graphql-python/graphene-sqlalchemy

Connect to multiple databases in a single schema

開放

#180 建立於 2018年12月20日

 (1 則留言) (3 個反應) (0 位負責人)Python (222 個分叉)batch import
enhancementhelp wanted

倉庫指標

星標
 (939 顆星)
PR 合併指標
 (30 天內沒有已合併 PR)

描述

I'm trying to use multiple SQL databases in a single GraphQL schema. Unfortunately I haven't figured out a good way to do this.

I know I can use the context argument when calling the execute method, as documented in the graphene docs:

# connect to database sources
metadata = create_engine('sqlite:///metadata.sqlite3')
users = create_engine('sqlite:///users.sqlite3')
session1 = scoped_session(sessionmaker(bind=metadata))
session2 = scoped_session(sessionmaker(bind=users))

# execute query
result = schema.execute('query { allUsers { edges { node { name } } } }', context={'session': session2})

But this doesn't make sense because I'd have to determine which database to use based on the query. I should be able to pass the query in and the schema will handle that.

Flask-SQLAlchemy handles this well by binding the data source name to the SQLAlchemy model. I'd recommend doing something similar, documentation here.

SQLALCHEMY_DATABASE_URI = 'postgres://localhost/main'
SQLALCHEMY_BINDS = {
    'users':        'mysqldb://localhost/users',
    'appmeta':      'sqlite:////path/to/appmeta.db'
}

class User(db.Model):
    __bind_key__ = 'users'  # this binds the User model to the 'users' mysql database
    id = db.Column(db.Integer, primary_key=True)
    username = db.Column(db.String(80), unique=True)

If you don't want to do that, the SQLAlchemyObjectType could also be a good place to map it.

class UserNode(SQLAlchemyObjectType):
    meta:
        model = User
        interfaces = (graphene.relay.Node,)
        bind_key = 'users'

Any thoughts on working this in, or is there already a way to connect to multiple databases in a single schema?

貢獻者指南