ruff_python_spo: add a plain-Python SPO harvest arm

[AdaWorldAPI]

Summary

ruff_python_spo's extract() gates on Odoo markers (models.Model / _name / _inherit), so it returns an empty graph on any plain-Python codebase. Measured, not assumed: models=0 over monarch-initiative/dismech's 90-file resolver package.

functions::analyze_method — the part that actually earns the harvest (reads / writes / raises / traverses / calls / guarded_writes) — is already language-generic. Only the is_model gate blocked it. This adds a second entry point that drops the gate and reuses analyze_method unmodified.

• extract_plain_from_source(source, module) -> ModelGraph
• extract_plain(root, namespace) -> ModelGraph — recursive *.py walk mirroring collect_py; module name from path (export/sepio_export.py → export.sepio_export, dot→underscore normalised as the Odoo arm does)

Every top-level class becomes a Model; bases become Model::inherits (reusing the frontend-agnostic carrier and the existing InheritsFrom lowering); module-level defs attach to a synthetic per-module Model so free functions keep has_function + their body facts; AnnAssign and simple literal assigns become Fields. Precedent: ruff_cpp_spo's walk_free_functions.

No new Predicate variants — the enum is closed and stays closed; expand.rs was read to confirm has_field/field_type/InheritsFrom lower generically rather than Odoo-coupled.

Measurement

$ cargo run -p ruff_python_spo --example plain_census -- /tmp/dismech-up/src/dismech
odoo baseline: models=0
plain: models=658 functions=1277 fields=3587 inherits_edges=546
  body facts: reads=1148 writes=635 raises=83 traverses=2 guarded_writes=6 calls=57

The models=0 baseline is measured in the same run over the same root, not asserted.

Genericity check on a second, unrelated corpus — /home/user/WoA/woa (Flask + SQLAlchemy, no Odoo, no relation to dismech), run untouched:

odoo baseline: models=0
plain: models=241 functions=1687 fields=426 inherits_edges=24
  body facts: reads=56 writes=23 raises=24 traverses=1 guarded_writes=1 calls=95

Different ORM, different framework, different shape (function-heavy rather than field-heavy, as Flask blueprints should be), no code change required.

Deliberately conservative, with the gaps named

• Field-type inference uses one rule (annotation head: Name, or Subscript's head) and falls to None for unions / X | None / forward-ref strings. dismech itself contains dataclasses, Enums, TypedDict-likes and heavy Optional[...] — the rule had to survive within-corpus variety before it ever met a second corpus.
• Bare-assign literals only become fields for str/int/float/bool/list/dict/set/tuple. Forced by real code: logger = logging.getLogger(__name__) must not become a field.
• Known gaps, not fixed here (outside the class/def shape this change scopes): module-level Assign constants are invisible — which notably means CURIE-shaped constants like sepio_export.py's HAS_PATHOPHYSIOLOGY = "dismech:has_pathophysiology" are not harvested — and nested classes aren't walked (same posture as the Odoo arm). Both are contained follow-ups.

Test plan

• cargo test -p ruff_python_spo — 66/66 (57 pre-existing + 9 new), independently re-run
• cargo clippy -p ruff_python_spo --all-targets -- -D warnings — clean
• cargo fmt -p ruff_python_spo — clean
• uvx prek run --from-ref <base> — clean (one typos auto-fix, committed separately, tests re-verified green)
• Census output above independently reproduced before opening this PR

New tests assert exact counts (==, not >=) and cover: dataclass fields, base + method quartet, module-level def, unparsable source, empty module, literal-inference boundary, subscript/complex-annotation boundary, module-name normalisation, tree walk.

---

Generated by Claude Code

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_913241e8-3a15-4aa4-a7bc-02c9aaccab04)