我试图模拟SQLAlchemy中机构参与者之间的等级和历史关系(即机构可以有父母/子女和前任/后继者)。到目前为止,我在SQLAlchemy文档中遵循directed graph example。现在,我希望能够访问字典中节点的左/右邻居,其中edge_type
为关键字节点,节点列表为值,如下所示:node.right_nodes['edge_type']
。
我认为这可以通过collection_class来完成,但使用collection_class=attribute_mapped_collection('edge_type')
只会产生一个键:值对而不是键:[值列表]。
实际结果:
>>> node.right_edges['edge_type']
<Edge object>
>>> node.right_nodes['edge_type']
<Node object>
预期结果:
>>> node.right_edges['edge_type']
[<Edge object>, <Edge object>]
>>> node.right_nodes['edge_type']
[<Node object>, <Node object>]
该模型如下所示:
from sqlalchemy import (Column, Integer, String, ForeignKey,
create_engine)
from sqlalchemy.orm import Session, relationship, backref
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.ext.associationproxy import association_proxy
from sqlalchemy.orm.collections import attribute_mapped_collection
Base = declarative_base()
class Node(Base):
__tablename__ = 'node'
id = Column(Integer, primary_key=True)
name = Column(String, nullable=False)
left_nodes = association_proxy('left_edges', 'left_node')
right_nodes = association_proxy('right_edges', 'right_node')
class Edge(Base):
__tablename__ = 'edge'
left_node_id = Column(Integer, ForeignKey('node.id'), primary_key=True)
right_node_id = Column(Integer, ForeignKey('node.id'), primary_key=True)
edge_type = Column(String)
left_node = relationship(
Node,
foreign_keys=left_node_id,
backref=backref(
'right_edges',
collection_class=attribute_mapped_collection('edge_type')
)
)
right_node = relationship(
Node,
foreign_keys=right_node_id,
backref=backref(
'left_edges',
collection_class=attribute_mapped_collection('edge_type')
)
)
像这样使用:
engine = create_engine('sqlite://', echo=True)
Base.metadata.create_all(engine)
session = Session(engine)
n1 = Node(name='LeftNode')
n2 = Node(name='RightNode1')
n3 = Node(name='RightNode2')
Edge(left_node=n1, right_node=n2, edge_type='hierarchy')
Edge(left_node=n1, right_node=n3, edge_type='hierarchy')
session.add_all([n1, n2, n3])
session.commit()
print(n1.right_nodes) # returns dict with 1 node as value
print(n1.right_nodes['hierarchy']) # returns 1 node
print(n1.right_edges) # returns dict with 1 edge as value
print(n1.right_edges['hierarchy']) # returns 1 edge
print(session.query(Edge).filter_by(left_node=n1).all()) # returns list of 2 edges
修改
以下不是我的问题的答案,而是记录我到目前为止的内容。
association_proxy
不仅将字段作为目标,还可以在目标类中定义属性:
class Node(Base):
# <snip>
left_nodes = association_proxy('left_edges', 'left_nodes_edge_type')
right_nodes = association_proxy('right_edges', 'right_nodes_edge_type')
class Edge(Base):
# <snip>
@property
def left_nodes_edge_type(self):
return {self.edge_type: self.left_node}
@property
def right_nodes_edge_type(self):
return {self.edge_type: self.left_node}
这当然不会产生所需的列表字典,而是列表dict:
>>> node.right_nodes
[{'edge_type': <Node object>}, {'edge_type': <Node object>}]
您也可以简单地在Node
类本身定义属性,而不是使用association_proxy
来代理Node
到Node
:
class Node(Base):
# <snip>
@property
def left_nodes(self):
d = defaultdict(list)
for edge in self.left_edges:
d[edge.edge_type].append(edge.left_node)
return d
@property
def right_nodes(self):
d = defaultdict(list)
for edge in self.right_edges:
d[edge.edge_type].append(edge.right_node)
return d
这允许将所需视图作为列表的字典:
>>> node.right_nodes
{'edge_type': [<Node object>, <Node object>]}
>>> node.right_nodes['edge_type']
[<Node object>, <Node object>]
但这只是一个方便的视图。