Merge pull request #282 from hotoo/alecgibson-group

Alecgibson group

README-us_EN.md

@@ -51,8 +51,12 @@ console.log(pinyin("中心", {
 }));                            // [ [ 'zhōng', 'zhòng' ], [ 'xīn' ] ]
 console.log(pinyin("中心", {
   heteronym: true,              // Enable heteronym mode.
-  segment: true                 // Enable Chinese words eegmentation, fix most heteronym problem.
+  segment: true                 // Enable Chinese words segmentation, fix most heteronym problem.
 }));                            // [ [ 'zhōng' ], [ 'xīn' ] ]
+console.log(pinyin("我喜欢你", {
+  segment: true,                // Enable segmentation. Needed for grouping.
+  group: true                   // Group pinyin segments
+}));                            // [ [ 'wǒ' ], [ 'xǐhuān' ], [ 'nǐ' ] ]
 console.log(pinyin("中心", {
   style: pinyin.STYLE_INITIALS, // Setting pinyin style.
   heteronym: true

README.md

@@ -59,6 +59,10 @@ console.log(pinyin("中心", {
   heteronym: true,              // 启用多音字模式
   segment: true                 // 启用分词，以解决多音字问题。
 }));                            // [ [ 'zhōng' ], [ 'xīn' ] ]
+console.log(pinyin("我喜欢你", {
+  segment: true,                // 启用分词
+  group: true                   // 启用词组
+}));                            // [ [ 'wǒ' ], [ 'xǐhuān' ], [ 'nǐ' ] ]
 console.log(pinyin("中心", {
   style: pinyin.STYLE_INITIALS, // 设置拼音风格
   heteronym: true

lib/index.js

@@ -2,6 +2,7 @@
 
 const assign = require("object-assign");
 const PINYIN_DICT = require("../data/dict-zi");
+const util = require("./util");
 const Pinyin = require("./pinyin");
 let jieba;
 let PHRASES_DICT;
@@ -34,10 +35,14 @@ class NodePinyin extends Pinyin {
           nohans = ""; // reset non-chinese words.
         }
 
-        if (words.length === 1) {
-          pys = pys.concat(super.convert(words, options));
+        const newPys = words.length === 1
+          ? super.convert(words, options)
+          : this.phrases_pinyin(words, options);
+
+        if (options.group) {
+          pys.push(groupPhrases(newPys));
         } else {
-          pys = pys.concat(this.phrases_pinyin(words, options));
+          pys = pys.concat(newPys);
         }
 
       } else {
@@ -96,6 +101,16 @@ function segment(hans) {
   return jieba.cutSmall(hans, 4);
 }
 
+function groupPhrases(phrases) {
+  if (phrases.length === 1) {
+    return phrases[0];
+  }
+
+  const grouped = util.combo(phrases);
+
+  return grouped;
+}
+
 const pinyin = new NodePinyin(PINYIN_DICT);
 
 module.exports = pinyin.convert.bind(pinyin);

lib/util.js

@@ -0,0 +1,56 @@
+"use strict";
+
+/**
+ * 组合 2 个拼音数组。
+ * @param {Array<String>} a1 第一个数组，形如 ["zhāo", "cháo"]
+ * @param {Array<String>} a2 字符串型数组。形如 ["yáng"]
+ * @return {Array<String>} 组合后的一维数组，如上可得 ["zhāoyáng", "cháoyáng"]
+ */
+function combo2array(a1, a2) {
+  const result = [];
+  if (!a1.length) {
+    return a2;
+  }
+  if (!a2.length) {
+    return a1;
+  }
+  for (let i = 0, l = a1.length; i < l; i++) {
+    for (let j = 0, m = a2.length; j < m; j++) {
+      result.push(a1[i] + a2[j]);
+    }
+  }
+  return result;
+  // arr.forEach(arrItem => {
+  //   resultItem = '';
+  //   arrItem.forEach(item => {
+  //     resultItem += item;
+  //   });
+  //   result.push(resultItem);
+  // });
+}
+
+/**
+ * 合并二维元祖。
+ * @param {Array<Array<String>>} arr 二维元祖 [["zhāo", "cháo"], ["yáng"], ["dōng"], ["shēng"]]
+ * @return {Array<String>} 返回二维字符串组合数组。形如
+ *  [
+ *    ["zhāoyáng"], ["dōng"], ["shēng"],
+ *    ["cháoyáng"], ["dōng"], ["shēng"]
+ *  ]
+ */
+function combo(arr) {
+  if (arr.length === 0) {
+    return [];
+  }
+  if (arr.length === 1) {
+    return arr[0];
+  }
+  let result = combo2array(arr[0], arr[1]);
+  for (let i = 2, l = arr.length; i < l; i++) {
+    result = combo2array(result, arr[i]);
+  }
+  return result;
+}
+
+exports.combo2array = combo2array;
+exports.combo = combo;

