wespy-ts 0.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +146 -0
- package/dist/cli/main.d.ts +7 -0
- package/dist/cli/main.d.ts.map +1 -0
- package/dist/cli/main.js +312 -0
- package/dist/cli/main.js.map +1 -0
- package/dist/converter/html-to-markdown.d.ts +9 -0
- package/dist/converter/html-to-markdown.d.ts.map +1 -0
- package/dist/converter/html-to-markdown.js +171 -0
- package/dist/converter/html-to-markdown.js.map +1 -0
- package/dist/converter/sanitize-html.d.ts +12 -0
- package/dist/converter/sanitize-html.d.ts.map +1 -0
- package/dist/converter/sanitize-html.js +22 -0
- package/dist/converter/sanitize-html.js.map +1 -0
- package/dist/core/errors.d.ts +17 -0
- package/dist/core/errors.d.ts.map +1 -0
- package/dist/core/errors.js +36 -0
- package/dist/core/errors.js.map +1 -0
- package/dist/core/result.d.ts +26 -0
- package/dist/core/result.d.ts.map +1 -0
- package/dist/core/result.js +26 -0
- package/dist/core/result.js.map +1 -0
- package/dist/core/types.d.ts +156 -0
- package/dist/core/types.d.ts.map +1 -0
- package/dist/core/types.js +29 -0
- package/dist/core/types.js.map +1 -0
- package/dist/fetcher/http-client.d.ts +31 -0
- package/dist/fetcher/http-client.d.ts.map +1 -0
- package/dist/fetcher/http-client.js +124 -0
- package/dist/fetcher/http-client.js.map +1 -0
- package/dist/index.d.ts +14 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +14 -0
- package/dist/index.js.map +1 -0
- package/dist/platforms/detector.d.ts +15 -0
- package/dist/platforms/detector.d.ts.map +1 -0
- package/dist/platforms/detector.js +30 -0
- package/dist/platforms/detector.js.map +1 -0
- package/dist/platforms/generic/generic-article.extractor.d.ts +25 -0
- package/dist/platforms/generic/generic-article.extractor.d.ts.map +1 -0
- package/dist/platforms/generic/generic-article.extractor.js +171 -0
- package/dist/platforms/generic/generic-article.extractor.js.map +1 -0
- package/dist/platforms/juejin/juejin-article.extractor.d.ts +20 -0
- package/dist/platforms/juejin/juejin-article.extractor.d.ts.map +1 -0
- package/dist/platforms/juejin/juejin-article.extractor.js +167 -0
- package/dist/platforms/juejin/juejin-article.extractor.js.map +1 -0
- package/dist/platforms/juejin/juejin.types.d.ts +13 -0
- package/dist/platforms/juejin/juejin.types.d.ts.map +1 -0
- package/dist/platforms/juejin/juejin.types.js +5 -0
- package/dist/platforms/juejin/juejin.types.js.map +1 -0
- package/dist/platforms/wechat/wechat-album.extractor.d.ts +25 -0
- package/dist/platforms/wechat/wechat-album.extractor.d.ts.map +1 -0
- package/dist/platforms/wechat/wechat-album.extractor.js +190 -0
- package/dist/platforms/wechat/wechat-album.extractor.js.map +1 -0
- package/dist/platforms/wechat/wechat-article.extractor.d.ts +20 -0
- package/dist/platforms/wechat/wechat-article.extractor.d.ts.map +1 -0
- package/dist/platforms/wechat/wechat-article.extractor.js +132 -0
- package/dist/platforms/wechat/wechat-article.extractor.js.map +1 -0
- package/dist/platforms/wechat/wechat.types.d.ts +17 -0
- package/dist/platforms/wechat/wechat.types.d.ts.map +1 -0
- package/dist/platforms/wechat/wechat.types.js +5 -0
- package/dist/platforms/wechat/wechat.types.js.map +1 -0
- package/dist/sdk/fetch-album-list.d.ts +10 -0
- package/dist/sdk/fetch-album-list.d.ts.map +1 -0
- package/dist/sdk/fetch-album-list.js +31 -0
- package/dist/sdk/fetch-album-list.js.map +1 -0
- package/dist/sdk/fetch-album.d.ts +24 -0
- package/dist/sdk/fetch-album.d.ts.map +1 -0
- package/dist/sdk/fetch-album.js +67 -0
- package/dist/sdk/fetch-album.js.map +1 -0
- package/dist/sdk/fetch-article.d.ts +24 -0
- package/dist/sdk/fetch-article.d.ts.map +1 -0
- package/dist/sdk/fetch-article.js +111 -0
- package/dist/sdk/fetch-article.js.map +1 -0
- package/dist/utils/fs.d.ts +16 -0
- package/dist/utils/fs.d.ts.map +1 -0
- package/dist/utils/fs.js +26 -0
- package/dist/utils/fs.js.map +1 -0
- package/dist/utils/text.d.ts +20 -0
- package/dist/utils/text.d.ts.map +1 -0
- package/dist/utils/text.js +96 -0
- package/dist/utils/text.js.map +1 -0
- package/dist/utils/url.d.ts +22 -0
- package/dist/utils/url.d.ts.map +1 -0
- package/dist/utils/url.js +63 -0
- package/dist/utils/url.js.map +1 -0
- package/package.json +64 -0
package/README.md
ADDED
|
@@ -0,0 +1,146 @@
|
|
|
1
|
+
# WeSpy (TypeScript)
|
|
2
|
+
|
|
3
|
+
文章抓取与 Markdown 转换工具 — TypeScript 版本。
|
|
4
|
+
|
|
5
|
+
## 支持平台
|
|
6
|
+
|
|
7
|
+
| 平台 | 单篇文章 | 专辑/合集 |
|
|
8
|
+
|------|---------|----------|
|
|
9
|
+
| 微信公众号 (`mp.weixin.qq.com`) | ✅ | ✅ |
|
|
10
|
+
| 掘金 (`juejin.cn`) | ✅ | — |
|
|
11
|
+
| 通用网页 | ✅ | — |
|
|
12
|
+
|
|
13
|
+
## 安装
|
|
14
|
+
|
|
15
|
+
### 全局安装
|
|
16
|
+
|
|
17
|
+
```bash
|
|
18
|
+
npm install -g wespy-ts
|
|
19
|
+
```
|
|
20
|
+
|
|
21
|
+
### 项目依赖
|
|
22
|
+
|
|
23
|
+
```bash
|
|
24
|
+
npm install wespy-ts
|
|
25
|
+
```
|
|
26
|
+
|
|
27
|
+
### 本地开发
|
|
28
|
+
|
|
29
|
+
```bash
|
|
30
|
+
