import from flickr -> mediagoblin

[Caleb Davis]

What does the glorious future hold for sucking the life out of existing centralized saas and then breathing it into free alternatives? Okay, specifically, how do we handle this use case:

• Hey, I just heard about MediaGoblin, and I'd like to try it out, but all my stuff is in Flickr. What's the easiest way to move things over?

[Elrond]

Oh well, that's a long run really.

I think, this asks for two things, basicly:

1. An API in mediagoblin, so that a user can remotely control mediagoblin in a programmable way
2. A standaline app, using flickr's and mediagloblin's API to transfer as much data as possible.

For the flickr api there is a python module, or maybe even two. I think, I even used one or both.

Is there a meta bug for "mediagoblin needs an API", that this bug can be linked to?

[Caleb Davis]

well, there's this one - http://bugs.foocorp.net/issues/293

[joar]

It's a feature

[joar]

Very much active. Wrong bug, sorry everyone.

[Christopher Allan Webber]

Yes, I think Elrond is right on with how we should approach this. I think API support will be one of the parts of the holy trinity of big features we kick off after 0.1.0 (along with multimedia, federation) ;)

[Elrond]

This looks quite Low Priority to me, really. Of course, if someone volunteers, it's great.

and to be honest, this sounds like something for the "Feature Ideas" wiki page.

[Greg Grossmeier]

I tried to flesh out some of the basics of this feature in the wiki (good idea, Elrond).

http://wiki.mediagoblin.org/Feature_Ideas/Flickr_Import

Let's edit that page when relevant to get the idea formalized.

[Will Kahn-Greene]

The original url for this bug was ​http://bugs.foocorp.net/issues/539 .
Relations:
#27: blocked

[warp]

I hacked together an import tool using phantomjs. It has quite a few problems, the ones I know about are documented in the README.

code: ​https://gitorious.org/phantomjs-mediagoblin-flickr-import/

[Elrond]

Some options:

1. We could test the above mentioned tool and link to it on the docs or wiki. That would partially close this bug.
2. We have an API by now. So in theory someone could write a (maybe cleaner) desktop tool to do the job. I am unsure, whether we can bring this forward without someone knowing the flickr api (or libraries for it) better.

I am putting this on the 0.4.0 queue, so that we decide on how to continue with this.