feat(pinyin): convert first of two third tones to second tone

packages/docs/en/.vitepress/theme/components/basic-params-table.vue

@@ -348,18 +348,20 @@ pinyin('赵钱孙李额', { pattern: 'first', toneType: 'none', type: 'array' })
     option: 'toneSandhi',
     type: 'boolean',
     description:
-      'whether to apply smart tone change to <code>一</code> and <code>不</code>, reference <a href=https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83 target="_blank">维基百科</a>',
+      'whether to apply smart tone change to <code>一</code> and <code>不</code>, reference <a href=https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83 target="_blank">维基百科</a>, and to two consecutive third-tone syllables (e.g. 你好 → ní hǎo), reference <a href=https://en.wikipedia.org/wiki/Tone_sandhi#Mandarin_Chinese target="_blank">Wikipedia</a>',
     default: 'true',
     children: [
       {
         value: 'true',
         desc: 'apply',
-        example: `pinyin('一旦被发现', { toneSandhi: true }); // 'yí dàn bèi fā xiàn'`,
+        example: `pinyin('一旦被发现', { toneSandhi: true }); // 'yí dàn bèi fā xiàn'
+pinyin('你好', { toneSandhi: true }); // 'ní hǎo'`,
       },
       {
         value: 'false',
         desc: 'not apply',
-        example: `pinyin('一旦被发现', { toneSandhi: false }); // 'yī dàn bèi fā xiàn'`,
+        example: `pinyin('一旦被发现', { toneSandhi: false }); // 'yī dàn bèi fā xiàn'
+pinyin('你好', { toneSandhi: false }); // 'nǐ hǎo'`,
       },
     ],
   },

packages/docs/zh/.vitepress/theme/components/basic-params-table.vue

@@ -346,18 +346,20 @@ pinyin('赵钱孙李额', { pattern: 'first', toneType: 'none', type: 'array' })
     option: 'toneSandhi',
     type: 'boolean',
     description:
-      '是否对<code>一</code>和<code>不</code>应用智能变调，参考<a href=https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83 target="_blank">维基百科</a>',
+      '是否应用智能变调，包括<code>一</code>和<code>不</code>（参考<a href=https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83 target="_blank">维基百科</a>），以及两个连续三声的变调（如 你好 → ní hǎo，参考<a href=https://en.wikipedia.org/wiki/Tone_sandhi#Mandarin_Chinese target="_blank">Wikipedia</a>）',
     default: 'true',
     children: [
       {
         value: 'true',
         desc: '应用',
-        example: `pinyin('一旦被发现', { toneSandhi: true }); // 'yí dàn bèi fā xiàn'`,
+        example: `pinyin('一旦被发现', { toneSandhi: true }); // 'yí dàn bèi fā xiàn'
+pinyin('你好', { toneSandhi: true }); // 'ní hǎo'`,
       },
       {
         value: 'false',
         desc: '不应用',
-        example: `pinyin('一旦被发现', { toneSandhi: false }); // 'yī dàn bèi fā xiàn'`,
+        example: `pinyin('一旦被发现', { toneSandhi: false }); // 'yī dàn bèi fā xiàn'
+pinyin('你好', { toneSandhi: false }); // 'nǐ hǎo'`,
       },
     ],
   },

packages/pinyin-pro/lib/core/pinyin/index.ts

@@ -34,7 +34,7 @@ export interface BasicOptions extends CommonOptions {
    */
   surname?: SurnameMode;
   /**
-   * @description 是否开启「一」和 「不」字的变调。默认开启。参考：https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83
+   * @description 是否开启变调。默认开启。包含「一」和「不」的变调（参考：https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83），以及两个连续三声的变调（如 你好 → ní hǎo，参考：https://en.wikipedia.org/wiki/Tone_sandhi#Mandarin_Chinese）
    * @value true：开启
    * @value false：不开启
    */
@@ -204,7 +204,7 @@ function pinyin(
     zhChars,
   );
 
-  // 一和不变调处理
+  // 变调处理
   list = middlewareToneSandhi(list, options.toneSandhi as boolean);
 
