chore: configure ruff lint rules and fix what they surfaced - #883

Description

The ruff and ruff-format pre-commit hooks were running with no configuration at
all, so only ruff's built-in defaults (E4/E7/E9 + F) were ever enforced. Adds a
[tool.ruff] section selecting the correctness-oriented groups: B, C4, PIE, PERF,
PLE, PLW, LOG, G, ASYNC and RUF. Formatting is unchanged (ruff-format defaults, which
the tree already matches).

Everything in select is currently clean, so the hook is enforceable as-is. Each entry
in ignore is a deliberate call with its reason inline; the notable ones:

• B905 (zip strict=) is a per-call-site behaviour change, turning a silent
  truncation into a runtime exception. Worth adopting deliberately, not in a lint sweep.
• RUF005, C408 and RUF007 are style-only rewrites of code that is already correct
  and readable (a + [b] → [*a, b], dict(a=1) → {"a": 1},
  zip(x[:-1], x[1:]) → itertools.pairwise(x)). Enforcing them means churning working
  call sites for no behaviour change, so they are left to author preference.
• RUF022 sorts __all__ alphabetically, which scrambles the semantic grouping
  comments the schema/package __init__ files rely on.
• RUF100 is evaluated against select, so it flags every noqa written for a rule not
  yet enabled (BLE001, PLC0415, N803, ...). Re-enable once those groups are adopted.

No live defects were found. Every rule fired on code that behaves correctly today;
what follows removes fragility, not bugs.

Correct today, fragile to a later change

• LOG014: _record_error passes exc_info=True, which reads the ambient
  sys.exc_info(). Both of its callers invoke it from inside an except block, so the
  traceback is logged correctly; the rule is lexical and fires because the logging call
  sits in a helper rather than in the handler itself. Passing the exception explicitly
  makes it independent of the caller's context.
• B023: a closure in the frame loop captured the loop variable phase by reference.
  It is called immediately in the same iteration, so the value was always correct.
  fill is hoisted above the loop and takes phase as a parameter, which also stops
  re-creating the function object every frame. Binding it as a default argument would
  satisfy the rule equally.
• B011: assert False in a test, which python -O strips, turning a failure into a
  silent pass. The suite is not run under -O today. Replaced by calling the
  constructor directly, so an exception fails the test with its own traceback.
• B017: a blind pytest.raises(Exception) that would also accept an unrelated
  failure. The test passes for the right reason today; naming the accepted exception set
  keeps it that way.
• RUF043: pytest match="libcloudxr.so" treats . as a regex wildcard where a literal
  filename was meant. The real message contains the literal, so the assertion passes
  correctly, but it is weaker than it reads. The remaining patterns are intentional
  regexes and are now raw strings.

Typing and explicitness

• RUF012: three mutable class attributes annotated ClassVar, one of them on the
  EnvConfig singleton.
• RUF013: implicit Optional spelled out as str | None.
• B904: raise ... from on re-raises, so the original cause is not lost.
• G004: log calls take %s arguments rather than eagerly formatted f-strings.
• PLW1510: subprocess.run calls that inspect returncode say check=False. The
  tree already spelled this out at 10 call sites, 8 of them in oob_teleop_adb.py; this
  covers the stragglers.

TRY004 (ValueError → TypeError in TeleopSessionConfig validation) was left
alone: it is a public API behaviour change, not a lint fix.

Python only — no C++ or CMake is touched. A companion PR adds the C++ warning set; the
two are independent and can land in either order.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Testing

Ubuntu 24.04 / x86_64, Python 3.12 —

• ruff check and ruff format --check both clean at v0.15.1, the version pinned in
  .pre-commit-config.yaml (268 files).
• SKIP=check-copyright-year pre-commit run --all-files: all hooks pass.
• ctest: 309/310. The one failure, cloudxr_test_launcher, is the missing CloudXR SDK
  (no NGC key on this host, so the download 404s and get_sdk_path() raises) — an
  environment gap, not a code failure.

Forward-compat note, not addressed here: at ruff 0.16 format --check wants to reflow
Python code blocks embedded in three Markdown files (examples/teleop_ros2/README.md,
src/plugins/oak/README.md, and one other) — newer ruff formats fenced code in Markdown,
which 0.15.1 does not. None of those files are touched by this PR, but it will surface
whenever the pre-commit rev is bumped past 0.16.

Checklist

• I have read and understood the contribution guidelines
• I have run the linter and formatter with SKIP=check-copyright-year pre-commit run --all-files
• I have made corresponding changes to the documentation
• I have added tests that prove my fix/feature works (or explained why not)
• I have signed off all my commits (git commit -s) per the DCO

Documentation: no user-facing behaviour changes, so no doc updates. Every non-obvious
ignore entry is documented inline in pyproject.toml.

Tests: no new tests — this is a lint-configuration change plus the fixes it surfaced,
and it is exercised by the existing suite passing.