refactor(backend): enable ruff SIM102 and collapse nested if statements

[ogenstad]

Why

SIM102 (collapsible-if) was suppressed in two places in pyproject.toml: the global ignore
list, and a backend/tests/fixtures/**.py per-file entry sitting under a "Refactor code and remove
the ignore rule" header. Both suppressions are now gone and the rule is fully enforced.

Goal: enable the rule with zero behavioral change.

Non-goals: no logic changes, no restructuring beyond what the rule requires, no changes to the
other rules still listed in either ignore block.

What changed

Behavioral changes: none. Every edit is structural.

• Removed "SIM102" from the global ignore list and from the
  backend/tests/fixtures/**.py per-file ignores. grep -c SIM102 pyproject.toml is now 0.
• Fixed the 38 violations that surfaced (36 in backend/infrahub/, 2 in fixtures) across 23 files.
  Nested if A: if B: becomes if A and B:, which short-circuits identically.
• Two sites in core/node/__init__.py where the collapsed form wrapped a multi-line await in a
  walrus inside an and were instead destructured into an assignment plus a plain if. That also
  satisfies the rule (the outer body then holds two statements) and matches the idiom already used
  a few lines above.
• Flattened redundant parenthesization the autofix left behind - (A and B and C) and D - in
  attribute_parameters.py and groups/parsers.py. No and-within-and or or-within-or
  nesting remains anywhere in the diff.

What stayed the same: no schema changes, no API changes, no dependency changes. Three comments
moved from inside an outer if to just above the merged condition; no comment was added, removed,
or reworded.

How to review

Most of the diff is mechanical and was produced by ruff check --fix. Worth actual attention:

1. backend/infrahub/core/node/__init__.py - the two destructured await blocks in delete(),
   and the multi-term guard chains in _collect_extra_filters / process_label.
2. backend/infrahub/core/schema/schema_branch.py - identify_required_object_templates, the one
   condition mixing or inside and.
3. Four sites where an elif was collapsed (graphql/app.py, graphql/field_extractor.py,
   core/diff/enricher/cardinality_one.py x2). Collapsing an elif is only sound when it is the
   terminal branch; each was verified to have no trailing else.

Two things reviewers should weigh in on:

• New idiom. Before this PR the backend had 110 walrus operators and zero used as a boolean
  operand. This PR introduces 14 across 8 files, 8 of which put the walrus in a later operand
  (if guard and (x := f()):), where the binding is conditional. It is correct - the name is only
  read inside the body, and the binding condition is unchanged from the nested original - but it is
  a convention arriving via a lint sweep. The alternative is destructuring those 8 the way the
  delete() blocks were. Happy to do that if preferred.
• Fixture drift. The two fixture edits live in local copies of infrahub-demo-edge. If that
  fixture is ever re-synced from upstream the violation returns, and because the per-file ignore is
  gone, CI will fail until it is re-applied. That is the intended tripwire rather than an
  oversight; a follow-up cleanup issue on opsmill/infrahub-demo-edge could remove the friction.

How to test

uv run ruff check .            # clean
uv run ruff format --check .   # clean
uv run mypy backend/infrahub/  # clean on changed files
uv run pytest backend/tests/unit -q

Beyond the standard checks, equivalence was proven mechanically rather than by eye. Both versions
of every changed file were canonicalised to a maximally-nested normal form (explode and-chains
into nested ifs; de-sugar sole-test walruses into assignment plus test; flatten same-operator
boolean nesting - all three behaviour-preserving), then compared as ASTs. All 23 files are
identical to their base version under that normal form, and comment token streams match exactly.

The canonicaliser was itself validated against 26 control cases, confirming it rejects and→or,
flipped comparisons, reordered operands, dropped negations, swallowed else branches, non-terminal
elif collapses, dropped sibling statements, and/or regrouping, and - specifically for the
destructured blocks - an assignment hoisted outside its guard.

The edited fixture check was additionally differential-tested: old and new validate() driven
through 122 enumerated input combinations, comparing recorded log_error calls. Zero mismatches,
with 112 cases producing non-empty output.

Impact & rollout

• Backward compatibility: no behavioral change; nothing to migrate.
• Performance: unaffected.
• Config/env changes: none at runtime; pyproject.toml lint config only.
• Deployment notes: safe to deploy.

Checklist

• Tests added/updated - n/a, no behavior change; existing suite covers the touched paths
• Changelog entry - n/a, not user-facing
• External docs updated - n/a
• Internal .md docs updated - n/a
• I have reviewed AI generated content

[codspeed-hq]

Merging this PR will not alter performance

✅ 13 untouched benchmarks

---

_{Comparing po-enable-ruff-sim102 (56aee23) with develop (e6e28c2)}

[cubic-dev-ai]

No issues found across 24 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

<sub>Shadow auto-approve: would auto-approve. Refactor enables ruff SIM102 lint rule by collapsing nested ifs into equivalent compound conditions with verified zero behavioral change across 24 files.

Re-trigger cubic</sub>