Repository navigation
feat: roll Python support to 3.11-3.15 and refresh dependencies - #241
Merged
Merged
Conversation
Keep pandas 2 on existing Python versions and use native pandas 3 defaults on Python 3.15 to avoid forcing disruptive NbExec image upgrades. Record that policy in ADR-001 and introduce lightweight decision documentation and Copilot guidance. Regenerate the full dependency lock to resolve known AnyIO advisories, update test tooling for compatibility, and exercise the optional Arrow extra in serialized integration CI. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
rbell517
commented
Oct 29, 2026
Move the decision index before lifecycle guidance, explain pandas-major-specific timestamp expectations, and keep internal decision and deployment context out of public user documentation. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
cameronwaterman
approved these changes
Oct 29, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What does this Pull Request accomplish?
>=3.11,<4.0).max-parallel: 1to avoid shared-service collisions.Why should this Pull Request be merged?
Python 3.10 is end of life; the supported stable CPython window is now 3.11-3.15. This keeps the library current, removes known vulnerable lock resolutions, and makes the compatibility and development policies discoverable.
Compatibility notice: Python 3.10 is no longer supported. Moving to Python 3.15 also adopts native pandas 3 inference, including possible string dtype, timestamp-unit, and missing-value representation differences. Existing Python 3.11-3.14 environments retain pandas 2; no NbExec image upgrades are part of this PR.
Administrator follow-up after merge: the maintainer has removed the obsolete required Python 3.10 check. The existing required Python 3.11/3.12/3.13 and documentation checks remain; the maintainer will add 3.14/3.15 as required checks after this PR merges. This PR does not change repository settings or automatically merge.
What testing has been done?
Local validation used CPython 3.15.2 and Poetry 2.2.1:
poetry lock --regeneratepoetry check --lockpoetry installpoetry install --extras pyarrowpoetry run pytest tests\product\test_product_dataframe_utilities.py tests\spec\test_spec_dataframe_utilitites.py tests\testmonitor\test_testmonitor_dataframe_utilities.py -q --tb=short --disable-warningspoetry run poe test -q --tb=short --disable-warningspoetry run poe checkpoetry run poe lintpoetry run poe typespoetry buildgit diff --checkAdditional assertions passed for the wheel's exact Python floor/classifiers, Python-specific pandas/Arrow dependency markers, Arrow remaining optional, both CI matrices, integration serialization/extra installation, and ADR metadata/index/local links.
The complete lock was enumerated into fully pinned audit inputs without environment markers, splitting alternate versions of a package into separate inputs. Both final
pip-auditcommands below passed. Coverage was checked against the final lock: all 60 package/version pairs, including main/dev/optional/platform/Python branches, zero known vulnerabilities and zero skipped packages. The baseline covered 61 pairs and found only the two AnyIO advisories above. Inputs/reports stay outside the repository.PR CI completed successfully in run 40928868327:
gh run watch 40928868327 --repo ni/nisystemlink-clients-python --interval 20 --exit-status- passed.build-lint-unit-testjobs (Python 3.11-3.15) - passed.integration-testjobs (Python 3.11-3.15), with the Arrow extra installed - passed. Each ran 249 passing tests, 52 skipped, 402 deselected. Completed job logs confirm that the DataFrame integration suite ran.The workflow's cloud API key argument was empty, so this is not evidence of credentialed Cloud coverage; the repository exposes an Enterprise API key but no repository-level
CLOUD_API_KEY. Existing service/credential skips are reported above rather than counted as tested. Local skips likewise are not counted as integration verification.Documentation/release tooling outside the Poetry lock is not represented as covered by the lock audit. That run completed with all emitted validation checks green. Normal review is still required; the obsolete required Python 3.10 check has now been removed.
Review follow-up in ed77cae moves the ADR index below its introduction, explains the pandas-major-specific timestamp expectations, and removes internal decision/deployment references from public user documentation. All three review threads have been replied to and resolved. The 23 targeted DataFrame tests, targeted Black/Flake8 checks, documentation assertions, and git diff --check passed locally. The new CI run is 40931290298; its results are pending.