Add different types and various cases to the dataset

[Jyuart]

It would be nice to have different types in the data set, first of all, to test dependents graphs in different combinations. It could be:

• indexes
• views (materialized views)
• triggers

Among cases there could be:

• self-referential tables
• cross-referential tables

[mathemancer]

Note to the implementer:

Consider simulating data akin to the 'academics' data set: https://github.com/centerofci/mathesar/blob/master/db/tests/resources/academics_create.sql, but with better types and some views, etc. I'd prefer using that one since it aligns with a reasonable use case, and would avoid complicating the library data set.