With this MR I'd like to introduce a reorganization of the current resources to be harvested. The proposals are as follows:
- to remove "Endpoint" as a base class, as it did not really add any benefit.
Sourcea pure file system abstraction, but no harvesting capabilities by themselves
- add a new class (I called them
FileSchemesbut I'm not too happy about that name) that are resources and use
Sourcesto create registerable items. Currently theres the "single file" thingy, which uses the filename to generate an identifier, and the STAC catalog harvester that recursively registers all items from a catalog stored on a
This MR shall be used as a discussion basis, and will remain a draft until we have a clear idea to where we want to go.