Skip to main content

Shard tests to support parallelism across multiple machines

Project description

PyPI version

Fork of pytest-shard

This is a fork of pytest-shard that seems to be abandon - it adds some features.

pytest-shard

Shards tests based on a hash of their test name enabling easy parallelism across machines, suitable for a wide variety of continuous integration services. Tests are split at the finest level of granularity, individual test cases, enabling parallelism even if all of your tests are in a single file (or even single parameterized test method).

Features

pytest-shard aims for simplicity. When installed, simply run:

$ pytest --shard-id=I --num-shards=N

where I is the index of this shard and N the total number of shards. For example, to split tests across two machines:

# On machine 1:
$ pytest --shard-id=0 --num-shards=2
# On machine 2:
$ pytest --shard-id=1 --num-shards=2

The intended use case is for continuous integration services that allow you to run jobs in parallel. For CircleCI, enable parallelism and then use:

pytest --shard-id=${CIRCLE_NODE_INDEX} --num-shards=${CIRCLE_NODE_TOTAL}

On Travis, you must define the environment variables explicitly, but can use a similar approach.

Shard by duration

To shard the tests based on their duration, first run all the tests and create a junitxml output:

$ pytest --junitxml=durations.xml

Then you can add --shard-by-duration to your pytest invocations:

$ pytest --shard-id=1 --num-shards=2 --shard-by-duration

Alternatives

pytest-xdist allows you to parallelize tests across cores on a single machine, and can also schedule tests on a remote machine. I use pytest-shard to split tests across CI workers, and pytest-xdist to parallelize across CPU cores within each worker.

pytest-shard does not take into account the run time of tests, which can lead to suboptimal allocations. pytest-circleci-parallelized uses test run time, but can only split at the granularity of classes, and is specific to CircleCI.

Please open a PR if there are other promising alternatives that I have overlooked.

Installation

You can install pytest-shard via pip:

$ pip install pytest-shard

Contributions

Contributions are welcome. Test may be run using tox.

License

This software is MIT licensed.

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_shard_fork-0.1.4.tar.gz (8.1 kB view details)

Uploaded Source

Built Distribution

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

pytest_shard_fork-0.1.4-py3-none-any.whl (6.2 kB view details)

Uploaded Python 3

File details

Details for the file pytest_shard_fork-0.1.4.tar.gz.

File metadata

  • Download URL: pytest_shard_fork-0.1.4.tar.gz
  • Upload date:
  • Size: 8.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.3

File hashes

Hashes for pytest_shard_fork-0.1.4.tar.gz
Algorithm Hash digest
SHA256 c8d7f0199d0dbb3c6bb3749b3c0b31abd4bf483c3c0e585257f9b308723063de
MD5 022db6f9a3acaae881fc961c17c79ed2
BLAKE2b-256 607a4bb7fd2b8d6cebf9853101a29e157eade0a55f83ffc64bcf400d2537051f

See more details on using hashes here.

File details

Details for the file pytest_shard_fork-0.1.4-py3-none-any.whl.

File metadata

File hashes

Hashes for pytest_shard_fork-0.1.4-py3-none-any.whl
Algorithm Hash digest
SHA256 b00f6a56f1db489fa39f6a1f12a69ae7a24c2fcf164124f50d44e749fbbbaef5
MD5 aa14c276047a96275d26f4959547d061
BLAKE2b-256 dffca964865f55262563e5531c0380bd08be4303e198c9cc6ce7338c9a9c5118

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