Struct ChatMemorySession 

pub struct ChatMemorySession { /* private fields */ }

High-level session helper for LLM chat comparison patterns.

Groups conversation turns under a single Dakera session so that:

• Every stored message is associated with session_id for scoped retrieval.
• recall queries the agent’s full memory — not just this session — so prior conversations inform the current exchange.

Create via ChatMemorySession::create; close via ChatMemorySession::close.

Implementations

impl ChatMemorySession

pub async fn create( client: Arc<DakeraClient>, agent_id: impl Into<String>, ) -> Result<ChatMemorySession>

Create a new Dakera session and return a ChatMemorySession.

Arguments

• client – Shared DakeraClient instance.
• agent_id – Identifier for the agent whose memory to use.

pub async fn create_with_metadata( client: Arc<DakeraClient>, agent_id: impl Into<String>, metadata: Value, ) -> Result<ChatMemorySession>

Create a session with attached metadata.

pub async fn store( &self, role: &str, content: &str, ) -> Result<StoreMemoryResponse>

Store a conversation turn in the session with default importance (0.6).

The role (e.g. "user" or "assistant") is appended to the memory’s tags automatically.

pub async fn store_with_opts( &self, role: &str, content: &str, importance: f32, extra_tags: &[&str], ) -> Result<StoreMemoryResponse>

Store a conversation turn with custom importance and additional tags.

pub async fn recall(&self, query: &str) -> Result<Vec<RecalledMemory>>

Recall up to 5 memories relevant to query for this agent.

Searches the agent’s full memory, not just the current session, so context from prior conversations is surfaced when relevant.

pub async fn recall_top_k( &self, query: &str, top_k: usize, ) -> Result<Vec<RecalledMemory>>

Recall up to top_k memories relevant to query.

pub async fn close(self) -> Result<()>

End the Dakera session.

pub fn session_id(&self) -> &str

The underlying Dakera session ID.

pub fn agent_id(&self) -> &str

The agent ID this session is bound to.