Send responses to HTTPX using pytest

[!NOTE]
Version 1.0.0 will be released once httpx is considered as stable (release of 1.0.0).

However, current state can be considered as stable.

Once installed, httpx_mock pytest fixture will make sure every httpx request will be replied to with user provided responses (unless some hosts are explicitly skipped).

• Add responses
  • JSON body
  • Custom body
  • Multipart body (files, ...)
  • HTTP status code
  • HTTP headers
  • HTTP/2.0
• Add dynamic responses
• Raising exceptions
• Check requests
• Configuration
  • Register more responses than requested
  • Register less responses than requested
  • Allow to register a response for more than one request
  • Do not mock some requests
• Migrating
  • responses
  • aioresponses

Add responses

You can register responses for both sync and async HTTPX requests.

import pytest
import httpx


def test_something(httpx_mock):
    httpx_mock.add_response()

    with httpx.Client() as client:
        response = client.get("https://test_url")


@pytest.mark.asyncio
async def test_something_async(httpx_mock):
    httpx_mock.add_response()

    async with httpx.AsyncClient() as client:
        response = await client.get("https://test_url")

If all registered responses are not sent back during test execution, the test case will fail at teardown (unless you turned assert_all_responses_were_requested option off).

Default response is a HTTP/1.1 200 (OK) without any body.

How response is selected

In case more than one response match request, the first one not yet sent (according to the registration order) will be sent.

In case all matching responses have been sent once, the request will not be considered as matched (unless you turned can_send_already_matched_responses option on).

You can add criteria so that response will be sent only in case of a more specific matching.

Matching on URL

url parameter can either be a string, a python re.Pattern instance or a httpx.URL instance.

Matching is performed on the full URL, query parameters included.

Order of parameters in the query string does not matter, however order of values do matter if the same parameter is provided more than once.

import httpx
import re
from pytest_httpx import HTTPXMock


def test_url(httpx_mock: HTTPXMock):
    httpx_mock.add_response(url="https://test_url?a=1&b=2")

    with httpx.Client() as client:
        response1 = client.delete("https://test_url?a=1&b=2")
        response2 = client.get("https://test_url?b=2&a=1")


def test_url_as_pattern(httpx_mock: HTTPXMock):
    httpx_mock.add_response(url=re.compile(".*test.*"))

    with httpx.Client() as client:
        response = client.get("https://test_url")


def test_url_as_httpx_url(httpx_mock: HTTPXMock):
    httpx_mock.add_response(url=httpx.URL("https://test_url", params={"a": "1", "b": "2"}))

    with httpx.Client() as client:
        response = client.get("https://test_url?a=1&b=2")

Matching on HTTP method

Use method parameter to specify the HTTP method (POST, PUT, DELETE, PATCH, HEAD) to reply to.

method parameter must be a string. It will be upper-cased, so it can be provided lower cased.

Matching is performed on equality.

import httpx
from pytest_httpx import HTTPXMock


def test_post(httpx_mock: HTTPXMock):
    httpx_mock.add_response(method="POST")

    with httpx.Client() as client:
        response = client.post("https://test_url")


def test_put(httpx_mock: HTTPXMock):
    httpx_mock.add_response(method="PUT")

    with httpx.Client() as client:
        response = client.put("https://test_url")


def test_delete(httpx_mock: HTTPXMock):
    httpx_mock.add_response(method="DELETE")

    with httpx.Client() as client:
        response = client.delete("https://test_url")


def test_patch(httpx_mock: HTTPXMock):
    httpx_mock.add_response(method="PATCH")

    with httpx.Client() as client:
        response = client.patch("https://test_url")


def test_head(httpx_mock: HTTPXMock):
    httpx_mock.add_response(method="HEAD")

    with httpx.Client() as client:
        response = client.head("https://test_url")
    

Matching on proxy URL

proxy_url parameter can either be a string, a python re.Pattern instance or a httpx.URL instance.

Matching is performed on the full proxy URL, query parameters included.

