Struct GeminiAdapter 

pub struct GeminiAdapter { /* private fields */ }

Google Gemini独立适配器，支持多模态AI服务

Google Gemini independent adapter for multimodal AI service

Gemini API is completely different from OpenAI format, requires independent adapter:

• Endpoint: /v1beta/models/{model}:generateContent
• Request body: contents array instead of messages
• Response: candidates[0].content.parts[0].text
• Authentication: URL parameter ?key=<API_KEY>

Implementations

impl GeminiAdapter

pub fn new() -> Result<Self, AiLibError>

pub fn with_transport_ref( transport: DynHttpTransportRef, api_key: String, base_url: String, ) -> Result<Self, AiLibError>

Construct using object-safe transport reference

pub fn with_transport_ref_and_metrics( transport: DynHttpTransportRef, api_key: String, base_url: String, metrics: Arc<dyn Metrics>, ) -> Result<Self, AiLibError>

Construct with an injected transport and metrics implementation

Trait Implementations

impl ChatApi for GeminiAdapter

fn chat_completion<'life0, 'async_trait>( &'life0 self, request: ChatCompletionRequest, ) -> Pin<Box<dyn Future<Output = Result<ChatCompletionResponse, AiLibError>> + Send + 'async_trait>>

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

Send chat completion request

fn chat_completion_stream<'life0, 'async_trait>( &'life0 self, _request: ChatCompletionRequest, ) -> Pin<Box<dyn Future<Output = Result<Box<dyn Stream<Item = Result<ChatCompletionChunk, AiLibError>> + Send + Unpin>, AiLibError>> + Send + 'async_trait>>

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

Streaming chat completion request

fn list_models<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<Vec<String>, AiLibError>> + Send + 'async_trait>>

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

Get list of supported models

fn get_model_info<'life0, 'life1, 'async_trait>( &'life0 self, model_id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<ModelInfo, AiLibError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Get model information

fn chat_completion_batch<'life0, 'async_trait>( &'life0 self, requests: Vec<ChatCompletionRequest>, concurrency_limit: Option<usize>, ) -> Pin<Box<dyn Future<Output = Result<Vec<Result<ChatCompletionResponse, AiLibError>>, AiLibError>> + Send + 'async_trait>>

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

批处理聊天完成请求