nemo-relay-instrument-typed-wrappers

от nvidia

Используйте этот навык при добавлении типизированных обёрток NeMo Relay, доменных типов или кодеков провайдера, сохраняя семантику JSON middleware и видимое вызывающему коду поведение.

npx skills add https://github.com/nvidia/skills --skill nemo-relay-instrument-typed-wrappers

Скачать ZIP GitHub

3.1k

Попробовать этот навыкСпонсировано

Use Typed Wrappers And Codecs

Use this skill when an application wants stronger domain types than raw JSON for tool or LLM integration. Keep typed boundaries explicit so middleware still sees predictable JSON.

Default Guidance

Prefer plain JSON first for initial adoption.
Reach for typed wrappers when the application already has stable domain models.
Keep in mind that middleware still operates on JSON, not typed objects.

Embedded Codec Model

A typed value codec is a pure boundary translator. It converts application-facing values to JSON before NeMo Relay emits events or runs middleware, then converts JSON back into the framework callback or caller type.
Python exposes JsonPassthrough, DataclassCodec, PydanticCodec, and BestEffortAnyCodec. Node.js exposes JsonPassthrough plus custom Codec<T> implementations.
Use BestEffortAnyCodec only at boundaries where strict schemas are not available. Prefer dataclass, Pydantic, or explicit Node.js codecs when the framework owns a stable schema.
Provider codecs are different from typed value codecs: they normalize provider-specific LLM requests and responses so middleware and subscribers can inspect messages, tools, model names, generation parameters, and response annotations.
Built-in provider codecs include OpenAIChatCodec, OpenAIResponsesCodec, and AnthropicMessagesCodec in Python, Node.js, and Rust. Choose the codec that matches the actual provider payload shape.
Response codecs annotate LLM end events with fields such as id, model, message, tool_calls, finish_reason, usage, provider-specific data, and extra unmodeled fields. They do not rewrite the caller-visible response.
Request codecs run before LLM request intercepts. Intercepts receive both the raw LLMRequest and optional annotated request; encode merges annotated edits back before execution intercepts and the provider callback run.

Key Rules

Typed wrappers are currently a first-class path for Python and Node.js; Rust uses codec traits directly
Request/response conversion belongs in codecs
Intercepts and guardrails see JSON values after encoding
Changes made by middleware survive into the decode step

Choose A Codec

JsonPassthrough for JSON-native values
DataclassCodec or PydanticCodec in Python when the models already exist
Custom codecs for domain-specific wire shapes
BestEffortAnyCodec only when broad flexibility is worth the looser contract
Provider codecs for LLM provider payloads, not application domain objects conversion

Validation Checklist

Codec output is JSON-compatible
Required fields survive toJson/fromJson or decode/encode
Middleware sees the expected serialized shape
Provider codecs preserve fields they do not understand
Response codec failures do not break the underlying LLM call
Request codec encode preserves original provider fields unless an intercept intentionally changes them