Module image_gen 

Image generation client: unified interface for multiple providers.

Supports two protocols:

• Openai: OpenAI-compatible Images API (POST /images/generations).
• Dashscope: DashScope native protocol (Wanxiang), with sync and async call modes. Async mode submits a task then polls until completion.

The model used is resolved from config (default_image_model) and is not exposed to callers - the client uses provider.model_id internally.

Structs

GeneratedImage

A single generated image, as returned by the provider API.

ImageGenClient

ImageGenErrorInfo

Structured error info extracted from an ImageGenError, for returning to the LLM in a machine-parseable form so it can decide whether to retry.

ImageGenRequest

Parameters for an image generation request.

Enums

ImageGenError