khmercut 0.1.5

A blazingly fast Khmer word segmentation tool written in Rust.
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

use crfs::{Attribute, Model};
use regex::Regex;
use std::vec;

pub fn text_tagger(input_str: &String) -> Vec<(&str, String)> {
    let grapheme_raw_regex =
        r"([\u1780-\u17FF](\u17d2[\u1780-\u17FF]|[\u17B6-\u17D1\u17D3\u17DD])*)";
    let grapheme_regex: Regex = Regex::new(grapheme_raw_regex).unwrap();
    let characters_regex = Regex::new(
        r"([\u1780-\u17d3]+)|([\u17d4-\u17dd]+)|([\u17e0-\u17e9]+)|(\s+)|([^\u1780-\u17ff\s]+)",
    )
    .unwrap();

    let mut outputs = vec![];
    for capture in characters_regex.captures_iter(&input_str) {
        let (data_type, text) = capture
            .iter()
            .enumerate()
            .skip(1)
            .find(|t| t.1.is_some())
            .map(|t| (t.0, t.1.unwrap().as_str()))
            .unwrap();

        if data_type != 1 {
            outputs.push((text, "NS".to_string()));
            continue;
        }

        for match_item in grapheme_regex.find_iter(text) {
            let grapheme = match_item.as_str();
            let size = grapheme.chars().count();
            let grapheme_type: String = if size == 1 {
                "C".to_string()
            } else {
                format!("K{}", size)
            };

            outputs.push((grapheme, grapheme_type));
        }
    }

    return outputs;
}

pub fn create_features(kccs: &Vec<(&str, String)>) -> Vec<Vec<Attribute>> {
    let mut chunks: Vec<Vec<Attribute>> = vec![];
    let size = kccs.len();

    for (i, el) in kccs.iter().enumerate() {
        // initial feature
        let mut items = vec![
            Attribute::new(format!("kcc:{}", el.0), 1.0),
            Attribute::new(format!("t:{}", el.1), 1.0),
            Attribute::new("ns".to_string(), if el.1 == "NS" { 1.0 } else { 0.0 }),
        ];

        // bos
        if i == 0 {
            items.push(Attribute::new("BOS", 1.0));
        }

        if i >= 1 {
            items.push(Attribute::new(format!("kcc[-1]:{}", kccs[i - 1].0), 1.0));
            items.push(Attribute::new(format!("kcc[-1]t:{}", kccs[i - 1].1), 1.0));
            items.push(Attribute::new(
                format!("kcc[-1:0]:{}{}", kccs[i - 1].0, el.0),
                1.0,
            ));
            let ns_1 = if kccs[i - 1].1 == "NS" { 1.0 } else { 0.0 };
            items.push(Attribute::new("ns-1".to_string(), ns_1));
        }

        if i >= 2 {
            items.push(Attribute::new(format!("kcc[-2]:{}", kccs[i - 2].0), 1.0));
            items.push(Attribute::new(format!("kcc[-2]t:{}", kccs[i - 2].1), 1.0));
            items.push(Attribute::new(
                format!("kcc[-2:-1]:{}{}", kccs[i - 2].0, kccs[i - 1].0),
                1.0,
            ));
            items.push(Attribute::new(
                format!("kcc[-2:0]:{}{}{}", kccs[i - 2].0, kccs[i - 1].0, kccs[i].0,),
                1.0,
            ));
        }

        if i >= 3 {
            items.push(Attribute::new(format!("kcc[-3]:{}", kccs[i - 3].0), 1.0));
            items.push(Attribute::new(format!("kcc[-3]t:{}", kccs[i - 3].1), 1.0));
            items.push(Attribute::new(
                format!(
                    "kcc[-3:0]:{}{}{}{}",
                    kccs[i - 3].0,
                    kccs[i - 2].0,
                    kccs[i - 1].0,
                    kccs[i].0
                ),
                1.0,
            ));
            items.push(Attribute::new(
                format!(
                    "kcc[-3:-1]:{}{}{}",
                    kccs[i - 3].0,
                    kccs[i - 2].0,
                    kccs[i - 1].0
                ),
                1.0,
            ));
            items.push(Attribute::new(
                format!("kcc[-3:-2]:{}{}", kccs[i - 3].0, kccs[i - 2].0),
                1.0,
            ));
        }

        if size >= 1 && i < size - 1 {
            items.push(Attribute::new(format!("kcc[+1]:{}", kccs[i + 1].0), 1.0));
            items.push(Attribute::new(format!("kcc[+1]t:{}", kccs[i + 1].1), 1.0));
            items.push(Attribute::new(
                format!("kcc[+1:0]t:{}{}", kccs[i].0, kccs[i + 1].0),
                1.0,
            ));
            items.push(Attribute::new(
                format!("ns+1"),
                if kccs[i + 1].1 == "NS" { 1.0 } else { 0.0 },
            ));
        }

        if size >= 2 && i < size - 2 {
            items.push(Attribute::new(format!("kcc[+2]:{}", kccs[i + 2].0), 1.0));
            items.push(Attribute::new(format!("kcc[+2]t:{}", kccs[i + 2].1), 1.0));
            items.push(Attribute::new(
                format!("kcc[+1:+2]:{}{}", kccs[i + 1].0, kccs[i + 2].0),
                1.0,
            ));
            items.push(Attribute::new(
                format!("kcc[0:+2]:{}{}{}", kccs[i].0, kccs[i + 1].0, kccs[i + 2].0),
                1.0,
            ));
            items.push(Attribute::new(
                "ns+2",
                if kccs[i + 2].1 == "NS" { 1.0 } else { 0.0 },
            ));
        }

        if size >= 3 && i < size - 3 {
            items.push(Attribute::new(format!("kcc[+3]:{}", kccs[i + 3].0), 1.0));
            items.push(Attribute::new(format!("kcc[+3]t:{}", kccs[i + 3].1), 1.0));
            items.push(Attribute::new(
                format!("kcc[+2:+3]t:{}{}", kccs[i + 2].0, kccs[i + 3].0),
                1.0,
            ));
            items.push(Attribute::new(
                format!(
                    "kcc[+1:+3]t:{}{}{}",
                    kccs[i + 1].0,
                    kccs[i + 2].0,
                    kccs[i + 3].0
                ),
                1.0,
            ));
            items.push(Attribute::new(
                format!(
                    "kcc[0:+3]t:{}{}{}{}",
                    kccs[i].0,
                    kccs[i + 1].0,
                    kccs[i + 2].0,
                    kccs[i + 3].0
                ),
                1.0,
            ));
        }

        if size >= 1 && i == size - 1 {
            items.push(Attribute::new("EOS", 1.0));
        }

        chunks.push(items);
    }

    return chunks;
}

pub fn tokenize(model: &Model, input_str: &String) -> Vec<String> {
    let normalized_text = input_str.replace("\u{200b}", "");
    let graphemes = text_tagger(&normalized_text);
    let features = create_features(&graphemes);
    let mut tagger = model.tagger().unwrap();
    let results: Vec<&str> = tagger.tag(&features).unwrap();
    let mut tokens = vec![];

    for (i, y) in results.iter().enumerate() {
        let (c, _) = graphemes.get(i).unwrap();
        let flag = y.parse::<i8>().unwrap();
        if flag == 1 || i == 0 {
            tokens.push(c.to_string());
            continue;
        }
        tokens.last_mut().unwrap().push_str(c);
    }
    tokens
}