Refactor new DB init and management

Summary

Currently the setup for a new DB (e.g. for tests) is done via calling two helper routines:

• upload_observatory_data
• database_initialization

The first tool creates and populates tables that require the content to be present. The second tool is creating tables and populates them with arbitrary data.

On top of this, another tool is used to upload the reference atmo models and called 3 times - can be refactored too.

What is the expected correct behavior?

The database tables should be analyzed and tools should be refactored such that:

1. Generalized code is used to create tables
2. No tables shall be populated with completely arbitrary data unless completely necessary.
3. Number of tools shall be kept to a minimum.

Relevant logs and/or screenshots