pinyin-pro - npm Package Compare versions

Comparing version 3.8.1 to 3.8.2

CHANGELOG.md

@@ -0,1 +1,10 @@
+## 3.8.2
+
+当前版本： 3.8.1 -> 3.8.2
+
+- 修复问题
+  - 修复双字节 unicode 编码的文字返回默认值不正确的问题
+
+点击查看 [版本更新文档](./CHANGELOG.md)
+
 ## 3.8.1
@@ -2,0 +11,0 @@

lib/handle.ts

@@ -9,3 +9,2 @@ import INITIAL_LIST from '../data/initial';
 import { getCustomDict } from './custom';
-import { getStringLength } from './utils';
 const dictArr = [{}, {}, DICT2, DICT3, DICT4, DICT5];
@@ -64,3 +63,3 @@
   if (length === 1) {
-    for (let i = 0; i < getStringLength(word); i++) {
+    for (let i = 0; i < word.length; i++) {
       const result = getSingleWordPinyin(word[i]);
@@ -85,8 +84,8 @@ const curIsChinese = result !== word[i];
       const left_pinyin = left_word
-        ? `${getPinyin(left_word, getStringLength(left_word), { nonZh })} `
+        ? `${getPinyin(left_word, left_word.length, { nonZh })} `
         : '';
       // 取出该词后右边拼音
-      const right_word = word.slice(index + getStringLength(key));
+      const right_word = word.slice(index + key.length);
       const right_pinyin = right_word
-        ? ` ${getPinyin(right_word, getStringLength(right_word), { nonZh })}`
+        ? ` ${getPinyin(right_word, right_word.length, { nonZh })}`
         : '';
@@ -124,3 +123,3 @@ // 取出的词的拼音
       const left_pinyin = left_word
-        ? `${getPinyin(left_word, getStringLength(left_word), {
+        ? `${getPinyin(left_word, left_word.length, {
             mode: 'surname',
@@ -131,5 +130,5 @@ nonZh,
       // 取出该词后右边拼音
-      const right_word = word.slice(index + getStringLength(key));
+      const right_word = word.slice(index + key.length);
       const right_pinyin = right_word
-        ? ` ${getPinyin(right_word, getStringLength(right_word), {
+        ? ` ${getPinyin(right_word, right_word.length, {
             mode: 'surname',
@@ -145,3 +144,3 @@ nonZh,
   // 若姓氏表中的词均为匹配成功，则使用常规匹配
-  return getPinyin(word, getStringLength(word), { nonZh });
+  return getPinyin(word, word.length, { nonZh });
 };
@@ -174,3 +173,3 @@
       const left_pinyin = left_word
-        ? `${getPinyin(left_word, getStringLength(left_word), {
+        ? `${getPinyin(left_word, left_word.length, {
             mode,
@@ -182,5 +181,5 @@ useCustomConfig: true,
       // 取出该词后右边拼音
-      const right_word = word.slice(index + getStringLength(key));
+      const right_word = word.slice(index + key.length);
       const right_pinyin = right_word
-        ? ` ${getPinyin(right_word, getStringLength(right_word), {
+        ? ` ${getPinyin(right_word, right_word.length, {
             mode,
@@ -196,3 +195,3 @@ useCustomConfig: true,
   }
-  return getPinyin(word, getStringLength(word), { mode, nonZh });
+  return getPinyin(word, word.length, { mode, nonZh });
 };
@@ -245,3 +244,3 @@
       if (_pinyin.startsWith(_initial)) {
-        const _final = _pinyin.slice(getStringLength(_initial));
+        const _final = _pinyin.slice(_initial.length);
         initial_arr.push(_initial);
@@ -248,0 +247,0 @@ final_arr.push(_final);

lib/match.ts

 import { pinyin as _pinyin } from './pinyin';
-import { getStringLength } from './utils';
 
