vibrato 0.5.2

Vibrato: viterbi-based accelerated tokenizer
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242

use std::io::BufRead;

use crate::trainer::{Corpus, Trainer, TrainerConfig};
use crate::utils;

const TRAIN_LEX_CSV: &[u8] = include_bytes!("./resources/train_lex.csv");
const CHAR_DEF: &[u8] = include_bytes!("./resources/char.def");
const TRAIN_UNK_DEF: &[u8] = include_bytes!("./resources/train_unk.def");
const REWRITE_DEF: &[u8] = include_bytes!("./resources/rewrite.def");
const FEATURE_DEF: &[u8] = include_bytes!("./resources/feature.def");
const CORPUS_TXT: &[u8] = include_bytes!("./resources/corpus.txt");
const USER_CSV: &[u8] = include_bytes!("./resources/user.csv");

#[test]
fn test_lexicon_format() {
    let config = TrainerConfig::from_readers(
        TRAIN_LEX_CSV,
        CHAR_DEF,
        TRAIN_UNK_DEF,
        FEATURE_DEF,
        REWRITE_DEF,
    )
    .unwrap();
    let corpus = Corpus::from_reader(CORPUS_TXT).unwrap();
    let trainer = Trainer::new(config).unwrap().max_iter(5);

    let mut lex = vec![];
    let mut matrix = vec![];
    let mut unk = vec![];
    let mut user_lex = vec![];
    let mut model = trainer.train(corpus).unwrap();
    model
        .write_dictionary(&mut lex, &mut matrix, &mut unk, &mut user_lex)
        .unwrap();

    // Retrieves the number of right and left connection IDs.
    let (right_len, left_len) = {
        let line = matrix.lines().next().unwrap().unwrap();
        let mut spl = line.split(' ');
        let right_len = spl.next().unwrap().parse::<usize>().unwrap();
        let left_len = spl.next().unwrap().parse::<usize>().unwrap();
        (right_len, left_len)
    };

    let seed_lex_lines: Vec<String> = TRAIN_LEX_CSV.lines().map(|line| line.unwrap()).collect();
    let result_lex_lines: Vec<String> = lex.lines().map(|line| line.unwrap()).collect();

    // Checks the number of lines
    assert_eq!(result_lex_lines.len(), 25);

    // The expected content of the lex file is too long to write here.
    for i in 0..25 {
        let seed_row = utils::parse_csv_row(&seed_lex_lines[i]);
        let result_row = utils::parse_csv_row(&result_lex_lines[i]);

        // surface
        assert_eq!(seed_row[0], result_row[0]);

        // scores
        assert!(result_row[1].parse::<usize>().unwrap() < left_len);
        assert!(result_row[2].parse::<usize>().unwrap() < right_len);
        assert!(result_row[3].parse::<i16>().is_ok());

        // features
        assert_eq!(seed_row[4..], result_row[4..]);
    }
}

#[test]
fn test_unk_format() {
    let config = TrainerConfig::from_readers(
        TRAIN_LEX_CSV,
        CHAR_DEF,
        TRAIN_UNK_DEF,
        FEATURE_DEF,
        REWRITE_DEF,
    )
    .unwrap();
    let corpus = Corpus::from_reader(CORPUS_TXT).unwrap();
    let trainer = Trainer::new(config).unwrap().max_iter(5);

    let mut lex = vec![];
    let mut matrix = vec![];
    let mut unk = vec![];
    let mut user_lex = vec![];
    let mut model = trainer.train(corpus).unwrap();
    model
        .write_dictionary(&mut lex, &mut matrix, &mut unk, &mut user_lex)
        .unwrap();

    // Retrieves the number of right and left connection IDs.
    let (right_len, left_len) = {
        let line = matrix.lines().next().unwrap().unwrap();
        let mut spl = line.split(' ');
        let right_len = spl.next().unwrap().parse::<usize>().unwrap();
        let left_len = spl.next().unwrap().parse::<usize>().unwrap();
        (right_len, left_len)
    };

    let result_unk_lines: Vec<String> = unk.lines().map(|line| line.unwrap()).collect();

    // Checks the number of lines
    assert_eq!(result_unk_lines.len(), 4);

