Skip to main content

Select a portion of the collected tests

Project description

pytest-portion

CI PyPI version Python versions Changelog docs

Select a portion of the collected tests, so you can run different parts of your test suite in different instances to scale horizontally.

Use case

Suppose you have a big, slow test suite, but you can trigger several CI workers to run different portions of it, in a sake lazy/simple way to parallelize it.

A basic, obvious way to do that is to explictily collect from different directories/modules:

  • worker1: pytest tests/a (100 tests, ~4 minutes to finish)
  • worker2: pytest tests/b (20 tests, ~1 minute to finish)
  • worker3: pytest tests/c tests/d (30 tests, ~1 minute to finish)

The problem is that directory tests/a may have a lot more tests that tests/c plus test/d, so worker1 takes a lot more to finish.

With pytest-portion you can still split the tests in different instances, but letting the extension makes the selection in a more balanced way.

  • worker1: pytest --portion 1/3 tests (first 50 tests, ~2 minutes)
  • worker2: pytest --portion 2/3 tests (next 50 tests, ~2 minutes)
  • worker3: pytest --portion 3/3 tests (last 50 tests, ~2 minutes)

In this case, the tests of all the directories are collected, but only a third (a different one!) of them will be actually executed on each worker.

Note this balance is by number of tests, so if there is very slow tests in a particular portion, the duration may not be expected.

For a fine tuning, you could pass the portion in a more explicit way:

  • worker1: pytest --portion 0:0.5 tests (first half, 1st to 75th test)
  • worker2: pytest --portion 0.5:0.8 tests (next 30%, from 76th to 125th)
  • worker3: pytest --portion 0.8:1 tests (last 20%)

Installation

You can add "pytest-portion" to your project from PyPI with uv.

uv add --dev pytetest-portion

Or via pip

pip install pytest-portion

Usage

There are three modes of operation: Test-level (default), File-level, and Function-level.

1. Test-level Slicing (Default)

Pytest collects all tests first, then pytest-portion filters them.

Pass --portion <i/n> where:

  • n is the total number of portions.
  • i is the i-th portion to select (1 <= i <= n).

Note: If the number of tests collected is not divisible by n, the last portion will contain the rest. For instance, if you have test_1, test_2 and test_3, --portion 1/2 will run the first one, and --portion 2/2 the last 2.

Alternatively, use --portion start:end where start and end are coefficients between 0 and 1.

2. File-level Slicing

For very large projects, collection itself can be slow. Use --portion-files to slice the list of discovered files before pytest starts collecting tests from within them. This can significantly reduce collection time in large repositories.

# Collect and run only the files belonging to the first half of the suite
pytest --portion 1/2 --portion-files tests/

3. Function-level Slicing

With multiple parametrized test functions (e.g. test_compare1[...], test_compare2[...]), the default splits the entire collected list. That can put all of one function in portion 1 and all of another in portion 2, so workers get uneven loads.

Use --portion-functions to split within each test function instead: each function’s cases are portioned separately, so every function contributes proportionally to the selected set.

# Half from test_compare1 and half from test_compare2
pytest --portion 1/2 --portion-functions tests/

Note: --portion-files and --portion-functions cannot be used together.

Contributing

Contributions are very welcome. Please ensure the coverage at least stays the same before you submit a pull request.

License

Distributed under the terms of the BSD-3 license, "pytest-portion" is free and open source software.

Issues

If you encounter any problems, please file an issue along with a detailed description.

Acknowledgements

I used cookiecutter along with @hackebrot's cookiecutter-pytest-plugin template for the boilerplate code of this package. Thanks!

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pytest_portion-0.3.0.tar.gz (60.6 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

pytest_portion-0.3.0-py3-none-any.whl (7.6 kB view details)

Uploaded Python 3

File details

Details for the file pytest_portion-0.3.0.tar.gz.

File metadata

  • Download URL: pytest_portion-0.3.0.tar.gz
  • Upload date:
  • Size: 60.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.10.8 {"installer":{"name":"uv","version":"0.10.8","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for pytest_portion-0.3.0.tar.gz
Algorithm Hash digest
SHA256 a1e7e8bd01fa8146bb33faff2030075546ee3be131c8573b4ad0aaaff43f2001
MD5 3895493b7736d4f43c7732347af5f3e5
BLAKE2b-256 202c7635034c170f1379ce47ab69a56cc2bd7f3d930c80ae1210be92eea5bcb7

See more details on using hashes here.

File details

Details for the file pytest_portion-0.3.0-py3-none-any.whl.

File metadata

  • Download URL: pytest_portion-0.3.0-py3-none-any.whl
  • Upload date:
  • Size: 7.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: uv/0.10.8 {"installer":{"name":"uv","version":"0.10.8","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for pytest_portion-0.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 020fb36dce7f0f52f5765cacec7bac34759492aece67f55fa3a50d4e110192b6
MD5 3880d63f0a9697e01a2014c13f043f25
BLAKE2b-256 98f41b2e017a5b6fd499451235f4c9eddc5aaf357e39949ddd7f336aec2546a8

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page