Research, test and document a mechanism for version controlling and auto-loading Superset datasources, metrics, charts and dashboards

To create a slice in superset, insert a row into the slices table with the correct key:value properties.

These are my findings so far about creating a datasource in superset:

1. Superset API has authorization errors when hitting the databaseview/api/create endpoint. Superset REST API documentation doesn't exist and from issues on Superset's repo, no one has been successful in interacting with the API.
2. Adding a datasource directly to superset's postgreSQL database has issues with hashing passwords. On the UI, the password on the SQL Alchemy URI is hashed before insertion.
3. Exporting a datasource as a .yaml file then importing it has the same password hashing issues

A workaround to both 2 and 3 is to click on edit the database and saving it without changing anything, this way the password in the SQL alchemy uri is hashed. Editing on NiFi returns an OK response but the database is not changed.
An upside to exporting the yaml files is all related tables to a datasource are exported as well.
A bug I just discovered on exporting a database with tables is that the maximum number of tables can only be 1.

@Antonate Maritim , can you document the various methods you tried with superset and failure points in this ticket?

The Stub of the documentation can be found at: https://openlmis.atlassian.net/wiki/spaces/OP/pages/413565018/Superset+API+Documentation

@Craig Appl I think we should try to fit it in for next Sprint if possible. I would prioritize the Scalyr ticket higher though, as this ticket is not required for SELV since we're doing the Superset build directly in the SELV instance of Superset