Data layer for the Bowery Framework.
There are a few intracies that I would like to abstract out.
- Query vs. Scan
- We provide _id to them and they can only find by id.
- Behind the scenes we've randomly chosen a field to be the rangeKey.
- Eventually Consistent vs Consistent
- There are only 6 types, we should do the casting behind the scenes
- 1MB Query size limit