Skip to main content

pytest-cdist

Like pytest-xdist, but for distributed environments - pytest-cdist allows to split a test suite into multiple parts and run them independently.

This is a work in progress

Why?

pytest-xdist can help to parallelize test execution, as long as you can scale horizontally. In many environments, such as GitHub actions with GitHub runners, this is only possible to a fairly limited degree, which can be an issue if your test suite grows large. pytest-cdist can help with this by allowing to execute individual chunks of your test suite in a deterministic order, so you can use multiple concurrent jobs to run each individual chunk.

How?

pytest --cdist-group=1/2  # will run the first half of the test suite
pytest --cdist-group=2/2  # will run the second half of the test suite

In a GitHub workflow

jobs:
  test:
    runs-on: ubuntu-latest
    matrix:
      strategy:
        cdist-groups: [1, 2, 3, 4]

    steps:
      - uses: actions/checkout@v4
      # set up environment here
      - name: Run pytest
        run: pytest --cdist-group=${{ matrix.cdist-group }}/4

Usage

Configuration

Pytest-cdist comes with several CLI and pytest-ini options:

CLI Ini Allowed values Default
--cdist-justify-items cdist-justify-items none, file, scope none
--cdist-group-steal --cdist-group-steal <target group>:<percentage> / <target group>:<percentage>:<source group> -
--cdist-report - - false
--cdist-report-dir cdist-report-dir .

Controlling how items are split up

By default, pytest-cdist will split up the items into groups as evenly as possible. Sometimes this may not be desired, for example if there's some costly fixtures requested by multiple tests, which should ideally only run once.

To solve this, the cdist-justify-items option can be used to configure how items are split up. It can take two possible values:

  • file: Ensure all items inside a file end up in the same group
  • scope: Ensure all items in the same pytest scope end up in the same group
[pytest]
cdist-justify-items=file
pytest --cdist-group=1/2 --cdist-justify-items=file

Skewing the group sizes

Normally, items are distributed evenly among groups, which is a good default, but there may be cases where this will result in an uneven execution time, if one group contains a number of slower tests than the other ones.

To work around this, the cdist-group-steal option can be used. It allows to specific a certain percentage of items a group will "steal" from other groups. For example --cdist-group-steal=2:30 will cause group 2 to steal 30% of items from all other groups.

[pytest]
cdist-group-steal=2:30
pytest --cdist-group=1/2 --cdist-group-steal=2:30

It is also possible to redistribute items between two specific groups, by specifying both as source and a target group. The following configuration would assign 50% of the items in group 1 to group 2:

pytest --cdist-group=1/3 --cdist-group-steal=1:50:2

With pytest-xdist

When running under pytest-xdist, pytest-cdist will honour tests marked with xdist_group, and group them together in the same cdist group.

With pytest-randomly

To use pytest-cdist with pytest-randomly's test reordering, a randomness seed that's consistent across the different pytest invocations needs to be specified, otherwise, test cases would be dropped, since pytest-cdist has to rely on a consistent collection order to split test cases among its groups.

To achieve this, a random source such as the current timestamp can be used:

jobs:
  setup_randomly_seed:
    runs-on: ubuntu-latest
    outputs:
      randomly_seed: ${{ steps.set-seed.outputs.randomly_seed }}
    steps:
      - name: Set seed
        id: set-seed
        run: echo "randomly_seed=$(date +%s)" >> $GITHUB_OUTPUT
        
  test:
    runs-on: ubuntu-latest
    needs: setup_randomly_seed
    matrix:
      strategy:
        cdist-groups: [1, 2, 3, 4]
    steps:
      - uses: actions/checkout@v4
      - name: Run pytest
        run: > 
          pytest 
          --cdist-group=${{ matrix.cdist-group }}/4 
          --randomly-seed=${{ needs.setup_randomly_seeds.outputs.randomly_seed }}

or a bit simpler (but less random), using the current git hash:

jobs:        
  test:
    runs-on: ubuntu-latest
    needs: setup_randomly_seed
    matrix:
      strategy:
        cdist-groups: [1, 2, 3, 4]
    steps:
      - uses: actions/checkout@v4
      - name: Run pytest
        run: > 
          pytest 
          --cdist-group=${{ matrix.cdist-group }}/4 
          --randomly-seed=$((16#$(git rev-parse HEAD)))

Metadata

Release files for pytest-cdist 0.5.2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for pytest-cdist 0.5.2
File Size Uploaded
pytest_cdist-0.5.2.tar.gz 31.0 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for pytest-cdist 0.5.2
File Interpreter ABI Platform
pytest_cdist-0.5.2-py3-none-any.whl Python 3 none any Details

Total release size: 39.3 kB

Release files / pytest_cdist-0.5.2.tar.gz

Download URL pytest_cdist-0.5.2.tar.gz
Size 31.0 kB
Tags Source
SHA-256 checksum
How to use checksums
4fca0c79d73869ba8293121786782f615bfb3e8ffe175428a716d193ef1ec484
BLAKE2b-256 checksum
How to use checksums
55b9bfbfe094302bfabd89a6879cda6c6d9b98b64cd2a9cc55e72b0b6203443d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Jul 7, 2026.

Transparency log

Release files / pytest_cdist-0.5.2-py3-none-any.whl

Download URL pytest_cdist-0.5.2-py3-none-any.whl
Size 8.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
f41c8b99dab83042473e64d43038c15ea2ea40c4589bae639b29dd92bd468f96
BLAKE2b-256 checksum
How to use checksums
393a7fdf321ac0828b60d65f3e7b6de5508a7e7425c06cbc0fb6eeb5fc358ce9
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Jul 7, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.5.2 This release

2 release files

0.5.1

2 release files

0.5.0

2 release files

0.4.0

2 release files

0.3.1

2 release files

0.3.0

2 release files

0.2.0

2 release files

0.1.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page