Conversation
The cap on the interpreter was the only thing keeping 3.14 out. No application code changes are needed: the suite passes unchanged on 3.14.7 with the locked dependency versions (1780 passed, 6 skipped), and the tree uses nothing that has been removed since 3.11 -- no distutils, pkg_resources, imp, utcnow(), or bare get_event_loop(). Two dev tools did need fixing, both for the same reason: they touch ast.Str, which no longer exists on 3.12+. black 22.12 (what the `^22.6.0` dev pin resolved to) aborts with "cannot format: module 'ast' has no attribute 'Str'" on five files, and flake8-implicit-str-concat 0.4.0 takes down the whole flake8 run from its multiprocessing pool. The latter is the pre-commit hook CI actually runs, so a 3.14 matrix entry would have gone red at the Linting step before reaching the tests. flake8-implicit-str-concat is capped at >=0.6.0, the first release that works on 3.14. 0.7.0 would also run, but it adds an ISC004 check that flags eight existing implicit concatenations in saude/databases.py, which is out of scope here. The lock moves only black, flake8, pycodestyle and pyflakes. Every runtime dependency keeps its resolved version; numpy already carried a python_version == "3.11" split and needs no new marker.
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #371 +/- ##
=======================================
Coverage ? 97.20%
=======================================
Files ? 180
Lines ? 23286
Branches ? 0
=======================================
Hits ? 22635
Misses ? 651
Partials ? 0 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
The previous commit used `>=0.6.0`, which resolves to 0.7.0. That version works on 3.14 too, but it adds an ISC004 check that flags eight existing implicit concatenations in api/saude/databases.py. Confirmed by running the lint under 3.11, 3.12, 3.13 and 3.14. Pin it exactly, matching the style of every other plugin in this file, so the lint result cannot shift underneath us the next time pre-commit autoupdate runs. 0.6.0 is the oldest release that does not crash on 3.14, and it reports no findings the newer ones would flag.
- Fix trailing/extra blank lines and comment indentation in `pysus/api/metadata/columns.py` (pre-existing whitespace from AlertaDengue#348) - Wrap long endpoint strings in `pysus/api/saude/databases.py` to satisfy flake8 E501 while keeping values identical and ISC004 clean. Also validates with black 24.2.
ISC004 findings were resolved (endpoint strings wrapped to avoid implicit concatenation in collections or made single-line as needed), so we can adopt 0.7.0 without new warnings.
This branch has not been deployed
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
2 participants
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.
Summary
The
<3.14cap was the only thing keeping 3.14 out. No application code changes are needed — the suite passes unchanged on all four interpreters, with the locked dependency versions.Identical counts across the matrix. The tree uses nothing removed since 3.11: no
distutils,pkg_resources,imp,utcnow(),utcfromtimestamp(), or bareget_event_loop().What changed
pyproject.tomlpython = ">=3.11,<3.14"→>=3.11,<3.15pyproject.toml(dev)black ^22.6.0→^24.2.0,flake8 ^5.0.4→^7.0.0.github/workflows/python-package.yml"3.11","3.12","3.13"→ adds"3.14".pre-commit-config.yamlflake8-implicit-str-concat==0.4.0→==0.6.0conda/dev.yamlpython>=3.11,<3.14→<3.15poetry.lockEvery runtime dependency keeps its resolved version (numpy 2.4.6, pandas 2.3.3, pyarrow 24.0.0 on all four). numpy already carried a
python_version == "3.11"split and needed no new marker.Two dev tools were broken on 3.12+
Both fail for the same reason: they touch
ast.Str, which no longer exists.black 22.12.0— what the^22.6.0dev pin resolved to — aborts withcannot format: module 'ast' has no attribute 'Str'on five files (api/types.py,api/metadata/columns.py,cli/configure.py,tests/api/test_databases.py,tests/api/ducklake/test_adapters.py). Already broken on 3.12/3.13 today; 3.14 just makes it universal.flake8-implicit-str-concat 0.4.0is the blocker for CI specifically. It raises the sameAttributeErrorout of flake8's multiprocessing pool, taking down the whole run. Since it is the hook theLintingstep runs, a 3.14 matrix entry would have gone red before reaching the tests. 0.6.0 is the first release that works.Why the plugin is pinned to
==0.6.00.7.0 also runs on 3.14, but it adds an
ISC004check that flags eight existing implicit concatenations inapi/saude/databases.py. Those are a separate concern, so the pin stays at the oldest release that unblocks 3.14 — and it is pinned exactly, like every other plugin in that file, so apre-commit autoupdatecannot silently shift the result.I verified the real CI lint command (
pre-commit run --files pysus/**/*) end to end on 3.14: all ten hooks pass, including black, flake8 and mypy.CI on this PR
All green, including both new 3.14 entries:
docker-testspassing matters too: theDockerfileis still on 3.12, and this confirms nothing in the dependency changes broke the image build or the suite inside it.Not included, deliberately
Dockerfilestays onpython:3.12-slim. 3.12 is still supported and I could not build the image here to verify. Worth a follow-up..readthedocs.yamlstays on 3.11 — docs builds do not need the newest interpreter.api/metadata/columns.pyis left as-is despite four pre-existing lint findings and ablackdiff, all from0048008(fix: improve metadata group discovery and loading #348).Follow-up worth filing separately
The CI lint step runs
pre-commit run --files pysus/**/*. Withoutglobstar, that glob expands to the files at depth 2 only — 96 paths, against 208 tracked.pyfiles. pre-commit discards the directories it matches and flake8 therefore never recurses into them, soapi/metadata/columns.pyandtests/cli/test_dadosgov.pyare never linted. That is whycolumns.pyhas sat inmainwith whitespace problems, andtest_dadosgov.pywith twoISC001hits, while CI stayed green.Fixing the glob would suddenly lint ~208 files and surface a backlog, so it deserves its own PR rather than riding along here.