HTTPretty 0.8.14

ChangeLog

disclaimer:

Due to big API incompatibility between python 3.3, 3.4 and 3.5, the author of HTTPretty is not supporting python3 officially. You will notice that the travis build for python 3 might be broken, and while pull requests fixing py3 support are most welcome, it is still not official at least for now.

Installing

Since you are interested in HTTPretty you should also be interested in speeding up your build. Replace pip with curdling and see your build running a lot faster.

You can use curdling to install not only HTTPretty but every dependency in your project and see the speed gains.

$ easy_install curdling
$ curd install HTTPretty

In a nutshell

Once upon a time a python developer wanted to use a RESTful api, everything was fine but until the day they needed to test the code that hits the RESTful API: what if the API server is down? What if its content has changed ?

Don't worry, HTTPretty is here for you:

import requests
from sure import expect
import httpretty


@httpretty.activate
def test_yipit_api_returning_deals():
    httpretty.register_uri(httpretty.GET, "http://api.yipit.com/v1/deals/",
                           body='[{"title": "Test Deal"}]',
                           content_type="application/json")

    response = requests.get('http://api.yipit.com/v1/deals/')

    expect(response.json()).to.equal([{"title": "Test Deal"}])

A more technical description

HTTPretty is a HTTP client mock library for Python 100% inspired on ruby's FakeWeb. If you come from ruby this would probably sound familiar :)

Usage

expecting a simple response body

import requests
import httpretty

def test_one():
    httpretty.enable()  # enable HTTPretty so that it will monkey patch the socket module
    httpretty.register_uri(httpretty.GET, "http://yipit.com/",
                           body="Find the best daily deals")

    response = requests.get('http://yipit.com')

    assert response.text == "Find the best daily deals"

    httpretty.disable()  # disable afterwards, so that you will have no problems in code that uses that socket module
    httpretty.reset()    # reset HTTPretty state (clean up registered urls and request history)

testing query strings

import requests
from sure import expect
import httpretty

def test_one():
    httpretty.enable()  # enable HTTPretty so that it will monkey patch the socket module
    httpretty.register_uri(httpretty.GET, "http://yipit.com/login",
                           body="Find the best daily deals")

    requests.get('http://yipit.com/login?email=user@github.com&password=foobar123')
    expect(httpretty.last_request()).to.have.property("querystring").being.equal({
        "email": "user@github.com",
        "password": "foobar123",
    })

    httpretty.disable()  # disable afterwards, so that you will have no problems in code that uses that socket module

ohhhh, really? can that be easier?

YES we've got a decorator

import requests
import httpretty

@httpretty.activate
def test_one():
    httpretty.register_uri(httpretty.GET, "http://yipit.com/",
                           body="Find the best daily deals")

    response = requests.get('http://yipit.com')
    assert response.text == "Find the best daily deals"

the @httpretty.activate is a short-hand decorator that wraps the decorated function with httpretty.enable() and then calls httpretty.disable() right after.

mocking the status code

import requests
from sure import expect
import httpretty

@httpretty.activate
def test_github_access():
    httpretty.register_uri(httpretty.GET, "http://github.com/",
                           body="here is the mocked body",
                           status=201)

    response = requests.get('http://github.com')
    expect(response.status_code).to.equal(201)

you can tell HTTPretty to return any HTTP headers you want

and all you need is to add keyword args in which the keys are always lower-cased and with underscores _ instead of dashes -

For example, let's say you want to mock that server returns content-type. To do so, use the argument content_type, all the keyword args are taken by HTTPretty and transformed in the RFC2616 equivalent name.

@httpretty.activate
def test_some_api():
    httpretty.register_uri(httpretty.GET, "http://foo-api.com/gabrielfalcao",
                           body='{"success": false}',
                           status=500,
                           content_type='text/json')

    response = requests.get('http://foo-api.com/gabrielfalcao')

    expect(response.json()).to.equal({'success': False})
    expect(response.status_code).to.equal(500)

Adding extra headers and forcing headers

You can pass the adding_headers argument as a dictionary and your headers will be united to the existing headers.

@httpretty.activate
def test_some_api():
    httpretty.register_uri(httpretty.GET, "http://foo-api.com/gabrielfalcao",
                           body='{"success": false}',
                           status=500,
                           content_type='text/json',
                           adding_headers={
                               'X-foo': 'bar'
                           })

    response = requests.get('http://foo-api.com/gabrielfalcao')

    expect(response.json()).to.equal({'success': False})
    expect(response.status_code).to.equal(500)

Although there are some situation where some headers line content-length will be calculated by HTTPretty based on the specified fake response body.