tests/test.js

@@ -268,3 +268,17 @@ describe("pinyin.compare", function() {
     expect(sortedData).to.eql("排我序要".split(""));
   });
 });
+
+describe("pinyin group", function() {
+  it("groups segments", function () {
+    const han = "我喜欢你";
+    const py = pinyin(han, {segment: true, group: true, heteronym: true});
+    expect(py).to.eql([["wǒ"], ["xǐhuān"], ["nǐ"]]);
+  });
+
+  it("groups segments with heteronyms", function() {
+    const han = "我都喜欢朝阳";
+    const py = pinyin(han, {segment: true, group: true, heteronym: true});
+    expect(py).to.eql([["wǒ"], ["dū", "dōu"], ["xǐhuān"], ["zhāoyáng", "cháoyáng"]]);
+  });
+});

tests/util.test.js

@@ -0,0 +1,70 @@
+"use strict";
+
+const expect = require("expect.js");
+const util = require("../lib/util");
+
+describe("test/util.test.js", function() {
+  describe("combo2array", function() {
+    it("combo2array([], [])", function() {
+      expect(util.combo2array([], [])).to.eql([]);
+    });
+
+    it("combo2array([a], [])", function() {
+      expect(util.combo2array(["a"], [])).to.eql(["a"]);
+    });
+
+    it("combo2array([], [1])", function() {
+      expect(util.combo2array([], ["1"])).to.eql(["1"]);
+    });
+
+    it("combo2array([a], [1])", function() {
+      expect(util.combo2array(["a"], ["1"])).to.eql(["a1"]);
+    });
+
+    it("combo2array([a,b], [1])", function() {
+      expect(util.combo2array(["a", "b"], ["1"])).to.eql(["a1", "b1"]);
+    });
+
+    it("combo2array([a], [1,2])", function() {
+      expect(util.combo2array(["a"], ["1", "2"])).to.eql(["a1", "a2"]);
+    });
+
+    it("combo2array([a,b], [1,2])", function() {
+      expect(util.combo2array(["a", "b"], ["1", "2"])).to.eql(["a1", "a2", "b1", "b2"]);
+    });
+
+    it("combo2array([a,b,c], [1,2,3])", function() {
+      expect(util.combo2array(["a", "b", "c"], ["1", "2", "3"])).to.eql(["a1", "a2", "a3", "b1", "b2", "b3", "c1", "c2", "c3"]);
+    });
+  });
+
+  describe("combo", function() {
+    it("combo([])", function() {
+      expect(util.combo([])).to.eql([]);
+    });
+
+    it("combo([[a]])", function() {
+      expect(util.combo([["a"]])).to.eql([["a"]]);
+    });
+
+    it("combo([[a,b]])", function() {
+      expect(util.combo([["a", "b"]])).to.eql([["a", "b"]]);
+    });
+
+    it("combo([[a,b],[1]])", function() {
+      expect(util.combo([["a", "b"], ["1"]])).to.eql([["a1", "b1"]]);
+    });
+
+    it("combo([[a,b],[1,2]])", function() {
+      expect(util.combo([["a", "b"], ["1", "2"]])).to.eql([["a1", "a2", "b1", "b2"]]);
+    });
+
+    it("combo([[a,b],[1,2],[A]])", function() {
+      expect(util.combo([["a", "b"], ["1", "2"], ["A"]])).to.eql([["a1A", "a2A", "b1A", "b2A"]]);
+    });
+
+    it("combo([[a,b],[1,2],[A,B]])", function() {
+      expect(util.combo([["a", "b"], ["1", "2"], ["A", "B"]])).to.eql([["a1A", "a1B", "a2A", "a2B", "b1A", "b1B", "b2A", "b2B"]]);
+    });
+  });
+});