Testing and Debugging - Aspect Build

This exercise uses pytest to add a simple test for the new library mylib you added. Create a new file mylib/say_test.py containing the basic syntax for a Python test:

from mylib import say

def test_saying():
    assert say.marvin("test") == "intentionally wrong"

Now you need to update the dependency graph, so run aspect gazelle. The mylib/BUILD file now has a py_test target named mylib_test which you can test with bazel test //mylib:mylib_test. Of course it fails, since the assertion was “intentionally wrong” but the function returns None - that’s good to know the test is running correctly! You can replace that with None to see the test pass.

from mylib import say

def test_saying():
    assert say.marvin("test") == None

Debugging

To debug this test, you want it to run correctly in the editor, which doesn’t require Bazel at all. You can use a virtualenv again. Find that venv target again:

% bazel query --output=label_kind mylib:all

py_library rule //mylib:__test__
_py_pytest_main rule //mylib:__test___template
determine_main rule //mylib:_mylib_test.find_main
_jq_rule rule //mylib:data
_run_binary rule //mylib:make_header
py_library rule //mylib:mylib
py_test rule //mylib:mylib_test
py_venv_rule rule //mylib:mylib_test.venv
pip_compile rule //mylib:pip_compile
alias rule //mylib:pip_compile.update
pip_compile_test rule //mylib:pip_compile_test
filegroup rule //mylib:requirements

//mylib:mylib_test.venv will create a virtualenv specific to this test target, with the right set of dependencies included. bazel run that target - the editor should offer to activate the new environment. You can open a new terminal in the editor which uses the new virtualenv and see that you can directly run the Bazel-managed pytest. That’s great since it means all tooling should just work here, including debuggers. For example, an engineer can just run pytest and forget about Bazel:

For Visual Studio Code, the steps to debug are:

open the “Testing” view, click “Configure Python tests” and choose “pytest”, then choose the “mylib” folder.

Set a breakpoint, maybe at line 11 of mylib/say.py which you know the test will call.
Choose the “debug test” icon for the file, and the IDE stops at the breakpoint:

​Debugging

Debugging