Order of parameters in the query string does not matter, however order of values do matter if the same parameter is provided more than once.

import httpx
from pytest_httpx import HTTPXMock


def test_proxy_url(httpx_mock: HTTPXMock):
    httpx_mock.add_response(proxy_url="http://test_proxy_url?b=1&a=2")

    with httpx.Client(proxy="http://test_proxy_url?a=2&b=1") as client:
        response = client.get("https://test_url")

Matching on HTTP headers

Use match_headers parameter to specify the HTTP headers (as a dict) to reply to.

Matching is performed on equality for each provided header.

import httpx
from pytest_httpx import HTTPXMock


def test_headers_matching(httpx_mock: HTTPXMock):
    httpx_mock.add_response(match_headers={'User-Agent': 'python-httpx/0.25.0'})

    with httpx.Client() as client:
        response = client.get("https://test_url")

Matching on HTTP body

Use match_content parameter to specify the full HTTP body (as bytes) to reply to.

Matching is performed on equality.

import httpx
from pytest_httpx import HTTPXMock


def test_content_matching(httpx_mock: HTTPXMock):
    httpx_mock.add_response(match_content=b"This is the body")

    with httpx.Client() as client:
        response = client.post("https://test_url", content=b"This is the body")

Matching on HTTP JSON body

Use match_json parameter to specify the JSON decoded HTTP body to reply to.

Matching is performed on equality. You can however use unittest.mock.ANY to do partial matching.

import httpx
from pytest_httpx import HTTPXMock
from unittest.mock import ANY

def test_json_matching(httpx_mock: HTTPXMock):
    httpx_mock.add_response(match_json={"a": "json", "b": 2})

    with httpx.Client() as client:
        response = client.post("https://test_url", json={"a": "json", "b": 2})

        
def test_partial_json_matching(httpx_mock: HTTPXMock):
    httpx_mock.add_response(match_json={"a": "json", "b": ANY})

    with httpx.Client() as client:
        response = client.post("https://test_url", json={"a": "json", "b": 2})

Note that match_content or match_files cannot be provided if match_json is also provided.

Matching on HTTP multipart body

Use match_files and match_data parameters to specify the full multipart body to reply to.

Matching is performed on equality.

import httpx
from pytest_httpx import HTTPXMock

def test_multipart_matching(httpx_mock: HTTPXMock):
    httpx_mock.add_response(match_files={"name": ("file_name", b"File content")}, match_data={"field": "value"})

    with httpx.Client() as client:
        response = client.post("https://test_url", files={"name": ("file_name", b"File content")}, data={"field": "value"})

Note that match_content or match_json cannot be provided if match_files is also provided.

Matching on extensions

Use match_extensions parameter to specify the extensions (as a dict) to reply to.

Matching is performed on equality for each provided extension.

import httpx
from pytest_httpx import HTTPXMock


def test_extensions_matching(httpx_mock: HTTPXMock):
    httpx_mock.add_response(match_extensions={'test': 'value'})

    with httpx.Client() as client:
        response = client.get("https://test_url", extensions={"test": "value"})

Matching on HTTP timeout(s)

Use match_extensions parameter to specify the timeouts (as a dict) to reply to.

Matching is performed on the full timeout dict equality.

import httpx
from pytest_httpx import HTTPXMock


def test_timeout_matching(httpx_mock: HTTPXMock):
    httpx_mock.add_response(match_extensions={'timeout': {'connect': 10, 'read': 10, 'write': 10, 'pool': 10}})

    with httpx.Client() as client:
        response = client.get("https://test_url", timeout=10)

Add JSON response

Use json parameter to add a JSON response using python values.

import httpx
from pytest_httpx import HTTPXMock


def test_json(httpx_mock: HTTPXMock):
    httpx_mock.add_response(json=[{"key1": "value1", "key2": "value2"}])

    with httpx.Client() as client:
        assert client.get("https://test_url").json() == [{"key1": "value1", "key2": "value2"}]
    

Note that the content-type header will be set to application/json by default in the response.

