global: move away from "experiment" paradigm to general "access scope"

[lukasheinrich]

This wi;l need some discussion, but maybe we can seed it here. Right now the notion of an "experiment" (as in ATLAS, CMS, etc..) is used in a lot of the code. In order to be generic probably it's best to define a number of "access scopes" and use that name across the code. If we want REANA e.g. be installed on-premises elsewhere, a "scope" could be a specific professor's lab, where various labs use the same reana instance but have their own secrets, etc.

[tiborsimko]

👍 Let's muse live about scope, tenant, etc.