@aiello/wechat-to-markdown 1.2.3

package/README.md

@@ -0,0 +1,105 @@
+# wechat-to-markdown
+
+Fork from https://github.com/LuckyRyan-web/wechat-to-markdown
+
+add some bug fix
+
+## description
+
+Enter the WeChat public address to convert it to markdown format
+
+English | [简体中文](README.zh-CN.md)
+
+## return options
+
+```ts
+interface TurnDownResult {
+    success: boolean
+    code: number
+    data?: {
+        title?: string
+        author?: string
+        content?: string
+    }
+    msg?: string
+}
+
+```
+
+## Basic Usage
+
+## CommonJs
+
+```javascript
+const transformHtml2Markdown = require('@ryan-liu/wechat-to-markdown').default
+
+setTimeout(async () => {
+    const articleData = await transformHtml2Markdown('https://mp.weixin.qq.com/s/9d5DWg7YdMHPvVl-2KLH2w')
+
+    const { title, author, content } = articleData.data
+
+    console.log('title', title)
+    console.log('author', author)
+    console.log('content', content)
+}, 0)
+```
+
+## vue3
+**index.ts**
+```javascript
+import transformHtml2Markdown from '@ryan-liu/wechat-to-markdown'
+
+setup() {
+    const getData = async () => {
+        const articleData = await transformHtml2Markdown(
+            '/api/s/9d5DWg7YdMHPvVl-2KLH2w'
+        )
+
+        const { title, author, content } = articleData.data
+
+        console.log('title', title)
+        console.log('author', author)
+        console.log('content', content)
+    }
+
+    getData()
+
+    return {}
+},
+```
+
+**vite.config.ts**
+
+```js
+...
+server: {
+    proxy: {
+        '/api': {
+            target: 'https://mp.weixin.qq.com',
+            changeOrigin: true,
+            rewrite: (path) => path.replace(/^\/api/, ''),
+        },
+    },
+},
+...
+```
+
+
+## Packages
+
+[axios](http://www.axios-js.com/)
+
+[cheerio](https://github.com/cheeriojs/cheerio)
+
+[turndown (html to markdown)](https://github.com/mixmark-io/turndown)
+
+[turndown-plugin-gfm (turndown parsing partial html element plugin)](https://github.com/mixmark-io/turndown-plugin-gfm)
+
+## Reference
+
+[html2md (A very well developed url to markdown project)](https://github.com/helloworld-Co/html2md)
+
+
+## TODO
+
+1. The markdown code that supports conversion can replace the image and upload it to the cdn

package/README.zh-CN.md

@@ -0,0 +1,98 @@
+## 说明
+
+输入微信公众号地址，将其转换为 markdown 格式
+
+[English](README.md) | 简体中文
+
+## 返回属性
+
+```ts
+interface TurnDownResult {
+    success: boolean
+    code: number
+    data?: {
+        title?: string
+        author?: string
+        content?: string
+    }
+    msg?: string
+}
+
+```
+## 基本用法
+
+## CommonJs
+
+```javascript
+const transformHtml2Markdown = require('@ryan-liu/wechat-to-markdown').default
+
+setTimeout(async () => {
+    const articleData = await transformHtml2Markdown('https://mp.weixin.qq.com/s/9d5DWg7YdMHPvVl-2KLH2w')
+
+    const { title, author, content } = articleData.data
+
+    console.log('title', title)
+    console.log('author', author)
+    console.log('content', content)
+}, 0)
+```
+
+## vue3
+**index.ts**
+
+```javascript
+import transformHtml2Markdown from '@ryan-liu/wechat-to-markdown'
+
+setup() {
+    const getData = async () => {
+        const articleData = await transformHtml2Markdown(
+            '/api/s/9d5DWg7YdMHPvVl-2KLH2w'
+        )
+
+        const { title, author, content } = articleData.data
+
+        console.log('title', title)
+        console.log('author', author)
+        console.log('content', content)
+    }
+
+    getData()
+
+    return {}
+},
+```
+
+**vite.config.ts**
+
+```js
+...
+server: {
+    proxy: {
+        '/api': {
+            target: 'https://mp.weixin.qq.com',
+            changeOrigin: true,
+            rewrite: (path) => path.replace(/^\/api/, ''),
+        },
+    },
+},
+...
+```
+
+
+## 使用到的包
+
+[axios (著名的请求库)](http://www.axios-js.com/)
+
+[cheerio (可以用 JQuery 方式操作 html 源码)](https://github.com/cheeriojs/cheerio)
+
+[turndown (html to markdown)](https://github.com/mixmark-io/turndown)
+
+[turndown-plugin-gfm (turndown 解析部分 html 元素插件)](https://github.com/mixmark-io/turndown-plugin-gfm)
+
+## 参考
+
+[html2md (一个非常完善的 url to markdown 项目)](https://github.com/helloworld-Co/html2md)
+
+## 待做事项
+
+1. 需要支持转化后的 markdown 代码可以替换图片地址到 cdn(微信的图片会有域名限制，非 weixin 域名不可以访问图片)

