feat(protocol): add decision model request and response types - #880
Conversation
Signed-off-by: nachiketb <nachiketb@nvidia.com>
|
|
Navigate logical layers of code changes, visualize relationships, and explore their blast radius. No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Repository: NVIDIA-NeMo/Switchyard/.coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (2)
Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review. WalkthroughThe protocol crate adds a public decision protocol. It validates request questions, response answers, numeric values, and probability distributions during construction and deserialization. ChangesDecision Protocol
Priority: ⬇️ Low Estimated code review effort: 3 (Moderate) | ~25 minutes Merge Risk: ⚪ Minimal · up to This adds standalone decision protocol types with validation at construction and deserialization. No concrete merge-blocking risk was identified in the supplied material. The author reports that formatting, check and clippy passed. No tests were added. 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
I’m a rabbit with a protocol to try, Comment |
Signed-off-by: nachiketb <nachiketb@nvidia.com>
Signed-off-by: nachiketb <nachiketb@nvidia.com>
Signed-off-by: nachiketb <nachiketb@nvidia.com>
Bring in NVIDIA-NeMo#794, NVIDIA-NeMo#878 and NVIDIA-NeMo#880 from main. None of them touch the escalation router, judge, prompts, or docs changed by this PR, and the merge applied without conflicts. Signed-off-by: Lin Jia <linj@nvidia.com> Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
What
Add public Decision Model types in
crates/protocol:ModelId, and sharedUsage.Why
Provide the Decision Model protocol for future clients and algorithms.
Notes for reviewers
All fields are public, with derived serialization and deserialization. No constructors, validation helpers, or new error type. Score levels remain in the request; responses preserve provider scores and probabilities.
Validation: formatting, protocol compilation, and targeted Clippy passed. No tests added or run, as requested.