relationship tables were created

n:m (many to many) relationships will be modeled with a third table holding the ids of the two related tables.
Naming conventions for such tables are described here:
https://stackoverflow.com/questions/1813321/what-should-i-name-a-table-that-maps-two-tables-together
This will result in the following tables:

• stationmeta_core_stationmeta_roles
• stationmeta_core_stationmeta_annotations
• timeseries_timeseries_roles
• timeseries_timeseries_annotations
• timeseries_timeseries_programmes

changed due date to June 03, 2020

added .doing label

changed the description

All tables have been created like this (example for stationmeta_core_stationmeta_roles):

CREATE TABLE stationmeta_core_stationmeta_roles
(
  station_id INT NOT NULL references stationmeta_core(id),
  role_id INT NOT NULL references stationmeta_roles(id),
  PRIMARY KEY(station_id, role_id)
);

The right-hand side tables of these relationships have then be altered to drop the reference to the left-hand side id (example for stationmeta_roles):

ALTER TABLE stationmeta_roles DROP COLUMN station_id;

mentioned in commit b6157ef9

Before writing new pytests, #4 (closed) has to be completed.
There have been problems in using string declarations (not working for pytest) and the explicit class names had to be used (see: https://groups.google.com/forum/#!topic/sqlalchemy/YjGhE4d6K4U).
With explicit class names, the association table did not work!

mentioned in issue #4 (closed)

The association tables are now also defined in models.py (and working). This was done in commit a4e58c65.

This issue will now be closed. Next steps will be undertaken in issue #4 (closed).

closed

added .done label and removed .doing label

added enhancement label