maintenance: select ruff rules explicitly so CI is deterministic

[igorsaevets]

CI is currently red on every pull request, and the cause is outside the changed code.

What happened

ruff 0.16.0 (released 2026-07-25) widened its default rule selection from 59 rules to 413. pyproject.toml asks for ruff >= 0.15 with no upper bound and does not configure [tool.ruff.lint] select, so CI installs whatever is newest and lints against the new defaults.

Reproduced against unmodified master at 3d89493, no other changes in the tree:

ruff	ruff check .
0.15.4	All checks passed!
0.16.0	Found 103 errors.

The 103 findings are spread over 31 files and come from rule families that were not previously on by default: BLE001 (22), I001 (18), FURB167 (12), UP031 (11), SIM102 (5), plus RUF, DTZ, S, PL, B, PIE, FLY and C4. None of them are new code.

Confirming it is only the selection and nothing else in the release:

ruff 0.16.0, master, --select E4,E7,E9,F  ->  All checks passed!

The change

One block in pyproject.toml selecting E4, E7, E9, F explicitly. Those are ruff's pre-0.16 defaults, so linting behaviour is unchanged from what the repository has been enforcing, and it no longer moves when ruff changes its mind about defaults.

This is the same reasoning already written down a few lines below for mypy in that file:

# explicit strictness — stable across mypy upgrades, unlike 'strict = true'

Verified after the change:

ruff 0.15.4  ruff check .                          ->  All checks passed!
ruff 0.16.0  ruff check .                          ->  All checks passed!
ruff 0.16.0  ruff format --check trafilatura tests ->  41 files already formatted

ruff format was never affected; only the lint step was.

Not included on purpose

Adopting any of the newly-default rules is a separate judgement call, and 103 mechanical edits do not belong in the same change as a CI repair. Roughly 43 are auto-fixable if you want them. Happy to open a follow-up per rule family, or to add an upper bound like ruff >= 0.15, < 0.17 instead if you prefer pinning the tool over pinning the selection. Your call and I will match whatever you pick.

Found this while checking why #891 was red. That PR touches only htmlprocessing.py, xml.py and tests/unit_tests.py and contributes none of the 103 findings.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.68%. Comparing base (3d89493) to head (ec98b36).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #892   +/-   ##
=======================================
  Coverage   99.68%   99.68%           
=======================================
  Files          21       21           
  Lines        4099     4099           
=======================================
  Hits         4086     4086           
  Misses         13       13           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[adbar]

Nice, thanks!