Backend and search engine for Nefertari

Next tasks:

• Find a way to specify/use multiple engines at once.

• Need a way to delegate search to es (or other secondary engine?). maybe just call get_collection on the generated es model.

• Need a way to generate an es model or mapping given an sql or mongo model. this shouldn't be too hard, since the field names are mostly the same.

• Try to re-use existing part of the es engine. for example, probably can re-use exiting base classes and fields in order to recreate something like the existing get_es_mapping.

• Will still need hooks in sql and mongo engines to update es models when models are changed. it might make sense to change these hooks to generate pyramid events. then the es engine can listen for these events, thus de-coupling the engines.

• Aggregations

• Multi-collection requests

• Rework es.py script in nefertari to work with nefertari-es (indexing all documents and missing documents)

• Update Ramses to work with multiple engines properly. ramses-example should work with nefertari-es and multiple engines.

• Docstrings

Advanced tasks:

• Unique constraint on PK fields?

• Make nefertari-es work with token auth

• Make nefertari-es work wirh nefertari-guards

• Caching of loaded ES documents

• Make relationships sync work on _update_many/_delete_many

• Implement ondelete/onupdate hooks for relationships