Struct DeepSeekProvider 

pub struct DeepSeekProvider { /* private fields */ }

Implementations

impl DeepSeekProvider

pub fn new( api_key: String, model: Option<String>, thinking: ThinkingMode, ) -> Self

Trait Implementations

impl LlmProvider for DeepSeekProvider

fn chat<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, messages: &'life1 [Message], tools: &'life2 [ToolDef], ) -> Pin<Box<dyn Future<Output = Result<LlmResponse>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Send messages + tool definitions, receive a response.

fn chat_streaming<'life0, 'life1, 'life2, 'life3, 'async_trait>( &'life0 self, messages: &'life1 [Message], tools: &'life2 [ToolDef], token_tx: &'life3 UnboundedSender<String>, ) -> Pin<Box<dyn Future<Output = Result<LlmResponse>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait,

Stream text tokens via token_tx while accumulating the full response. The default falls back to chat() and delivers the entire text as one chunk.

fn supports_streaming(&self) -> bool

Whether this provider supports true token-by-token streaming. Providers that override chat_streaming with SSE should return true.

fn name(&self) -> &str

Human-readable name for this provider.

fn context_window(&self) -> u64

Context window size in tokens.

fn default_model(&self) -> &str

Default model name.