Struct vaporetto::Dataset

pub struct Dataset<'a> { /* fields omitted */ }

This is supported on crate feature train only.

Dataset manager.

Implementations

impl<'a> Dataset<'a>

pub fn new<D, P>(
    char_ngram_size: usize,
    char_window_size: usize,
    type_ngram_size: usize,
    type_window_size: usize,
    dictionary: D,
    dict_word_max_size: usize
) -> Result<Self> where
    D: AsRef<[P]>,
    P: AsRef<[u8]> + AsRef<str>, 

Creates a new dataset manager.

Arguments

• char_ngram_size - The character n-gram length.
• char_window_size - The character window size.
• type_ngram_size - The character type n-gram length.
• type_window_size - The character type window size.
• dictionary - A word dictionary (must be alphabetical ascending order).
• dict_word_max_size - Dictionary words greater than this value will be grouped together.

Returns

A dataset manager.

Errors

If invalid parameters are given, an error variant will be returned.

pub fn push_sentence(&mut self, s: &'a Sentence)

Adds a sentence to the dataset.

Arguments

• s - A sentence.

pub fn n_features(&self) -> usize

Gets the number of features.

Returns

The number of features.