@@ -13,3 +12,3 @@ /**
   let currentPinyin = pinyin;
-  for (let i = 0; i < getStringLength(words); i++) {
+  for (let i = 0; i < words.length; i++) {
     // 当前字的多音字拼音
@@ -42,3 +41,3 @@ const ps = _pinyin(words[i], {
   let length = 0;
-  for (let i = 0; i < getStringLength(pinyin1); i++) {
+  for (let i = 0; i < pinyin1.length; i++) {
     if (pinyin1[i] === pinyin2[length]) {
@@ -45,0 +44,0 @@ length++;

lib/pinyin.ts

@@ -10,3 +10,3 @@ import {
 } from './handle';
-import { getStringLength } from './utils';
+import { getStringLength, isZhChar } from './utils';
 import { hasCustomConfig } from './custom';
@@ -47,3 +47,3 @@
  * @param {string} word 要转换的汉语字符串
- * @param {OptionsReturnString} options 配置项
+ * @param {OptionsReturnString=} options 配置项
  * @return {string | string[]} options 配置项中的 type 为 string 时，返回字符串，中间用空格隔开；为 array 时，返回拼音字符串数组
@@ -56,3 +56,3 @@ */
  * @param {string} word 要转换的汉语字符串
- * @param {OptionsReturnArray} options 配置项
+ * @param {OptionsReturnArray=} options 配置项
  * @return {string | string[]} options 配置项中的 type 为 string 时，返回字符串，中间用空格隔开；为 array 时，返回拼音字符串数组
@@ -65,3 +65,3 @@ */
  * @param {string} word 要转换的汉语字符串
- * @param {CompleteOptions} options 配置项
+ * @param {CompleteOptions=} options 配置项
  * @return {string | string[]} options 配置项中的 type 为 string 时，返回字符串，中间用空格隔开；为 array 时，返回拼音字符串数组
@@ -77,6 +77,5 @@ */
     let str = '';
-    for (let i = 0; i < getStringLength(word); i++) {
+    for (let i = 0; i < word.length; i++) {
       const char = word[i];
-      let code = char.charCodeAt(0);
-      if (code >= 19968 && code <= 40869) {
+      if (isZhChar(char)) {
         str += char;
@@ -92,9 +91,81 @@ }
 
-  // 获取原始拼音
-  let pinyin = getPinyin(word, getStringLength(word), {
-    mode: options.mode || 'normal',
-    nonZh: options.nonZh || 'spaced',
-    useCustomConfig: hasCustomConfig(),
-  });
+  let pinyin = '';
+  if (options.removeNonZh || options.nonZh === 'removed') {
+    pinyin = getPinyin(word, getStringLength(word), {
+      mode: options.mode || 'normal',
+      nonZh: options.nonZh,
+      useCustomConfig: hasCustomConfig(),
+    });
+  } else {
+    // 针对双精度unicode编码字符处理
+    let i = 0;
+    let lastIndex = 0;
+    let items: {
+      val: string;
+      isDouble: boolean;
+      firstNonZh?: boolean;
+      lastNonZh?: boolean;
+    }[] = [];
+    while (i < word.length) {
+      const currentWord = word.substring(i, i + 2);
+      if (getStringLength(currentWord) !== currentWord.length) {
+        // 双精度unicode编码字符
+        if (lastIndex !== i) {
+          const leftWord = word.substring(lastIndex, i);
+          const leftPinyin = getPinyin(leftWord, getStringLength(leftWord), {
+            mode: options.mode || 'normal',
+            nonZh: options.nonZh || 'spaced',
+            useCustomConfig: hasCustomConfig(),
+          });
+          items.push({
+            val: leftPinyin,
+            isDouble: false,
+            firstNonZh: !isZhChar(leftWord[0]),
+            lastNonZh: !isZhChar(leftWord[leftWord.length - 1]),
+          });
+        }
+        items.push({ val: currentWord, isDouble: true });
+        lastIndex = i + 2;
+        i = i + 2;
+      } else {
+        i++;
+      }
+    }
+    if (lastIndex !== i) {
+      const remainedWord = word.substring(lastIndex, i);
+      const remainedPinyin = getPinyin(
+        remainedWord,
+        getStringLength(remainedWord),
+        {
+          mode: options.mode || 'normal',
+          nonZh: options.nonZh || 'spaced',
+          useCustomConfig: hasCustomConfig(),
+        }
+      );
+      items.push({
+        val: remainedPinyin,
+        isDouble: false,
+        firstNonZh: !isZhChar(remainedWord[0]),
+        lastNonZh: !isZhChar(remainedWord[remainedWord.length - 1]),
+      });
+    }
 
