[regression] properly support more than 1 dbt project under scope

[bastienboutonnet]

#Rationale Originally we wanted to provide support for more than one dbt project under a given syrup config. However, this would slow our initial development but it is a regression as we strongly believe dbt-sugar was built with the main objective in mind to keep column definitions consistent across all of a user's tables.

We discussed a few points in the following issues:

• https://github.com/bitpicky/dbt-sugar/issues/38
• https://github.com/bitpicky/dbt-sugar/issues/13

But we made a conscious choice in the code to only support one dbt project for now to allow us to have an MVP out sooner (https://github.com/bitpicky/dbt-sugar/blob/38ba2aff80514c8eacebb4d2ee64a7dd38f2acc1/dbt_sugar/core/main.py#L124-L128) --well actually we are sketchy about it now and we should correct that and throw a proper error.

Challenge/Potential Solutions

Achieving this means that we need to be able to represent and store the models information that we parse from the schema.yml files in a one-level more nested dict that also contains which dbt_project it's in so that we can target and use the right profiles (i.e. credentials and db info) to use when we query information schemas.