Allow the AIDB engine to set up the configuration and metadata tables about aidb HOT 4 CLOSED

https://github.com/ddkang/aidb-new/tree/conrevo/setup

In setup folder, we can construct table schema from the old config.yaml via a single command. At this time we should separately run setup once. I don't think we should implement this in aidb engine since it is user's responsibility.

from aidb.

We should use small set of CSVs (like 1k rows of the object detection CSV, count, and color or something) instead of copying a bunch of code to run the setup.

from aidb.

In that way, I think a config.yaml is still needed, otherwise we have no idea what is primary key, what is foreign key and so on. I think the only thing to add is to insert the csv into the database.

from aidb.

A yaml is fine, but there's many ways to go about doing this that don't involve huge amounts of code copy. For example, you could implement serialization and deserialization of the new config. Let's keep the code clean.

from aidb.