Struct ConversationTokenBufferMemory 

pub struct ConversationTokenBufferMemory { /* private fields */ }

A memory strategy that keeps messages within a token budget.

Uses a simple estimator (~4 chars per token) to approximate token counts. On load, removes the oldest messages until the total estimated tokens fit within max_tokens.

Implementations

impl ConversationTokenBufferMemory

pub fn new(store: Arc<dyn MemoryStore>, max_tokens: usize) -> Self

Create a new token buffer memory wrapping the given store.

pub fn estimate_tokens(text: &str) -> usize

Estimate the number of tokens in a text string.

Uses the simple heuristic of ~4 characters per token, with a minimum of 1.

Trait Implementations

impl MemoryStore for ConversationTokenBufferMemory

fn append<'life0, 'life1, 'async_trait>( &'life0 self, session_id: &'life1 str, message: Message, ) -> Pin<Box<dyn Future<Output = Result<(), SynapticError>> + Send + 'async_trait>>

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

fn load<'life0, 'life1, 'async_trait>( &'life0 self, session_id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Vec<Message>, SynapticError>> + Send + 'async_trait>>

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

fn clear<'life0, 'life1, 'async_trait>( &'life0 self, session_id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<(), SynapticError>> + Send + 'async_trait>>

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