git clone git@github.com:Cuimc/WeSpy-TS.git
|
|
31
|
+
cd WeSpy-TS
|
|
32
|
+
npm install
|
|
33
|
+
npm run build
|
|
34
|
+
```
|
|
35
|
+
|
|
36
|
+
## CLI 使用
|
|
37
|
+
|
|
38
|
+
全局安装后可直接使用 `wespy` 命令,也可以通过 `npx` 运行:
|
|
39
|
+
|
|
40
|
+
```bash
|
|
41
|
+
# 获取单篇文章(默认输出 Markdown)
|
|
42
|
+
wespy fetch-article https://mp.weixin.qq.com/s/xxxxx
|
|
43
|
+
|
|
44
|
+
# 使用 npx 运行
|
|
45
|
+
npx wespy fetch-article https://mp.weixin.qq.com/s/xxxxx
|
|
46
|
+
|
|
47
|
+
# 指定输出格式和目录
|
|
48
|
+
wespy fetch-article https://mp.weixin.qq.com/s/xxxxx -o ./output --all
|
|
49
|
+
|
|
50
|
+
# 批量下载微信专辑
|
|
51
|
+
wespy fetch-album https://mp.weixin.qq.com/mp/appmsgalbum?... --max 5
|
|
52
|
+
|
|
53
|
+
# 智能模式(自动判断 URL 类型)
|
|
54
|
+
wespy https://mp.weixin.qq.com/s/xxxxx -o ./output --html --json
|
|
55
|
+
```
|
|
56
|
+
|
|
57
|
+
### CLI 参数
|
|
58
|
+
|
|
59
|
+
| 参数 | 说明 | 默认值 |
|
|
60
|
+
|------|------|--------|
|
|
61
|
+
| `-o, --output <dir>` | 输出目录 | `articles` |
|
|
62
|
+
| `--html` | 同时保存 HTML 文件 | false |
|
|
63
|
+
| `--json` | 同时保存 JSON 文件 | false |
|
|
64
|
+
| `--all` | 保存所有格式 | false |
|
|
65
|
+
| `--max <n>` | 专辑最大下载数 | 10 |
|
|
66
|
+
| `--timeout <ms>` | 请求超时时间 | 30000 |
|
|
67
|
+
| `--album-only` | 仅获取文章列表 | false |
|
|
68
|
+
|
|
69
|
+
## SDK API
|
|
70
|
+
|
|
71
|
+
```typescript
|
|
72
|
+
import { fetchArticle, fetchAlbum } from 'wespy-ts'
|
|
73
|
+
|
|
74
|
+
// 获取单篇文章
|
|
75
|
+
const result = await fetchArticle({
|
|
76
|
+
url: 'https://mp.weixin.qq.com/s/xxxxx',
|
|
77
|
+
format: ['markdown', 'json'],
|
|
78
|
+
outputDir: './output',
|
|
79
|
+
})
|
|
80
|
+
|
|
81
|
+
if (result.ok) {
|
|
82
|
+
console.log(result.article.title)
|
|
83
|
+
console.log(result.article.markdown)
|
|
84
|
+
console.log(result.artifacts)
|
|
85
|
+
} else {
|
|
86
|
+
console.error(result.error.code, result.error.message)
|
|
87
|
+
}
|
|
88
|
+
|
|
89
|
+
// 批量获取专辑
|
|
90
|
+
const albumResult = await fetchAlbum({
|
|
91
|
+
url: 'https://mp.weixin.qq.com/mp/appmsgalbum?...',
|
|
92
|
+
maxArticles: 10,
|
|
93
|
+
format: ['markdown'],
|
|
94
|
+
})
|
|
95
|
+
|
|
96
|
+
if (albumResult.ok) {
|
|
97
|
+
console.log(`获取了 ${albumResult.articles.length} 篇文章`)
|
|
98
|
+
}
|
|
99
|
+
```
|
|
100
|
+
|
|
101
|
+
## 输出结构
|
|
102
|
+
|
|
103
|
+
### ArticleDraft
|
|
104
|
+
|
|
105
|
+
```typescript
|
|
106
|
+
interface ArticleDraft {
|
|
107
|
+
platform: 'wechat' | 'juejin' | 'generic'
|
|
108
|
+
url: string
|
|
109
|
+
title: string
|
|
110
|
+
author?: string
|
|
111
|
+
publishTime?: string
|
|
112
|
+
contentHtml?: string
|
|
113
|
+
contentText?: string
|
|
114
|
+
markdown?: string
|
|
115
|
+
metadata: Record<string, unknown>
|
|
116
|
+
fetchedAt: string
|
|
117
|
+
warnings: string[]
|
|
118
|
+
}
|
|
119
|
+
```
|
|
120
|
+
|
|
121
|
+
### FetchArticleResult
|
|
122
|
+
|
|
123
|
+
```typescript
|
|
124
|
+
// 成功
|
|
125
|
+
{ ok: true, article: ArticleDraft, artifacts: OutputArtifact[], warnings: string[] }
|
|
126
|
+
|
|
127
|
+
// 失败
|
|
128
|
+
{ ok: false, error: { code: string, message: string, retryable: boolean } }
|
|
129
|
+
```
|
|
130
|
+
|
|
131
|
+
## 测试
|
|
132
|
+
|
|
133
|
+
```bash
|
|
134
|
+
npm test
|
|
135
|
+
```
|
|
136
|
+
|
|
137
|
+
## 开发
|
|
138
|
+
|
|
139
|
+
```bash
|
|
140
|
+
npm run dev # watch 模式
|
|
141
|
+
npm run lint # 类型检查
|
|
142
|
+
```
|
|
143
|
+
|
|
144
|
+
## Python 原版
|
|
145
|
+
|
|
146
|
+
Python 原版代码保留在 `../WeSpy/` 目录中,未被删除。详见 [MIGRATION.md](./MIGRATION.md)。
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"main.d.ts","sourceRoot":"","sources":["../../src/cli/main.ts"],"names":[],"mappings":";AACA;;;GAGG"}
|
package/dist/cli/main.js
ADDED
|
@@ -0,0 +1,312 @@
|
|
|
1
|
+
#!/usr/bin/env node
|
|
2
|
+
/**
|
|
3
|
+
* WeSpy CLI 入口
|
|
4
|
+
* 使用 commander 实现,调用 SDK API
|
|
5
|
+
*/
|
|
6
|
+
import { createInterface } from 'node:readline';
|
|
7
|
+
import { readFileSync } from 'node:fs';
|
|
8
|
+
import { join, dirname } from 'node:path';
|
|
9
|
+
import { fileURLToPath } from 'node:url';
|
|
10
|
+
import { Command } from 'commander';
|
|
11
|
+
import { fetchArticle } from '../sdk/fetch-article.js';
|
|
12
|
+
import { fetchAlbum } from '../sdk/fetch-album.js';
|
|
13
|
+
import { fetchAlbumList } from '../sdk/fetch-album-list.js';
|
|
14
|
+
import { isWechatAlbumUrl } from '../platforms/detector.js';
|
|
15
|
+
import { writeJsonSafe, ensureDir } from '../utils/fs.js';
|
|
16
|
+
// 从 package.json 读取版本号,避免硬编码
|
|
17
|
+
const __dirname = dirname(fileURLToPath(import.meta.url));
|
|
18
|
+
const pkg = JSON.parse(readFileSync(join(__dirname, '../../package.json'), 'utf-8'));
|
|
19
|
+
function addCommonOptions(cmd) {
|
|
20
|
+
return cmd
|
|
21
|
+
.option('-o, --output <dir>', '输出目录', 'articles')
|
|
22
|
+
.option('--html', '同时保存 HTML 文件')
|
|
23
|
+
.option('--json', '同时保存 JSON 文件')
|
|
24
|
+
.option('--all', '保存所有格式 (HTML + JSON + Markdown)')
|
|
25
|
+
