mocket /mɔˈkɛt/

A socket mock framework

for all kinds of socket animals, web-clients included - with gevent/asyncio/SSL support

...and then MicroPython's urequests (mocket >= 3.9.1)

Outside GitHub

Mocket packages are available for Arch Linux, openSUSE, NixOS, ALT Linux, NetBSD, and of course from PyPI.

Versioning

Starting from 3.7.0, Mocket major version will follow the same numbering pattern as Python's and therefore indicate the most recent Python version that is supported.

FYI: the last version compatible with Python 2.7 is 3.9.4, bugfixing or backporting of features introduced after that release will only be available as commercial support.

Support it

Star the project on GitHub, Buy Me a Coffee clicking the button below or, even better, contribute with patches or documentation.

How to use it

Read the following blog posts if you want to have a big picture of what Mocket is capable of:

• https://medium.com/p/mocket-is-alive-and-is-fighting-with-us-b2810d52597a
• https://hackernoon.com/make-development-great-again-faab769d264e
• https://hackernoon.com/httpretty-now-supports-asyncio-e310814704c6
• https://medium.com/@mindflayer/how-to-make-your-tests-fail-when-they-try-to-access-the-network-python-eb80090a6d24
• https://medium.com/@mindflayer/testing-in-an-asyncio-world-a9a0ad41b0c5

The starting point to understand how to use Mocket to write a custom mock is the following example:

• https://github.com/mindflayer/mocketoy

As next step, you are invited to have a look at the implementation of both the mocks it provides:

• HTTP mock (similar to HTTPretty) - https://github.com/mindflayer/python-mocket/blob/master/mocket/mockhttp.py
• Redis mock (basic implementation) - https://github.com/mindflayer/python-mocket/blob/master/mocket/mockredis.py

Please also have a look at the huge test suite:

• Tests module at https://github.com/mindflayer/python-mocket/tree/master/tests

Installation

Using pip:

$ pip install mocket

Speedups

Mocket uses xxhash when available instead of hashlib.md5 for creating hashes, you can install it as follows:

$ pip install mocket[speedups]

Issues

When opening an Issue, please add few lines of code as failing test, or -better- open its relative Pull request adding this test to our test suite.

Example of how to mock an HTTP[S] call

Let's create a new virtualenv with all we need:

$ python3 -m venv example
$ source example/bin/activate
$ pip install pytest requests mocket

As second step, we create an example.py file as the following one:

import json

from mocket import mocketize
from mocket.mockhttp import Entry
import requests
import pytest


@pytest.fixture
def response():
    return {
        "integer": 1,
        "string": "asd",
        "boolean": False,
    }


@mocketize  # use its decorator
def test_json(response):
    url_to_mock = 'https://testme.org/json'

    Entry.single_register(
        Entry.GET,
        url_to_mock,
        body=json.dumps(response),
        headers={'content-type': 'application/json'}
    )

    mocked_response = requests.get(url_to_mock).json()

    assert response == mocked_response

# OR use its context manager
from mocket import Mocketizer

def test_json_with_context_manager(response):
    url_to_mock = 'https://testme.org/json'

    Entry.single_register(
        Entry.GET,
        url_to_mock,
        body=json.dumps(response),
        headers={'content-type': 'application/json'}
    )

    with Mocketizer():
        mocked_response = requests.get(url_to_mock).json()

    assert response == mocked_response

Let's fire our example test:

$ py.test example.py

How to make Mocket fail when it tries to write to a real socket?

NEW!!! Sometimes you just want your tests to fail when they attempt to use the network.

with Mocketizer(strict_mode=True):
    with pytest.raises(StrictMocketException):
        requests.get("https://duckduckgo.com/")

# OR

@mocketize(strict_mode=True)
def test_get():
    with pytest.raises(StrictMocketException):
        requests.get("https://duckduckgo.com/")

You can specify exceptions as a list of hosts or host-port pairs.

with Mocketizer(strict_mode=True, strict_mode_allowed=["localhost", ("intake.ourmetrics.net", 443)]):
    ...

# OR

@mocketize(strict_mode=True, strict_mode_allowed=["localhost", ("intake.ourmetrics.net", 443)])
def test_get():
    ...

How to be sure that all the Entry instances have been served?

Add this instruction at the end of the test execution:

Mocket.assert_fail_if_entries_not_served()

Example of how to fake socket errors

It's very important that we test non-happy paths.