Reply with custom body

Use text parameter to reply with a custom body by providing UTF-8 encoded string.

import httpx
from pytest_httpx import HTTPXMock


def test_str_body(httpx_mock: HTTPXMock):
    httpx_mock.add_response(text="This is my UTF-8 content")

    with httpx.Client() as client:
        assert client.get("https://test_url").text == "This is my UTF-8 content"

Use content parameter to reply with a custom body by providing bytes.

import httpx
from pytest_httpx import HTTPXMock


def test_bytes_body(httpx_mock: HTTPXMock):
    httpx_mock.add_response(content=b"This is my bytes content")

    with httpx.Client() as client:
        assert client.get("https://test_url").content == b"This is my bytes content"
    

Use html parameter to reply with a custom body by providing UTF-8 encoded string.

import httpx
from pytest_httpx import HTTPXMock


def test_html_body(httpx_mock: HTTPXMock):
    httpx_mock.add_response(html="<body>This is <p> HTML content</body>")

    with httpx.Client() as client:
        assert client.get("https://test_url").text == "<body>This is <p> HTML content</body>"

Reply by streaming chunks

Use stream parameter (as httpx.SyncByteStream or httpx.AsyncByteStream) to stream chunks that you specify.

Note that pytest_httpx.IteratorStream can be used to provide an iterable.

import httpx
import pytest
from pytest_httpx import HTTPXMock, IteratorStream

def test_sync_streaming(httpx_mock: HTTPXMock):
    httpx_mock.add_response(stream=IteratorStream([b"part 1", b"part 2"]))

    with httpx.Client() as client:
        with client.stream(method="GET", url="https://test_url") as response:
            assert list(response.iter_raw()) == [b"part 1", b"part 2"]


@pytest.mark.asyncio
async def test_async_streaming(httpx_mock: HTTPXMock):
    httpx_mock.add_response(stream=IteratorStream([b"part 1", b"part 2"]))

    async with httpx.AsyncClient() as client:
        async with client.stream(method="GET", url="https://test_url") as response:
            assert [part async for part in response.aiter_raw()] == [b"part 1", b"part 2"]
    

Add multipart response

Use the httpx MultipartStream via the stream parameter to send a multipart response.

Reach out to httpx developers if you need this publicly exposed as this is not a standard use case.

import httpx
from httpx._multipart import MultipartStream
from pytest_httpx import HTTPXMock


def test_multipart_body(httpx_mock: HTTPXMock):
    httpx_mock.add_response(stream=MultipartStream(data={"key1": "value1"}, files={"file1": b"content of file 1"}, boundary=b"2256d3a36d2a61a1eba35a22bee5c74a"))

    with httpx.Client() as client:
        assert client.get("https://test_url").text == '''--2256d3a36d2a61a1eba35a22bee5c74a\r
Content-Disposition: form-data; name="key1"\r
\r
value1\r
--2256d3a36d2a61a1eba35a22bee5c74a\r
Content-Disposition: form-data; name="file1"; filename="upload"\r
Content-Type: application/octet-stream\r
\r
content of file 1\r
--2256d3a36d2a61a1eba35a22bee5c74a--\r
'''
    

Add non 200 response

Use status_code parameter to specify the HTTP status code (as an int) of the response.

import httpx
from pytest_httpx import HTTPXMock


def test_status_code(httpx_mock: HTTPXMock):
    httpx_mock.add_response(status_code=404)

    with httpx.Client() as client:
        assert client.get("https://test_url").status_code == 404

Reply with custom headers

Use headers parameter to specify the extra headers of the response.

