Discover and Read RSS and Atom Feeds

by Matthew Weier O'Phinney

Remember RSS and Atom feeds?

Chances are, you may have discovered this book because it was announced on a feed:

• A number of Twitter services poll feeds and send links when new entries are discovered.
• Some of you may be using feed readers such as Feedly¹.
• Many news aggregator services, including tools such as Google Now, use RSS and Atom feeds as sources.

An interesting fact: Atom itself is often used as a data transfer format for REST services, particularly content management platforms! As such, being familiar with feeds and having tools to work with them is an important skill for a web developer!

In this first of a two part series on feeds, we'll look at feed discovery, as well as reading, using zend-feed's Reader subcomponent.

Getting started

First, of course, you need to install zend-feed:

$ composer require zendframework/zend-feed

As of version 2.6.0, the component has a very minimal set of dependencies: it only requires zendframework/zend-escaper and zendframework/zend-stdlib in order to work. It has a number of additional, optional requirements depending on features you want to opt-in to:

• psr/http-message and/or zend-http, to allow polling pages for feeds, feeds themselves, or PubSubHubbub services.
• zendframework/zend-cache, to allow caching feeds between requests.
• zendframework/zend-db, which is used when using the PubSubHubbub subcomponent, in order for PuSH subscribers to store updates.
• zendframework/zend-validator, for validating addresses used in Atom feeds and entries when using the Writer subcomponent.

For our examples, we will need an HTTP client in order to fetch pages. For the sake of simplicity, we'll go ahead and use zendframework/zend-http; if you are already using Guzzle in your application, you can create a wrapper for it following instructions in the zend-feed manual².

$ composer require zendframework/zend-http

Now that we have these pieces in place, we can move on to link discovery!

Link discovery

The Reader subcomponent contains facilities for finding Atom and RSS links within an HTML page. Let's try this now:

// In discovery.php:

use Zend\Feed\Reader\Reader;

require 'vendor/autoload.php';

$feedUrls  = [];
$feedLinks = Reader::findFeedLinks('https://framework.zend.com');

foreach ($feedLinks as $link) {
    switch ($link['type']) {
        case 'application/atom+xml':
            $feedUrls[] = $link['href'];
            break;
        case 'application/rss+xml':
            $feedUrls[] = $link['href'];
            break;
    }
}

var_export($feedUrls);

If you run the above, you should get a list like the following (at the time of writing):

array (
  0 => 'https://framework.zend.com/security/feed',
  1 => 'https://framework.zend.com/blog/feed-atom.xml',
  2 => 'https://framework.zend.com/blog/feed-rss.xml',
  3 => 'https://framework.zend.com/releases/atom.xml',
  4 => 'https://framework.zend.com/releases/rss.xml',
)

That's rather useful! We can poll a page to discover links, and then follow them!

Internally, the returned $feedLinks is a Zend\Feed\Reader\FeedSet instance, which is really just an ArrayObject where each item it composes is itself a FeedSet with specific attributes set (including the type, href, and rel, usually). It only returns links that are known feed types; any other type of link is ignored.

Reading a feed

Now that we know where some feeds are, we can read them.

To do that, we pass a URL for a feed to the reader, and then pull data from the returned feed:

// In reader.php:

use Zend\Feed\Reader\Reader;

require 'vendor/autoload.php';

$feed = Reader::import('https://framework.zend.com/releases/rss.xml');

printf(
    "[%s](%s): %s\n",
    $feed->getTitle(),
    $feed->getLink(),
    $feed->getDescription()
);

The above will result in:

[Zend Framework Releases](https://github.com/zendframework): Zend Framework and zfcampus releases

The above is considered the feed channel data; it's information about the feed itself. Most likely, though, we want to know what entries are in the feed!

Getting feed entries

The feed returned by Reader::import() is itself iterable, which each item of iteration being an entry. At its most basic:

foreach ($feed as $entry) {
    printf(
        "[%s](%s): %s\n",
        $entry->getTitle(),
        $entry->getLink(),
        $entry->getDescription()
    );
}

This will loop through each entry, listing the title, the canonical link to the item, and a description of the entry.

The above will work across any type of feed. However, feed capabilities vary based on type. RSS and Atom feed entries will have different data available; in fact, Atom is considered an extensible protocol, which means that such entries can potentially expose quite a lot of additional data!

You may want to read up on what's available; follow the footnotes to find relevant links:

• RSS entry properties³
• Atom entries⁴

Until next time

zend-feed's Reader subcomponent offers a number of other capabilities, including:

• Importing actual feed strings (versus fetching via an HTTP client)
• The ability to utilize alternate HTTP clients.
• The ability to extend the Atom protocol in order to access additional data.

The zend-feed component has extensive documentation⁵, which will answer most questions you may have at this point.

We hope this quick primer gets you started consuming feeds!

Footnotes

¹. https://feedly.com ↩

². https://docs.zendframework.com/zend-feed/psr7-clients/ ↩

³. https://docs.zendframework.com/zend-feed/consuming-rss/#get-properties ↩

⁴. https://docs.zendframework.com/zend-feed/consuming-atom/ ↩

⁵. https://docs.zendframework.com/zend-feed/ ↩