feat: 只根据核心字符比较

feat: 处理列表和标题等不必紧跟换行符的格式

aio/tools/translator/translate.spec.ts

@@ -1,21 +1,30 @@
 import { expect } from 'chai';
 import { dirs } from './dirs';
-import { dict, lookup, translate } from './translate';
+import { dict, kernelText, lookup, normalizeLines, translate } from './translate';
 
 
 describe('根据字典进行翻译', () => {
   it('忽略明显错误的条目', function () {
     expect(dict.filter(entry => /^<div/.test(entry.original))).eql([]);
   });
+
+  it('抽取核心字符', function () {
+    expect(kernelText(' # Forms   ABC ')).eql('# Forms ABC');
+  });
+
   it('查字典', () => {
     expect(lookup('# Forms')[0].translation).eql('# 表单');
   });
 
+  it('把“- * 1. #”等处理成空行分隔的格式，以便处理', function () {
+    const lines = normalizeLines('1. abc\n11. def\n');
+    expect(lines).eql('1. abc\n\n11. def\n');
+  });
+
   it('自动根据字典翻译单个文件', function () {
     const fs = require('fs');
-    const content = fs.readFileSync(dirs.content + 'guide/forms.md', 'utf-8');
+    const content = fs.readFileSync(__dirname + '/../../../../content-en/' + 'guide/forms.md', 'utf-8');
     const result = translate(content);
     fs.writeFileSync(dirs.content + 'guide/forms.md', result.join('\n\n'), 'utf-8');
   });
-
 });

aio/tools/translator/translate.ts

@@ -4,22 +4,29 @@ import { DictEntry } from './dict-entry';
 const dict1 = require('./dict-1.json') as DictEntry[];
 const dict2 = require('./dict-2.json') as DictEntry[];
 const dict3 = require('./dict-3.json') as DictEntry[];
-export const dict = dict1.concat(dict2).concat(dict3).filter(entry => !/^\s*<div/.test(entry.original));
+export const dict = dict1.concat(dict2).concat(dict3)
+  .filter(entry => !/^\s*<div/.test(entry.original));
 
 export function lookup(english: string, filename: RegExp = /.*/): DictEntry[] {
-  return _.uniqBy(dict.filter(entry => filename.test(entry.sourceFile)).filter(entry => entry.original === english), 'translation');
+  let entries = dict3
+    .filter(entry => filename.test(entry.sourceFile))
+    .filter(entry => kernelText(entry.original) === kernelText(english));
+  return _.uniqBy(entries, 'translation');
+}
+
+export function kernelText(text: string): string {
+  return text.replace(/[\s\n]+/g, ' ').trim();
 }
 
 export function translate(content: string): string[] {
-  const lines = content.split(/\n+\s*\n+/);
+  const lines = normalizeLines(content)
+    .split(/\n+\s*\n+/);
   return lines
-    .map(splitList)
-    .map(flatten)
     .map(line => {
       if (!line.trim()) {
         return line;
       }
-      const translations = lookup(line.trim(), /forms.md$/);
+      const translations = lookup(line, /forms.md$/);
       const indent = indentOf(line);
       const padding = repeat(indent);
       if (translations.length === 0) {
@@ -53,12 +60,6 @@ function repeat(indent: number): string {
   return result;
 }
 
-function splitList(line: string): string[] {
-  const subLines = line.split(/\n/);
-  // 把 markdown 里面不必空行的元素都拆开
-  return [line];
-}
-
-export function flatten<T>(arr: T[]): T {
-  return Array.prototype.concat.apply([], arr);
+export function normalizeLines(text: string): string {
+  return text.replace(/(?=\n *(\d+\.|-|\*|#|<) )\n/g, '\n\n');
 }