Skip to content
No description, website, or topics provided.
Python
Branch: master
Clone or download
Latest commit 7ddfaeb Jan 3, 2020

README.md

Build Status Coverage Status

graphene-tornado

A project for running Graphene on top of Tornado in Python 2 and 3. The codebase is a port of graphene-django.

Getting started

Create a Tornado application and add the GraphQL handlers:

import tornado.web
from tornado.ioloop import IOLoop

from graphene_tornado.schema import schema
from graphene_tornado.tornado_graphql_handler import TornadoGraphQLHandler


class ExampleApplication(tornado.web.Application):

    def __init__(self):
        handlers = [
            (r'/graphql', TornadoGraphQLHandler, dict(graphiql=True, schema=schema)),
            (r'/graphql/batch', TornadoGraphQLHandler, dict(graphiql=True, schema=schema, batch=True)),
            (r'/graphql/graphiql', TornadoGraphQLHandler, dict(graphiql=True, schema=schema))
        ]
        tornado.web.Application.__init__(self, handlers)

if __name__ == '__main__':
    app = ExampleApplication()
    app.listen(5000)
    IOLoop.instance().start()

When writing your resolvers, decorate them with either Tornado's @coroutine decorator for Python 2.7:

@gen.coroutine
def resolve_foo(self, info):
  foo = yield db.get_foo()
  raise Return(foo)

Or use the async / await pattern in Python 3:

async def resolve_foo(self, info):
  foo = await db.get_foo()
  return foo

Extensions

graphene-tornado supports server-side extensions like Apollo Server. The extensions go a step further than Graphene middleware to allow for finer grained interception of request processing. The canonical use case is for tracing; see graphene_tornado/apollo_engine_reporting/engine_agent.py for an example.

Extensions are experimental and most likely will change in future releases as they should be extensions provided by graphql-server-core.

Apollo Engine Reporting

You can integrate with Apollo Engine Reporting by enabling the extension.

$ pip install graphene-tornado[apollo-engine-reporting]
engine_options = EngineReportingOptions()
agent = EngineReportingAgent(engine_options, generate_schema_hash(schema))


class ExampleEngineReportingApplication(tornado.web.Application):

    def __init__(self):
        engine_extension = lambda: EngineReportingExtension(engine_options, agent.add_trace)
        handlers = [
            (r'/graphql', TornadoGraphQLHandler, dict(graphiql=True, schema=schema, extensions=[engine_extension])),
            (r'/graphql/batch', TornadoGraphQLHandler, dict(graphiql=True, schema=schema, batch=True, 
                                                            extensions=[engine_extension])),
            (r'/graphql/graphiql', TornadoGraphQLHandler, dict(graphiql=True, schema=schema, 
                                                               extensions=[engine_extension]))
        ]
        tornado.web.Application.__init__(self, handlers)
ENGINE_API_KEY=<your engine API key here> python -m examples.apollo_engine_reporting_example

Then visit http://localhost:5000/graphql/graphiql, make some queries, and view the results in Apollo Engine.

OpenCensus

You can also use OpenCensus for tracing:

$ pip install graphene-tornado[opencensus]
class ExampleOpenCensusApplication(tornado.web.Application):

    def __init__(self):
        extension = lambda: OpenCensusExtension()
        handlers = [
            (r'/graphql', TornadoGraphQLHandler, dict(graphiql=True, schema=schema, extensions=[extension])),
            (r'/graphql/batch', TornadoGraphQLHandler, dict(graphiql=True, schema=schema, batch=True)),
        ]
        tornado.web.Application.__init__(self, handlers)
You can’t perform that action at this time.