.option('--download-images', '下载图片到本地(当前迁移版本暂未实现)')
|
|
26
|
+
.option('--timeout <ms>', '请求超时时间 (毫秒)', '30000')
|
|
27
|
+
.option('-v, --verbose', '显示详细信息');
|
|
28
|
+
}
|
|
29
|
+
/** 打印文章信息,缺失字段不显示 */
|
|
30
|
+
function printArticleInfo(article) {
|
|
31
|
+
console.log(` 标题: ${article.title}`);
|
|
32
|
+
if (article.author)
|
|
33
|
+
console.log(` 作者: ${article.author}`);
|
|
34
|
+
if (article.publishTime)
|
|
35
|
+
console.log(` 时间: ${article.publishTime}`);
|
|
36
|
+
}
|
|
37
|
+
function resolveFormats(options) {
|
|
38
|
+
if (options.all)
|
|
39
|
+
return ['markdown', 'html', 'json'];
|
|
40
|
+
const formats = ['markdown'];
|
|
41
|
+
if (options.html)
|
|
42
|
+
formats.push('html');
|
|
43
|
+
if (options.json)
|
|
44
|
+
formats.push('json');
|
|
45
|
+
return formats;
|
|
46
|
+
}
|
|
47
|
+
function printVerbose(options, url, extra) {
|
|
48
|
+
const formats = resolveFormats(options);
|
|
49
|
+
console.log('\n[verbose] 配置信息:');
|
|
50
|
+
console.log(` URL: ${url}`);
|
|
51
|
+
console.log(` 输出目录: ${options.output}`);
|
|
52
|
+
console.log(` 输出格式: HTML=${formats.includes('html')}, JSON=${formats.includes('json')}, Markdown=${formats.includes('markdown')}`);
|
|
53
|
+
if (extra) {
|
|
54
|
+
for (const [k, v] of Object.entries(extra)) {
|
|
55
|
+
console.log(` ${k}: ${v}`);
|
|
56
|
+
}
|
|
57
|
+
}
|
|
58
|
+
}
|
|
59
|
+
// ── 交互模式 ──────────────────────────────────────────────
|
|
60
|
+
async function promptInput(question, defaultVal) {
|
|
61
|
+
const rl = createInterface({ input: process.stdin, output: process.stdout });
|
|
62
|
+
return new Promise((resolve) => {
|
|
63
|
+
rl.question(question, (answer) => {
|
|
64
|
+
rl.close();
|
|
65
|
+
resolve(answer.trim() || defaultVal);
|
|
66
|
+
});
|
|
67
|
+
});
|
|
68
|
+
}
|
|
69
|
+
async function runInteractiveMode() {
|
|
70
|
+
console.log('文章获取工具');
|
|
71
|
+
console.log('='.repeat(40));
|
|
72
|
+
const url = await promptInput('请输入文章 URL: ', '');
|
|
73
|
+
if (!url) {
|
|
74
|
+
console.error('URL 不能为空!');
|
|
75
|
+
process.exit(1);
|
|
76
|
+
}
|
|
77
|
+
const outputDir = await promptInput("输出目录 (回车使用默认 'articles'): ", 'articles');
|
|
78
|
+
console.log('\n输出格式选择:');
|
|
79
|
+
console.log('1. 仅 Markdown (默认)');
|
|
80
|
+
console.log('2. Markdown + HTML');
|
|
81
|
+
console.log('3. Markdown + JSON');
|
|
82
|
+
console.log('4. 全部格式 (HTML + JSON + Markdown)');
|
|
83
|
+
const choice = await promptInput('请选择 (1-4, 回车使用默认1): ', '1');
|
|
84
|
+
let formats = ['markdown'];
|
|
85
|
+
if (choice === '2')
|
|
86
|
+
formats = ['markdown', 'html'];
|
|
87
|
+
else if (choice === '3')
|
|
88
|
+
formats = ['markdown', 'json'];
|
|
89
|
+
else if (choice === '4')
|
|
90
|
+
formats = ['markdown', 'html', 'json'];
|
|
91
|
+
const maxArticles = 10;
|
|
92
|
+
// 执行
|
|
93
|
+
if (isWechatAlbumUrl(url)) {
|
|
94
|
+
console.log('\n检测到微信专辑 URL,将批量下载...');
|
|
95
|
+
const result = await fetchAlbum({
|
|
96
|
+
url,
|
|
97
|
+
outputDir,
|
|
98
|
+
format: formats,
|
|
99
|
+
maxArticles,
|
|
100
|
+
timeoutMs: 30000,
|
|
101
|
+
downloadImages: false,
|
|
102
|
+
});
|
|
103
|
+
if (result.ok) {
|
|
104
|
+
console.log(`\n✅ 批量下载完成! 成功: ${result.articles.length} 篇`);
|
|
105
|
+
}
|
|
106
|
+
else {
|
|
107
|
+
console.error(`\n❌ ${result.error.message}`);
|
|
108
|
+
process.exit(1);
|
|
109
|
+
}
|
|
110
|
+
}
|
|
111
|
+
else {
|
|
112
|
+
const result = await fetchArticle({ url, outputDir, format: formats, timeoutMs: 30000, downloadImages: false });
|
|
113
|
+
if (result.ok) {
|
|
114
|
+
if ('article' in result) {
|
|
115
|
+
console.log('\n✅ 成功获取文章!');
|
|
116
|
+
printArticleInfo(result.article);
|
|
117
|
+
}
|
|
118
|
+
else {
|
|
119
|
+
console.log('\n✅ 批量下载完成!');
|
|
120
|
+
console.log(` 成功: ${result.articles.length} 篇`);
|
|
121
|
+
if (result.failedCount)
|
|
122
|
+
console.log(` 失败: ${result.failedCount} 篇`);
|
|
123
|
+
}
|
|
124
|
+
if (result.artifacts.length > 0) {
|
|
125
|
+
console.log(' 产物:');
|
|
126
|
+
for (const a of result.artifacts) {
|
|
127
|
+
console.log(` - [${a.type}] ${a.path}`);
|
|
128
|
+
}
|
|
129
|
+
}
|
|
130
|
+
}
|
|
131
|
+
else {
|
|
132
|
+
console.error(`\n❌ 获取失败: ${result.error.message}`);
|
|
133
|
+
process.exit(1);
|
|
134
|
+
}
|
|
135
|
+
}
|
|
136
|
+
}
|
|
137
|
+
// ── album-only 处理 ──────────────────────────────────────
|
|
138
|
+
async function handleAlbumOnly(url, options) {
|
|
139
|
+
const maxArticles = parseInt(options.max, 10);
|
|
140
|
+
const timeoutMs = parseInt(options.timeout, 10);
|
|
141
|
+
console.log('仅获取专辑文章列表...');
|
|
142
|
+
// 使用 fetchAlbumList —— 仅 API 列表,不下载文章(与 Python 一致)
|
|
143
|
+
const result = await fetchAlbumList({ url, maxArticles, timeoutMs });
|
|
144
|
+
if (result.ok) {
|
|
145
|
+
console.log(`\n获取到 ${result.articles.length} 篇文章:`);
|
|
146
|
+
for (let i = 0; i < result.articles.length; i++) {
|
|
147
|
+
const a = result.articles[i];
|
|
148
|
+
