fix some pinyin issues

index.js

@@ -1,12 +1,13 @@
 var pinyinify = require("./src/pinyinify"),
     segment = require("./src/segment"),
+    tag = require('./src/tag'),
     { simplify, traditionalize } = require("./src/simplify");
-let nodejieba = require("nodejieba");
+
 
 module.exports = {
     pinyinify,
     segment,
     simplify,
     traditionalize,
-    tag: nodejieba.tag
+    tag
 };

spec/spec.js

@@ -41,14 +41,19 @@ describe("Pinyinify", () => {
         expect("我说得快。").becomes("wǒ shuō de kuài.");
         expect("你们觉得我画得怎么样？").becomes("nǐ​men jué​de wǒ huà de zěn​me​yàng?");
         expect("喂，睡得还好吗？").becomes("wèi, shuì de hái​hǎo ma?");
+        expect("你得做个决定。").becomes("nǐ děi zuò gè jué​dìng.");
+        expect("你得在这儿休息。").becomes("nǐ děi zài zhè​r xiū​xi.");
         // 还
         expect("我有钱了就还你。").becomes("wǒ yǒu​qián le jiù huán nǐ.");
         expect("我还给你。").becomes("wǒ huán​gěi nǐ.");
         expect("你还喜欢她吗？").becomes("nǐ hái xǐ​huan tā ma?");
         expect("你得还我的东西。").becomes("nǐ děi huán wǒ de dōng​xi.");
         expect("他还会把钱还律师吗？").becomes("tā hái huì bǎ qián huán lǜ​shī ma?");
+        expect("好吧，我至少还有些朋友。").becomes("hǎo ba, wǒ zhì​shǎo hái yǒu​xiē péng​you.");
         // 只
         expect("他是一只鸟。").becomes("tā shì yī zhī niǎo.");
+        // 长
+        expect("她长着一张圆脸和一双明亮的眼睛。").becomes("tā zhǎng zhe yī zhāng yuán liǎn hé yī shuāng míng​liàng de yǎn​jing.");
 
         expect("行了吗？").becomes("xíng le ma?");
         expect("人要是行干一行行一行。").becomes("rén yào​shi xíng gàn yī háng xíng yī háng.");

src/pinyinDict.js

@@ -100337,7 +100337,7 @@ let dict = [
     "重音|||重音|zhòng​yīn",
     "重音节|||重音節|zhòng​yīn​jié",
     "重头戏|||重頭戲|zhòng​tóu​xì",
-    "重点|||重點|chóng​diǎn",
+    "重点|||重點|zhòngdiǎn",
     "野人|||野人|yě​rén",
     "野兔|||野兔|yě​tù",
     "野叟曝言|||野叟曝言|Yě​sǒu​Pù​yán",

src/pinyinify.js

@@ -2,7 +2,8 @@ let segment = require("./segment"),
     pinyin = require("pinyin"),
     { pinyinDict } = require("./pinyinDict"),
     { normalizeEnglish } = require("./punctuation"),
-    tag = require("nodejieba").tag;
+    tag = require("./tag"),
+    { isCharacterText } = require("./util");
 
 function pinyinify(text, isDetailed) {
     let segments = segment(text);
@@ -62,6 +63,9 @@ function decideAmbiguousChar(char, cuts, cutIndex) {
             return "jué";
         case "长":
         case "長":
+            nextTags = tag(afterText.join(""));
+            if (nextTags && nextTags.length && nextTags[0].tag === "uz")
+                return "zhǎng";
             // zhǎng has higher frequency due to compond words, 
             // but cháng is more common as an individual character.
             return "cháng";
@@ -80,7 +84,7 @@ function decideAmbiguousChar(char, cuts, cutIndex) {
                         return "děi";
                     }
                 }
-                if (afterTag[0] === "t" || afterTag[0] === "v") {
+                if (afterTag[0] === "t" || afterTag[0] === "v" || afterTag[0] === "p") {
                     return "děi";
                 }
             }
@@ -90,6 +94,7 @@ function decideAmbiguousChar(char, cuts, cutIndex) {
             nextTags = tag(afterText.join(""));
             if (nextTags && nextTags.length) {
                 let afterTag = nextTags[0].tag;
+                if (afterText[0][0] === "有") break;
                 if (afterTag[0] === "r" || afterTag[0] === "n") {
                     return "huán";
                 }
@@ -111,11 +116,6 @@ function decideAmbiguousChar(char, cuts, cutIndex) {
     }
 }
 
-function isCharacterText(text) {
-    // https://stackoverflow.com/questions/21109011/javascript-unicode-string-chinese-character-but-no-punctuation
-    return /^([\u4E00-\u9FCC\u3400-\u4DB5\uFA0E\uFA0F\uFA11\uFA13\uFA14\uFA1F\uFA21\uFA23\uFA24\uFA27-\uFA29]|[\ud840-\ud868][\udc00-\udfff]|\ud869[\udc00-\uded6\udf00-\udfff]|[\ud86a-\ud86c][\udc00-\udfff]|\ud86d[\udc00-\udf34\udf40-\udfff]|\ud86e[\udc00-\udc1d])+$/.test(String(text));
-}
-
 function shouldPutSpaceBetween(word1, word2) {
     if (!word2) return false;
     if (isCharacterText(word1) && isCharacterText(word2)) return true;

src/tag.js

@@ -0,0 +1,18 @@
+let nodejieba = require("nodejieba"),
+    { isCharacterText } = require("./util");
+
+function tag(text) {
+    let tokens = nodejieba.tag(text);
+    let outTokens = [];
+    for (let { word, tag } of tokens) {
+        if (word.length > 1 && tag === "x") {
+            for (let char of word) {
+                outTokens.push(nodejieba.tag(char)[0]);
+            }
+        }
+        else outTokens.push({ word, tag });
+    }
+    return outTokens;
+}
+
+module.exports = tag;

src/util.js

@@ -0,0 +1,6 @@
+module.exports = {
+    isCharacterText: (text) => {
+        // https://stackoverflow.com/questions/21109011/javascript-unicode-string-chinese-character-but-no-punctuation
+        return /^([\u4E00-\u9FCC\u3400-\u4DB5\uFA0E\uFA0F\uFA11\uFA13\uFA14\uFA1F\uFA21\uFA23\uFA24\uFA27-\uFA29]|[\ud840-\ud868][\udc00-\udfff]|\ud869[\udc00-\uded6\udf00-\udfff]|[\ud86a-\ud86c][\udc00-\udfff]|\ud86d[\udc00-\udf34\udf40-\udfff]|\ud86e[\udc00-\udc1d])+$/.test(String(text));
+    }
+}