TODOs

In no particular order

User defined default settings for all mappings
Refactor, methods too big, objects missing
Name things better
Full adaptorization of the database, refactor some Sequel specific things into the adapter

Candidate features to consider

Column aliasing
Callbacks e.g. after_save, after_insert as functions defined in mapping
Database generated IDs and Timestamps (perhaps implemented as callbacks)
When possible optimise blocks given to AssociationProxy#select with Sequel's #where with block querying API
#eager_load! that raises an error when traversing outside the eagerly loaded data

Hopefully done

Persistence

Efficient saving
- Part one, if it wasn't loaded it wasn't modified, check identity map
- Part two, dirty tracking
Support swapping out DB for in memory datasets

Associations

Eager loading

Querying

Querying API, what would a repository with some arbitrary queries look like?
- e.g. an association on post called burger_comments that finds comments with the word burger in them

Configuration

Automatic config generation based on schema, foreign keys etc
Config to take either a classes or callable factory

Not happening (at least for now)

Associations

Read only associations
- Loaded objects would be immutable
- Collection proxy would have no #push or #remove
- Skipped when dumping
Associations defined with a join
Composable associations