Struct ConversationContext 

pub struct ConversationContext {
    pub config: ContextConfig,
    pub messages: VecDeque<Message>,
    pub total_tokens: usize,
}

Multi-turn conversation context with automatic token budget enforcement.

Fields

config: ContextConfig

Configuration (strategy, budgets, system prompt).

messages: VecDeque<Message>

Ordered message history (front = oldest).

total_tokens: usize

Running total of estimated tokens across all messages.

Implementations

impl ConversationContext

pub fn new(config: ContextConfig) -> Self

Create a new context. If the config has a system_prompt, it is added as the first message immediately.

pub fn add_message(&mut self, role: Role, content: &str)

Append a new message and enforce the token budget.

pub fn enforce_budget(&mut self)

Apply the configured TruncationStrategy if the effective budget is exceeded.

pub fn messages_for_api(&self) -> Vec<&Message>

Return the messages that should be sent to the LLM API.

pub fn token_utilization(&self) -> f64

Token utilisation ratio in [0.0, 1.0].

pub fn summary(&self) -> ContextSummary

Return a lightweight summary snapshot.