Struct BaseTokenizer

pub struct BaseTokenizer<T: Vocab> { /* private fields */ }

Base tokenizer

Base tokenizer performing:

• whitespace tokenization
• splitting on special characters
• splitting on punctuation
• splitting on CJK characters
• (optional) lower casing
• (optional) accent stripping

This tokenizer is used as a pre-tokenizer step in the BERT and GPT tokenizers.

Implementations

impl<T: Vocab + Sync> BaseTokenizer<T>

pub fn from_file_with_special_token_mapping<P: AsRef<Path>, S: AsRef<Path>>( path: P, lower_case: bool, strip_accents: bool, special_token_mapping_path: S, ) -> Result<BaseTokenizer<T>, TokenizerError>

Create a new instance of a BaseTokenizer Expects a vocabulary flat-file and special token mapping file as inputs.

Parameters

• path (&str): path to the vocabulary file (only used for special character splitting)
• lower_case (bool): flag indicating if the text should be lower-cased as part of the tokenization
• strip_accents (bool): flag indicating if accents should be stripped from the text
• special_token_mapping_path (&str): path to a special token mapping file to overwrite default special tokens

Example

use rust_tokenizers::tokenizer::{BaseTokenizer, Tokenizer};
use rust_tokenizers::vocab::BaseVocab;
let strip_accents = false;
let lower_case = false;
let tokenizer: BaseTokenizer<BaseVocab> = BaseTokenizer::from_file_with_special_token_mapping(
    "path/to/vocab/file",
    lower_case,
    strip_accents,
    "path/to/special/token/mapping/file",
)
.unwrap();

pub fn from_file<P: AsRef<Path>>( path: P, lower_case: bool, strip_accents: bool, ) -> Result<BaseTokenizer<T>, TokenizerError>

Create a new instance of a BaseTokenizer Expects a vocabulary flat-file as an input.

Parameters

• path (&str): path to the vocabulary file (only used for special character splitting)
• lower_case (bool): flag indicating if the text should be lower-cased as part of the tokenization
• strip_accents (bool): flag indicating if accents should be stripped from the text

Example

use rust_tokenizers::tokenizer::{BaseTokenizer, Tokenizer};
use rust_tokenizers::vocab::BaseVocab;
let strip_accents = false;
let lower_case = false;
let tokenizer: BaseTokenizer<BaseVocab> =
    BaseTokenizer::from_file("path/to/vocab/file", lower_case, strip_accents).unwrap();

pub fn from_existing_vocab( vocab: T, lower_case: bool, strip_accents: bool, ) -> BaseTokenizer<T>

Create a new instance of a BaseTokenizer from an existing vocabulary

Parameters

• vocab (Vocab): Thread-safe reference to a vocabulary
• lower_case (bool): flag indicating if the text should be lower-cased as part of the tokenization
• strip_accents (bool): flag indicating if accents should be stripped from the text

Example

use rust_tokenizers::tokenizer::{BaseTokenizer, Tokenizer};
use rust_tokenizers::vocab::{BaseVocab, Vocab};
let strip_accents = false;
let lower_case = false;
let base_vocab = BaseVocab::from_file("path/to/vocab/file").unwrap();

let tokenizer = BaseTokenizer::from_existing_vocab(base_vocab, lower_case, strip_accents);

Trait Implementations

impl<T: Vocab + Sync + Send> MultiThreadedTokenizer<T> for BaseTokenizer<T>

fn vocab(&self) -> &T

returns a reference to the tokenizer vocabulary

fn tokenize_list_with_offsets<S>( &self, text_list: &[S], ) -> Vec<TokensWithOffsets>

where S: AsRef<str> + Sync,

Tokenize a list of strings (with multithreading), where each corresponds to for example a sentence, returns a vector of TokensWithOffsets containing the tokens and their offset information. This calls tokenize_with_offsets on the list provided.

fn tokenize_list<S>(&self, text_list: &[S]) -> Vec<Vec<String>>

where S: AsRef<str> + Sync,

Multithreaded tokenization of a list of strings, returning tokens with offset information

fn encode_list<S>( &self, text_list: &[S], max_len: usize, truncation_strategy: &TruncationStrategy, stride: usize, ) -> Vec<TokenizedInput>

where S: AsRef<str> + Sync,

Multithreaded encoding of a sequence of string-like texts (tokenization followed by encoding). Not that in contrast with encode optional second text, each text provided is encoded independently.