   // nonZh 参数及 removeNonZh 参数

packages/pinyin-pro/lib/core/pinyin/middlewares.ts

@@ -235,6 +235,34 @@ export const middlewareType = (
   return list.map((item) => item.result).join(options.separator);
 };
 
+const thirdToneToSecondToneMap = {
+  ǎ: "á",
+  ǒ: "ó",
+  ě: "é",
+  ǐ: "í",
+  ǔ: "ú",
+  ǚ: "ǘ",
+  ň: "ń",
+  "m̌": "ḿ",
+  "ê̌": "ế",
+};
+
+// 轻量三声判断：避免在默认热路径上调用完整的 getNumOfTone
+const thirdTonePattern = /ǎ|ǒ|ě|ǐ|ǔ|ǚ|ň|m̌|ê̌/;
+
+const isThirdTone = (item: SingleWordResult) =>
+  item.isZh && thirdTonePattern.test(item.result);
+
+const convertThirdToneToSecondTone = (pinyin: string) => {
+  return pinyin.replace(
+    thirdTonePattern,
+    (thirdTone) =>
+      thirdToneToSecondToneMap[
+        thirdTone as keyof typeof thirdToneToSecondToneMap
+      ],
+  );
+};
+
 // 是否开启变调
 export const middlewareToneSandhi = (
   list: SingleWordResult[],
@@ -248,6 +276,27 @@ export const middlewareToneSandhi = (
         item.result = item.originPinyin = "bù";
       }
     });
+    return list;
+  }
+
+  for (let start = 0; start < list.length; ) {
+    if (!isThirdTone(list[start])) {
+      start += 1;
+      continue;
+    }
+
+    let end = start + 1;
+    while (end < list.length && isThirdTone(list[end])) {
+      end += 1;
+    }
+
+    if (end - start === 2) {
+      const pinyin = convertThirdToneToSecondTone(list[start].result);
+      list[start].result = pinyin;
+      list[start].originPinyin = pinyin;
+    }
+
+    start = end;
   }
   return list;
 };

packages/pinyin-pro/lib/core/segment/index.ts

@@ -145,7 +145,7 @@ export function segment(word: string, options?: SegmentCompleteOptions) {
     zhChars,
   );
 
-  // 一和不变调处理
+  // 变调处理
   list = middlewareToneSandhi(list, options.toneSandhi as boolean);
 
   // nonZh

packages/pinyin-pro/test/basic.test.js

@@ -23,7 +23,7 @@ describe('basic', () => {
 
   it('[basic]好好', () => {
     const result = pinyin('好好学习');
-    expect(result).to.be.equal('hǎo hǎo xué xí');
+    expect(result).to.be.equal('háo hǎo xué xí');
   });
 
   it('[basic]拼音+非汉字数组', () => {

packages/pinyin-pro/test/segment.test.js

@@ -346,7 +346,7 @@ describe("segment surname mode", () => {
         format: OutputFormat.PinyinString
       }
     );
-    expect(result).to.be.equal("zēng xiǎo xián nǐhǎo")
+    expect(result).to.be.equal("zēng xiǎo xián níhǎo")
   });
 });
 
@@ -379,4 +379,4 @@ describe("segment final nonZh", () => {
       "xiǎo míng shuòshì bìyè yú zhōngguókēxuéyuàn jìsuànsuǒ ， hòu zài rìběnjīngdūdàxué shēnzào 。"
     );
   });
-});
+});

packages/pinyin-pro/test/special.test.js

@@ -232,7 +232,7 @@ describe("tone sandhi for “不”", () => {
 
   it("[special tone sandhi]要不你走", () => {
     const result = pinyin("要不你走");
-    expect(result).to.be.equal("yào bù nǐ zǒu");
+    expect(result).to.be.equal("yào bù ní zǒu");
   });
 
