Enum ChunkingStrategy 

pub enum ChunkingStrategy {
    NoChunking,
    ChunkAndAverage {
        chunk_size: usize,
        chunk_overlap: usize,
    },
    UsePrefix(usize),
    UseSuffix(usize),
}

Text chunk-based embedding strategy

Defines how to handle long text embeddings: splitting, averaging, using specific parts, etc.

Variants

NoChunking

Maximize use of model context window, no chunking needed

ChunkAndAverage

Split long text into chunks, generate embeddings for each chunk, then average

Fields

chunk_size: usize

chunk_overlap: usize

UsePrefix(usize)

Only use the first N tokens of the text

UseSuffix(usize)

Only use the last N tokens of the text

Trait Implementations

impl Default for ChunkingStrategy

fn default() -> ChunkingStrategy

Returns the “default value” for a type.