pytest-tornado

A py.test plugin providing fixtures and markers to simplify testing of asynchronous tornado applications.

Installation

pip install pytest-tornado

Example

import pytest
import tornado.web

class MainHandler(tornado.web.RequestHandler):
    def get(self):
        self.write("Hello, world")

application = tornado.web.Application([
    (r"/", MainHandler),
])

@pytest.fixture
def app():
    return application

@pytest.mark.gen_test
def test_hello_world(http_client, base_url):
    response = yield http_client.fetch(base_url)
    assert response.code == 200

Running tests

py.test

Fixtures

io_loop: creates an instance of the tornado.ioloop.IOLoop for each test case
http_port: get a port used by the test server
base_url: Get an absolute base url for the test server, for example http://localhost:59828. Can also be used in a test with HTTPS fixture and will then return a corresponding url, for example http://localhost:48372.
http_server: start a tornado HTTP server, you must create an app fixture, which returns the tornado.web.Application to be tested
http_client: get an asynchronous HTTP client

There is also the possibility to test applications with HTTPS. For running a server with HTTPS you need a certificate.

https_port: Get a port used by the test server.
https_server: Start a tornado HTTPS server. You must create an app fixture, which returns the tornado.web.Application to be tested, and an ssl_options fixture which returns the SSL options for the tornado.httpserver.HTTPServer.
https_client: Get an asynchronous HTTP client. In case your test uses an self-signed certificate you can set verify=False on the fetch method.

Show fixtures provided by the plugin:

py.test --fixtures

Markers

A gen_test marker lets you write a coroutine-style tests used with the tornado.gen module:

@pytest.mark.gen_test
def test_tornado(http_client):
    response = yield http_client.fetch('http://www.tornadoweb.org/')
    assert response.code == 200

This marker supports writing tests with async/await syntax as well:

@pytest.mark.gen_test
async def test_tornado(http_client):
    response = await http_client.fetch('http://www.tornadoweb.org/')
    assert response.code == 200

Marked tests will time out after 5 seconds. The timeout can be modified by setting an ASYNC_TEST_TIMEOUT environment variable, --async-test-timeout command line argument or a marker argument.

@pytest.mark.gen_test(timeout=5)
def test_tornado(http_client):
    yield http_client.fetch('http://www.tornadoweb.org/')

The mark can also receive a run_sync flag, which if turned off will, instead of running the test synchronously, will add it as a coroutine and run the IOLoop (until the timeout). For instance, this allows to test things on both a client and a server at the same time.

@pytest.mark.gen_test(run_sync=False)
def test_tornado(http_server, http_client):
    response = yield http_client.fetch('http://localhost:5555/my_local_server_test/')
    assert response.body == 'Run on the same IOLoop!'

Show markers provided by the plugin: