Atlas is an open source, open API that gathers metadata on audio and video from across the web and makes it accessible through a powerful and rich query interface. Basically, Atlas does the legwork needed so that you can quickly and simply build cool stuff with content.
Have a look at the API Examples for an idea of the kind of queries you can make, and at Sources for a list of the 10 or so content sources currently indexed.
http://atlasapi.org/2.0/:type.:format?:query
What you get back
Items contain all the important information about a piece of media. Episodes extend Items as episodic content containing series and episode number, along with a specific Brand.
Playlists contain lists of Items and other Playlists. Brands are Playlists but represent established brands by the broadcasters. For example, The Wire or Spaced are Brands that contain episodes while the current BBC Hotlist is a Playlist.
The format
We support loads of output formats. Chances are, you’ll want to go with the JSONP or XML output, but the HTML output is useful for playing about with the API. You might also be interested in the ATOM feeds or the super simple OEMBED format.
For the linked-data minded, we have an illustrious RDF history with a fully linked up RDF output and published ontology.
Query parameters
Atlas supports a rich set of potential query parameters to help you filter your results down to what you actually want. You can filter by pretty much anything in the data-model but there are a number of short-cuts and defaults that will make your life easier. The basic format of a parameter entityName.attributeName-operator=value. To make this clearer a here are a few examples:
- Find ‘Eastenders’ by URI: uri=http://bit.ly/8abiYZ
- Find ‘Peep Show’ on Channel4 by title: title=peepshow&publisher=channel4.com
- The first 5 episodes of Peep Show: brand.uri=c4:peep-show&seasonPosition=1&position-lessThan=5
- Downloadable items from blip.tv: publisher=blip.tv&transportType=download&limit=2
Shortcuts and defaults…
Rather than always asking for /brands.json?brand.uri=… when you want a brand, and /items.json?item.uri=… when you want a specific item, you can just ask for /brands.json?uri=… and it will work out what you’re after. If, however, you are filtering based on an attribute in a different data type then you will have to be more specific. This would be the case when asking for a brand’s episodes: /items.json?brand.uri=….
Similarly, /items.json?location.available=true can simply be shortened to /items.json?available=true. However, as we figured you’d usually be after media that you could actually use, this filter is on by default.
We generally expect you to be looking for information on content that is available and in a webby format, if you’re looking for media for an iPhone, you’ll want to explicitly specify /items.json?dataContainerFormat=video/mp4.
Title searches
The title parameter works in a similar way to /items.json?item.uri=… However, title is a bit special as it actually performs a search so you don’t have to get it exactly right (if you know exactly what you’re looking for ask for then it’s best to ask for it it by URI or CURIE).
So, /brands.json?title=east will return