Make the CKAN Datastore an independent library

It would be great to be able to use the CKAN Datastore as a completely separate library from CKAN.

This would allow more focused development on the datastore, make it easier for someone to create an open source service using it, easier to plug into scalable systems, and easier for other CMSs to share open data and still be able to parse and share delimited data with an API.

There are a number of Datastore as a service companies popping up as well as a couple of other datastore libraries that I know of. Making the CKAN Datastore independent would allow for a much more versatile and competitive within this ecosystem.

The issue proposes making the CKAN Datastore independent from CKAN, but names no files, tests, or entry points. Start by reviewing how the Datastore is currently integrated with CKAN and define the library boundary and supported use cases. Done would mean the Datastore can be used separately as a library and remains able to parse and share delimited data through an API.

Written by the indexing model from the issue text.