Auto sync storage profiles #378

[UmmulkiramR]

Score-client will now dynamically load storage profile specific implementations (currently S3 or Azure) based on the profile score-server is running on. This will eliminate the need for users to ensure that score-client is running on a profile compatible with score-server.

This change contains

• a new endpoint to fetch the profiles currently active on score-server (current storage profiles on score-server are azure, amazon, collaboratory).
• configuration changes to ensure the right storage implementations are injected at runtime for uploads and downloads based on the score-server profiles fetched.
• defaulting behaviour when used with older score-server versions where the profile endpoint isn't available. There the score-client will use the default (S3) storage profile specific implementation.

[UmmulkiramR]

Thanks for the quick turn around on this!

Most important thing I'm requesting here is the response data in the /profiles endpoint. I think we need to be returning information strictly about the storage type for the server, and not exposing all our runtime profiles. The client will then need to correctly map the storage type to its own corresponding profiles/beans.

One additional request: Can we please update the documentation in the score-client README.md . Currently there is text about using a profile to connect with azure, we now replace this with text that explains that the client will request the storage profile from the server and use the correct functionality, but when the score server is version {current version} or lower then the Azure profile is necessary.

yes, I forgot to mention that the API only returns information regarding the storage type. I have added a new profile configuration (in the server) that will handle that part.