Expose observation metadata for RLSSM producers

[AlexanderFengler]

Depends on #340.

Scope

• expose semantic observation metadata from ssms.rl.ModelConfig through the generic producer accessor
• derive the two existing RL response layouts only:
  • rt, response for RT-based models
  • response for choice-only models
• take categorical labels from the ordered raw choices tuple, independently of response_to_choice
• require an explicit ordered v1 schema for every non-standard response layout
• snapshot resolved metadata on AssembledModel while returning a fresh descriptor on access

Compatibility boundary

This remains an out-of-band metadata interface. It does not change RL simulation or predictive panels, dummy RT behavior, context fields, participant input order, learning updates, preset metadata, or to_hssm_config_dict().

The new ModelConfig inputs are keyword-only. AssembledModel retains its exact constructor, pattern-matching, dataclass-field, serialization, representation, and equality surfaces. Its private schema snapshot is not a dataclass field.

The one intentional tightening is that an assembled model with a non-standard response layout must now declare that layout explicitly. Standard layouts cannot override their derived schemas, so RT and response semantics continue to have one source of truth.

Evidence

• focused RL metadata contract: 15 passed
• complete tests/rl suite: 309 passed
• complete package suite: 1,258 passed, 134 skipped
• repository-wide Ruff check and format check: passed
• focused affected-file type check: passed after isolating accepted repository-wide missing-stub and pre-existing optional-narrowing categories
• live HSSM bridge canaries against this worktree: 3 passed
• independent seeded parity checks for RT and 2-, 3-, and 4-choice panels, preset info, bridge dictionaries, omission handling, and learning trajectories: unchanged

Deferred

Cross-repository consumer matrices and contributor documentation remain isolated to 6A3d. No downstream consumer is switched to normalized structured results in this PR.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 8079d103-a376-4ae7-9a56-7d30b1fc327a

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[codecov]

Codecov Report

❌ Patch coverage is 97.36842% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
ssms/rl/config.py	96.55%	1 Missing ⚠️

Flag	Coverage Δ
unittests	94.42% <97.36%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
ssms/rl/assembled.py	95.58% <100.00%> (+0.18%)	⬆️
ssms/rl/config.py	94.17% <96.55%> (+0.26%)	⬆️

🚀 New features to boost your workflow:

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