Struct AzureOpenAIResponsesClient 

pub struct AzureOpenAIResponsesClient { /* private fields */ }

An Azure OpenAI Responses API chat client (POST {endpoint}/openai/v1/responses).

See the module docs for the URL/api-version rationale.

Implementations

impl AzureOpenAIResponsesClient

pub fn new( endpoint: impl Into<String>, deployment: impl Into<String>, api_key: impl Into<String>, ) -> Self

Create a client authenticating with a static API key (api-key header).

pub fn with_token_credential( endpoint: impl Into<String>, deployment: impl Into<String>, credential: Arc<dyn TokenCredential>, ) -> Self

Create a client authenticating via a TokenCredential (Authorization: Bearer <token>, e.g. Microsoft Entra ID).

pub fn from_env() -> Result<Self>

Build an API-key-authenticated client from AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY, AZURE_OPENAI_RESPONSES_DEPLOYMENT_NAME, and optional AZURE_OPENAI_API_VERSION/AZURE_OPENAI_BASE_URL — the same generic AZURE_OPENAI_* variables AzureOpenAIClient::from_env reads, except for the Responses-specific deployment variable (mirrors upstream’s responses_deployment_name settings field, distinct from Chat Completions’ AZURE_OPENAI_CHAT_DEPLOYMENT_NAME, so a resource with differently named deployments per API surface works; azure/_shared.py:102-103, docstring at azure/_responses_client.py:53-56).

pub fn with_api_version(self, api_version: impl Into<String>) -> Self

Override the API version (default "preview").

pub fn without_api_version(self) -> Self

Send no api-version query parameter at all.

Microsoft’s v1 Responses examples call the bare {endpoint}/openai/v1/responses URL, and the v1 lifecycle removes dated api-version parameters — GA v1 resources and some OpenAI-compatible gateways reject or misroute requests carrying one. The "preview" default mirrors upstream Python (_responses_client.py:112); use this to produce the documented query-less URL instead. (Also reachable by setting the AZURE_OPENAI_API_VERSION env var to an empty string with from_env.)

pub fn without_implicit_encrypted_reasoning(self) -> Self

Stop adding reasoning.encrypted_content to a stateless request’s include on the caller’s behalf.

The implicit add is right for OpenAI and Azure OpenAI, where a store: false tool loop has to replay the reasoning item itself. Azure AI Foundry does not want it unless it is asked for by name, so FoundryChatClient builds its transport with this set (upstream #7536).

A caller that puts reasoning.encrypted_content in its own include (via ChatOptions::additional_properties) still gets it — this governs only what the client adds unprompted.

pub fn with_base_url(self, base_url: impl Into<String>) -> Self

Override the full base URL used to build requests, taking precedence over the endpoint-derived /openai/v1/ route. Mirrors upstream’s base_url parameter, which “must end with /openai/v1/” (azure/_responses_client.py:60-62) — e.g. for a differently-shaped gateway or proxy in front of Azure OpenAI.

pub fn deployment(&self) -> &str

The deployment name this client targets (sent as the request body’s model field; see the module docs).

pub fn api_version(&self) -> Option<&str>

The API version this client sends, or None when the query parameter is omitted (see without_api_version).

pub fn base_url(&self) -> Option<&str>

The effective base URL requests are built against: the explicit with_base_url override when set, otherwise None (the endpoint-derived /openai/v1/ route is computed lazily by url).

Trait Implementations

impl ChatClient for AzureOpenAIResponsesClient

fn get_response<'life0, 'async_trait>( &'life0 self, messages: Vec<Message>, options: ChatOptions, ) -> Pin<Box<dyn Future<Output = Result<ChatResponse>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Get a complete (non-streaming) response.

fn get_streaming_response<'life0, 'async_trait>( &'life0 self, messages: Vec<Message>, options: ChatOptions, ) -> Pin<Box<dyn Future<Output = Result<ChatStream>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Get a streaming response as a sequence of updates.

fn model(&self) -> Option<&str>

The default model id for this client, if any.

impl Clone for AzureOpenAIResponsesClient

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for AzureOpenAIResponsesClient

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.