Add chunking to serializing models into Store

[bjester]

Overview

Serializing huge amounts of data into the Store could cause large increases in memory usage. Currently _serialize_into_store processes batches by model, but loads all of the model's data into memory before saving the bulk writes to the Store model.

Deliverables

• Relocate and split up _serialize_into_store into smaller functions
• Have the serialization process the model objects in chunks, or via a queryset iterator() (ref)