39 lines
2.5 KiB
JavaScript
39 lines
2.5 KiB
JavaScript
|
// Copyright 2018 the V8 project authors. All rights reserved.
|
|||
|
|
// Use of this source code is governed by a BSD-style license that can be
|
|||
|
|
// found in the LICENSE file.
|
|||
|
|
|
|||
|
|
// Flags: --harmony-intl-segmenter
|
|||
|
|
|
|||
|
|
const seg = new Intl.Segmenter([], {granularity: "word"})
|
|||
|
|
for (const text of [
|
|||
|
|
"Hello world!", // English
|
|||
|
|
" Hello world! ", // English with space before/after
|
|||
|
|
" Hello world? Foo bar!", // English
|
|||
|
|
"Jedovatou mambu objevila žena v zahrádkářské kolonii.", // Czech
|
|||
|
|
"Việt Nam: Nhất thể hóa sẽ khác Trung Quốc?", // Vietnamese
|
|||
|
|
"Σοβαρές ενστάσεις Κομισιόν για τον προϋπολογισμό της Ιταλίας", // Greek
|
|||
|
|
"Решение Индии о покупке российских С-400 расценили как вызов США", // Russian
|
|||
|
|
"הרופא שהציל נשים והנערה ששועבדה ע", // Hebrew,
|
|||
|
|
"ترامب للملك سلمان: أنا جاد للغاية.. عليك دفع المزيد", // Arabic
|
|||
|
|
"भारत की एस 400 मिसाइल के मुकाबले पाक की थाड, जानें कौन कितना ताकतवर", // Hindi
|
|||
|
|
"ரெட் அலர்ட் எச்சரிக்கை; புதுச்சேரியில் நாளை அரசு விடுமுறை!", // Tamil
|
|||
|
|
"'ఉత్తర్వులు అందే వరకు ఓటర్ల తుది జాబితాను వెబ్సైట్లో పెట్టవద్దు'", // Telugu
|
|||
|
|
"台北》抹黑柯P失敗？朱學恒酸：姚文智氣pupu嗆大老闆", // Chinese
|
|||
|
|
"วัดไทรตีระฆังเบาลงช่วงเข้าพรรษา เจ้าอาวาสเผยคนร้องเรียนรับผลกรรมแล้ว", // Thai
|
|||
|
|
"九州北部の一部が暴風域に入りました(日直予報士 2018年10月06日) - 日本気象協会 tenki.jp", // Japanese
|
|||
|
|
"법원 “다스 지분 처분권·수익권 모두 MB가 보유”", // Korean
|
|||
|
|
]) {
|
|||
|
|
const iter = seg.segment(text);
|
|||
|
|
let prev = 0;
|
|||
|
|
let segments = [];
|
|||
|
|
while (!iter.following()) {
|
|||
|
|
assertTrue(["word", "none"].includes(iter.breakType), iter.breakType);
|
|||
|
|
assertTrue(iter.position >= 0);
|
|||
|
|
assertTrue(iter.position <= text.length);
|
|||
|
|
assertTrue(iter.position > prev);
|
|||
|
|
segments.push(text.substring(prev, iter.position));
|
|||
|
|
prev = iter.position;
|
|||
|
|
}
|
|||
|
|
assertEquals(text, segments.join(""));
|
|||
|
|
}
|