Media Web Services module

Posted by iaine on August 31, 2011 at 4:45pm

Afternoon,

Due to a need at the company I work for, I am starting to put together the use cases for a web services front end to the Media module.

The idea is that we will be able to perform CRUD operations on a site using the Media module as a sort of asset management system but to also index any items that are stored in the system (hoped to be done through hooking into taxonomy). The hope is that this module would sit in front of Media and also the Derivatives API so that various other systems can retrieve media without re-implementing the work in other modules. The services will be provided by the Services module.

I am working on some skeleton code at the moment as a concept for work but hope to be able to post a stub shortly. I might wait for the big media sprint which is being organised to avoid using older hooks and having to refactor after the sprint.

I am trying to start out fairly small and to build up so any uses cases that might be of interest will be welcomed.

Thanks, Iain

Comments

Maybe I'm missing something,

Posted by kreynen on August 31, 2011 at 5:47pm

Maybe I'm missing something, but I'm not really sure what you are proposing. Are you trying to develop ZenCoder or Kaltura in Drupal?

Having tried to do something like that, I can tell you that Drupal is not the best hammer for that nail. While there's been a lot of discussion lately about small core and Drupal as a framework, right now Drupal is still a CMS product at this point and the APIs, hooks, and UI are all designed at least in part to facilitate publishing content.

If that's not it, there are already several modules that consume and display media from different providers/services that sit in front of/on top of media. The Derivatives API provides the structure for moving files between different providers while keeping track of the fact that it is the same file.

Can you be more specific about what you are working on?

Indexing media asset locations

Posted by iaine on August 31, 2011 at 8:31pm

At the moment I see it more as indexing assets in a particular repository created by the Media module than transcoding assets / streams on the fly.

Essentially I am trying to write a lightweight web service API that I can index and direct a user or system to the description or location of an asset (image, audio, or video) through XML or JSON. It would then initially given the stored location of the asset and its metadata to load from (or a base 64 encoded version of the file if it is a static image but I'm not fully set on that notion) that particular repository.

We may have a need to create or streams from our own systems as we do some video work already in another part of the company but this is something that will be dealt with as we come to that issue. A long term idea is to have some way of using the service to perform full CRUD operations but that does open the transcoding can of worms that you allude to. Derivatives is something that we will require if we make copies or transcode any assets.

I'm trying to design it to sit on top of existing streams, hooks and API implementations rather than try to recreate these as I already use module such as Media:YouTube and so on. One of my issues is that I have non-Drupal systems requiring access to some of these assets so I need to write something to allow discovery from one API and then hooking the consumption into that is the next step but will have to be on a case by case basis.