Use Github API to extract license info

[andreww]

Github scans repos for licence files, and if it finds them and understands the contents it annotates repositories with the licence. This is exposed via the API: https://developer.github.com/v3/repos/

We could (and probably should) extract this information via the API call rather than trying to scrape the content of the repo. However, this leaves us with an question: what do we do if there is a file but github didn't understand it. We should probably do a fallback check for this and tag such resources as having an "unknown" licence (rather than having putting "no licence" in the output).

While we are at it, it would be worth seeing if there is any other information we can use. For example, the API exposes parents of forked repos.

[andreww]

This is also where we will find the last repository update time.

[longr]

Just had a tinker with this earlier. Needs some playing with as it is not clear which field is the best to use in the return field.

[andreww]

The spdx_id value I think.

[andreww]

Hum - as far as I can see the update to pygithub to expose this bit of the API only got merged three days ago (PyGithub/PyGithub#734), so if this gets used we'll need an up-to-date pygithub!