docs: document omit sentinel behavior on provider wires (carry of #2432) - #3603
Conversation
Co-authored-by: mdwsk88 <11055210+mdwsk88@users.noreply.github.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: ASSERTIVE Plan: Team Run ID: 📒 Files selected for processing (9)
📝 WalkthroughWalkthroughThe change documents the reserved ChangesReasoning omission documentation
Estimated code review effort: 1 (Trivial) | ~5 minutes Suggested reviewers: ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
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. Comment |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9a6582c4d7
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| | `reasoningEffortMap?` | `Record<string, string>` | Provider-wide wire aliases for reasoning labels. Map a label to `"__omit__"` to drop the reasoning field from the upstream request entirely: `reasoning_effort` on an OpenAI-compatible wire, and Ollama's native `think` field on the Ollama native adapter (#2356). | | ||
| | `modelReasoningEffortMap?` | `Record<string, Record<string, string>>` | Per-model wire aliases for reasoning labels. Map a label to `"__omit__"` to drop the reasoning field from the upstream request entirely. | |
There was a problem hiding this comment.
For a non-forward openai-responses route with a configured reasoning ladder and a label mapped to "__omit__", mapRoutedResponsesReasoningEffort in src/adapters/openai-responses.ts:632-634 receives undefined and returns the original body, leaving body.reasoning.effort on the upstream request. The unconditional statement that the reasoning field is dropped entirely therefore gives users the wrong result on a documented provider wire; either remove the Responses effort when the sentinel is selected or explicitly scope this documentation to the Chat-compatible and native Ollama paths that implement omission.
AGENTS.md reference: docs-site/AGENTS.md:L7-L8
Useful? React with 👍 / 👎.
| | `annotateEmptyToolOutputs?` | `boolean` | Remplace un résultat d’outil présent mais vide par un court marqueur avant qu’il n’atteigne le modèle, afin qu’un résultat vide ne soit pas interprété comme manquant. S’applique aux chaînes vides et aux tableaux de parties contenant uniquement du texte ; les parties d’image, de fichier et chiffrées ne sont jamais modifiées. La valeur par défaut issue du registre intégré est `true` pour DeepSeek ; dans les autres cas, elle n’est pas définie. Définissez `false` pour exclure un fournisseur : une valeur `false` explicite est conservée lors des modifications ultérieures qui omettent ce champ. `PATCH /api/providers?name=<provider>` accepte `true`, `false` ou `null` pour effacer le remplacement et revenir au comportement par défaut du registre. | | ||
| | `reasoningEffortMap?` | `Record<string, string>` | Alias de fil à l’échelle du fournisseur pour les étiquettes de raisonnement. | | ||
| | `modelReasoningEffortMap?` | `Record<string, Record<string, string>>` | Alias de fil par modèle pour les étiquettes de raisonnement. | | ||
| | `reasoningEffortMap?` | `Record<string, string>` | Alias de fil à l'échelle du fournisseur pour les étiquettes de raisonnement. Mappez une étiquette à `"__omit__"` pour supprimer complètement le champ de raisonnement de la requête en amont (par exemple pour les modèles Ollama dont le gabarit de conversation exige l'omission de `reasoning_effort` pour activer le mode de raisonnement approfondi). | |
There was a problem hiding this comment.
think field in every translation
This localized text says Ollama deep mode requires omitting reasoning_effort, while the canonical English page and src/adapters/ollama-native.ts identify the native field as think. The same contradictory wording was added to the Japanese, Korean, Russian, Turkish, Simplified Chinese, and Traditional Chinese pages, so localized readers are given a different wire contract; update all seven translations to describe omission of native think.
AGENTS.md reference: docs-site/AGENTS.md:L9-L9
Useful? React with 👍 / 👎.
|
✅ Deterministic PR hygiene checks passed. |
github-actions
Bot
added
the
documentation
2 participants
Summary
Carries #2432 and corrects the provider field comments: the literal omit suppresses reasoning_effort on OpenAI-compatible wires and think on the native Ollama adapter. The documentation references the canonical sentinel owner instead of implying the config interface exports it.
Verification
Checklist
Co-authored-by: mdwsk88 11055210+mdwsk88@users.noreply.github.com
Summary by CodeRabbit
"__omit__".