* Add RAG button to a message containing a fullsheet * * @param {number} messageId - Message ID

* Add RAG buttons to all existing messages

* Delete specific hashes from a vector collection

* Get saved hashes for a collection (checks if collection exists)

* Insert vector items into a collection

* Apply automatic cross-section linking based on keyword frequency thresholds * @param {Array} chunks - Array of chunks to process

* Create a backup of a lorebook before modification

* Split fullsheet content into semantic chunks respecting SECTION 1-8 structure. * @param {string} content * @param {string} characterName * @pa

* Extract keywords using hybrid approach: * 1. Title/topic words (language-agnostic) * 2. Frequency analysis (language-agnostic) * 3. Semantic m

* Generate chunk previews based on current settings

* Check if a string is a valid regex (using ST's official validator)

* Vectorize a fullsheet from message content * * @param {string} characterName - Character name * @param {string} content - Fullsheet content

* Delete an entire vector collection

* Query a vector collection

* Attach handlers for chunk-specific interactions

* Attach event handlers to modal elements

* Auto-vectorize a message if it contains a fullsheet * * @param {number} messageId - Message ID

* Build linked chunks section for a chunk

* Build query context from recent chat messages * * @param {number} messageCount - Number of recent messages to include * @returns {string} Quer

* Calculate keyword weight boost for a chunk based on query keywords * This applies custom weights from the visualizer to boost semantic scores *