[Idea] Allow cleoc to either use the GraphQL API or access the DB directly

The idea being that on a local system you would not need to start the web UI in order to add another forecasts or similar operations. It would not be too hard to do probably by just moving parts of cleosrv into a common library. On the other hand, what are the advantages? Why not have cleosrv run continously as a service?

Start by reading the cleoc and cleosrv entry points and comparing the GraphQL API with direct DB access. The issue does not name files or tests; done would require an agreed choice between a shared library, direct database access, or continuously running cleosrv, along with a defined implementation scope.

Written by the indexing model from the issue text.