Any valid httpx headers type is supported, you can submit headers as a dict (str or bytes), a list of 2-tuples (str or bytes) or a httpx.Header instance.

import httpx
from pytest_httpx import HTTPXMock


def test_headers_as_str_dict(httpx_mock: HTTPXMock):
    httpx_mock.add_response(headers={"X-Header1": "Test value"})

    with httpx.Client() as client:
        assert client.get("https://test_url").headers["x-header1"] == "Test value"


def test_headers_as_str_tuple_list(httpx_mock: HTTPXMock):
    httpx_mock.add_response(headers=[("X-Header1", "Test value")])

    with httpx.Client() as client:
        assert client.get("https://test_url").headers["x-header1"] == "Test value"


def test_headers_as_httpx_headers(httpx_mock: HTTPXMock):
    httpx_mock.add_response(headers=httpx.Headers({b"X-Header1": b"Test value"}))

    with httpx.Client() as client:
        assert client.get("https://test_url").headers["x-header1"] == "Test value"

Reply with cookies

Cookies are sent in the set-cookie HTTP header.

You can then send cookies in the response by setting the set-cookie header with the value following key=value format.

import httpx
from pytest_httpx import HTTPXMock


def test_cookie(httpx_mock: HTTPXMock):
    httpx_mock.add_response(headers={"set-cookie": "key=value"})

    with httpx.Client() as client:
        response = client.get("https://test_url")
    assert dict(response.cookies) == {"key": "value"}


def test_cookies(httpx_mock: HTTPXMock):
    httpx_mock.add_response(headers=[("set-cookie", "key=value"), ("set-cookie", "key2=value2")])

    with httpx.Client() as client:
        response = client.get("https://test_url")
    assert dict(response.cookies) == {"key": "value", "key2": "value2"}

Add HTTP/2.0 response

Use http_version parameter to specify the HTTP protocol version (as a string) of the response.

import httpx
from pytest_httpx import HTTPXMock


def test_http_version(httpx_mock: HTTPXMock):
    httpx_mock.add_response(http_version="HTTP/2.0")

    with httpx.Client() as client:
        assert client.get("https://test_url").http_version == "HTTP/2.0"

Add callbacks

You can perform custom manipulation upon request reception by registering callbacks.

Callback should expect one parameter, the received httpx.Request.

If all callbacks are not executed during test execution, the test case will fail at teardown (unless you turned assert_all_responses_were_requested option off).

Note that callbacks are considered as responses, and thus are selected the same way. Meaning that you can transpose httpx_mock.add_response calls in the related examples into httpx_mock.add_callback.

Dynamic responses

Callback should return a httpx.Response instance.

import httpx
from pytest_httpx import HTTPXMock


def test_dynamic_response(httpx_mock: HTTPXMock):
    def custom_response(request: httpx.Request):
        return httpx.Response(
            status_code=200, json={"url": str(request.url)},
        )

    httpx_mock.add_callback(custom_response)

    with httpx.Client() as client:
        response = client.get("https://test_url")
        assert response.json() == {"url": "https://test_url"}

Alternatively, callbacks can also be asynchronous.

As in the following sample simulating network latency on some responses only.

import asyncio
import httpx
import pytest
from pytest_httpx import HTTPXMock


@pytest.mark.asyncio
async def test_dynamic_async_response(httpx_mock: HTTPXMock):
    async def simulate_network_latency(request: httpx.Request):
        await asyncio.sleep(1)
        return httpx.Response(
            status_code=200, json={"url": str(request.url)},
        )

    httpx_mock.add_callback(simulate_network_latency)
    httpx_mock.add_response()

    async with httpx.AsyncClient() as client:
        responses = await asyncio.gather(
            # Response will be received after one second
            client.get("https://test_url"),
            # Response will instantly be received (1 second before the first request)
            client.get("https://test_url")
        )

Raising exceptions

You can simulate HTTPX exception throwing by raising an exception in your callback or use httpx_mock.add_exception with the exception instance.

This can be useful if you want to assert that your code handles HTTPX exceptions properly.

import httpx
import pytest
from pytest_httpx import HTTPXMock


def test_exception_raising(httpx_mock: HTTPXMock):
    httpx_mock.add_exception(httpx.ReadTimeout("Unable to read within timeout"))
    
    with httpx.Client() as client:
        with pytest.raises(httpx.ReadTimeout):
            client.get("https://test_url")

In case no response can be found

The default behavior is to instantly raise a httpx.TimeoutException in case no matching response can be found.

