fix(parser): persist Python class decorators

[tirth8205]

Focused current-main replacement for the remaining class-decorator gap from #332, with Gideon Zenz attribution.

Overlap reconciliation:

• maint: reconcile verified local and community fixes #624 already persists decorated functions and class annotations for Java, Kotlin, and C#.
• Python class_definition nodes are children of decorated_definition, so the existing modifiers-child helper cannot see their decorators.
• This PR ports only that missing behavior; it does not replay the cumulative source branch.

Behavior:

• captures stacked Python class decorators in source order;
• stores the comma-joined contract in NodeInfo.modifiers;
• stores the list contract in NodeInfo.extra[decorators];
• leaves undecorated classes unchanged;
• proves parsed metadata reaches framework-entry detection and find_dead_code.

Verification:

• 1,538 passed, 2 xpassed (full local suite)
• focused parser and end-to-end dead-code regressions pass
• repository lint scope passes for code_review_graph/parser.py
• git diff --check passes

[github-actions]

code-review-graph review

Overall risk: 0.40 (MEDIUM) — 7 changed function(s)/class(es), 0 affected flow(s), 5 test gap(s)

Risk-scored changes

Risk	Level	Symbol	Location	Tested
0.40	medium	code_review_graph/parser.py::CodeParser	code_review_graph/parser.py:983	no
0.35	low	code_review_graph/parser.py::_python_decorator_names	code_review_graph/parser.py:901	no
0.35	low	code_review_graph/parser.py::CodeParser._extract_classes	code_review_graph/parser.py:4904	no
0.30	low	tests/test_parser.py::TestCodeParser.test_parse_python_class_decorators_persisted	tests/test_parser.py:182	(test)
0.20	low	tests/test_parser.py::TestCodeParser	tests/test_parser.py:11	no
0.20	low	tests/test_refactor.py::TestFindDeadCode.test_find_dead_code_excludes_parsed_python_decorated_class	tests/test_refactor.py:426	(test)
0.10	low	tests/test_refactor.py::TestFindDeadCode	tests/test_refactor.py:109	no

Test gaps

• code_review_graph/parser.py::_python_decorator_names (code_review_graph/parser.py:901)
• code_review_graph/parser.py::CodeParser (code_review_graph/parser.py:983)
• code_review_graph/parser.py::CodeParser._extract_classes (code_review_graph/parser.py:4904)
• tests/test_parser.py::TestCodeParser (tests/test_parser.py:11)
• tests/test_refactor.py::TestFindDeadCode (tests/test_refactor.py:109)

Token savings: this graph-backed report used ~112,151 fewer tokens (~98%) than reading every changed file in full (estimated, chars/4 approximation).

---

Powered by code-review-graph — local-first analysis; no code leaves the CI runner.