Flickr Integration - API support

[Fitzgerald Steele]

I just tried out the Flickr translator...very neat idea, guys! It
lead me to two questions:

1) Does the translator screen scrape, or use the Flickr API? I'm
guessing that it scrapes because I noticed that when I imported a
picture, Zotero created a tag "Add a Tag." I'll submit a ticket for
that as soon as my trac access somes through.

2) More generally, this brings up the issue of importing metadata via
various web services and APIs. I imagine people will want to import
from the increasing variety of sources that expose an API (Youtube
comes to mind, as well as all the flickr clones, connotea and clones,
etc). It seems if the service has an API, Zotero should be able to
grab the data in a very reliable fashion. Has there been any
discussion on standards or conventions for developers to map API
points to Zotero fields?

Thanks,

Fitzgerald Steele
PhD Student, University of Iowa

[Raymond Yee]

Hi Fitzgerald,

I'm very keen on doing more work on using Zotero to access various
APIs. (I teach a course on "mixing and remixing information" at
Berkeley and am finishing a book on the subject.) I don't have much
time to put into the subject over the next couple of months but hope
to jump in deeply in January/February. In the meantime, I'd be glad
to work with you and others who want to pursue this topic.

A few comments about your points:

1) I would imagine that the Flickr translator scrapes rather than uses
the API. It would be interesting to think about how to use APIs --
you immediately run into a number of issues. One is that you have to
have an API key. Two, you would have to authenticate the user to
access private photos, etc. Scraping, though fragile, gets around
those issues.

2) I'd be really happy to participate in any discussions around
standards or conventions for developers to map API data to Zotero
fields. Do you have a specific problem in mind to give some focus to
our discussion? Interested in building out the Flickr integration?

-Raymond

On Nov 16, 9:15 am, Fitzgerald Steele <fitzgeraldste...@gmail.com>
wrote:
> I just tried out theFlickrtranslator...very neat idea, guys! It

> lead me to two questions:
>

> 1) Does the translator screen scrape, or use theFlickrAPI? I'm

> guessing that it scrapes because I noticed that when I imported a
> picture, Zotero created a tag "Add a Tag." I'll submit a ticket for
> that as soon as my trac access somes through.
>
> 2) More generally, this brings up the issue of importing metadata via
> various web services and APIs. I imagine people will want to import
> from the increasing variety of sources that expose an API (Youtube

> comes to mind, as well as all theflickrclones, connotea and clones,

[Sean Takats]

In general we try to use APIs rather than screen scrape. For example,
we have been using the API at arXiv.org and for Amazon.com for over a
year.

The initial version of the Flickr translator did indeed scrape, but
now it uses the API and is considerably more reliable. Thanks to the
API it can also now capture the highest resolution version available
of the image. We have also recently released a YouTube translator
which also uses their API.

Raymond's point about authentication is important, and I will look
into how we might incorporate authentication into the translator,
which might not be too difficult. For now, however, I believe that
the increased reliability of the API version outweighs authentication
considerations. As for the Flickr API key, I simply created one for
Zotero.

A discussion about API mapping sounds great. One place where we have
made some progress in mapping items to Zotero is through our
modifications to our COinS translator, which will support the import
of any Zotero item type.

Regards,
Sean