We need a general workflow to be the following:
- there is some schema / definition of a dataset, and corresponding flies
- we create dynamic code that will generate the in-memory database
- entrypoints must allow for interaction / search ,etc.
Need to think more about this, namely what kind of dataset to start with (possibly something simple, without an actual schema, like a folder of files) and then move toward things like BIDS, schema.org, etc.