104 lines
3.2 KiB
JavaScript
104 lines
3.2 KiB
JavaScript
const fs = require('fs');
|
||
const path = require('path');
|
||
|
||
const DataFile = path.join(__dirname, './data.json')
|
||
const DataFile2 = path.join(__dirname, './data-friendly.json')
|
||
const DataFile3 = path.join(__dirname, './data-single-word.txt')
|
||
let data = JSON.parse(fs.readFileSync(DataFile, 'utf8'))
|
||
|
||
function structure() {
|
||
return {
|
||
// 汉字
|
||
word: "",
|
||
// 拼音
|
||
pinyin: [],
|
||
// 部首
|
||
radical: [],
|
||
// 笔画
|
||
stroke: 0,
|
||
// 类型 0-未标注 1-常用字 2-生僻字
|
||
type: 0,
|
||
// 来源（从original文件夹中哪个来源录入的）
|
||
from: [],
|
||
// 是否人工确认过
|
||
confirm: false,
|
||
}
|
||
}
|
||
|
||
function saveData() {
|
||
const ObjSort = function (arys) {
|
||
var newkey = Object.keys(arys).sort((a, b) => a.localeCompare(b));
|
||
var newObj = {}; //创建一个新的对象，用于存放排好序的键值对
|
||
for (var i = 0; i < newkey.length; i++) {
|
||
newObj[newkey[i]] = arys[newkey[i]];
|
||
}
|
||
return newObj; //返回排好序的新对象
|
||
}
|
||
data = ObjSort(data)
|
||
fs.writeFileSync(DataFile, JSON.stringify(data), 'utf8')
|
||
fs.writeFileSync(DataFile2, JSON.stringify(data, null, 4), 'utf8')
|
||
fs.writeFileSync(DataFile3, Object.keys(data).join(''), 'utf8')
|
||
}
|
||
|
||
function addData(wordStruct, saveData = true) {
|
||
let word = wordStruct.word
|
||
if (Object.keys(data).includes(word)) {
|
||
// 已经存在
|
||
let struct = data[word]
|
||
let isNeedUpdate = false
|
||
|
||
// 合并拼音
|
||
for (let py of wordStruct.pinyin) {
|
||
if (!struct.pinyin.includes(py)) {
|
||
struct.pinyin.push(py)
|
||
isNeedUpdate = true
|
||
}
|
||
}
|
||
struct.pinyin.sort()
|
||
|
||
// 合并笔画
|
||
if (wordStruct.stroke > 0 && wordStruct.stroke != struct.stroke) {
|
||
if (struct.stroke != 0) {
|
||
console.log(`${wordStruct.word} 字的笔画(stroke)出现歧义 [${struct.stroke}, ${wordStruct.stroke}]，当前保存 ${wordStruct.stroke}`)
|
||
}
|
||
struct.stroke = wordStruct.stroke
|
||
}
|
||
|
||
// 合并类型
|
||
if (wordStruct.type > 0 && wordStruct.type != struct.type) {
|
||
if (struct.type != 0) {
|
||
// 同时在生僻字和常见字出现 记为常见字
|
||
console.log(`${wordStruct.word} 字的类型(type)出现歧义 [${struct.type}, ${wordStruct.type}]，当前保存 1`)
|
||
struct.type = 1
|
||
}
|
||
struct.type = wordStruct.type
|
||
}
|
||
|
||
// 合并来源
|
||
for (let fr of wordStruct.from) {
|
||
if (!struct.from.includes(fr)) {
|
||
struct.from.push(fr)
|
||
isNeedUpdate = true
|
||
}
|
||
}
|
||
struct.from.sort()
|
||
|
||
if (isNeedUpdate) {
|
||
data[word] = struct
|
||
saveData && saveData()
|
||
}
|
||
} else {
|
||
// 还不存在
|
||
data[wordStruct.word] = wordStruct
|
||
isNeedUpdate = true
|
||
|
||
saveData && saveData()
|
||
}
|
||
|
||
}
|
||
|
||
module.exports = {
|
||
structure,
|
||
addData,
|
||
saveData
|
||
} |