jan/miniflux-legacy
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Add new grabber rules: degroupnews.com and sitepoint.com

Browse Source
This commit is contained in:
Frédéric Guillot 2014-10-28 18:40:13 -04:00
parent c0ce32094e
commit ab406e8eaa
4 changed files with 175 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
vendor/PicoFeed/Filter.php vendored
View File

@ -110,9 +110,13 @@ class Filter
*/
public static function stripWhiteSpace($value)
{
$value = str_replace("\r", "", $value);
$value = str_replace("\t", "", $value);
$value = str_replace("\n", "", $value);
$value = str_replace("\r", ' ', $value);
$value = str_replace("\t", ' ', $value);
$value = str_replace("\n", ' ', $value);
// Break UTF-8 strings (TODO: find a better way)
// $value = preg_replace('/\s+/', ' ', $value);
return trim($value);
}

149
vendor/PicoFeed/Parser.php vendored
View File

@ -2,6 +2,7 @@
namespace PicoFeed;
use SimpleXMLElement;
use DateTime;
use DateTimeZone;
@ -467,4 +468,152 @@ abstract class Parser
{
$this->grabber_ignore_urls = $urls;
}
/**
* Find the feed url
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findFeedUrl(SimpleXMLElement $xml, Feed $feed);
/**
* Find the feed title
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findFeedTitle(SimpleXMLElement $xml, Feed $feed);
/**
* Find the feed description
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findFeedDescription(SimpleXMLElement $xml, Feed $feed);
/**
* Find the feed language
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findFeedLanguage(SimpleXMLElement $xml, Feed $feed);
/**
* Find the feed id
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findFeedId(SimpleXMLElement $xml, Feed $feed);
/**
* Find the feed date
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findFeedDate(SimpleXMLElement $xml, Feed $feed);
/**
* Find the feed logo url
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findFeedLogo(SimpleXMLElement $xml, Feed $feed);
/**
* Get the path to the items XML tree
*
* @access public
* @param SimpleXMLElement $xml Feed xml
* @return SimpleXMLElement
*/
public abstract function getItemsTree(SimpleXMLElement $xml);
/**
* Find the item author
*
* @access public
* @param SimpleXMLElement $xml Feed
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
*/
public abstract function findItemAuthor(SimpleXMLElement $xml, SimpleXMLElement $entry, Item $item);
/**
* Find the item URL
*
* @access public
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
*/
public abstract function findItemUrl(SimpleXMLElement $entry, Item $item);
/**
* Find the item title
*
* @access public
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
*/
public abstract function findItemTitle(SimpleXMLElement $entry, Item $item);
/**
* Genereate the item id
*
* @access public
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findItemId(SimpleXMLElement $entry, Item $item, Feed $feed);
/**
* Find the item date
*
* @access public
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
*/
public abstract function findItemDate(SimpleXMLElement $entry, Item $item);
/**
* Find the item content
*
* @access public
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
*/
public abstract function findItemContent(SimpleXMLElement $entry, Item $item);
/**
* Find the item enclosure
*
* @access public
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findItemEnclosure(SimpleXMLElement $entry, Item $item, Feed $feed);
/**
* Find the item language
*
* @access public
* @param SimpleXMLElement $entry Feed item
* @param \PicoFeed\Item $item Item object
* @param \PicoFeed\Feed $feed Feed object
*/
public abstract function findItemLanguage(SimpleXMLElement $entry, Item $item, Feed $feed);
}

10
vendor/PicoFeed/Rules/degroupnews.com.php vendored Normal file
View File

@ -0,0 +1,10 @@
<?php
return array(
'test_url' => 'http://www.degroupnews.com/medias/vodsvod/amazon-concurrence-la-chromecast-de-google-avec-fire-tv-stick',
'body' => array(
'//div[@class="contenu"]',
),
'strip' => array(
'//div[contains(@class, "a2a")]'
),
);

9
vendor/PicoFeed/Rules/sitepoint.com.php vendored Normal file
View File

@ -0,0 +1,9 @@
<?php
return array(
'test_url' => 'http://www.sitepoint.com/creating-hello-world-app-swift/',
'body' => array(
'//section[@class="article_body"]',
),
'strip' => array(
),
);