Models for open data simplified database

[Luisella21]

TP2000-1740 Open data database

There is a desire to use TAP data in open data analysis. The data currently published contain all the historical information, and analysis are complex.

This PR create a set of 'current' data that can be exported every day and used to create reports within TAP.

It creates a shadow model of each tracked model in TAP. The shadow model contains only the approved and current version.
Where feasible, models have been merged: for instance, descriptions are in the described model, not in a different one.

The tables supporting the models are created in a different partition of the database, so they can easily be excluded from the daily back up of the database.

There is a mechanism to populate the data : this will eventually happen daily, but it is not part of the current PR.
Some raw sql is used to copy the data (to improve performance). Even so, the copy process on localhost takes a couple of hours. This may or may not be a problem, it is too early to know.

• Requires migrations: Yes
• Requires dependency updates - No

[TomMacca]

Really nice work @Luisella21

One possible addition could be to save the date of the simplified snapshot somewhere in the new reporting schema. It might help that we pass it along from the schema itself to the guys who are going to be making reports so its easier to find out if somethings gone wrong with the transfer/ snapshot. Maybe:

• A snapshot_date column on all the tables?
• Or it could be in the table names themselves?
• Or a metadata table?
  🤔

What do you think?
Another q, I take it the plan is to overwrite the reporting schema each day in TAP?

[TomMacca]

Approved from me, looks great @Luisella21