console.log(`${String(i + 1).padStart(2)}. ${a.title}`);
|
|
149
|
+
console.log(` URL: ${a.url}`);
|
|
150
|
+
if (a.createTime)
|
|
151
|
+
console.log(` 时间: ${a.createTime}`);
|
|
152
|
+
if (i < result.articles.length - 1)
|
|
153
|
+
console.log();
|
|
154
|
+
}
|
|
155
|
+
// 保存文章列表到 JSON(包含完整元数据,与 Python 一致)
|
|
156
|
+
await ensureDir(options.output);
|
|
157
|
+
const listFile = join(options.output, `album_articles_${Math.floor(Date.now() / 1000)}.json`);
|
|
158
|
+
await writeJsonSafe(listFile, result.articles);
|
|
159
|
+
console.log(`\n文章列表已保存到: ${listFile}`);
|
|
160
|
+
}
|
|
161
|
+
else {
|
|
162
|
+
console.error(`\n❌ 获取失败: ${result.error.message}`);
|
|
163
|
+
process.exit(1);
|
|
164
|
+
}
|
|
165
|
+
}
|
|
166
|
+
// ── album 批量下载处理 ────────────────────────────────────
|
|
167
|
+
async function handleAlbumDownload(url, options, formats) {
|
|
168
|
+
const maxArticles = parseInt(options.max, 10);
|
|
169
|
+
const timeoutMs = parseInt(options.timeout, 10);
|
|
170
|
+
console.log('检测到微信专辑 URL,将批量下载...');
|
|
171
|
+
const result = await fetchAlbum({
|
|
172
|
+
url,
|
|
173
|
+
outputDir: options.output,
|
|
174
|
+
format: formats,
|
|
175
|
+
maxArticles,
|
|
176
|
+
timeoutMs,
|
|
177
|
+
downloadImages: options.downloadImages,
|
|
178
|
+
});
|
|
179
|
+
if (result.ok) {
|
|
180
|
+
const failed = result.failedCount ?? 0;
|
|
181
|
+
console.log(`\n✅ 批量下载完成!`);
|
|
182
|
+
console.log(` 成功: ${result.articles.length} 篇`);
|
|
183
|
+
if (failed > 0)
|
|
184
|
+
console.log(` 失败: ${failed} 篇`);
|
|
185
|
+
console.log(` 文章保存在: ${options.output}`);
|
|
186
|
+
if (result.summaryFile) {
|
|
187
|
+
console.log(` 汇总: ${result.summaryFile}`);
|
|
188
|
+
}
|
|
189
|
+
}
|
|
190
|
+
else {
|
|
191
|
+
console.error(`\n❌ 下载失败: ${result.error.message}`);
|
|
192
|
+
process.exit(1);
|
|
193
|
+
}
|
|
194
|
+
}
|
|
195
|
+
// ── CLI 程序 ──────────────────────────────────────────────
|
|
196
|
+
const program = new Command();
|
|
197
|
+
program
|
|
198
|
+
.name('wespy')
|
|
199
|
+
.description('文章抓取与 Markdown 转换工具')
|
|
200
|
+
.version(pkg.version);
|
|
201
|
+
// ── fetch-article 子命令 ──────────────────────────────────
|
|
202
|
+
addCommonOptions(program
|
|
203
|
+
.command('fetch-article')
|
|
204
|
+
.description('获取单篇文章')
|
|
205
|
+
.argument('<url>', '文章 URL')).action(async (url, options) => {
|
|
206
|
+
const formats = resolveFormats(options);
|
|
207
|
+
const timeoutMs = parseInt(options.timeout, 10);
|
|
208
|
+
if (options.verbose) {
|
|
209
|
+
printVerbose(options, url);
|
|
210
|
+
}
|
|
211
|
+
const result = await fetchArticle({
|
|
212
|
+
url,
|
|
213
|
+
outputDir: options.output,
|
|
214
|
+
format: formats,
|
|
215
|
+
timeoutMs,
|
|
216
|
+
downloadImages: options.downloadImages,
|
|
217
|
+
});
|
|
218
|
+
if (result.ok) {
|
|
219
|
+
if ('article' in result) {
|
|
220
|
+
console.log('\n✅ 成功获取文章!');
|
|
221
|
+
printArticleInfo(result.article);
|
|
222
|
+
}
|
|
223
|
+
else {
|
|
224
|
+
console.log('\n✅ 批量下载完成!');
|
|
225
|
+
console.log(` 成功: ${result.articles.length} 篇`);
|
|
226
|
+
if (result.failedCount)
|
|
227
|
+
console.log(` 失败: ${result.failedCount} 篇`);
|
|
228
|
+
}
|
|
229
|
+
if (result.artifacts.length > 0) {
|
|
230
|
+
console.log(' 产物:');
|
|
231
|
+
for (const a of result.artifacts) {
|
|
232
|
+
console.log(` - [${a.type}] ${a.path}`);
|
|
233
|
+
}
|
|
234
|
+
}
|
|
235
|
+
}
|
|
236
|
+
else {
|
|
237
|
+
console.error(`\n❌ 获取失败: ${result.error.message}`);
|
|
238
|
+
console.error(` 错误码: ${result.error.code}`);
|
|
239
|
+
process.exit(1);
|
|
240
|
+
}
|
|
241
|
+
});
|
|
242
|
+
// ── fetch-album 子命令 ────────────────────────────────────
|
|
243
|
+
addCommonOptions(program
|
|
244
|
+
.command('fetch-album')
|
|
245
|
+
.description('批量获取微信专辑文章')
|
|
246
|
+
.argument('<url>', '专辑 URL')
|
|
247
|
+
.option('--max <n>', '最大下载文章数', '10')
|
|
248
|
+
.option('--album-only', '仅获取文章列表,不下载内容')).action(async (url, options) => {
|
|
249
|
+
if (options.verbose) {
|
|
250
|
+
printVerbose(options, url, {
|
|
251
|
+
最大文章数量: parseInt(options.max, 10),
|
|
252
|
+
仅获取列表: options.albumOnly ?? false,
|
|
253
|
+
});
|
|
254
|
+
}
|
|
255
|
+
if (options.albumOnly) {
|
|
256
|
+
await handleAlbumOnly(url, options);
|
|
257
|
+
return;
|
|
258
|
+
}
|
|
259
|
+
const formats = resolveFormats(options);
|
|
260
|
+
await handleAlbumDownload(url, options, formats);
|
|
261
|
+
});
|
|
262
|
+
// ── 默认命令(智能判断 URL 类型)──────────────────────────
|
|
263
|
+
addCommonOptions(program
|
|
264
|
+
.argument('[url]', '文章或专辑 URL')
|
|
265
|
+
.option('--max <n>', '专辑最大下载文章数', '10')
|
|
266
|
+
.option('--album-only', '仅获取文章列表,不下载内容')).action(async (url, options) => {
|
|
267
|
+
// 无 URL → 交互模式
|
|
268
|
+
if (!url) {
|
|
269
|
+
await runInteractiveMode();
|
|
270
|
+
return;
|
|
271
|
+
}
|
|
272
|
+
const formats = resolveFormats(options);
|
|
273
|
+
const timeoutMs = parseInt(options.timeout, 10);
