Struct llm_weaver::TapestryFragment

pub struct TapestryFragment<T: Config> {
    pub context_tokens: <T::PromptModel as Llm<T>>::Tokens,
    pub context_messages: Vec<ContextMessage<T>>,
}

Represents a single part of a conversation containing a list of messages along with other metadata.

LLM can only hold a limited amount of tokens in a the entire message history/context. The total number of context_tokens is tracked when Loom::weave is executed and if it exceeds the maximum number of tokens allowed for the current GPT Config::PromptModel, then a summary is generated and a new TapestryFragment instance is created.

Fields

context_tokens: <T::PromptModel as Llm<T>>::Tokens

Total number of GPT tokens in the context_messages.

context_messages: Vec<ContextMessage<T>>

List of ContextMessages that represents the message history.

Trait Implementations

impl<T: Clone + Config> Clone for TapestryFragment<T>

where T::PromptModel: Clone,

fn clone(&self) -> TapestryFragment<T>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<T: Debug + Config> Debug for TapestryFragment<T>

where T::PromptModel: Debug,

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<T: Default + Config> Default for TapestryFragment<T>

where T::PromptModel: Default,

fn default() -> TapestryFragment<T>

Returns the “default value” for a type.

impl<T> Serialize for TapestryFragment<T>

where T: Serialize + Config,

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.