Use global UUIDs for pipelines & algorithms

[mihow]

Currently we use slugs & keys for matching pipelines between Antenna and the Processing Service / ML Backends. However we don't want to miss-match pipelines across projects that happen to have the same slug (e.g. something like "default" or "moths"). And we do want to match pipelines and algorithms that are truly the same, even if they are in both a private Processing Service and a shared Processing Service.

More TDB @mihow