The exception message will display the request and every registered responses to help you identify any possible mismatch.

import httpx
import pytest
from pytest_httpx import HTTPXMock


@pytest.mark.httpx_mock(assert_all_requests_were_expected=False)
def test_timeout(httpx_mock: HTTPXMock):
    with httpx.Client() as client:
        with pytest.raises(httpx.TimeoutException):
            client.get("https://test_url")

Check sent requests

The best way to ensure the content of your requests is still to use the match_headers and / or match_content parameters when adding a response. In the same spirit, ensuring that no request was issued does not necessarily require any code (unless you turned assert_all_requests_were_expected option off).

In any case, you always have the ability to retrieve the requests that were issued.

As in the following samples:

import httpx
from pytest_httpx import HTTPXMock


def test_many_requests(httpx_mock: HTTPXMock):
    httpx_mock.add_response()

    with httpx.Client() as client:
        response1 = client.get("https://test_url")
        response2 = client.get("https://test_url")

    requests = httpx_mock.get_requests()


def test_single_request(httpx_mock: HTTPXMock):
    httpx_mock.add_response()

    with httpx.Client() as client:
        response = client.get("https://test_url")

    request = httpx_mock.get_request()


def test_no_request(httpx_mock: HTTPXMock):
    assert not httpx_mock.get_request()

How requests are selected

You can add criteria so that requests will be returned only in case of a more specific matching.

Note that requests are selected the same way as responses. Meaning that you can transpose httpx_mock.add_response calls in the related examples into httpx_mock.get_requests or httpx_mock.get_request.

Configuring httpx_mock

The httpx_mock marker is available and can be used to change the default behavior of the httpx_mock fixture.

Refer to available options for an exhaustive list of options that can be set per test, per module or even on the whole test suite.

Per test

import pytest

@pytest.mark.httpx_mock(assert_all_responses_were_requested=False)
def test_something(httpx_mock):
    ...

Per module

import pytest

pytestmark = pytest.mark.httpx_mock(assert_all_responses_were_requested=False)

For the whole test suite

This should be set in the root conftest.py file.

import pytest

def pytest_collection_modifyitems(session, config, items):
    for item in items:
        item.add_marker(pytest.mark.httpx_mock(assert_all_responses_were_requested=False))

[!IMPORTANT]
Note that there currently is a bug in pytest where pytest_collection_modifyitems will actually add the marker AFTER its module and class registration.

Meaning the order is currently: module -> class -> test suite -> test

instead of: test suite -> module -> class -> test

Available options

Allow to register more responses than what will be requested

By default, pytest-httpx will ensure that every response was requested during test execution.

If you want to add an optional response, you can use the is_optional parameter when registering a response or a callback.

def test_fewer_requests_than_expected(httpx_mock):
    # Even if this response never received a corresponding request, the test will not fail at teardown
    httpx_mock.add_response(is_optional=True)

If you don't have control over the response registration process (shared fixtures), and you want to allow fewer requests than what you registered responses for, you can use the httpx_mock marker assert_all_responses_were_requested option.

[!CAUTION] Use this option at your own risk of not spotting regression (requests not sent) in your code base!

import pytest

@pytest.mark.httpx_mock(assert_all_responses_were_requested=False)
def test_fewer_requests_than_expected(httpx_mock):
    # Even if this response never received a corresponding request, the test will not fail at teardown
    httpx_mock.add_response()

Note that the is_optional parameter will take precedence over the assert_all_responses_were_requested option. Meaning you can still register a response that will be checked for execution at teardown even if assert_all_responses_were_requested was set to False.

import pytest

@pytest.mark.httpx_mock(assert_all_responses_were_requested=False)
def test_force_expected_request(httpx_mock):
    # Even if the assert_all_responses_were_requested option is set, the test will fail at teardown if this is not matched
    httpx_mock.add_response(is_optional=False)

Allow to not register responses for every request

By default, pytest-httpx will ensure that every request that was issued was expected.

