Ephemeral relations, Refactor & Cleanup

[MMagueta]

Context

In this branch I did a major refactor to enforce a clear segregation between iterators and pure relational operations.

Implementations

Ephemeral Relations

Ephemeral relations are like views in the SQL standard, except that here they are evanescent objects generated out of running any query whatsoever. In an analogy, they are more closely related to the output of a query.

Provenance and lineage at the relation level

It didn't make sense to store the provenance and lineage at the tuple level. A lonely tuple is not bound by any relational scope and practically speaking, causes horrendous duplication of information on rendering for no reason. So I moved it all up to the relation record itself.

Segregation between iterators and operators

The distinction was not clear, so I made pure relational operations to spawn an iterator on demand as to generate its tuples. This is used by both ephemeral and persistent relations. Iterators just pull data, operations organize tuples.

Documentation

Purged old dummy AI-created docs

Example

operations:setup(), 
DB = operations:create_database(test),
{DB1, _} = operations:create_relation(DB, employees, #{id => integer, name => string, dept_id => integer}),
{DB2, _} = operations:create_relation(DB1, departments, #{id => integer, dept_name => string}), 
Employees = [#{id => I, name => "Emp" ++ integer_to_list(I), dept_id => (I rem 3) + 1} || I <- lists:seq(1, 10)],
Depts = [#{dept_id => 1, dept_name => "Eng"}, #{dept_id => 2, dept_name => "Sales"}, #{dept_id => 3, dept_name => "HR"}],
{DB3, E} = operations:create_all_tuples(DB2, employees, Employees),
{DB4, D} = operations:create_all_tuples(DB3, departments, Depts),
operations:collect_all(((relational_operators:join(E, D, dept_id))#relation.generator)(#{})).