Does anyone know of a document store than can run in memory like HsqlDB or configured to run on little resources (similar to postgres)?

I have written a dockerised set of Java services, most of them have fixed data structures and so I have used Hibernate with HsqlDB/Postgres for them, however…

The idea is to construct dynamic data processing pipelines which each add microformats for analysis if this takes place.

This means the resulting objects can be quite different, normally I would use a document store like MonogoDB or Elastic Search but…

I want developers/data scientists to be able to spin everything up on their development machine and machine learning models are often insanely RAM intensive.

So I want something I can use that keeps its memory/cpu footprint low, it doesn’t need to be performant since its for development and the production instance can use something better.