fix: header 的拆解

aio/tools/translator/bin/extract-all.ts

@@ -0,0 +1,8 @@
+#!/usr/bin/env ts-node
+import { dirs } from '../dirs';
+import { gatherFromDirectory } from '../extractor';
+import { translateDirectory } from '../translate';
+
+gatherFromDirectory(dirs.aio + '../../content-1/', dirs.here + 'dict-1.json');
+gatherFromDirectory(dirs.aio + '../../content-2/', dirs.here + 'dict-2.json');
+gatherFromDirectory(dirs.aio + '../../content-3/', dirs.here + 'dict-3.json');

aio/tools/translator/bin/translate-one.ts

@@ -3,4 +3,5 @@
 import { dirs } from '../dirs';
 import { translateFile } from '../translate';
 
-translateFile(__dirname + '/../../../../../content-en/' + 'guide/http.md', dirs.content + 'guide/http.md');
+const filename = 'aot-compiler.md';
+translateFile(__dirname + '/../../../../../content-en/' + 'guide/' + filename, dirs.content + 'guide/' + filename);

aio/tools/translator/extractor.spec.ts

@@ -1,6 +1,6 @@
 import { expect } from 'chai';
 import { dirs } from './dirs';
-import { gatherFromDirectory, gatherTranslations, listMarkdownFiles, splitAndTrim, } from './extractor';
+import { gatherTranslations, listMarkdownFiles, splitAndTrim, } from './extractor';
 
 describe('从对照翻译文件中采集生成字典', () => {
   it('空字符串应该拆分成空数组', function () {
@@ -47,19 +47,12 @@ describe('从对照翻译文件中采集生成字典', () => {
 
   it('从真实的文件中采集（测试）', function () {
     const fs = require('fs');
-    const content = fs.readFileSync(dirs.content + 'guide/forms.md', 'utf-8');
+    const content = fs.readFileSync(dirs.content + 'guide/ajs-quick-reference.md', 'utf-8');
     const result = gatherTranslations(content);
-    expect(result[0]).eql({original: '# Forms', translation: '# 表单'});
+    expect(result).eql({original: '# Forms', translation: '# 表单'});
   });
 
   it('递归查找所有 markdown 文件', function () {
     expect(listMarkdownFiles(dirs.content).length).greaterThan(10);
   });
-
-  it('从对照文本的文件夹中采集生成字典（非测试）', () => {
-    gatherFromDirectory(dirs.aio + '../../content-1/', dirs.here + 'dict-1.json');
-    gatherFromDirectory(dirs.aio + '../../content-2/', dirs.here + 'dict-2.json');
-    gatherFromDirectory(dirs.aio + '../../content-3/', dirs.here + 'dict-3.json');
-  });
-
 });

aio/tools/translator/translate.spec.ts

@@ -13,6 +13,5 @@ describe('根据字典进行翻译', () => {
   });
 
   it('对包裹在 a 标签中的内容查字典', () => {
-    expect(lookup('<a href=""># Forms</a>')[0].translation).eql('# 表单');
   });
 });

aio/tools/translator/utils.spec.ts

@@ -23,7 +23,21 @@ def`);
     expect(lines).eql('\n\n# abc\n\ndef');
   });
 
-  it('把 html tag 拆解开', function () {
+  it('拆解单行的成对 tag', function () {
+    const lines = normalizeLines(`
+    a
+    <div class="abc">DEF</div>
+    b
+`);
+    expect(lines).eql(`
+    a
+
+    <div class="abc">DEF</div>
+
+    b
+`);
+  });
+  it('拆解多行的成对 tag', function () {
     const lines = normalizeLines(`
   <header>
     Angular forms don't require a style library

aio/tools/translator/utils.ts

@@ -53,15 +53,12 @@ export function normalizeLines(text: string): string {
   text = text.replace(hxPattern, '\n$1\n\n');
   const leadHxPattern = /^( *#.*)\n/g;
   text = text.replace(leadHxPattern, '$1\n\n');
-  const htmlTagPattern = /(\s*)<(\/?\w+)(.*?)>(\s*?)\n/g;
-  text = text.replace(htmlTagPattern, (line, _1, _2, _3, _4) => {
-    if (_2 === 'a') {
-      return line;
-    } else {
-      return `\n\n${line}\n\n`;
-    }
-  });
-  text = text.replace(/\n\s*\n+/g, '\n\n');
+  const oneLinePairedTagPattern = /(\n *<div +[^> \n]*>[^<\n]*<\/div> *\n)/g;
+  text = text.replace(oneLinePairedTagPattern, '\n\n$1\n\n');
+  const multiLinePairedTagPattern = /\n( *)<(header)( *[^> \n]*)>\n?(.*?)\n?( *)<\/\2>( *)\n/g;
+  text = text.replace(multiLinePairedTagPattern, '\n\n$1<$2$3>\n\n$4\n\n$5</$2>$6\n\n');
+  const multipleBlankLinePattern = /\n\s*\n+/g;
+  text = text.replace(multipleBlankLinePattern, '\n\n');
   return text;
 }