Struct ResponsesCompletionModel 

pub struct ResponsesCompletionModel<H = Client> {
    pub model: String,
    pub tools: Vec<ResponsesToolDefinition>,
    pub strict_tools: bool,
    /* private fields */
}

Fields

model: String

tools: Vec<ResponsesToolDefinition>

strict_tools: bool

Implementations

impl<H> ResponsesCompletionModel<H>

where Client<ChatGPTExt, H>: HttpClientExt + Clone + Debug + 'static, H: Clone + Default + Debug + WasmCompatSend + WasmCompatSync + 'static,

pub fn new( client: Client<ChatGPTExt, H>, model: impl Into<String>, ) -> ResponsesCompletionModel<H>

pub fn with_strict_tools(self) -> ResponsesCompletionModel<H>

Enable strict mode for function tool schemas.

pub fn with_tool( self, tool: impl Into<ResponsesToolDefinition>, ) -> ResponsesCompletionModel<H>

pub fn with_tools<I, Tool>(self, tools: I) -> ResponsesCompletionModel<H>

where I: IntoIterator<Item = Tool>, Tool: Into<ResponsesToolDefinition>,

pub async fn raw_completion( &self, completion_request: CompletionRequest, ) -> Result<CompletionResponse, CompletionError>

Execute a ChatGPT completion and return the Responses API’s own wire response.

This is the escape hatch for fields rig does not normalize, and it issues the same single request the normalized path does.

One caveat is specific to this provider: /responses answers with an SSE body even for a non-streaming request, so the value returned here is reassembled from the terminal response.completed event. That event sometimes carries an empty output, in which case the assistant content exists only in the preceding events and completion::CompletionModel::completion rebuilds it from them. When you need the provider’s events in full fidelity rather than just its terminal record, use ResponsesCompletionModel::raw_stream.

impl<H> ResponsesCompletionModel<H>

where Client<ChatGPTExt, H>: HttpClientExt + Clone + Debug + 'static, H: Clone + Default + Debug + WasmCompatSend + WasmCompatSync + 'static,

pub async fn stream( &self, completion_request: CompletionRequest, ) -> Result<StreamingCompletionResponse, CompletionError>

Open a stream normalized to rig’s terminal record.

Delegates to ResponsesCompletionModel::raw_stream — one request either way.

pub async fn raw_stream( &self, completion_request: CompletionRequest, ) -> Result<Pin<Box<dyn Stream<Item = Result<RawStreamingChoice<StreamingCompletionResponse>, CompletionError>> + Send>>, CompletionError>

Open a stream whose terminal record stays the Responses API’s own type.

Trait Implementations

impl<H> Clone for ResponsesCompletionModel<H>

where H: Clone,

fn clone(&self) -> ResponsesCompletionModel<H>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<H> CompletionModel for ResponsesCompletionModel<H>

where Client<ChatGPTExt, H>: HttpClientExt + Clone + Debug + 'static, H: Clone + Default + Debug + WasmCompatSend + WasmCompatSync + 'static,

async fn completion( &self, completion_request: CompletionRequest, ) -> Result<CompletionResponse, CompletionError>

Generates a completion response for the given completion request.

async fn stream( &self, completion_request: CompletionRequest, ) -> Result<StreamingCompletionResponse, CompletionError>

Streams a completion response for the given completion request.

fn completion_request( &self, prompt: impl Into<Message>, ) -> CompletionRequestBuilder<Self>

where Self: Sized + Clone,

Generates a completion request builder for the given prompt.

fn capabilities(&self) -> ProviderCapabilities

Provider behavior a runtime should account for when preparing requests.

impl<H> ConstructCompletionModel<Client<ChatGPTExt, H>> for ResponsesCompletionModel<H>

where Client<ChatGPTExt, H>: HttpClientExt + Clone + Debug + 'static, H: Clone + Default + Debug + WasmCompatSend + WasmCompatSync + 'static,

fn construct( client: &Client<ChatGPTExt, H>, model: String, ) -> ResponsesCompletionModel<H>

Build this model from its provider client and a model identifier.