❓How it works?

pytest-remote-response works on top of interceptors specific for each library.

In a nutshell, interceptors wrap the library calls, thereby attaching additional mock/capture logic.

Note

For some nerds, interceptors uses MonkeyPatch to monkey patch the original library with a wrapped one.

This approach certainly has some benefits and limitations; for instance, it’s surprisingly easy to write new interceptors as it’s only a wrapper method but it’s only applicable when that method is called explicitly; for countering this, pytest-remote-response ships with two more deep interceptors _urllib and _urllib3.

Warning

_urllib and _urllib3 are more low-level but are plagued with threading issues; use them carefully!

🕸️Interceptor

Note

Any custom intercept must have at-least install and uninstall methods and should return a BaseMockResponse

Example

from functools import warps

# Import `response` instance to get access to :class:`~pytest.MonkeyPatch` and
# :class:`~pytest_response.database.ResponseDB`
from pytest_response import response

# Library wrapper
def wrapper(func):
    @wraps(func)
    def inner_func(url, *args, **kwargs):
        # Check if response is `True`, if `True` spoof the request from the database
        if response.response:
            status, data, headers = response.get(url=url)
            return MockResponse(status, data, headers)

        # Actual library call
        lib_response = func(url, *args, **kwargs)

        # Check if capture is `True`, if `True` then extract and dump the response into the database.
        if not response.capture:
            return _

        # Extract data, headers and status from `lib_response`

        response.insert(url=url, response=data, headers=dict(headers), status=_.status)
        return _
    return inner_func

class MockResponse(BaseMockResponse):
    # A basic Mock Response for spoofing data, headers and status
    def __init__(self, status, data, headers={}):
        super().__init__(status, data, headers)
    pass


def install():
    # A basic install method for Monkey Patching the library
    lib_call = library.call
    wrapped_lib_call = urlopen_wrapper(lib_call)
    response.mpatch.setattr("library.call", wrapped_lib_call)
    return


def uninstall():
    # A basic uninstall method
    response.mpatch.undo()

📁Database

pytest-remote-response has ResponseDB to facilitate talking to the database. Internally, its actually a wrapper for sqlite3.

Primary communication methods are insert() and get().

Note

Some fields such as headers and data are compressed using zlib to reduce the over-all 🦶footprint.

Database fields
Fields	Dumped as
url	`str` – `base64.b64encode()` serialized
cache_date	`str`
status	`str`
headers	`str` – `base64.b64encode()` serialized after `zlib.compress()`
response	`str` – `base64.b64encode()` serialized after `zlib.compress()`

Example

# Import the database interface :class:`pytest_response.database.ResponseDB`
from base64 import b64encode

from pytest_response.database import ResponseDB

# Setup the database
db = ResponseDB(path="database.json")

# Insert new element
url = "https://www.python.org"
status = 200
headers = {}
data = b""
db.insert(url=url, status=status, response=data, headers=headers)

# Verify the index
assert b64encode(url.encode()).decode() in db.index()

# Query for an URL
url = "https://www.python.org"
get_status, get_data, get_headers = db.get(url)

# Verify the data is unchanged
assert status == get_status
assert data == get_data
assert headers == get_headers