fix(test): pin consolidator fallback test to the config model

[yyiilluu]

Why

main is red. test_consolidator_calls_litellm_with_fallback_configured has been
failing since #372 landed:

E   ValueError: Structured-output fallback models must use the same transport
    strategy as 'gpt-test' (pydantic_passthrough); incompatible fallbacks: ['gpt-5.4-mini']

Root cause

_build_real_client_consolidator patched the generation site var to a fake
"gpt-test", so the LiteLLMConfig(model=...) each test passed never reached the
request — PlaybookConsolidator reads its request model from the site var
(deduplication_utils.py:191), not from the client config, and passes it as
model=self.model_name (components/consolidator.py:743).

#372 added _validate_structured_fallback_strategies, which resolves a transport
strategy per model and rejects a ladder whose fallbacks disagree with the primary.
Measured strategies:

model	provider	strategy
gpt-test	(unresolvable)	pydantic_passthrough
gpt-5.4-mini	openai	native_json_schema
minimax/MiniMax-M3	minimax	native_json_schema

The fake model made the pairing artificially incompatible. The guard itself is
correct — the test was asserting against a model it never intended to use.

Fix

Pin the patched site var to config.model. The fallback-configured case now
exercises a realistic production pairing (MiniMax primary + OpenAI fallback, both
native_json_schema). The second test (claude-code/..., no fallbacks) never
reaches the guard and is unaffected.

Also included: a style: commit applying ruff format to client/client.py,
the only file in the package ruff format --check flagged. No behavior change.

Verification

• pytest tests/server/services/playbook/test_playbook_consolidator_integration.py — 14 passed (1 failed before)
• ruff format --check reflexio tests — 821 files already formatted
• ruff check + pyright clean on both touched files

Summary by CodeRabbit

• Bug Fixes

  • Improved filtering of outdated operation status updates, helping prevent stale results from being processed.

• Tests

  • Updated integration coverage to validate structured-output fallback behavior using the configured primary model.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 0693db7f-fc5f-41f9-be89-727123704123

📥 Commits

Reviewing files that changed from the base of the PR and between 5227fb1 and 834c213.

📒 Files selected for processing (2)

• reflexio/client/client.py
• tests/server/services/playbook/test_playbook_consolidator_integration.py

---

📝 Walkthrough

Walkthrough

The client’s stale-operation condition is reformatted without behavioral change. The playbook consolidator integration helper now derives its default generation model from the supplied LiteLLMConfig.

Changes

Polling and fallback alignment

Layer / File(s)	Summary
Stale-operation filter expression reflexio/client/client.py	The stale-operation boolean guard is reformatted while preserving its existing condition.
Fallback integration model configuration tests/server/services/playbook/test_playbook_consolidator_integration.py	The patched site-variable lookup uses config.model as the default generation model and updates the helper documentation.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Possibly related PRs

• ReflexioAI/reflexio#100: Updates the same stale-operation filtering logic in ReflexioClient._poll_operation_status.
• ReflexioAI/reflexio#365: Modifies the stale-operation threshold logic in reflexio/client/client.py.

Suggested reviewers: yilu331

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: pinning the consolidator fallback test to the configured model.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/consolidator-fallback-test-model

---

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.}