A graph based queryable object persistance framework built on top of Neo4j.
In addition to objects, Kaiso also stores the class information in the graph. This allows us to use cypher to query instance information, but also to answer questions about our types.
Let's define some basic classes
from kaiso.attributes import Integer, Outgoing, String, Uuid
from kaiso.types import Entity, Relationship
# define a simple type hierarchy
class Knows(Relationship):
pass
class Animal(Entity):
id = Uuid(unique=True)
name = String()
knows = Outgoing(Knows)
class Carnivore(Animal):
pass
class Herbivore(Animal):
pass
class Penguin(Herbivore):
favourite_ice_cream = String()
class Lion(Carnivore):
n_siblings = Integer()As with any orm, we can make some instances and persist them in our graph
from kaiso.persistence import Manager
manager = Manager("http://localhost:7474/db/data/")
# create some instances
fred = Penguin(name="Fred")
tom = Lion(name="Tom")
relation = Knows(fred, tom)
manager.save(fred)
manager.save(tom)
manager.save(relation)Using the Neo4j web interface to explore our graph, we find Tom and Fred:
However, in addition, we can also see the type information in the graph:
We can make use of the type information in our queries, e.g. to find all herbivores who know a carnivore
MATCH
(Herbivore:PersistableType {id: "Herbivore"}),
(Carnivore:PersistableType {id: "Carnivore"}),
Carnivore <-[:ISA*]-()<-[:INSTANCEOF]-(carnivore),
Herbivore <-[:ISA*]-()<-[:INSTANCEOF]-(herbivore),
(herbivore)-[:KNOWS]->(carnivore)
RETURN
"The herbivore",
herbivore.name,
"knows the carnivore",
carnivore.name;==> +---------------------------------------------------------------------+
==> | "The herbivore" | "Fred" | "knows the carnivore" | "Tom" |
==> +---------------------------------------------------------------------+