Module completions 

Vector completion client and types. Vector completion client and supporting types.

This module provides the client for creating vector completions, which orchestrate multiple LLM chat completions for voting on response options.

Modules

cache

Vote caching client for the global ObjectiveAI cache. Vote cache retrieval client.

cache_vote_fetcher

Fetcher for retrieving votes from the global cache. Fetcher for votes from the global ObjectiveAI cache.

completion_votes_fetcher

Fetcher for retrieving votes from historical completions. Fetcher for votes from historical vector completions.

usage_handler

Usage tracking for vector completions. Usage tracking for vector completions.

vector_responses

Vector response transformation utilities. Vector response transformation for LLM prompts.

Structs

Client

Client for creating vector completions.

PfxData

Prefix data for a specific LLM, including tree and regex patterns.

ResponseKey

Parsed response key from LLM structured output.

Enums

Error

Errors that can occur during vector completion creation.

Pfx

Single-character prefix labels A-T.

PfxTree

A tree structure for generating unique prefix keys.

Functions

get_vote

Extracts a vote from an LLM choice.

response_id

Generates a unique response ID for a vector completion.