Finding the URL for podcast feeds from an iTunes id. (iTMS API)

Question

I'm look at a way of turning an iTunes podcast id into the RSS feed that the podcast producer serves.

I'm aware of the RSS generator, which can be used to generate a feed of links to podcasts, but these links are to HTML pages.

If you have iTunes open, you can manually export the list of podcasts by exporting to OPML, so we can surmise that iTunes eventually knows how to decode them (i.e. they're not exclusively going through an iTMS host).

I have looked at the Affiliate API document which gives you some nice JSON back. This gives you a collectionViewUrl which is the same as the ones given in the RSS generator, and incidentally, the iTunes Link Generator. It also give you the id, and a whole load of other things including a preview audio file which is not hosted on the phobos.

At this point, I'm looking for anything that would help me solve this question, including any language, unofficial or not.

(in actual fact, I'd prefer something vaguely supported, and in Java, that didn't involve HTML scraping).

Answer 1

Through a combination of answers from these two questions, I have found a way to do what I want.

Example of finding podcasts

First: grab a list of podcasts from iTunes, using the RSS generator. I'm not sure how the query parameters work yet, but here is an RSS feed for top tech podcasts in the US.

http://ax.itunes.apple.com/WebObjects/MZStoreServices.woa/ws/RSS/toppodcasts/sf=143441/limit=25/genre=1318/xml

• sf relates to country, and is optional. I would guess that this defaults to global if absent.
• genre relates to genre, and is optional. I would guess that this defaults to "all genres" is absent.
• limit is optional, and seems to default to 9.

This gives you an Atom feed of podcasts. You'll need to do some sperlunking with XPath to get to the ITMS id of podcast, but you're looking for the numeric id contained in the URL found at the following XPath:

/atom:feed/atom:entry/atom:link[@rel='alernate']/@href

For example, the excellent JavaPosse has an id of 81157308.

The Answer to the Question

Once you have that id, you can get another document which will tell you the last episode, and the original feed URL. The catch here is that you need to use an iTunes user-agent to get this document.

e.g.

wget --user-agent iTunes/7.4.1 \
     --no-check-certificate \ 
     "https://buy.itunes.apple.com/WebObjects/MZFinance.woa/wa/com.apple.jingle.app.finance.DirectAction/subscribePodcast?id=81157308&wasWarnedAboutPodcasts=true"

This is a plist containing some metadata about the podcast, including the feed URL.

<key>feedURL</key><string>http://feeds.feedburner.com/javaposse</string>

The XPath for this could be something like:

//key[@text='feedURL']/following-sibling::string/text()

Disclaimer

Not entirely sure how stable any of this is, or how legal it is. YMMV.