use crate::error::TokenizerError;
use crate::vocab::base_vocab::swap_key_values;
use crate::vocab::Vocab;
use std::collections::HashMap;
use std::fs::File;
use std::io::BufReader;
#[derive(Debug, Clone)]
pub struct MarianVocab {
pub values: HashMap<String, i64>,
pub indices: HashMap<i64, String>,
pub unknown_value: &'static str,
pub special_values: HashMap<String, i64>,
pub special_indices: HashMap<i64, String>,
}
impl MarianVocab {
pub fn pad_value() -> &'static str {
"<pad>"
}
pub fn eos_value() -> &'static str {
"</s>"
}
}
impl Vocab for MarianVocab {
fn unknown_value() -> &'static str {
"<unk>"
}
fn get_unknown_value(&self) -> &'static str {
"<unk>"
}
fn values(&self) -> &HashMap<String, i64> {
&self.values
}
fn indices(&self) -> &HashMap<i64, String> {
&self.indices
}
fn special_values(&self) -> &HashMap<String, i64> {
&self.special_values
}
fn special_indices(&self) -> &HashMap<i64, String> {
&self.special_indices
}
fn from_file(path: &str) -> Result<MarianVocab, TokenizerError> {
let f = File::open(path).map_err(|e| {
TokenizerError::FileNotFound(format!("{} vocabulary file not found :{}", path, e))
})?;
let br = BufReader::new(f);
let values: HashMap<String, i64> = match serde_json::from_reader(br) {
Ok(value) => value,
Err(e) => {
return Err(TokenizerError::VocabularyParsingError(e.to_string()));
}
};
let mut special_values = HashMap::new();
let unknown_value = MarianVocab::unknown_value();
MarianVocab::_register_as_special_value(unknown_value, &values, &mut special_values)?;
let pad_value = MarianVocab::pad_value();
MarianVocab::_register_as_special_value(pad_value, &values, &mut special_values)?;
let eos_value = MarianVocab::eos_value();
MarianVocab::_register_as_special_value(eos_value, &values, &mut special_values)?;
let indices = swap_key_values(&values);
let special_indices = swap_key_values(&special_values);
Ok(MarianVocab {
values,
indices,
unknown_value,
special_values,
special_indices,
})
}
fn token_to_id(&self, token: &str) -> i64 {
self._token_to_id(
token,
&self.values,
&self.special_values,
self.unknown_value,
)
}
fn id_to_token(&self, id: &i64) -> String {
self._id_to_token(id, &self.indices, &self.special_indices, self.unknown_value)
}
}