Skip to main content

ChunkSizer

Trait ChunkSizer 

Source
pub trait ChunkSizer {
    // Required method
    fn size(&self, chunk: &str) -> usize;
}
Expand description

Determines the size of a given chunk.

Required Methods§

Source

fn size(&self, chunk: &str) -> usize

Determine the size of a given chunk to use for validation

Dyn Compatibility§

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementations on Foreign Types§

Source§

impl ChunkSizer for CoreBPE

Available on crate feature tiktoken-rs only.
Source§

fn size(&self, chunk: &str) -> usize

Returns the number of tokens in a given text after tokenization.

Source§

impl ChunkSizer for Tokenizer

Available on crate feature tokenizers only.
Source§

fn size(&self, chunk: &str) -> usize

Returns the number of tokens in a given text after tokenization.

If the tokenizer has truncation enabled, the returned size may be truncated by the tokenizer. Disable truncation on the tokenizer before constructing a splitter if you need chunk sizes to reflect the full input text.

§Panics

Will panic if you don’t have a byte-level tokenizer and the splitter encounters text it can’t tokenize.

Source§

impl<T> ChunkSizer for &T
where T: ChunkSizer,

Source§

fn size(&self, chunk: &str) -> usize

Source§

impl<T> ChunkSizer for Arc<T>
where T: ChunkSizer,

Source§

fn size(&self, chunk: &str) -> usize

Source§

impl<T> ChunkSizer for Box<T>
where T: ChunkSizer,

Source§

fn size(&self, chunk: &str) -> usize

Source§

impl<T> ChunkSizer for Cow<'_, T>
where T: ChunkSizer + ToOwned + ?Sized, <T as ToOwned>::Owned: ChunkSizer,

Source§

fn size(&self, chunk: &str) -> usize

Source§

impl<T> ChunkSizer for Rc<T>
where T: ChunkSizer,

Source§

fn size(&self, chunk: &str) -> usize

Source§

impl<T> ChunkSizer for Ref<'_, T>
where T: ChunkSizer,

Source§

fn size(&self, chunk: &str) -> usize

Source§

impl<T> ChunkSizer for RefMut<'_, T>
where T: ChunkSizer,

Source§

fn size(&self, chunk: &str) -> usize

Implementors§