* Traverse the AABB tree for the closest hit * @param {Ray} ray - The ray to cast * @param {number} node_index - Index of the current node

* Traverse the AABB tree for all hits * @param {Ray} ray - The ray to cast * @param {number} node_index - Index of the current node * @

* Handles keyboard events from UIContext instead of document events. * This method should be called once per frame (e.g., in update()).

* Update the maximum depth of the tree * @param {number} start_node_index - The index of the node to start from

* Uses a free block for allocation, potentially splitting it if it's larger than needed. * @param {number} page_index - The index of the page conta

* Add a coordinate axes visualization * @param {Array} origin - Origin position [x, y, z] * @param {number} size - Size of the axes * @param

* Add a circle outline * @param {Array} center - Circle center [x, y, z] * @param {number} radius - Circle radius * @param {Array} normal - C

* Add gradients together * * @param {Object} grad1 - First gradient * @param {Object} grad2 - Second gradient * @returns {Object} Combined

* Add multiple child nodes to a parent node. * @param {*} parent_data - The data of the parent node to add to. If null, creates roots. * @param

* Adds a training batch for a model running in batch training mode. * @param {string} model_id - The model to which the training batch should be ad

* Adds an element to the end of the allocator * @param {*} value - The value to add * @returns {number} The index of the added element

* Automatically loads all fonts from registered font paths by scanning their font manifests. * Iterates through each registered font path and calls

* Begins a new frame in the render graph. * This method resets the render graph state and prepares it for a new frame of rendering. * * @retu

* Calculate a fat margin for an AABB * @param {Array<number>} min_point - The minimum point of the AABB * @param {Array<number>} max_point - The

* Find space or create new chunks * @param {number} [instance_count=1] - Number of instances to allocate * @returns {Object} - Object containing

* Optionally, clean up any non-persistent tensors in the entire pool. * This walks through the pool and disposes tensors not marked persistent.

* Clears the dirty flag for this chunk. Called after flushing.

* Clears all dirty flags for all chunks.

* Combine multiple inputs for a layer * * @param {Layer} layer - The layer * @param {Array} inputs - The inputs to combine * @returns {Obj

* Compile all enabled post-process passes for a specific view into the render graph

* Check if the queue contains a value. * @param {number} value - The value to check for. * @returns {boolean} True if the value is in the queue,

* Utility to copy a tensor. * @param {Tensor} a - the tensor to copy. * @param {number} offset - the offset to copy the tensor from. * @param

* Copies the GPU buffers to the CPU. * @param {GPUCommandEncoder} encoder - The command encoder.

* Enqueue GPU→CPU copies for all CPU-readback buffers via the given encoder.

* Deallocates an object from the free list allocator. * @param {object} object - The object to deallocate * @throws {Error} If the object is

* Optional compaction when holes accumulate

* Destroy an entity and release its resources. * * @param {EntityHandle} handle - The handle of the entity to destroy

* Disconnects a layer from all its parents and children * * @param {number} id - The ID of the layer * @returns {boolean} True if disconnecti

* End the current collection and return to the default collection

* Toggles the display of the ML stats panel.

* Flushes the GPU buffers.