* Manual trigger for character consistency processing

* Automatically wrap existing tag libraries on load (if not already wrapped)

* Section-based chunking with intelligent subsection handling * This is the OLD chunkFullsheet logic, now renamed for clarity

* Delete an entire collection from the vector database

* Finalize chunks and save to library

* Get all lorebook connections for a character * @param {number} chid - Character ID * @returns {Array} Array of connection objects

* Get the stem/root of a word by stripping common suffixes * Examples: "psychological" -> "psych", "psychology" -> "psych", "psyche" -> "psych"

* Initialize the RAG system