An experimental framework to test your library against type checkers, allowing
to formulate type expectations and expected typechecker errors. Its purpose is
the same as the one of
pytest-mypy-plugins. While
pytest-mypy-plugins
requires .yaml
files for specifying the tests, typest
test cases are python files, expectations are formulated in comments:
from mylibrary import some_function
result = some_function()
reveal_type(result) # expect-type: int
Besides expressing type expectations, you can also specify to expect an error from the typechecker:
string: str = "not a number"
number: int = string # expect-error
You can also specify to expect a mismatch error, i.e. an error where an assigned type is mismatching the actual type:
string: str = "not a number"
number: int = string # expect-mismatch: int <> str
mypy
pyright
typest
is available at pypi. You can install it through pip:
pip install typest
python -m typest [PATH] [TYPECHECKERS]
If PATH is a directory, all python files under that directory (including subdirectories) are going to be checked. If PATH points to a file, it has to be a python file.
TYPECHECKERS is an optional argument, a comma separated list of names of
typecheckers you want to run your tests against. Currently, mypy
and pyright
are suppported.
You can add more typecheckers by subclassing
typest.typecheckers.base.TypeChecker
and importing your new class in
typest/typecheckers/__init__.py
.