbr> is an XML format used to exchange multiple feeds between aggregators. A single OPML document may contain a collection of many feeds. Drupal can parse such a file and import all feeds at once, saving you the effort of adding them manually. You may either upload a local file from your computer or enter a URL where Drupal can download it.') . '</p>';
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Implements hook_theme().
|
|
*/
|
|
function aggregator_theme() {
|
|
return array(
|
|
'aggregator_feed_source' => array(
|
|
'variables' => array('aggregator_feed' => NULL, 'view_mode' => NULL),
|
|
'file' => 'aggregator.theme.inc',
|
|
'template' => 'aggregator-feed-source',
|
|
),
|
|
'aggregator_block_item' => array(
|
|
'variables' => array('item' => NULL, 'feed' => 0),
|
|
'file' => 'aggregator.theme.inc',
|
|
'template' => 'aggregator-block-item',
|
|
),
|
|
'aggregator_summary_items' => array(
|
|
'variables' => array('summary_items' => NULL, 'source' => NULL),
|
|
'file' => 'aggregator.theme.inc',
|
|
'template' => 'aggregator-summary-items',
|
|
),
|
|
'aggregator_summary_item' => array(
|
|
'variables' => array('aggregator_item' => NULL, 'view_mode' => NULL),
|
|
'file' => 'aggregator.theme.inc',
|
|
'template' => 'aggregator-summary-item',
|
|
),
|
|
'aggregator_item' => array(
|
|
'variables' => array('aggregator_item' => NULL, 'view_mode' => NULL),
|
|
'file' => 'aggregator.theme.inc',
|
|
'template' => 'aggregator-item',
|
|
),
|
|
'aggregator_page_opml' => array(
|
|
'variables' => array('feeds' => NULL),
|
|
'file' => 'aggregator.theme.inc',
|
|
),
|
|
'aggregator_page_rss' => array(
|
|
'variables' => array('feeds' => NULL),
|
|
'file' => 'aggregator.theme.inc',
|
|
),
|
|
);
|
|
}
|
|
|
|
/**
|
|
* Implements hook_menu_link_defaults().
|
|
*/
|
|
function aggregator_menu_link_defaults() {
|
|
$links = array();
|
|
$links['aggregator.admin_overview'] = array(
|
|
'link_title' => 'Feed aggregator',
|
|
'description' => "Configure which content your site aggregates from other sites, how often it polls them, and how they're categorized.",
|
|
'route_name' => 'aggregator.admin_overview',
|
|
'parent' => 'system.admin_config_services',
|
|
'weight' => 10,
|
|
);
|
|
$links['aggregator.page_last'] = array(
|
|
'link_title' => 'Feed aggregator',
|
|
'weight' => 5,
|
|
'route_name' => 'aggregator.page_last',
|
|
);
|
|
$links['aggregator.sources'] = array(
|
|
'link_title' => 'Sources',
|
|
'route_name' => 'aggregator.sources',
|
|
);
|
|
|
|
return $links;
|
|
}
|
|
|
|
/**
|
|
* Implements hook_permission().
|
|
*/
|
|
function aggregator_permission() {
|
|
return array(
|
|
'administer news feeds' => array(
|
|
'title' => t('Administer news feeds'),
|
|
),
|
|
'access news feeds' => array(
|
|
'title' => t('View news feeds'),
|
|
),
|
|
);
|
|
}
|
|
|
|
/**
|
|
* Implements hook_cron().
|
|
*
|
|
* Queues news feeds for updates once their refresh interval has elapsed.
|
|
*/
|
|
function aggregator_cron() {
|
|
$result = db_query('SELECT fid FROM {aggregator_feed} WHERE queued = 0 AND checked + refresh < :time AND refresh <> :never', array(
|
|
':time' => REQUEST_TIME,
|
|
':never' => AGGREGATOR_CLEAR_NEVER
|
|
));
|
|
$queue = \Drupal::queue('aggregator_feeds');
|
|
foreach ($result->fetchCol() as $fid) {
|
|
$feed = aggregator_feed_load($fid);
|
|
if ($queue->createItem($feed)) {
|
|
// Add timestamp to avoid queueing item more than once.
|
|
$feed->setQueuedTime(REQUEST_TIME);
|
|
$feed->save();
|
|
}
|
|
}
|
|
|
|
// Delete queued timestamp after 6 hours assuming the update has failed.
|
|
db_update('aggregator_feed')
|
|
->fields(array('queued' => 0))
|
|
->condition('queued', REQUEST_TIME - (3600 * 6), '<')
|
|
->execute();
|
|
}
|
|
|
|
/**
|
|
* Implements hook_queue_info().
|
|
*/
|
|
function aggregator_queue_info() {
|
|
$queues['aggregator_feeds'] = array(
|
|
'title' => t('Aggregator refresh'),
|
|
'worker callback' => 'aggregator_refresh',
|
|
'cron' => array(
|
|
'time' => 60,
|
|
),
|
|
);
|
|
return $queues;
|
|
}
|
|
|
|
/**
|
|
* Checks a news feed for new items.
