wimbd 0.3.0

A CLI for inspecting and analyzing large text datasets.
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

//! Tokenizer classes and functions.

use anyhow::{anyhow, Result};
use tokenizers::tokenizer::Tokenizer;
use unicode_segmentation::UnicodeSegmentation;

/// Tokenize a string using a basic unicode tokenizer.
pub fn tokenize(s: &str) -> impl Iterator<Item = &str> {
    s.split_word_bounds().filter(|w| {
        for c in w.chars() {
            if !c.is_whitespace() {
                return true;
            }
        }
        false
    })
}

/// A wrapper class for HuggingFace tokenizers.
#[derive(Debug, Clone)]
pub struct PretrainedTokenizer(Tokenizer);

impl PretrainedTokenizer {
    pub fn tokenize(&self, text: &str) -> Result<Vec<String>> {
        Ok(self
            .0
            .encode(text, false)
            .map_err(|err| anyhow!("{}", err))?
            .get_tokens()
            .to_vec())
    }

    /// Initialize a new pretrained tokenizer from a path or identifier on HuggingFace.
    pub fn new(name: &str) -> Result<Self> {
        Ok(PretrainedTokenizer(
            Tokenizer::from_pretrained(name, None)
                .map_err(|err| anyhow!("Failed to load pretrained tokenizer {} - {}", name, err))?,
        ))
    }

    pub fn decode(&self, tokens: &[String]) -> Result<String> {
        let ids: Vec<u32> = tokens
            .iter()
            .filter_map(|t| self.0.token_to_id(t))
            .collect();
        self.0.decode(&ids, true).map_err(|err| anyhow!("{}", err))
    }
}

#[cfg(test)]
mod tests {
    use super::tokenize;
    use crate::ngrams::Ngram;

    #[test]
    fn test_tokenize_and_ngrams() {
        let s = "You can follow any responses to this entry through the RSS 2.0 feed";
        let tokens = tokenize(s).collect::<Vec<&str>>();
        assert_eq!(
            tokens,
            vec![
                "You",
                "can",
                "follow",
                "any",
                "responses",
                "to",
                "this",
                "entry",
                "through",
                "the",
                "RSS",
                "2.0",
                "feed"
            ]
        );

        let ngrams = tokenize(s).ngrams(10).collect::<Vec<Vec<&str>>>();
        assert_eq!(
            ngrams,
            vec![
                vec![
                    "You",
                    "can",
                    "follow",
                    "any",
                    "responses",
                    "to",
                    "this",
                    "entry",
                    "through",
                    "the",
                ],
                vec![
                    "can",
                    "follow",
                    "any",
                    "responses",
                    "to",
                    "this",
                    "entry",
                    "through",
                    "the",
                    "RSS",
                ],
                vec![
                    "follow",
                    "any",
                    "responses",
                    "to",
                    "this",
                    "entry",
                    "through",
                    "the",
                    "RSS",
                    "2.0",
                ],
                vec![
                    "any",
                    "responses",
                    "to",
                    "this",
                    "entry",
                    "through",
                    "the",
                    "RSS",
                    "2.0",
                    "feed",
                ],
            ]
        );
    }
}