Struct MultimodalChunker 

pub struct MultimodalChunker { /* private fields */ }

Turns ordered MediaBlocks into modality-tagged Documents.

Implementations

impl MultimodalChunker

pub fn new() -> Self

Creates a chunker with default config (one document per text block).

pub fn with_config(config: MultimodalChunkConfig) -> Self

Creates a chunker with explicit config.

pub fn chunk(&self, blocks: &[MediaBlock]) -> Vec<Document>

Chunks the blocks into Documents, preserving input order.

• Empty/whitespace text blocks are skipped;
• text blocks are optionally split into UTF-8-safe windows;
• images always become one document each; an image without a caption still embeds (its content is empty), but RetrieverError is never raised at chunk time — validation happens at embedding time.

pub fn text_document(&self, content: impl Into<String>) -> Document

Builds a tagged text document directly.

pub fn image_document(&self, asset: &ImageAsset) -> Document

Builds a tagged image document directly.

Trait Implementations

impl Clone for MultimodalChunker

fn clone(&self) -> MultimodalChunker

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MultimodalChunker

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

Formats the value using the given formatter.

impl Default for MultimodalChunker

fn default() -> MultimodalChunker

Returns the “default value” for a type.