This directory contains several tests:

tests/mypy_test.py tests the stubs with mypy
tests/pytype_test.py tests the stubs with pytype.
tests/pyright_test.py tests the stubs with pyright.
tests/regr_test.py runs mypy against the test cases for typeshed's stubs, guarding against accidental regressions.
tests/check_consistent.py checks certain files in typeshed remain consistent with each other.
tests/stubtest_stdlib.py checks standard library stubs against the objects at runtime.
tests/stubtest_third_party.py checks third-party stubs against the objects at runtime.
tests/typecheck_typeshed.py runs mypy against typeshed's own code in the tests and scripts directories.

To run the tests, follow the setup instructions in the CONTRIBUTING.md document. In particular, you have to run with Python 3.9+.

In order for pytype_test and pyright_test to work correctly, some third-party stubs may require extra dependencies external to typeshed to be installed in your virtual environment prior to running the test. You can list or install all of a stubs package's external dependencies using the following script:

(.venv3)$ python tests/get_external_stub_requirements.py <third_party_stub>  # List external dependencies for <third_party_stub>
(.venv3)$ python tests/get_external_stub_requirements.py <third_party_stub1> <third_party_stub2>  # List external dependencies for <third_party_stub1> and <third_party_stub2>
(.venv3)$ python tests/get_external_stub_requirements.py  # List external dependencies for all third-party stubs in typeshed
# Install external dependencies for all third-party stubs in typeshed
(.venv3)$ DEPENDENCIES=$(python tests/get_external_stub_requirements.py)
(.venv3)$ if [ -n "$DEPENDENCIES" ]; then pip install $DEPENDENCIES; fi

Run all tests for a specific stub

Run using:

(.venv3)$ python3 scripts/runtests.py <stdlib-or-stubs>/<stub-to-test>

This script will run all tests below for a specific typeshed directory. If a test supports multiple python versions, the oldest supported by typeshed will be selected. A summary of the results will be printed to the terminal.

You must provide a single argument which is a path to the stubs to test, like so: stdlib/os or stubs/requests.

Run python scripts/runtests.py --help for information on the various configuration options for this script.

mypy_test.py

Run using:

(.venv3)$ python3 tests/mypy_test.py

The test has two parts: running mypy on the stdlib stubs, and running mypy on the third-party stubs.

This test is shallow — it verifies that all stubs can be imported but doesn't check whether stubs match their implementation (in the Python standard library or a third-party package).

Run python tests/mypy_test.py --help for information on the various configuration options for this script.

pytype_test.py

Note: this test cannot be run on Windows systems unless you are using Windows Subsystem for Linux.

Run using:

(.venv3)$ python3 tests/pytype_test.py

This test works similarly to mypy_test.py, except it uses pytype.

pyright_test.py

This test requires Node.js to be installed. Although typeshed runs pyright in CI, it does not currently use this script. However, this script uses the same pyright version and configuration as the CI.

(.venv3)$ python3 tests/pyright_test.py                                # Check all files
(.venv3)$ python3 tests/pyright_test.py stdlib/sys.pyi                 # Check one file
(.venv3)$ python3 tests/pyright_test.py -p pyrightconfig.stricter.json # Check with the stricter config.

pyrightconfig.stricter.json is a stricter configuration that enables additional checks that would typically fail on incomplete stubs (such as Unknown checks). In typeshed's CI, pyright is run with these configuration settings on a subset of the stubs in typeshed (including the standard library).

regr_test.py

This test runs mypy against the test cases for typeshed's stdlib and third-party stubs. See the README in the test_cases directory for more information about what these test cases are for and how they work. Run python tests/regr_test.py --help for information on the various configuration options.

check_consistent.py

Run using:

$ python3 tests/check_consistent.py

stubtest_stdlib.py

Run using

(.venv3)$ python3 tests/stubtest_stdlib.py

This test compares the stdlib stubs against the objects at runtime. Because of this, the output depends on which version of Python and on what kind of system it is run. As such, if you run this test locally, it may complain about system-specific differences (in e.g, socket) that the type system cannot capture or our stubtest settings in CI do not account for. If you run into this issue, consider opening a draft PR and letting CI test it automatically (or running the test via Github Actions on your typeshed fork).

If you need a specific version of Python to repro a CI failure, pyenv can also help.

Due to its dynamic nature, you may run into false positives. In this case, you can add to the allowlists for each affected Python version in tests/stubtest_allowlists. Please file issues for stubtest false positives at mypy.

stubtest_third_party.py

⚠️ This script downloads and executes arbitrary code from PyPI. Only run this script locally if you know you can trust the packages you're running stubtest on.

Run using

(.venv3)$ python3 tests/stubtest_third_party.py

Similar to stubtest_stdlib.py, but tests the third party stubs. By default, it checks all third-party stubs, but you can provide the distributions to check on the command line:

(.venv3)$ python3 tests/stubtest_third_party.py Pillow toml  # check stubs/Pillow and stubs/toml

If you have the runtime package installed in your local virtual environment, you can also run stubtest directly, with

(.venv3)$ MYPYPATH=<path-to-module-stubs> python3 -m mypy.stubtest \
  --custom-typeshed-dir <path-to-typeshed> \
  <third-party-module>

For each distribution, stubtest ignores definitions listed in a @tests/stubtest_allowlist.txt file, relative to the distribution. Additional packages that are needed to run stubtest for a distribution can be added to tool.stubtest.stubtest_requirements in METADATA.toml.

Using stubtest to find objects missing from the stubs

By default, stubtest emits an error if a public object is present at runtime but missing from the stub. However, this behaviour can be disabled using the --ignore-missing-stub option.

If a distribution has ignore_missing_stub = true in the [tool.stubtest] section of its tests/METADATA.toml file, stubtest_third_party.py will test that distribution with the --ignore-missing-stub option. This indicates that the stubs for this distribution are considered "incomplete".

You can help make typeshed's stubs more complete by removing ignore_missing_stub = true from the tests/METADATA.toml file for a third-party stubs distribution, running stubtest, and then adding things that stubtest reports to be missing to the stub. However, note that not everything that stubtest reports to be missing should necessarily be added to the stub. For some implementation details, it is often better to add allowlist entries for missing objects rather than trying to match the runtime in every detail.

typecheck_typeshed.py

Run using

(.venv3)$ python3 tests/typecheck_typeshed.py

This is a small wrapper script that uses mypy to typecheck typeshed's own code in the scripts and tests directories. Run python tests/typecheck_typeshed.py --help for information on the various configuration options.

Uh oh!