|
|
274
|
+
const maxArticles = parseInt(options.max, 10);
|
|
275
|
+
if (options.verbose) {
|
|
276
|
+
printVerbose(options, url, { 最大文章数量: maxArticles, 仅获取列表: options.albumOnly ?? false });
|
|
277
|
+
}
|
|
278
|
+
if (isWechatAlbumUrl(url)) {
|
|
279
|
+
if (options.albumOnly) {
|
|
280
|
+
await handleAlbumOnly(url, options);
|
|
281
|
+
return;
|
|
282
|
+
}
|
|
283
|
+
await handleAlbumDownload(url, options, formats);
|
|
284
|
+
}
|
|
285
|
+
else {
|
|
286
|
+
const result = await fetchArticle({
|
|
287
|
+
url,
|
|
288
|
+
outputDir: options.output,
|
|
289
|
+
format: formats,
|
|
290
|
+
timeoutMs,
|
|
291
|
+
downloadImages: options.downloadImages,
|
|
292
|
+
});
|
|
293
|
+
if (result.ok) {
|
|
294
|
+
if ('article' in result) {
|
|
295
|
+
console.log('\n✅ 成功获取文章!');
|
|
296
|
+
printArticleInfo(result.article);
|
|
297
|
+
}
|
|
298
|
+
else {
|
|
299
|
+
console.log('\n✅ 批量下载完成!');
|
|
300
|
+
console.log(` 成功: ${result.articles.length} 篇`);
|
|
301
|
+
if (result.failedCount)
|
|
302
|
+
console.log(` 失败: ${result.failedCount} 篇`);
|
|
303
|
+
}
|
|
304
|
+
}
|
|
305
|
+
else {
|
|
306
|
+
console.error(`\n❌ ${result.error.message}`);
|
|
307
|
+
process.exit(1);
|
|
308
|
+
}
|
|
309
|
+
}
|
|
310
|
+
});
|
|
311
|
+
program.parse();
|
|
312
|
+
//# sourceMappingURL=main.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"main.js","sourceRoot":"","sources":["../../src/cli/main.ts"],"names":[],"mappings":";AACA;;;GAGG;AAEH,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAA;AAC/C,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAA;AACtC,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AACzC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAA;AACxC,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAA;AACnC,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAA;AACtD,OAAO,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAA;AAClD,OAAO,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAA;AAC3D,OAAO,EAAE,gBAAgB,EAAE,MAAM,0BAA0B,CAAA;AAE3D,OAAO,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAA;AAEzD,6BAA6B;AAC7B,MAAM,SAAS,GAAG,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAA;AACzD,MAAM,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,SAAS,EAAE,oBAAoB,CAAC,EAAE,OAAO,CAAC,CAAwB,CAAA;AAgB3G,SAAS,gBAAgB,CAAC,GAAY;IACpC,OAAO,GAAG;SACP,MAAM,CAAC,oBAAoB,EAAE,MAAM,EAAE,UAAU,CAAC;SAChD,MAAM,CAAC,QAAQ,EAAE,cAAc,CAAC;SAChC,MAAM,CAAC,QAAQ,EAAE,cAAc,CAAC;SAChC,MAAM,CAAC,OAAO,EAAE,iCAAiC,CAAC;SAClD,MAAM,CAAC,mBAAmB,EAAE,qBAAqB,CAAC;SAClD,MAAM,CAAC,gBAAgB,EAAE,aAAa,EAAE,OAAO,CAAC;SAChD,MAAM,CAAC,eAAe,EAAE,QAAQ,CAAC,CAAA;AACtC,CAAC;AAED,qBAAqB;AACrB,SAAS,gBAAgB,CAAC,OAAiE;IACzF,OAAO,CAAC,GAAG,CAAC,UAAU,OAAO,CAAC,KAAK,EAAE,CAAC,CAAA;IACtC,IAAI,OAAO,CAAC,MAAM;QAAE,OAAO,CAAC,GAAG,CAAC,UAAU,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IAC3D,IAAI,OAAO,CAAC,WAAW;QAAE,OAAO,CAAC,GAAG,CAAC,UAAU,OAAO,CAAC,WAAW,EAAE,CAAC,CAAA;AACvE,CAAC;AAED,SAAS,cAAc,CAAC,OAAmB;IACzC,IAAI,OAAO,CAAC,GAAG;QAAE,OAAO,CAAC,UAAU,EAAE,MAAM,EAAE,MAAM,CAAC,CAAA;IACpD,MAAM,OAAO,GAAmB,CAAC,UAAU,CAAC,CAAA;IAC5C,IAAI,OAAO,CAAC,IAAI;QAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;IACtC,IAAI,OAAO,CAAC,IAAI;QAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;IACtC,OAAO,OAAO,CAAA;AAChB,CAAC;AAED,SAAS,YAAY,CAAC,OAAmB,EAAE,GAAW,EAAE,KAA+B;IACrF,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAA;IAChC,OAAO,CAAC,GAAG,CAAC,UAAU,GAAG,EAAE,CAAC,CAAA;IAC5B,OAAO,CAAC,GAAG,CAAC,WAAW,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IACxC,OAAO,CAAC,GAAG,CAAC,gBAAgB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC,UAAU,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC,cAAc,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAC,EAAE,CAAC,CAAA;IACnI,IAAI,KAAK,EAAE,CAAC;QACV,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;YAC3C,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,EAAE,CAAC,CAAA;QAC7B,CAAC;IACH,CAAC;AACH,CAAC;AAED,yDAAyD;AAEzD,KAAK,UAAU,WAAW,CAAC,QAAgB,EAAE,UAAkB;IAC7D,MAAM,EAAE,GAAG,eAAe,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IAC5E,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE;QAC7B,EAAE,CAAC,QAAQ,CAAC,QAAQ,EAAE,CAAC,MAAM,EAAE,EAAE;YAC/B,EAAE,CAAC,KAAK,EAAE,CAAA;YACV,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,IAAI,UAAU,CAAC,CAAA;QACtC,CAAC,CAAC,CAAA;IACJ,CAAC,CAAC,CAAA;AACJ,CAAC;AAED,KAAK,UAAU,kBAAkB;IAC/B,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;IACrB,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAA;IAE3B,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,aAAa,EAAE,EAAE,CAAC,CAAA;IAChD,IAAI,CAAC,GAAG,EAAE,CAAC;QACT,OAAO,CAAC,KAAK,CAAC,WAAW,CAAC,CAAA;QAC1B,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;IAED,MAAM,SAAS,GAAG,MAAM,WAAW,CAAC,4BAA4B,EAAE,UAAU,CAAC,CAAA;IAE7E,OAAO,CAAC,GAAG,CAAC,WAAW,CAAC,CAAA;IACxB,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAA;IACjC,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAA;IACjC,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAA;IACjC,OAAO,CAAC,GAAG,CAAC,kCAAkC,CAAC,CAAA;IAE/C,MAAM,MAAM,GAAG,MAAM,WAAW,CAAC,sBAAsB,EAAE,GAAG,CAAC,CAAA;IAE7D,IAAI,OAAO,GAAmB,CAAC,UAAU,CAAC,CAAA;IAC1C,IAAI,MAAM,KAAK,GAAG;QAAE,OAAO,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;SAC7C,IAAI,MAAM,KAAK,GAAG;QAAE,OAAO,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAA;SAClD,IAAI,MAAM,KAAK,GAAG;QAAE,OAAO,GAAG,CAAC,UAAU,EAAE,MAAM,EAAE,MAAM,CAAC,CAAA;IAE/D,MAAM,WAAW,GAAG,EAAE,CAAA;IAEtB,KAAK;IACL,IAAI,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,OAAO,CAAC,GAAG,CAAC,wBAAwB,CAAC,CAAA;QACrC,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;YAC9B,GAAG;YACH,SAAS;YACT,MAAM,EAAE,OAAO;YACf,WAAW;YACX,SAAS,EAAE,KAAK;YAChB,cAAc,EAAE,KAAK;SACtB,CAAC,CAAA;QACF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;YACd,OAAO,CAAC,GAAG,CAAC,mBAAmB,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;QAC5D,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,KAAK,CAAC,OAAO,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;YAC5C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACjB,CAAC;IACH,CAAC;SAAM,CAAC;QACN,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC,EAAE,GAAG,EAAE,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,SAAS,EAAE,KAAK,EAAE,cAAc,EAAE,KAAK,EAAE,CAAC,CAAA;QAC/G,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;YACd,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;gBACxB,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAClC,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;gBACjD,IAAI,MAAM,CAAC,WAAW;oBAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;YACvE,CAAC;YACD,IAAI,MAAM,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAChC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;gBACrB,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;oBACjC,OAAO,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;gBAC7C,CAAC;YACH,CAAC;QACH,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;YAClD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACjB,CAAC;IACH,CAAC;AACH,CAAC;AAED,0DAA0D;AAE1D,KAAK,UAAU,eAAe,CAAC,GAAW,EAAE,OAAmB;IAC7D,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC,CAAA;IAC7C,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAE/C,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,CAAA;IAE3B,mDAAmD;IACnD,MAAM,MAAM,GAAG,MAAM,cAAc,CAAC,EAAE,GAAG,EAAE,WAAW,EAAE,SAAS,EAAE,CAAC,CAAA;IAEpE,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,SAAS,MAAM,CAAC,QAAQ,CAAC,MAAM,OAAO,CAAC,CAAA;QACnD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAChD,MAAM,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAA;YAC7B,OAAO,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,CAAC,CAAA;YACvD,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC,GAAG,EAAE,CAAC,CAAA;YACjC,IAAI,CAAC,CAAC,UAAU;gBAAE,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,UAAU,EAAE,CAAC,CAAA;YACzD,IAAI,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC;gBAAE,OAAO,CAAC,GAAG,EAAE,CAAA;QACnD,CAAC;QAED,oCAAoC;QACpC,MAAM,SAAS,CAAC,OAAO,CAAC,MAAM,CAAC,CAAA;QAC/B,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,kBAAkB,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,OAAO,CAAC,CAAA;QAC7F,MAAM,aAAa,CAAC,QAAQ,EAAE,MAAM,CAAC,QAAQ,CAAC,CAAA;QAC9C,OAAO,CAAC,GAAG,CAAC,eAAe,QAAQ,EAAE,CAAC,CAAA;IACxC,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;QAClD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;AACH,CAAC;AAED,uDAAuD;AAEvD,KAAK,UAAU,mBAAmB,CAAC,GAAW,EAAE,OAAmB,EAAE,OAAuB;IAC1F,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC,CAAA;IAC7C,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAE/C,OAAO,CAAC,GAAG,CAAC,sBAAsB,CAAC,CAAA;IACnC,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC;QAC9B,GAAG;QACH,SAAS,EAAE,OAAO,CAAC,MAAM;QACzB,MAAM,EAAE,OAAO;QACf,WAAW;QACX,SAAS;QACT,cAAc,EAAE,OAAO,CAAC,cAAc;KACvC,CAAC,CAAA;IAEF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;QACd,MAAM,MAAM,GAAG,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;QACtC,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;QAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;QACjD,IAAI,MAAM,GAAG,CAAC;YAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,IAAI,CAAC,CAAA;QACjD,OAAO,CAAC,GAAG,CAAC,aAAa,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;QAC1C,IAAI,MAAM,CAAC,WAAW,EAAE,CAAC;YACvB,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,EAAE,CAAC,CAAA;QAC7C,CAAC;IACH,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;QAClD,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;AACH,CAAC;AAED,2DAA2D;AAE3D,MAAM,OAAO,GAAG,IAAI,OAAO,EAAE,CAAA;AAE7B,OAAO;KACJ,IAAI,CAAC,OAAO,CAAC;KACb,WAAW,CAAC,qBAAqB,CAAC;KAClC,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;AAEvB,0DAA0D;AAE1D,gBAAgB,CACd,OAAO;KACJ,OAAO,CAAC,eAAe,CAAC;KACxB,WAAW,CAAC,QAAQ,CAAC;KACrB,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC,CAC/B,CAAC,MAAM,CAAC,KAAK,EAAE,GAAW,EAAE,OAAmB,EAAE,EAAE;IAClD,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAE/C,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;QACpB,YAAY,CAAC,OAAO,EAAE,GAAG,CAAC,CAAA;IAC5B,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC;QAChC,GAAG;QACH,SAAS,EAAE,OAAO,CAAC,MAAM;QACzB,MAAM,EAAE,OAAO;QACf,SAAS;QACT,cAAc,EAAE,OAAO,CAAC,cAAc;KACvC,CAAC,CAAA;IAEF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;QACd,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;YACxB,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;YAC1B,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;QAClC,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;YAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;YACjD,IAAI,MAAM,CAAC,WAAW;gBAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;QACvE,CAAC;QACD,IAAI,MAAM,CAAC,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAChC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAA;YACrB,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;gBACjC,OAAO,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;YAC7C,CAAC;QACH,CAAC;IACH,CAAC;SAAM,CAAC;QACN,OAAO,CAAC,KAAK,CAAC,aAAa,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;QAClD,OAAO,CAAC,KAAK,CAAC,WAAW,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAA;QAC7C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACjB,CAAC;AACH,CAAC,CAAC,CAAA;AAEF,0DAA0D;AAE1D,gBAAgB,CACd,OAAO;KACJ,OAAO,CAAC,aAAa,CAAC;KACtB,WAAW,CAAC,YAAY,CAAC;KACzB,QAAQ,CAAC,OAAO,EAAE,QAAQ,CAAC;KAC3B,MAAM,CAAC,WAAW,EAAE,SAAS,EAAE,IAAI,CAAC;KACpC,MAAM,CAAC,cAAc,EAAE,eAAe,CAAC,CAC3C,CAAC,MAAM,CAAC,KAAK,EAAE,GAAW,EAAE,OAAmB,EAAE,EAAE;IAClD,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;QACpB,YAAY,CAAC,OAAO,EAAE,GAAG,EAAE;YACzB,MAAM,EAAE,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC;YACjC,KAAK,EAAE,OAAO,CAAC,SAAS,IAAI,KAAK;SAClC,CAAC,CAAA;IACJ,CAAC;IAED,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;QACtB,MAAM,eAAe,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;QACnC,OAAM;IACR,CAAC;IAED,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,MAAM,mBAAmB,CAAC,GAAG,EAAE,OAAO,EAAE,OAAO,CAAC,CAAA;AAClD,CAAC,CAAC,CAAA;AAEF,iDAAiD;AAEjD,gBAAgB,CACd,OAAO;KACJ,QAAQ,CAAC,OAAO,EAAE,WAAW,CAAC;KAC9B,MAAM,CAAC,WAAW,EAAE,WAAW,EAAE,IAAI,CAAC;KACtC,MAAM,CAAC,cAAc,EAAE,eAAe,CAAC,CAC3C,CAAC,MAAM,CAAC,KAAK,EAAE,GAAuB,EAAE,OAAmB,EAAE,EAAE;IAC9D,eAAe;IACf,IAAI,CAAC,GAAG,EAAE,CAAC;QACT,MAAM,kBAAkB,EAAE,CAAA;QAC1B,OAAM;IACR,CAAC;IAED,MAAM,OAAO,GAAG,cAAc,CAAC,OAAO,CAAC,CAAA;IACvC,MAAM,SAAS,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAA;IAC/C,MAAM,WAAW,GAAG,QAAQ,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,CAAC,CAAA;IAE7C,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;QACpB,YAAY,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE,MAAM,EAAE,WAAW,EAAE,KAAK,EAAE,OAAO,CAAC,SAAS,IAAI,KAAK,EAAE,CAAC,CAAA;IACxF,CAAC;IAED,IAAI,gBAAgB,CAAC,GAAG,CAAC,EAAE,CAAC;QAC1B,IAAI,OAAO,CAAC,SAAS,EAAE,CAAC;YACtB,MAAM,eAAe,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;YACnC,OAAM;QACR,CAAC;QACD,MAAM,mBAAmB,CAAC,GAAG,EAAE,OAAO,EAAE,OAAO,CAAC,CAAA;IAClD,CAAC;SAAM,CAAC;QACN,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC;YAChC,GAAG;YACH,SAAS,EAAE,OAAO,CAAC,MAAM;YACzB,MAAM,EAAE,OAAO;YACf,SAAS;YACT,cAAc,EAAE,OAAO,CAAC,cAAc;SACvC,CAAC,CAAA;QAEF,IAAI,MAAM,CAAC,EAAE,EAAE,CAAC;YACd,IAAI,SAAS,IAAI,MAAM,EAAE,CAAC;gBACxB,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,gBAAgB,CAAC,MAAM,CAAC,OAAO,CAAC,CAAA;YAClC,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;gBAC1B,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAA;gBACjD,IAAI,MAAM,CAAC,WAAW;oBAAE,OAAO,CAAC,GAAG,CAAC,UAAU,MAAM,CAAC,WAAW,IAAI,CAAC,CAAA;YACvE,CAAC;QACH,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,KAAK,CAAC,OAAO,MAAM,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAA;YAC5C,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACjB,CAAC;IACH,CAAC;AACH,CAAC,CAAC,CAAA;AAEF,OAAO,CAAC,KAAK,EAAE,CAAA"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"html-to-markdown.d.ts","sourceRoot":"","sources":["../../src/converter/html-to-markdown.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAOH;;GAEG;AACH,wBAAgB,cAAc,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAInD"}
|
|
@@ -0,0 +1,171 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* HTML → Markdown 转换器
|
|
3
|
+
* 基于 cheerio 解析 + 递归遍历,忠实还原 Python 版行为
|
|
4
|
+
*/
|
|
5
|
+
import * as cheerio from 'cheerio';
|
|
6
|
+
import { buildProxyImageUrl } from '../utils/url.js';
|
|
7
|
+
import { detectCodeLanguage, cleanCodeContent } from '../utils/text.js';
|
|
8
|
+
/**
|
|
9
|
+
* 将 HTML 字符串转换为 Markdown
|
|
10
|
+
*/
|
|
11
|
+
export function htmlToMarkdown(html) {
|
|
12
|
+
if (!html)
|
|
13
|
+
return '';
|
|
14
|
+
const $ = cheerio.load(html, { xmlMode: false });
|
|
15
|
+
return convertNode($, $.root());
|
|
16
|
+
}
|
|
17
|
+
function convertNode($, node) {
|
|
18
|
+
let md = '';
|
|