So you might want to "force" those headers:

@httpretty.activate
def test_some_api():
    httpretty.register_uri(httpretty.GET, "http://foo-api.com/gabrielfalcao",
                           body='{"success": false}',
                           status=500,
                           content_type='text/json',
                           forcing_headers={
                               'content-length': '100'
                           })

    response = requests.get('http://foo-api.com/gabrielfalcao')

    expect(response.json()).to.equal({'success': False})
    expect(response.status_code).to.equal(500)

You should, though, be careful with it. The HTTP client is likely to rely on the content length to know how many bytes of response payload should be loaded. Forcing a content-length that is bigger than the action response body might cause the HTTP client to hang because it is waiting for data. Read more in the "caveats" session on the bottom.

rotating responses

Same URL, same request method, the first request return the first httpretty.Response, all the subsequent ones return the last (status 202).

Notice that the responses argument is a list and you can pass as many responses as you want.

import requests
from sure import expect


@httpretty.activate
def test_rotating_responses():
    httpretty.register_uri(httpretty.GET, "http://github.com/gabrielfalcao/httpretty",
                           responses=[
                               httpretty.Response(body="first response", status=201),
                               httpretty.Response(body='second and last response', status=202),
                            ])

    response1 = requests.get('http://github.com/gabrielfalcao/httpretty')
    expect(response1.status_code).to.equal(201)
    expect(response1.text).to.equal('first response')

    response2 = requests.get('http://github.com/gabrielfalcao/httpretty')
    expect(response2.status_code).to.equal(202)
    expect(response2.text).to.equal('second and last response')

    response3 = requests.get('http://github.com/gabrielfalcao/httpretty')

    expect(response3.status_code).to.equal(202)
    expect(response3.text).to.equal('second and last response')

streaming responses

Mock a streaming response by registering a generator response body.

import requests
from sure import expect
import httpretty

# mock a streaming response body with a generator
def mock_streaming_tweets(tweets):
    from time import sleep
    for t in tweets:
        sleep(.5)
        yield t

@httpretty.activate
def test_twitter_api_integration(now):
    twitter_response_lines = [
        '{"text":"If @BarackObama requests to follow me one more time I\'m calling the police."}\r\n',
        '\r\n',
        '{"text":"Thanks for all your #FollowMe1D requests Directioners! We\u2019ll be following 10 people throughout the day starting NOW. G ..."}\r\n'
    ]

    TWITTER_STREAMING_URL = "https://stream.twitter.com/1/statuses/filter.json"

    # set the body to a generator and set `streaming=True` to mock a streaming response body
    httpretty.register_uri(httpretty.POST, TWITTER_STREAMING_URL,
                           body=mock_streaming_tweets(twitter_response_lines),
                           streaming=True)

    # taken from the requests docs
    # http://docs.python-requests.org/en/latest/user/advanced/#streaming-requests
    response = requests.post(TWITTER_STREAMING_URL, data={'track':'requests'},
                            auth=('username','password'), prefetch=False)

    #test iterating by line
    line_iter = response.iter_lines()
    for i in xrange(len(twitter_response_lines)):
        expect(line_iter.next().strip()).to.equal(twitter_response_lines[i].strip())

dynamic responses through callbacks

Set a callback to allow for dynamic responses based on the request.

import requests
from sure import expect
import httpretty

@httpretty.activate
def test_response_callbacks():

    def request_callback(request, uri, headers):
        return (200, headers, "The {} response from {}".format(request.method, uri))

    httpretty.register_uri(
        httpretty.GET, "https://api.yahoo.com/test",
        body=request_callback)

    response = requests.get('https://api.yahoo.com/test')

    expect(response.text).to.equal('The GET response from https://api.yahoo.com/test')

Dynamic responses can also be used when you have to work with badly designed APIs where, for example, the same uri and method are used to handle different requests based on request body which contains xml.

import requests
import httpretty

@httpretty.activate
def test_response_callbacks():

    def request_callback(request, uri, headers):
        # parse_xml() extracts important data from request
        data = parse_xml(request.body)
        # response based on that data
        if data.something_important:
            return (200, headers, "relevant data")
        else:
            return (400, headers, "panic mode!")

    httpretty.register_uri(
        httpretty.GET, "https://api.brilliant-api.com/",
        body=request_callback)

    response = requests.get('https://api.brilliant-api.com/')

matching regular expressions

You can register a compiled regex and it will be matched against the requested urls.

