Python SqlalChemy同时过滤父对象和子对象

我得到了这个多对多的设计

class Parent(Base):
    __tablename__ = "parent"
    id = Column(Integer, primary_key=True)
    parent_number = Column(String)
    start_date = Column(DateTime, primary_key=True)
    end_date = Column(DateTime)
    children= relationship("Child", lazy="selectin")


class Child(Base):
    __tablename__ = "child"
    id = Column(Integer, primary_key=True)
    start_date = Column(DateTime, primary_key=True)
    end_date = Column(DateTime)
    parent_id = Column(Integer, ForeignKey("parent.id"))

这两个模型都有一个start_date和end_date来注册每一行的有效日期范围，从而产生一个复合主键.

我想要做的是能够构造满足start_date和end_date上的条件的Parent对象以及子array.

我try 了一个简单的查询，比如

session.query(Parent)
.filter(Parent.parent_number == '1234')
.filter(Parent.start_date > given_date & Parent.end_date < given_date)

这会导致两个查询

SELECT *
FROM Parent
WHERE parent_number = '1234' AND start_date > '2024-02-14' AND end_date < '2024-02-14'
 
SELECT Parent.id, Parent.start_date, Child.*
FROM Parent
JOIN Child ON Child.parent_id = Parent.id
WHERE (Parent.id, Parent.start_date) IN ((id_of_previous_parent, start_date_of_previous_parent))

第一个查询很好，但对于第二个查询，我希望类似于

SELECT Parent.id, Parent.start_date, Child.*
FROM Parent
JOIN Child ON Child.parent_id = Parent.id
WHERE (Parent.id, Parent.start_date) IN ((id_of_previous_parent, start_date_of_previous_parent)) AND Child.start_date > '2024-02-14' AND Child.end_date < '2024-02-14'

我试过这个天真的问题

session.query(Parent)
.filter(Parent.parent_number == '1234')
.filter(Parent.start_date > given_date & Parent.end_date < given_date)
.filter(Child.start_date > given_date & Child.end_date < given_date)

但它只改变了第一个查询.

我还try 加入如下查询

session.query(Parent)
.join(Child)
.filter(Parent.parent_number == '1234')
.filter(Parent.start_date > given_date & Parent.end_date < given_date)
.filter(Child.start_date > given_date & Child.end_date < given_date)

没有结果.