saku 0.1.6

A simple yet efficient rule-based Japanese Sentence 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

use crate::{CharTable, ControlFlow, SentenceTokenizerBuilder};
use std::borrow::Cow;

#[derive(Clone, Debug)]
pub struct SentenceTokenizer {
    pub eos: char,
    pub(crate) eos_size: usize,
    pub(crate) num_parens: u8,
    pub(crate) char_table: CharTable,
}

impl Default for SentenceTokenizer {
    #[inline(always)]
    fn default() -> Self {
        SentenceTokenizerBuilder::new().build()
    }
}

impl SentenceTokenizer {
    #[inline(always)]
    fn char_to_control_flow(&self, ch: &char) -> Option<&ControlFlow> {
        // self.chmap.get(ch)
        self.char_table.get(*ch)
    }

    #[inline(always)]
    pub fn tokenize<'a>(&self, document: &'a str) -> Vec<Cow<'a, str>> {
        self.tokenize_ignore_line_breaks(document)
    }

    #[inline(always)]
    pub fn process_line_breaks<'a>(
        &self,
        sentence: &mut String,
        start: &mut usize,
        i: usize,
        document: &'a str,
    ) {
        sentence.push_str(&document[*start..i]);
        *start = i + 1;
    }

    #[inline(always)]
    pub fn process_left_parens<'a>(
        &self,
        flags: &mut Vec<u8>,
        nest_count: &mut u8,
        flag_id: usize,
    ) {
        flags[flag_id] += 1;
        *nest_count += 1;
    }

    #[inline(always)]
    pub fn process_right_parens<'a>(
        &self,
        flags: &mut Vec<u8>,
        nest_count: &mut u8,
        flag_id: usize,
    ) {
        if flags[flag_id] > 0 {
            flags[flag_id] -= 1;
            *nest_count -= 1;
        }
    }

    #[inline(always)]
    pub fn tokenize_ignore_line_breaks<'a>(&self, document: &'a str) -> Vec<Cow<'a, str>> {
        let document: &'a str = document.trim();
        let mut start: usize = 0;
        let mut sentences: Vec<Cow<'a, str>> = Vec::new();
        let mut sentence: String = String::new();
        let mut flags: Vec<u8> = vec![0; self.num_parens as usize];
        let mut nest_count: u8 = 0;

        for (i, ch) in document.char_indices() {
            if let Some(flow) = self.char_to_control_flow(&ch) {
                match *flow {
                    ControlFlow::LineBreaks => {
                        self.process_line_breaks(&mut sentence, &mut start, i, &document)
                    }
                    ControlFlow::LeftParens(flag_id) => {
                        self.process_left_parens(&mut flags, &mut nest_count, flag_id as usize)
                    }
                    ControlFlow::RightParens(flag_id) => {
                        self.process_right_parens(&mut flags, &mut nest_count, flag_id as usize)
                    }
                    ControlFlow::Eos => {
                        if nest_count > 0 {
                            continue;
                        }
                        if sentence.is_empty() {
                            sentences.push(Cow::Borrowed(&document[start..i + self.eos_size]));
                        } else {
                            sentence.push_str(&document[start..i + self.eos_size]);
                            sentences.push(Cow::Owned(sentence));
                            sentence = String::new();
                        }
                        start = i + self.eos_size;
                    }
                }
            }
        }
        if start < document.len() {
            if sentence.is_empty() {
                sentences.push(Cow::Borrowed(&document[start..]));
            } else {
                sentence.push_str(&document[start..]);
                sentences.push(Cow::Owned(sentence));
            }
        }

        sentences
    }

    // This function returns a vector of references of sentences, considering line breaks as the terminator.
    // `tokenize` and `tokenize_raw` have a different return type to improve performance.
    #[inline]
    pub fn tokenize_raw<'a>(&self, document: &'a str) -> Vec<&'a str> {
        let document = document.trim();
        let mut start: usize = 0;
        let mut sentences: Vec<&'a str> = Vec::new();
        let mut flags: Vec<u8> = vec![0; self.num_parens as usize];
        let mut nest_count: u8 = 0;

        for (i, ch) in document.char_indices() {
            if let Some(flow) = self.char_to_control_flow(&ch) {
                match *flow {
                    ControlFlow::LineBreaks => {
                        if i != start {
                            sentences.push(&document[start..i]);
                        }
                        nest_count = 0;
                        start = i + 1;
                    }
                    ControlFlow::LeftParens(flag_id) => {
                        self.process_left_parens(&mut flags, &mut nest_count, flag_id as usize)
                    }
                    ControlFlow::RightParens(flag_id) => {
                        self.process_right_parens(&mut flags, &mut nest_count, flag_id as usize)
                    }
                    ControlFlow::Eos => {
                        if nest_count > 0 {
                            continue;
                        }
                        sentences.push(&document[start..i + self.eos_size]);
                        start = i + self.eos_size;
                    }
                }
            }
        }
        if start < document.len() {
            sentences.push(&document[start..]);
        }

        sentences
    }
}