|
|
*
|
|
* @param \Drupal\aggregator\FeedInterface $feed
|
|
* An object describing the feed to be refreshed.
|
|
*/
|
|
function aggregator_refresh(FeedInterface $feed) {
|
|
// Store feed URL to track changes.
|
|
$feed_url = $feed->getUrl();
|
|
|
|
$config = \Drupal::config('aggregator.settings');
|
|
// Fetch the feed.
|
|
$fetcher_manager = \Drupal::service('plugin.manager.aggregator.fetcher');
|
|
try {
|
|
$success = $fetcher_manager->createInstance($config->get('fetcher'))->fetch($feed);
|
|
}
|
|
catch (PluginException $e) {
|
|
$success = FALSE;
|
|
watchdog_exception('aggregator', $e);
|
|
}
|
|
|
|
// Retrieve processor manager now.
|
|
$processor_manager = \Drupal::service('plugin.manager.aggregator.processor');
|
|
// Store instances in an array so we dont have to instantiate new objects.
|
|
$processor_instances = array();
|
|
foreach ($config->get('processors') as $processor) {
|
|
try {
|
|
$processor_instances[$processor] = $processor_manager->createInstance($processor);
|
|
}
|
|
catch (PluginException $e) {
|
|
watchdog_exception('aggregator', $e);
|
|
}
|
|
}
|
|
|
|
// We store the hash of feed data in the database. When refreshing a
|
|
// feed we compare stored hash and new hash calculated from downloaded
|
|
// data. If both are equal we say that feed is not updated.
|
|
$hash = hash('sha256', $feed->source_string);
|
|
|
|
if ($success && ($feed->getHash() != $hash)) {
|
|
// Parse the feed.
|
|
$parser_manager = \Drupal::service('plugin.manager.aggregator.parser');
|
|
try {
|
|
if ($parser_manager->createInstance($config->get('parser'))->parse($feed)) {
|
|
if ($feed->getWebsiteUrl()) {
|
|
$feed->setWebsiteUrl($feed->getUrl());
|
|
}
|
|
$feed->setHash($hash);
|
|
// Update feed with parsed data.
|
|
$feed->save();
|
|
|
|
// Log if feed URL has changed.
|
|
if ($feed->getUrl() != $feed_url) {
|
|
watchdog('aggregator', 'Updated URL for feed %title to %url.', array('%title' => $feed->label(), '%url' => $feed->getUrl()));
|
|
}
|
|
|
|
watchdog('aggregator', 'There is new syndicated content from %site.', array('%site' => $feed->label()));
|
|
drupal_set_message(t('There is new syndicated content from %site.', array('%site' => $feed->label())));
|
|
|
|
// If there are items on the feed, let enabled processors process them.
|
|
if (!empty($feed->items)) {
|
|
foreach ($processor_instances as $instance) {
|
|
$instance->process($feed);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
catch (PluginException $e) {
|
|
watchdog_exception('aggregator', $e);
|
|
}
|
|
}
|
|
else {
|
|
drupal_set_message(t('There is no new syndicated content from %site.', array('%site' => $feed->label())));
|
|
}
|
|
|
|
// Regardless of successful or not, indicate that this feed has been checked.
|
|
$feed->setLastCheckedTime(REQUEST_TIME);
|
|
$feed->setQueuedTime(0);
|
|
$feed->save();
|
|
|
|
// Processing is done, call postProcess on enabled processors.
|
|
foreach ($processor_instances as $instance) {
|
|
$instance->postProcess($feed);
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Loads an aggregator feed.
|
|
*
|
|
* @param int $fid
|
|
* The feed id.
|
|
*
|
|
* @return \Drupal\aggregator\FeedInterface
|
|
* An object describing the feed.
|
|
*/
|
|
function aggregator_feed_load($fid) {
|
|
return entity_load('aggregator_feed', $fid);
|
|
}
|
|
|
|
/**
|
|
* Renders the HTML content safely, as allowed.
|
|
*
|
|
* @param $value
|
|
* The content to be filtered.
|
|
*
|
|
* @return
|
|
* The filtered content.
|
|
*/
|
|
function aggregator_filter_xss($value) {
|
|
return filter_xss($value, preg_split('/\s+|<|>/', \Drupal::config('aggregator.settings')->get('items.allowed_html'), -1, PREG_SPLIT_NO_EMPTY));
|
|
}
|
|
|
|
/**
|
|
* Implements hook_preprocess_HOOK() for block templates.
|
|
*/
|
|
function aggregator_preprocess_block(&$variables) {
|
|
if ($variables['configuration']['module'] == 'aggregator') {
|
|
$variables['attributes']['role'] = 'complementary';
|
|
}
|
|
}
|