fn encode_pair_list<S>( &self, text_list: &[(S, S)], max_len: usize, truncation_strategy: &TruncationStrategy, stride: usize, ) -> Vec<TokenizedInput>

where S: AsRef<str> + Sync,

Multithreaded ncoding of a sequence of string-like text pairs (tokenization followed by encoding). This combines with encode with the list processing of encode_list.

fn decode_list( &self, token_ids_list: &[Vec<i64>], skip_special_tokens: bool, clean_up_tokenization_spaces: bool, ) -> Vec<String>

Multithreaded conversion a list of sequence of ids (integer) into a string, using the tokenizer and vocabulary with options to remove special tokens and clean up tokenization spaces. This calls decode for each provided sequence of ids

impl<T: Vocab + Sync + Send> Tokenizer<T> for BaseTokenizer<T>

fn vocab(&self) -> &T

returns a reference to the tokenizer vocabulary

fn vocab_mut(&mut self) -> &mut T

returns a mutable reference to the tokenizer vocabulary

fn tokenize_to_tokens(&self, initial_token: TokenRef<'_>) -> Vec<Token>

Tokenize a TokenRef, returning a sequence of tokens

fn tokenize(&self, text: &str) -> Vec<String>

Tokenize a string, returns a vector of tokens as strings. Use tokenize_with_offsets or tokenize_to_tokens to return offset information.

fn tokenize_with_offsets(&self, text: &str) -> TokensWithOffsets

Tokenize a string, returning tokens with offset information

fn tokenize_list<S>(&self, text_list: &[S]) -> Vec<Vec<String>>

where S: AsRef<str>,

Tokenize a list of strings, returning tokens with offset information

fn tokenize_list_with_offsets<S>( &self, text_list: &[S], ) -> Vec<TokensWithOffsets>

where S: AsRef<str>,

Tokenize a list of strings, where each corresponds to for example a sentence, returns a vector of TokensWithOffsets containing the tokens and their offset information. This calls tokenize_with_offsets on the list provided.

fn convert_tokens_to_ids<S>(&self, tokens: &[S]) -> Vec<i64>

where S: AsRef<str>,

Convert a slice of string-like to a vector ot token indices

fn encode( &self, text_1: &str, text_2: Option<&str>, max_len: usize, truncation_strategy: &TruncationStrategy, stride: usize, ) -> TokenizedInput

Encode a string-like (tokenization followed by encoding)

fn encode_list<S>( &self, text_list: &[S], max_len: usize, truncation_strategy: &TruncationStrategy, stride: usize, ) -> Vec<TokenizedInput>

where S: AsRef<str>,

Encode a sequence of string-like texts (tokenization followed by encoding). Not that in contrast with encode optional second text, each text provided is encoded independently.

fn encode_pair_list<S>( &self, text_list: &[(S, S)], max_len: usize, truncation_strategy: &TruncationStrategy, stride: usize, ) -> Vec<TokenizedInput>

where S: AsRef<str>,

Encode a sequence of string-like text pairs (tokenization followed by encoding). This combines with encode with the list processing of encode_list.

fn decode_to_vec( &self, token_ids: &[i64], skip_special_tokens: bool, ) -> Vec<String>

Decode a sequence of token indices to a sequence of Strings, optionally skipping special indices

fn decode( &self, token_ids: &[i64], skip_special_tokens: bool, clean_up_tokenization_spaces: bool, ) -> String

Converts a sequence of ids (integer) into a string, using the tokenizer and vocabulary with options to remove special tokens and clean up tokenization spaces.

fn convert_tokens_to_string(&self, tokens: Vec<String>) -> String

Converts a sequence of strings into a single string. This will clean-up artifacts from tokenization (for example sub ##word) and generate a single output string

fn clean_up_tokenization(&self, input_string: String) -> String

Cleans-up tokenization artifacts (for example whitespace before punctuation)

fn decode_list( &self, token_ids_list: &[Vec<i64>], skip_special_tokens: bool, clean_up_tokenization_spaces: bool, ) -> Vec<String>

Converts a list of sequence of ids (integer) into a string, using the tokenizer and vocabulary with options to remove special tokens and clean up tokenization spaces. This calls decode for each provided sequence of ids

fn build_input_with_special_tokens( &self, tokens_ids_with_offsets_1: TokenIdsWithOffsets, tokens_ids_with_offsets_2: Option<TokenIdsWithOffsets>, ) -> TokenIdsWithSpecialTokens

Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens.

fn add_tokens(&mut self, tokens: &[&str])

Add arbitrary tokens to the vocabulary.

fn add_extra_ids(&mut self, num_extra_ids: i64)

Add arbitrary tokens to the vocabulary.