pytest-asyncio-httpx-mocking

Awaiting a mocked httpx AsyncClient method raises a TypeError because standard MagicMock returns a non-awaitable object. You must use unittest.mock.AsyncMock for async methods like get and post to ensure proper coroutine behavior and avoid await errors.

To patch httpx AsyncClient post and get methods in pytest with AsyncMock, apply unittest.mock.patch to the specific methods. This ensures the mocked methods return awaitable responses, preserving the coroutine context needed for realistic asynchronous behavior during test execution.

The difference between MagicMock and AsyncMock when testing asyncio httpx calls is that AsyncMock produces awaitable coroutines for async methods, while MagicMock returns non-awaitable sync objects. AsyncMock is required to prevent TypeError when awaiting mocked httpx responses in pytest.

To correctly wrap return values for AsyncMock in pytest httpx tests, configure the return_value or side_effect on the AsyncMock object. This ensures the awaited response behaves as expected, allowing you to test data fetchers that process httpx AsyncClient responses without flakiness.

Yes, you can use unittest.mock.patch to test an async data fetcher awaiting httpx responses without flakiness. By patching httpx.AsyncClient methods with AsyncMock, you prevent non-awaitable return values and maintain realistic asynchronous behavior throughout your pytest test scenarios.