Derivatives are extremely simple, and don't need the app environment loaded to do their work. It would be nice if we could pull all their code out of OD2 and put it into a separate project that is dedicated to derivative processing. The derivative servers won't see any significant gain from such a move, but the mental overhead of the core project would be reduced (and dev environments might be simpler).
There are a few ways to go about this. We could go with a totally generic derivative file server that does nothing but convert files from one type to another based on very simple rules / instructions we send it. Or we could go all-in with another Hydra app that handles our use case exactly as we want it, and updates Fedora / solr as necessary. There are pros and cons to both. Discuss, experiment with balancing the two approaches, and see what works best.
Derivatives are extremely simple, and don't need the app environment loaded to do their work. It would be nice if we could pull all their code out of OD2 and put it into a separate project that is dedicated to derivative processing. The derivative servers won't see any significant gain from such a move, but the mental overhead of the core project would be reduced (and dev environments might be simpler).
There are a few ways to go about this. We could go with a totally generic derivative file server that does nothing but convert files from one type to another based on very simple rules / instructions we send it. Or we could go all-in with another Hydra app that handles our use case exactly as we want it, and updates Fedora / solr as necessary. There are pros and cons to both. Discuss, experiment with balancing the two approaches, and see what works best.