Handle authentication schemes

[almet]

The SPORE specification talks about authentication, but I don't think they really talk about any authentication scheme, probably because this has to be handled by the challenge on the website itself.

We need to add support for that in our code. Also, some resources may need authentication, but could work without. I'm not sure SPORE handles this properly, we'll need to dig a bit more to see what we can do about that.

[Natim]

Related to #1

[Natim]

Actually, how do you provide the authentication material with respire. That is the real question.

About some SPORE API description I found this: https://github.com/SPORE/api-description/tree/master/services

[Natim]

For github it is easy, you just need to provide the access_token with the data:

GET https://api.github.com/user?access_token=...

[Natim]

For ducksboard, it is HTTP_AUTH header so we need to be able to give some header at client_from_api creation that will be use.

I think of a list of middleware (callbacks) that will manipulates the dict of requests.method arguments.

[almet]

I think this is handled by the http spec, see http://pretty-rfc.herokuapp.com/RFC2616#header.www-authenticate

I guess that we should try to get the authentication scheme this way if we don't know how to deal with it.