Wallace

Wallace is a Python API for modeling data. Use it to construct, manage, and manipulate domain models for the PostgreSQL (psycopg), Redis (redispy), and MongoDB (pymongo) database adaptors. It was designed to sit in the space between direct database access and industrial strength ORMs, without abstracting away performance considerations or making assumptions about underlying data storage.

Please note: version 0.9.* is a breaking change, freeze 0.0.9 in your pip reqs file if your code relies on it. wallace==0.0.9

Basic SQL Example

Start using Wallace by configuring at least one database connection:

>>> from wallace.config import App
>>>
>>> app = App()
>>> app.add_postgres_connection(<dbname>, <host>, <port>, name="my_pg_conn")

As an additional step for Postgres, a representation of the table is required. (Likewise for Mongo collections.)

>>> from wallace import PostgresTable
>>>
>>> class UserTable(PostgresTable):
>>>
>>>     db_name = "my_pg_conn"
>>>     table_name = "user"

Postgres models require at least one attribute be labeled as a primary key field with pk=True. Note that the table defined above is tied in as well; calls to the database will be proxied through it.

>>> from wallace import PostgresModel
>>> from wallace import Integer, String
>>>
>>> class User(PostgresModel):
>>>
>>>     table = UserTable
>>>
>>>     first_name = String()
>>>     last_name = String()
>>>     email = String(pk=True)  # primary key
>>>     age = Integer()

Use construct to create a new model instance and save to persist it:

>>> user = User.construct(
>>>     first_name="john", last_name="cleese",
>>>     email="[email protected]", age=50)
>>>
>>> user.save()

Retrieve an existing row with fetch, passing in the primary key fields:

>>> user = User.fetch(email="[email protected]")
>>> user.first_name, user.last_name
('john', 'cleese')

Update, find, delete:

>>> user = User.fetch(email="[email protected]")
>>> user.age += 1
>>> user.save()
>>>
>>> [u.email for u in User.find_all(first_name="john")]
['[email protected]']
>>>
>>> user.delete()

Consistency

The same connection registration, type descriptors, etc. are used for all the database drivers wrapped by Wallace. Compare a Redis model:

>>> import time
>>> import uuid
>>>
>>> from wallace import ExpiringRedisHash
>>> from wallace import Integer, Moment, Now, UUID
>>> from wallace.config import get_app
>>>
>>> app = get_app()
>>> app.add_redis_connection("0.0.0.0", port=6379, name="my_redis_conn")
>>>
>>> class WebSession(ExpiringRedisHash):
>>>
>>>     db_name = "my_redis_conn"
>>>     ttl = 60*60
>>>
>>>     session_id = UUID(key=True, default=lambda: uuid.uuid4())
>>>     created_at = Now()
>>>     last_authed_at = Moment()
>>>     user_id = Integer(default=None)
>>>
>>>     def login(self, user_id):
>>>         self.user_id = user_id
>>>         self.last_authed_at = int(time.time())
>>>         self.save()

Create a custom type

Wallace "types" need not map directly to Python primitives. Build new ones ad hoc, particularly to improve readability:

>>> from wallace import RedisHash, Boolean, String
>>>
>>> suits = ("hearts", "spades", "diamonds", "clubs")
>>>
>>>
>>> def validate_cardrank(cardrank):
>>>     if cardrank.isdigit():
>>>         cardrank = int(cardrank)
>>>         return cardrank > 1 and cardrank < 10
>>>     return cardrank in ("J", "Q", "K", "A")
>>>
>>>
>>> class CardRank(String):
>>>
>>>     default = None
>>>     validators = (validate_cardrank,)
>>>
>>>
>>> class PlayingCard(RedisHash):
>>>
>>>     # validators can also be passed directly into the attribute
>>>     suit = String(validators=( lambda val: val in suits, ))
>>>     rank = CardRank()
>>>     is_joker = Boolean()
>>>
>>>     @property
>>>     def key(self):
>>>         return "joker" if self.is_joker else "{}-of-{}".format(self.rank, self.suit)

Download and Install

The latest stable release is always on PyPI. pip install wallace

Enjoy your data.