    {
        let result_row = utils::parse_csv_row(&result_unk_lines[0]);
        assert_eq!(result_row[0], "DEFAULT");
        assert!(result_row[1].parse::<usize>().unwrap() < left_len);
        assert!(result_row[2].parse::<usize>().unwrap() < right_len);
        assert!(result_row[3].parse::<i16>().is_ok());
        assert_eq!(result_row[4..], ["補助記号", "一般", "*", "*"]);
    }
    // ALPHA is defined earlier than KANJI in unk.def, but KANJI is defined earlier than ALPHA
    // in char.def.
    // The trainer sorts results in the order defined in char.def.
    {
        let result_row = utils::parse_csv_row(&result_unk_lines[1]);
        assert_eq!(result_row[0], "KANJI");
        assert!(result_row[1].parse::<usize>().unwrap() < left_len);
        assert!(result_row[2].parse::<usize>().unwrap() < right_len);
        assert!(result_row[3].parse::<i16>().is_ok());
        assert_eq!(result_row[4..], ["名詞", "普通名詞", "一般", "*"]);
    }
    {
        let result_row = utils::parse_csv_row(&result_unk_lines[2]);
        assert_eq!(result_row[0], "ALPHA");
        assert!(result_row[1].parse::<usize>().unwrap() < left_len);
        assert!(result_row[2].parse::<usize>().unwrap() < right_len);
        assert!(result_row[3].parse::<i16>().is_ok());
        assert_eq!(result_row[4..], ["名詞", "普通名詞", "一般", "*"]);
    }
    {
        let result_row = utils::parse_csv_row(&result_unk_lines[3]);
        assert_eq!(result_row[0], "KANJINUMERIC");
        assert!(result_row[1].parse::<usize>().unwrap() < left_len);
        assert!(result_row[2].parse::<usize>().unwrap() < right_len);
        assert!(result_row[3].parse::<i16>().is_ok());
        assert_eq!(result_row[4..], ["名詞", "数", "*", "*"]);
    }
}

#[test]
fn test_matrix_format() {
    let config = TrainerConfig::from_readers(
        TRAIN_LEX_CSV,
        CHAR_DEF,
        TRAIN_UNK_DEF,
        FEATURE_DEF,
        REWRITE_DEF,
    )
    .unwrap();
    let corpus = Corpus::from_reader(CORPUS_TXT).unwrap();
    let trainer = Trainer::new(config).unwrap().max_iter(5);

    let mut lex = vec![];
    let mut matrix = vec![];
    let mut unk = vec![];
    let mut user_lex = vec![];
    let mut model = trainer.train(corpus).unwrap();
    model
        .write_dictionary(&mut lex, &mut matrix, &mut unk, &mut user_lex)
        .unwrap();

    let mut matrix_it = matrix.lines();

    let firstline = matrix_it.next().unwrap().unwrap();

    // Retrieves the number of right and left connection IDs.
    let (right_len, left_len) = {
        let mut spl = firstline.split(' ');
        let right_len = spl.next().unwrap().parse::<usize>().unwrap();
        let left_len = spl.next().unwrap().parse::<usize>().unwrap();
        assert!(spl.next().is_none());
        (right_len, left_len)
    };

    for line in matrix_it {
        let line = line.unwrap();
        let mut spl = line.split(' ');
        // right ID, left ID, weight
        assert!(spl.next().unwrap().parse::<usize>().unwrap() < right_len);
        assert!(spl.next().unwrap().parse::<usize>().unwrap() < left_len);
        assert!(spl.next().unwrap().parse::<i16>().is_ok());
        assert!(spl.next().is_none());
    }
}

#[test]
fn test_user_lex_format() {
    let config = TrainerConfig::from_readers(
        TRAIN_LEX_CSV,
        CHAR_DEF,
        TRAIN_UNK_DEF,
        FEATURE_DEF,
        REWRITE_DEF,
    )
    .unwrap();
    let corpus = Corpus::from_reader(CORPUS_TXT).unwrap();
    let trainer = Trainer::new(config).unwrap().max_iter(5);

    let mut lex = vec![];
    let mut matrix = vec![];
    let mut unk = vec![];
    let mut user_lex = vec![];
    let mut model = trainer.train(corpus).unwrap();

    model.read_user_lexicon(USER_CSV).unwrap();

    model
        .write_dictionary(&mut lex, &mut matrix, &mut unk, &mut user_lex)
        .unwrap();

    let result_user_lines: Vec<String> = user_lex.lines().map(|line| line.unwrap()).collect();

    // Checks the number of lines
    assert_eq!(result_user_lines.len(), 3);

    {
        let result_row = utils::parse_csv_row(&result_user_lines[0]);
        assert_eq!(result_row[0], "京都東京都");
        assert_eq!(result_row[1].parse::<usize>().unwrap(), 6);
        assert_eq!(result_row[2].parse::<usize>().unwrap(), 8);
        assert_eq!(result_row[3].parse::<i16>().unwrap(), -1000);
        assert_eq!(result_row[4..], ["カスタム名詞"]);
    }
    {
        let result_row = utils::parse_csv_row(&result_user_lines[1]);
        assert_eq!(result_row[0], "kampersanda");
        assert_eq!(result_row[1].parse::<usize>().unwrap(), 7);
        assert_eq!(result_row[2].parse::<usize>().unwrap(), 7);
        assert_eq!(result_row[3].parse::<i16>().unwrap(), -2000);
        assert_eq!(result_row[4..], ["カスタム名詞"]);
    }
    {
        let result_row = utils::parse_csv_row(&result_user_lines[2]);
        assert_eq!(result_row[0], "ヴェネツィア");
        assert_ne!(result_row[1].parse::<usize>().unwrap(), 0);
        assert_ne!(result_row[2].parse::<usize>().unwrap(), 0);
        assert!(result_row[3].parse::<i16>().is_ok());
        assert_eq!(result_row[4..], ["名詞", "固有名詞", "地名", "一般"]);
    }
}