19
|
+
node.contents().each((_, child) => {
|
|
20
|
+
if (child.type === 'text') {
|
|
21
|
+
const text = $(child).text().trim();
|
|
22
|
+
if (text)
|
|
23
|
+
md += text;
|
|
24
|
+
return;
|
|
25
|
+
}
|
|
26
|
+
if (child.type !== 'tag')
|
|
27
|
+
return;
|
|
28
|
+
const el = $(child);
|
|
29
|
+
const tag = child.tagName?.toLowerCase();
|
|
30
|
+
switch (tag) {
|
|
31
|
+
case 'br':
|
|
32
|
+
md += '\n';
|
|
33
|
+
break;
|
|
34
|
+
case 'p':
|
|
35
|
+
case 'div':
|
|
36
|
+
case 'section': {
|
|
37
|
+
const content = convertNode($, el).trim();
|
|
38
|
+
if (content)
|
|
39
|
+
md += '\n\n' + content + '\n';
|
|
40
|
+
break;
|
|
41
|
+
}
|
|
42
|
+
case 'h1':
|
|
43
|
+
case 'h2':
|
|
44
|
+
case 'h3':
|
|
45
|
+
case 'h4':
|
|
46
|
+
case 'h5':
|
|
47
|
+
case 'h6': {
|
|
48
|
+
const level = parseInt(tag[1], 10);
|
|
49
|
+
const content = convertNode($, el).trim();
|
|
50
|
+
if (content)
|
|
51
|
+
md += '\n' + '#'.repeat(level) + ' ' + content + '\n';
|
|
52
|
+
break;
|
|
53
|
+
}
|
|
54
|
+
case 'strong':
|
|
55
|
+
case 'b': {
|
|
56
|
+
const content = convertNode($, el).trim();
|
|
57
|
+
if (content)
|
|
58
|
+
md += '**' + content + '**';
|
|
59
|
+
break;
|
|
60
|
+
}
|
|
61
|
+
case 'em':
|
|
62
|
+
case 'i': {
|
|
63
|
+
const content = convertNode($, el).trim();
|
|
64
|
+
if (content)
|
|
65
|
+
md += '*' + content + '*';
|
|
66
|
+
break;
|
|
67
|
+
}
|
|
68
|
+
case 'img': {
|
|
69
|
+
const src = el.attr('data-src') ?? el.attr('src') ?? '';
|
|
70
|
+
const alt = el.attr('alt') ?? '';
|
|
71
|
+
if (src) {
|
|
72
|
+
const proxySrc = buildProxyImageUrl(src);
|
|
73
|
+
md += `\n\n`;
|
|
74
|
+
}
|
|
75
|
+
break;
|
|
76
|
+
}
|
|
77
|
+
case 'a': {
|
|
78
|
+
const href = el.attr('href') ?? '';
|
|
79
|
+
const text = convertNode($, el).trim();
|
|
80
|
+
if (href && text) {
|
|
81
|
+
md += `[${text}](${href})`;
|
|
82
|
+
}
|
|
83
|
+
else if (text) {
|
|
84
|
+
md += text;
|
|
85
|
+
}
|
|
86
|
+
break;
|
|
87
|
+
}
|
|
88
|
+
case 'ul':
|
|
89
|
+
case 'ol': {
|
|
90
|
+
const listContent = convertList($, el, tag === 'ol');
|
|
91
|
+
if (listContent)
|
|
92
|
+
md += '\n' + listContent + '\n';
|
|
93
|
+
break;
|
|
94
|
+
}
|
|
95
|
+
case 'code': {
|
|
96
|
+
// 行内 code(如果父元素是 pre,由 pre 处理)
|
|
97
|
+
const parent = child.parent;
|
|
98
|
+
if (parent && parent.type === 'tag' && parent.tagName?.toLowerCase() === 'pre')
|
|
99
|
+
break;
|
|
100
|
+
const codeContent = el.text().trim();
|
|
101
|
+
if (codeContent)
|
|
102
|
+
md += '`' + codeContent + '`';
|
|
103
|
+
break;
|
|
104
|
+
}
|
|
105
|
+
case 'pre': {
|
|
106
|
+
const codeContent = extractCodeFromPre(el);
|
|
107
|
+
const language = detectPreLanguage($, el);
|
|
108
|
+
if (codeContent) {
|
|
109
|
+
if (language) {
|
|
110
|
+
md += `\n\`\`\`${language}\n${codeContent}\n\`\`\`\n`;
|
|
111
|
+
}
|
|
112
|
+
else {
|
|
113
|
+
md += `\n\`\`\`\n${codeContent}\n\`\`\`\n`;
|
|
114
|
+
}
|
|
115
|
+
}
|
|
116
|
+
break;
|
|
117
|
+
}
|
|
118
|
+
default: {
|
|
119
|
+
md += convertNode($, el);
|
|
120
|
+
}
|
|
121
|
+
}
|
|
122
|
+
});
|
|
123
|
+
return md;
|
|
124
|
+
}
|
|
125
|
+
function convertList($, listEl, ordered) {
|
|
126
|
+
let md = '';
|
|
127
|
+
let index = 1;
|
|
128
|
+
listEl.children('li').each((_, item) => {
|
|
129
|
+
const content = convertNode($, $(item)).trim();
|
|
130
|
+
if (content) {
|
|
131
|
+
if (ordered) {
|
|
132
|
+
md += `${index}. ${content}\n`;
|
|
133
|
+
index++;
|
|
134
|
+
}
|
|
135
|
+
else {
|
|
136
|
+
md += `- ${content}\n`;
|
|
137
|
+
}
|
|
138
|
+
}
|
|
139
|
+
});
|
|
140
|
+
return md;
|
|
141
|
+
}
|
|
142
|
+
function extractCodeFromPre(preEl) {
|
|
143
|
+
const codeEl = preEl.find('code');
|
|
144
|
+
const raw = codeEl.length > 0 ? codeEl.text() : preEl.text();
|
|
145
|
+
return cleanCodeContent(raw);
|
|
146
|
+
}
|
|
147
|
+
function detectPreLanguage(_$, preEl) {
|
|
148
|
+
// 从 pre 的 class 检测
|
|
149
|
+
const preClasses = (preEl.attr('class') ?? '').split(/\s+/).filter(Boolean);
|
|
150
|
+
let lang = detectCodeLanguage(preClasses);
|
|
151
|
+
if (lang)
|
|
152
|
+
return lang;
|
|
153
|
+
// 从内部 code 的 class 检测
|
|
154
|
+
const codeEl = preEl.find('code');
|
|
155
|
+
if (codeEl.length > 0) {
|
|
156
|
+
const codeClasses = (codeEl.attr('class') ?? '').split(/\s+/).filter(Boolean);
|
|
157
|
+
lang = detectCodeLanguage(codeClasses);
|
|
158
|
+
if (lang)
|
|
159
|
+
return lang;
|
|
160
|
+
// data-language 属性
|
|
161
|
+
const dataLang = codeEl.attr('data-language') ?? codeEl.attr('lang');
|
|
162
|
+
if (dataLang)
|
|
163
|
+
return dataLang.toLowerCase();
|
|
164
|
+
}
|
|
165
|
+
// pre 的 data-language 属性
|
|
166
|
+
const preDataLang = preEl.attr('data-language') ?? preEl.attr('lang');
|
|
167
|
+
if (preDataLang)
|
|
168
|
+
return preDataLang.toLowerCase();
|
|
169
|
+
return null;
|
|
170
|
+
}
|
|
171
|
+
//# sourceMappingURL=html-to-markdown.js.map
|