+    if (options.nonZh === 'consecutive') {
+      for (let i = 0; i < items.length; i++) {
+        if (i === 0) {
+          pinyin += items[0].val;
+        } else {
+          // 当前字符起始是否为非中文
+          const currentNonZh = items[i].isDouble || items[i].firstNonZh;
+          // 上一个字符结束是否为非中文
+          const preNonZh = items[i - 1].isDouble || items[i - 1].lastNonZh;
+          pinyin =
+            pinyin + (currentNonZh && preNonZh ? '' : ' ') + items[i].val;
+        }
+      }
+    } else {
+      pinyin = items.map((item) => item.val).join(' ');
+    }
+  }
+
   // 对multiple进行处理
@@ -101,0 +172,0 @@ if (getStringLength(word) === 1 && options.multiple) {

lib/utils.ts

@@ -6,1 +6,6 @@ var regex = /[\uD800-\uDBFF][\uDC00-\uDFFF]/g;
 }
+
+export function isZhChar(char: string) {
+  let code = char.charCodeAt(0);
+  return code >= 19968 && code <= 40869;
+}

package.json

 {
   "name": "pinyin-pro",
-  "version": "3.8.1",
+  "version": "3.8.2",
   "description": "汉字转拼音库。获取中文拼音、韵母、声母、声调、首字母，支持拼音匹配",
@@ -5,0 +5,0 @@ "main": "./dist/index.cjs.js",

README.md

@@ -26,3 +26,3 @@ [![pinyin-pro Logo](https://i.ibb.co/26fJ5vF/pinyin-logo.png)](https://github.com/zh-lx/pinyin-pro)
 
-当前版本： 3.8.0 -> 3.8.1
+当前版本： 3.8.1 -> 3.8.2
 
@@ -29,0 +29,0 @@ - 修复问题

test/dist.test.js

@@ -415,2 +415,7 @@ const { pinyin, customPinyin, match } = require('../dist/index.cjs.js');
   });
+
+  it('match9', () => {
+    const result = match('𧒽测试', 'cs');
+    expect(result).to.deep.equal([2, 3]);
+  });
 });
@@ -513,2 +518,21 @@
   });
+
+  it('double unicode', () => {
+    const result1 = pinyin('𧒽');
+    expect(result1).to.be.equal('𧒽');
+
+    const result2 = pinyin('𧒽测试');
+    expect(result2).to.be.equal('𧒽 cè shì');
+
+    const result3 = pinyin('𧒽测试𧒽测试', {});
+    expect(result3).to.be.equal('𧒽 cè shì 𧒽 cè shì');
+
+    const result4 = pinyin('𧒽测试', { nonZh: 'consecutive' });
+    expect(result4).to.be.equal('𧒽 cè shì');
+
+    const result6 = pinyin('测试a𧒽𧒽a测试a𧒽𧒽a测试', {
+      nonZh: 'consecutive',
+    });
+    expect(result6).to.be.equal('cè shì a𧒽𧒽a cè shì a𧒽𧒽a cè shì');
+  });
 });

types/lib/pinyin.d.ts

@@ -18,3 +18,3 @@ interface BasicOptions {
  * @param {string} word 要转换的汉语字符串
- * @param {OptionsReturnString} options 配置项
+ * @param {OptionsReturnString=} options 配置项
  * @return {string | string[]} options 配置项中的 type 为 string 时，返回字符串，中间用空格隔开；为 array 时，返回拼音字符串数组
@@ -26,3 +26,3 @@ */
  * @param {string} word 要转换的汉语字符串
- * @param {OptionsReturnArray} options 配置项
+ * @param {OptionsReturnArray=} options 配置项
  * @return {string | string[]} options 配置项中的 type 为 string 时，返回字符串，中间用空格隔开；为 array 时，返回拼音字符串数组
@@ -29,0 +29,0 @@ */

types/lib/utils.d.ts

 export declare function getStringLength(string: string): number;
+export declare function isZhChar(char: string): boolean;

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

2276324

increased by0.22%

Lines of code

54818

increased by0.18%

No dependency changes