Function withRequestContextStream

withRequestContextStream<TInput, TOutput>(
req: AIRequest<TInput>,
fn: (req: AIRequest<TInput>) => AsyncGenerator<AIResponseChunk<TOutput>>,
): AsyncGenerator<AIResponseChunk<TOutput>>
Wraps a streaming provider call with request-scoped tracing metadata.

This helper:
- Generates a unique requestId
- Captures start time
- Injects request context into the AIRequest
- Attaches timing and request metadata to each streamed chunk
Designed for streaming chat, image generation, or any provider API that emits incremental results.
Type Parameters
- TInput
  The input payload type sent to the provider.
- TOutput
  The output payload type for each streamed chunk.
Parameters
- req: AIRequest<TInput>
  Unified AI request containing input, options, and optional context.
- fn: (req: AIRequest<TInput>) => AsyncGenerator<AIResponseChunk<TOutput>>
  Provider streaming execution function to wrap.
Returns AsyncGenerator<AIResponseChunk<TOutput>>
An async generator emitting response chunks with metadata attached.
- Defined in src/core/utils/WithRequestContext.ts:74