diff options
author | Cashew <52880648+cashewnuttynuts@users.noreply.github.com> | 2024-06-22 03:24:21 +0700 |
---|---|---|
committer | GitHub <noreply@github.com> | 2024-06-21 20:24:21 +0000 |
commit | d724b403f94b7fd1ecec3f6d2e4f5a1ed805c6ec (patch) | |
tree | 1f696742b75b4f8377fe3b952863af141ff6494e /ext/js/language/language-descriptors.js | |
parent | 1c609d972ae76f8779190d7a3621f77a664a6dec (diff) |
Add diacritics normalization preprocessors for Vietnamese (#1107)
* add viet diacritics normalization
* move regexp construction outside of function
* fix eslint
* add 'off' option
* fix lint
* fix type
Diffstat (limited to 'ext/js/language/language-descriptors.js')
-rw-r--r-- | ext/js/language/language-descriptors.js | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/ext/js/language/language-descriptors.js b/ext/js/language/language-descriptors.js index 2a8762e9..517c908c 100644 --- a/ext/js/language/language-descriptors.js +++ b/ext/js/language/language-descriptors.js @@ -36,6 +36,7 @@ import {removeRussianDiacritics, yoToE} from './ru/russian-text-preprocessors.js import {oldIrishTransforms} from './sga/old-irish-transforms.js'; import {removeSerboCroatianAccentMarks} from './sh/serbo-croatian-text-preprocessors.js'; import {albanianTransforms} from './sq/albanian-transforms.js'; +import {normalizeDiacritics} from './vi/viet-text-preprocessors.js'; import {capitalizeFirstLetter, decapitalize, removeAlphabeticDiacritics} from './text-processors.js'; import {isStringPartiallyChinese} from './zh/chinese.js'; @@ -261,7 +262,10 @@ const languageDescriptors = [ iso: 'vi', name: 'Vietnamese', exampleText: 'đọc', - textPreprocessors: capitalizationPreprocessors, + textPreprocessors: { + ...capitalizationPreprocessors, + normalizeDiacritics, + }, }, { iso: 'yue', |