You can use the httpx_mock marker assert_all_requests_were_expected option to allow more requests than what you registered responses for.

[!CAUTION] Use this option at your own risk of not spotting regression (unexpected requests) in your code base!

import pytest
import httpx

@pytest.mark.httpx_mock(assert_all_requests_were_expected=False)
def test_more_requests_than_expected(httpx_mock):
    with httpx.Client() as client:
        # Even if this request was not expected, the test will not fail at teardown
        with pytest.raises(httpx.TimeoutException):
            client.get("https://test_url")

Allow to register a response for more than one request

By default, pytest-httpx will ensure that every request that was issued was expected.

If you want to add a response once, while allowing it to match more than once, you can use the is_reusable parameter when registering a response or a callback.

import httpx

def test_more_requests_than_responses(httpx_mock):
    httpx_mock.add_response(is_reusable=True)
    with httpx.Client() as client:
        client.get("https://test_url")
        # Even if only one response was registered, the test will not fail at teardown as this request will also be matched
        client.get("https://test_url")

If you don't have control over the response registration process (shared fixtures), and you want to allow multiple requests to match the same registered response, you can use the httpx_mock marker can_send_already_matched_responses option.

With this option, in case all matching responses have been sent at least once, the last one (according to the registration order) will be sent.

[!CAUTION] Use this option at your own risk of not spotting regression (requests issued more than the expected number of times) in your code base!

import pytest
import httpx

@pytest.mark.httpx_mock(can_send_already_matched_responses=True)
def test_more_requests_than_responses(httpx_mock):
    httpx_mock.add_response()
    with httpx.Client() as client:
        client.get("https://test_url")
        # Even if only one response was registered, the test will not fail at teardown as this request will also be matched
        client.get("https://test_url")

Do not mock some requests

By default, pytest-httpx will mock every request.

But, for instance, in case you want to write integration tests with other servers, you might want to let some requests go through.

To do so, you can use the httpx_mock marker should_mock option and provide a callable expecting the httpx.Request as parameter and returning a boolean.

Returning True will ensure that the request is handled by pytest-httpx (mocked), False will let the request pass through (not mocked).

import pytest
import httpx

@pytest.mark.httpx_mock(should_mock=lambda request: request.url.host != "www.my_local_test_host")
def test_partial_mock(httpx_mock):
    httpx_mock.add_response()

    with httpx.Client() as client:
        # This request will NOT be mocked
        response1 = client.get("https://www.my_local_test_host/sub?param=value")
        # This request will be mocked
        response2 = client.get("https://test_url")

Migrating to pytest-httpx

Here is how to migrate from well-known testing libraries to pytest-httpx.

From responses

Add a response or a callback

Undocumented parameters means that they are unchanged between responses and pytest-httpx. Below is a list of parameters that will require a change in your code.

Sample adding a response with responses:

from responses import RequestsMock

def test_response(responses: RequestsMock):
    responses.add(
        method=responses.GET,
        url="https://test_url",
        body=b"This is the response content",
        status=400,
    )

Sample adding the same response with pytest-httpx:

from pytest_httpx import HTTPXMock

def test_response(httpx_mock: HTTPXMock):
    httpx_mock.add_response(
        method="GET",
        url="https://test_url",
        content=b"This is the response content",
        status_code=400,
    )

From aioresponses

Add a response or a callback

Undocumented parameters means that they are unchanged between responses and pytest-httpx. Below is a list of parameters that will require a change in your code.

Sample adding a response with aioresponses:

import pytest
from aioresponses import aioresponses


@pytest.fixture
def mock_aioresponse():
    with aioresponses() as m:
        yield m


def test_response(mock_aioresponse):
    mock_aioresponse.get(
        url="https://test_url",
        body=b"This is the response content",
        status=400,
    )

Sample adding the same response with pytest-httpx:

def test_response(httpx_mock):
    httpx_mock.add_response(
        method="GET",
        url="https://test_url",
        content=b"This is the response content",
        status_code=400,
    )