Mechanism to store the data inside FINALES into the BIGMAP Archive

[ramirezfranciscof]

It is still not clear if this should be done via a separate Tenant or by the server itself.

• Advantages of doing it via a tenant: would help with modularization and separation of concerns (specially since the people who developing FINALES and the Archive are in different groups). It is also easier to do in a way in which FINALES remains case agnostic (and could be connected to a different type of long term storage by just switching tenants).
• Disadvantages: would involve extra unnecessary transfers of data, since all of the data of FINALES would have to be sent to the Tenant and then re-sent to the Archive. It also complicates things with the RESTapi interface since it was not thought for exposing data for long term storage but rather operational communication.

Independently of whether this is done by the server or a tenant, there is also the discussion of how the data should be formatted for storage: just a dump of the database? or should it take the shape of the objects that are exposed through the RESTapi?

Here are some resources for when considering the step of dumping the database:

• The backup function of SQLAlchemy: https://docs.python.org/3/library/sqlite3.html#sqlite3.Connection.backup
• The way in which the objectstore deals with SQLite connections: https://github.com/aiidateam/disk-objectstore/blob/main/disk_objectstore/database.py
• Instructions on creating a backup of SQLite using python: https://www.geeksforgeeks.org/how-to-create-a-backup-of-a-sqlite-database-using-python/

[SKS94]

This has been implemented via a tenant. Sufficient endpoints for exposing the necessary information has been added, along with a dump endpoint of the entire database.