@mocketize
def test_raise_exception(self):
    url = "http://github.com/fluidicon.png"
    Entry.single_register(Entry.GET, url, exception=socket.error())
    with self.assertRaises(requests.exceptions.ConnectionError):
        requests.get(url)

Example of how to record real socket traffic

You probably know what VCRpy is capable of, that's the mocket's way of achieving it:

@mocketize(truesocket_recording_dir=tempfile.mkdtemp())
def test_truesendall_with_recording_https():
    url = 'https://httpbin.org/ip'

    requests.get(url, headers={"Accept": "application/json"})
    resp = requests.get(url, headers={"Accept": "application/json"})
    assert resp.status_code == 200

    dump_filename = os.path.join(
        Mocket.get_truesocket_recording_dir(),
        Mocket.get_namespace() + '.json',
    )
    with io.open(dump_filename) as f:
        response = json.load(f)

    assert len(response['httpbin.org']['443'].keys()) == 1

HTTPretty compatibility layer

Mocket HTTP mock can work as HTTPretty replacement for many different use cases. Two main features are missing:

• URL entries containing regular expressions;
• response body from functions (used mostly to fake errors, mocket doesn't need to do it this way).

Two features which are against the Zen of Python, at least imho (mindflayer), but of course I am open to call it into question.

Example:

import json

import aiohttp
import asyncio
from unittest import TestCase

from mocket.plugins.httpretty import httpretty, httprettified


class AioHttpEntryTestCase(TestCase):
    @httprettified
    def test_https_session(self):
        url = 'https://httpbin.org/ip'
        httpretty.register_uri(
            httpretty.GET,
            url,
            body=json.dumps(dict(origin='127.0.0.1')),
        )

        async def main(l):
            async with aiohttp.ClientSession(
                loop=l, timeout=aiohttp.ClientTimeout(total=3)
            ) as session:
                async with session.get(url) as get_response:
                    assert get_response.status == 200
                    assert await get_response.text() == '{"origin": "127.0.0.1"}'

        loop = asyncio.new_event_loop()
        loop.set_debug(True)
        loop.run_until_complete(main(loop))

What about the other socket animals?

Using Mocket with asyncio based clients:

$ pip install aiohttp

Example:

# `aiohttp` creates SSLContext instances at import-time
# that's why Mocket would get stuck when dealing with HTTP
# Importing the module while Mocket is in control (inside a
# decorated test function or using its context manager would
# be enough for making it work), the alternative is using a
# custom TCPConnector which always return a FakeSSLContext
# from Mocket like this example is showing.
import aiohttp
import pytest

from mocket import async_mocketize
from mocket.mockhttp import Entry
from mocket.plugins.aiohttp_connector import MocketTCPConnector


@pytest.mark.asyncio
@async_mocketize
async def test_aiohttp():
    """
    The alternative to using the custom `connector` would be importing
    `aiohttp` when Mocket is already in control (inside the decorated test).
    """

    url = "https://bar.foo/"
    data = {"message": "Hello"}

    Entry.single_register(
        Entry.GET,
        url,
        body=json.dumps(data),
        headers={"content-type": "application/json"},
    )

    async with aiohttp.ClientSession(
        timeout=aiohttp.ClientTimeout(total=3), connector=MocketTCPConnector()
    ) as session, session.get(url) as response:
        response = await response.json()
        assert response == data

Works well with others

Using Mocket as pook engine:

$ pip install mocket[pook]

Example:

import pook
from mocket.plugins.pook_mock_engine import MocketEngine

pook.set_mock_engine(MocketEngine)

pook.on()

url = 'http://twitter.com/api/1/foobar'
status = 404
response_json = {'error': 'foo'}

mock = pook.get(
    url,
    headers={'content-type': 'application/json'},
    reply=status,
    response_json=response_json,
)
mock.persist()

requests.get(url)
assert mock.calls == 1

resp = requests.get(url)
assert resp.status_code == status
assert resp.json() == response_json
assert mock.calls == 2

First appearance

EuroPython 2013, Florence

• Video: https://www.youtube.com/watch?v=-LvXbl5d02U
• Slides: https://prezi.com/tmuiaugamsti/mocket/
• Slides as PDF: https://ep2013.europython.eu/media/conference/slides/mocket-a-socket-mock-framework.pdf