aiorethink

aiorethink is a fairly comprehensive but easy-to-use asyncio-enabled Object Document Mapper for RethinkDB. It is currently in development.

Documentation: http://aiorethink.readthedocs.org (very early stages)

Source: https://github.com/lars-tiede/aiorethink

Simple example

import aiorethink as ar

class Hero(ar.Document):
    name = ar.Field(ar.StringValueType(), indexed = True)

That's all you need to start out with your own documents. More than that, actually: declaring and "typing" fields is entirely optional.

Obviously, you need a RethinkDB instance running, and you need a database including the tables for your Document classes. aiorethink can't help you with the RethinkDB instance, but the DB setup can be done like so (assuming a RethinkDB instance runs on localhost):

ar.configure_db_connection(db = "my_db")
await ar.init_app_db()

Let's make a document:

spiderman = Hero(name = "Spiderma")

# declared fields can be accessed by attribute or dict interface
spiderman.name = "Spierman"
spiderman["name"] = "Spiderman" # third time's the charm

# with the dict interface, we can make and access undeclared fields
spiderman["nickname"] = "Spidey"

Validate and save to DB:

try:
    await spiderman.save() # calls spiderman.validate()
except ar.ValidationError as e:
    print("validation failed, doc not saved: {}".format(e))

# if we don't declare a primary key field, RethinkDB makes us an 'id' field
doc_id = spiderman.id

Load a document from the DB:

spidey = Hero.load(doc_id) # using primary key

spidey = Hero.from_query(  # using arbitrary query
    Hero.cq(). # "class query prefix", basically rethinkdb.table("Heros")
        get_all("Spiderman", index = "name").nth(0)
)

Iterate over a document's RethinkDB changefeed:

async for spidey, changed_keys, change_msg in await spidey.aiter_changes():
    if "name" in changed_keys:
        print("what, a typo again? {}?".format(spidey.name))

    # change_msg is straight from the rethinkdb changes() query

Features

The following features are either fully or partially implemented:

• optional schema: declare fields in Document classes and get serialization and validation magic much like you know it from other ODMs / ORMs. Or don't declare fields and just use them with the dictionary interface. Or use a mix of declared and undeclared fields.
• schema for complex fields such as lists, dicts, or even "sub-documents" with named and typed fields just like documents.
• dict interface that works for both declared and undeclared fields.
• all I/O is is asynchronous, done with async def / await style coroutines, using asyncio.
• lazy-loading and caching (i.e. "awaitable" fields), for example references to other documents.
• asynchronous changefeeds using async for, on documents and document classes. aiorethink can also assist with Python object creation on just about any other changefeed.

Planned features:

• maybe explicit relations between document classes (think "has_many" etc.)
• maybe schema migrations

Philosophy

aiorethink aims to do the following two things very well:

• make translations between database documents and Python objects easy and convenient
• help with schema and validation

Other than that, aiorethink tries not to hide RethinkDB under a too thick abstraction layer. RethinkDB's excellent Python driver, and certainly its awesome query language, are never far removed and always easy to access. Custom queries on document objects should be easy. Getting document objects out of vanilla rethinkdb queries, including changefeeds, should also be easy.

Status

aiorethink is in development. The API is not complete and not stable yet, although the most important features are present now.