@httpretty.activate
def test_httpretty_should_allow_registering_regexes():
    u"HTTPretty should allow registering regexes"

    httpretty.register_uri(
        httpretty.GET,
        re.compile("api.yipit.com/v2/deal;brand=(\w+)"),
        body="Found brand",
    )

    response = requests.get('https://api.yipit.com/v2/deal;brand=GAP')
    expect(response.text).to.equal('Found brand')
    expect(httpretty.last_request().method).to.equal('GET')
    expect(httpretty.last_request().path).to.equal('/v1/deal;brand=GAP')

By default, the regexp you register will match the requests without looking at the querystring. If you want the querystring to be considered, you can set match_querystring=True when calling register_uri.

expect for a response, and check the request got by the "server" to make sure it was fine.

import requests
from sure import expect
import httpretty


@httpretty.activate
def test_yipit_api_integration():
    httpretty.register_uri(httpretty.POST, "http://api.yipit.com/foo/",
                           body='{"repositories": ["HTTPretty", "lettuce"]}')

    response = requests.post('http://api.yipit.com/foo',
                            '{"username": "gabrielfalcao"}',
                            headers={
                                'content-type': 'text/json',
                            })

    expect(response.text).to.equal('{"repositories": ["HTTPretty", "lettuce"]}')
    expect(httpretty.last_request().method).to.equal("POST")
    expect(httpretty.last_request().headers['content-type']).to.equal('text/json')

checking whether a request was made or not

import httpretty
import requests

def order_pizza(user, home_delivery=True):
    check_number = make_pizza()
    if home_delivery:
        requests.post('http://api.pizzas.com/deliveries/', {'address': user.address, 'check_number': check_number})
    else:
        # for pick up.
        pass
    return check_number

@httpretty.activate
def test_pizza_delivery():
    httpretty.register_uri(httpretty.POST, 'http://api.pizzas.com/deliveries/', body='OK')

    order_pizza(some_user)
    expect(httpretty.has_request()).to.be.true

    httpretty.reset()
    order_pizza(some_user, home_delivery=False)
    expect(httpretty.has_request()).to.be.false

checking if is enabled

httpretty.enable()
httpretty.is_enabled().should.be.true

httpretty.disable()
httpretty.is_enabled().should.be.false

raising an error if an unregistered endpoint is requested

import urllib2
import httpretty

httpretty.enable()
httpretty.HTTPretty.allow_net_connect = False

httpretty.register_uri(httpretty.GET, 'http://www.google.com', body='OK')

urllib2.urlopen('http://www.google.com')
urllib2.urlopen.when.called_with('http://www.reddit.com').should.have.raised(httpretty.errors.UnmockedError)

Motivation

When building systems that access external resources such as RESTful webservices, XMLRPC or even simple HTTP requests, we stumble in the problem:

"I'm gonna need to mock all those requests"

It brings a lot of hassle, you will need to use a generic mocking tool, mess with scope and so on.

The idea behind HTTPretty (how it works)

HTTPretty monkey patches Python's socket core module, reimplementing the HTTP protocol, by mocking requests and responses.

As for it works in this way, you don't need to worry what http library you're gonna use.

HTTPretty will mock the response for you :) (and also give you the latest requests so that you can check them)

Acknowledgements

caveats with the requests library

forcing_headers + Content-Length

if you use the forcing_headers options make sure to add the header Content-Length otherwise the requests will try to load the response endlessly

Officially supported libraries

Because HTTPretty works in the socket level it should work with any HTTP client libraries, although it is battle tested against:

• requests
• httplib2
• urllib2

Hacking on HTTPretty

create a virtual env

you will need virtualenvwrapper

mkvirtualenv --distribute --no-site-packages HTTPretty

install the dependencies

pip install -r requirements.txt

next steps:

1. run the tests with make:

make unit functional

2. hack at will
3. commit, push etc
4. send a pull request

License

<HTTPretty - HTTP client mock for Python>
Copyright (C) <2011-2015>  Gabriel Falcão <gabriel@nacaolivre.org>

Permission is hereby granted, free of charge, to any person
obtaining a copy of this software and associated documentation
files (the "Software"), to deal in the Software without
restriction, including without limitation the rights to use,
copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following
conditions:

The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
OTHER DEALINGS IN THE SOFTWARE.

Main contributors

There folks made remarkable contributions to HTTPretty:

• Steve Pulec ~> @spulec
• Hugh Saunders ~> @hughsaunders
• Matt Luongo ~> @mhluongo
• James Rowe ~> @JNRowe