   // 变调为二声
@@ -294,7 +294,7 @@ describe('[special tone sandhi]绕口令', () => {
 
   it('[special 々]々', () => {
     const result = pinyin('天々向上，好々学习');
-    expect(result).to.be.equal('tiān tiān xiàng shàng ， hǎo hǎo xué xí');
+    expect(result).to.be.equal('tiān tiān xiàng shàng ， háo hǎo xué xí');
 
     const result1 = pinyin('々々');
     expect(result1).to.be.equal('tóng tóng');

packages/pinyin-pro/test/toneSandhi.test.js

@@ -1,10 +1,75 @@
-import { pinyin } from '../lib/index';
-import { expect, describe, it } from 'vitest';
+import { clearCustomDict, customPinyin, pinyin } from '../lib/index';
+import { afterEach, expect, describe, it } from 'vitest';
 
 describe("toneSandhi", () => {
+  afterEach(() => {
+    clearCustomDict('pinyin');
+  });
+
   it("[toneSandhi]不", () => {
     const result = pinyin("不是", { toneSandhi: false });
     expect(result).to.be.equal("bù shì");
   });
 
+  it("changes the first of two consecutive third tones", () => {
+    expect(pinyin("你好", { toneSandhi: true })).to.be.equal("ní hǎo");
+    expect(pinyin("你高", { toneSandhi: true })).to.be.equal("nǐ gāo");
+  });
+
+  it("supports each tone output format", () => {
+    expect(
+      pinyin("你好", { toneSandhi: true, toneType: "symbol" }),
+    ).to.be.equal("ní hǎo");
+    expect(
+      pinyin("你好", { toneSandhi: true, toneType: "num" }),
+    ).to.be.equal("ni2 hao3");
+    expect(
+      pinyin("你好", { toneSandhi: true, toneType: "none" }),
+    ).to.be.equal("ni hao");
+  });
+
+  it("reproduces the issue 286 example", () => {
+    expect(
+      pinyin("你好", {
+        toneSandhi: true,
+        type: "array",
+        toneType: "num",
+      }),
+    ).toEqual(["ni2", "hao3"]);
+  });
+
+  it("keeps both third tones when tone sandhi is disabled", () => {
+    expect(pinyin("你好", { toneSandhi: false })).to.be.equal("nǐ hǎo");
+  });
+
+  it("does not apply across non-Chinese boundaries", () => {
+    expect(pinyin("你，好", { toneSandhi: true })).to.be.equal("nǐ ， hǎo");
+    expect(pinyin("你a好", { toneSandhi: true })).to.be.equal("nǐ a hǎo");
+  });
+
+  it("handles separate pairs between punctuation boundaries", () => {
+    expect(pinyin("你好，很好", { toneSandhi: true })).to.be.equal(
+      "ní hǎo ， hén hǎo",
+    );
+  });
+
+  it("keeps runs longer than two third tones unchanged", () => {
+    expect(pinyin("我很好", { toneSandhi: true })).to.be.equal("wǒ hěn hǎo");
+    expect(pinyin("我想很好", { toneSandhi: true })).to.be.equal(
+      "wǒ xiǎng hěn hǎo",
+    );
+  });
+
+  it("uses custom pinyin results for third-tone sandhi", () => {
+    customPinyin({
+      你好: "wǒ hǎo",
+    });
+
+    expect(pinyin("你好", { toneSandhi: true })).to.be.equal("wó hǎo");
+  });
+
+  it("preserves existing tone sandhi for 一 and 不", () => {
+    expect(pinyin("一把", { toneSandhi: true })).to.be.equal("yì bǎ");
+    expect(pinyin("不想", { toneSandhi: true })).to.be.equal("bù xiǎng");
+  });
 });

packages/pinyin-pro/types/core/pinyin/index.d.ts

@@ -21,7 +21,7 @@ export interface BasicOptions extends CommonOptions {
      */
     surname?: SurnameMode;
     /**
-     * @description 是否开启「一」和 「不」字的变调。默认开启。参考：https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83
+     * @description 是否开启变调。默认开启。包含「一」和「不」的变调（参考：https://zh.wiktionary.org/wiki/Appendix:%E2%80%9C%E4%B8%80%E2%80%9D%E5%8F%8A%E2%80%9C%E4%B8%8D%E2%80%9D%E7%9A%84%E5%8F%98%E8%B0%83），以及两个连续三声的变调（如 你好 → ní hǎo，参考：https://en.wikipedia.org/wiki/Tone_sandhi#Mandarin_Chinese）
      * @value true：开启
      * @value false：不开启
      */