1 |
|
2 |
|
3 |
|
4 |
|
5 |
|
6 | import { SubSModuleTokenizer, ISubTokenizerCreate } from '../mod';
|
7 | import { Segment, IWord } from '../Segment';
|
8 | import { IWordDebugInfo } from '../util/index';
|
9 | export declare class ForeignTokenizer extends SubSModuleTokenizer {
|
10 | name: string;
|
11 | |
12 |
|
13 |
|
14 | _REGEXP_SPLIT_1: RegExp;
|
15 | |
16 |
|
17 |
|
18 | _REGEXP_SPLIT_2: RegExp;
|
19 | _cache(): void;
|
20 | |
21 |
|
22 |
|
23 |
|
24 |
|
25 |
|
26 | split(words: IWord[]): IWord[];
|
27 | |
28 |
|
29 |
|
30 |
|
31 |
|
32 | splitForeign2(text: string, cur?: number): IWord[];
|
33 | |
34 |
|
35 |
|
36 |
|
37 |
|
38 |
|
39 |
|
40 | splitForeign(text: string, cur?: number): IWord[];
|
41 | createForeignToken(word: IWord, lasttype?: number, attr?: IWordDebugInfo): Segment.IWord;
|
42 | }
|
43 | export declare const init: ISubTokenizerCreate<ForeignTokenizer, SubSModuleTokenizer>;
|
44 | export default ForeignTokenizer;
|