package/dist/index.cjs

@@ -0,0 +1,182 @@
+var __create = Object.create;
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __getProtoOf = Object.getPrototypeOf;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __markAsModule = (target) => __defProp(target, "__esModule", { value: true });
+var __export = (target, all) => {
+  __markAsModule(target);
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __reExport = (target, module2, desc) => {
+  if (module2 && typeof module2 === "object" || typeof module2 === "function") {
+    for (let key of __getOwnPropNames(module2))
+      if (!__hasOwnProp.call(target, key) && key !== "default")
+        __defProp(target, key, { get: () => module2[key], enumerable: !(desc = __getOwnPropDesc(module2, key)) || desc.enumerable });
+  }
+  return target;
+};
+var __toModule = (module2) => {
+  return __reExport(__markAsModule(__defProp(module2 != null ? __create(__getProtoOf(module2)) : {}, "default", module2 && module2.__esModule && "default" in module2 ? { get: () => module2.default, enumerable: true } : { value: module2, enumerable: true })), module2);
+};
+
+// src/index.ts
+__export(exports, {
+  Status: () => Status,
+  default: () => transformHtml2Markdown
+});
+
+// node_modules/tsup/assets/cjs_shims.js
+var importMetaUrlShim = typeof document === "undefined" ? new (require("url")).URL("file:" + __filename).href : document.currentScript && document.currentScript.src || new URL("main.js", document.baseURI).href;
+
+// src/index.ts
+var import_axios = __toModule(require("axios"));
+var import_cheerio2 = __toModule(require("cheerio"));
+
+// src/error.ts
+var errObj = {
+  "400": "\u5185\u5BB9\u89E3\u6790\u5931\u8D25"
+};
+
+// src/type.ts
+var Status;
+(function(Status2) {
+  Status2[Status2["Success"] = 200] = "Success";
+  Status2[Status2["Fail"] = 400] = "Fail";
+})(Status || (Status = {}));
+
+// src/turndownCode.ts
+var import_turndown = __toModule(require("turndown"));
+var import_turndown_plugin_gfm = __toModule(require("turndown-plugin-gfm"));
+
+// src/formatHtml.ts
+var import_cheerio = __toModule(require("cheerio"));
+function formatCode(htmlStr) {
+  let code = htmlStr;
+  code = code.replace(/<br>/gi, "\n");
+  code = code.replace(/&nbsp;/gi, " ");
+  code = code.replace(/&lt;/gi, "<");
+  code = code.replace(/&gt;/gi, ">");
+  code = code.replace(/&amp;/gi, "&");
+  code = code.replace(/&quot;/gi, '"');
+  code = code.replace(/&apos;/gi, "\u2018");
+  code = code.replace(/&times;/gi, "*");
+  code = code.replace(/&divide;/gi, "%");
+  const $ = import_cheerio.default.load(code);
+  return $.text();
+}
+function figure2markdown(figureHTML) {
+  const imgRegex = /<img.*?data-src=['"](.*?)['"]/;
+  const descRegex = /\<figcaption .*?>(.+)<\/figcaption>/;
+  const imgArr = figureHTML.match(imgRegex);
+  const descArr = figureHTML.match(descRegex);
+  let imgUrl = "";
+  let desc = "";
+  if (Array.isArray(imgArr)) {
+    imgUrl = imgArr[1];
+  }
+  if (Array.isArray(descArr)) {
+    desc = descArr[1];
+  }
+  if (imgUrl) {
+    return `
+
+ ![${desc}](${imgUrl}) 
+
+`;
+  }
+  return;
+}
+
+// src/turndownCode.ts
+var turndownService = new import_turndown.default({
+  codeBlockStyle: "fenced",
+  hr: ""
+});
+turndownService.use(import_turndown_plugin_gfm.gfm);
+turndownService.addRule("pre2Code", {
+  filter: ["pre"],
+  replacement(content, node) {
+    const len = content.length;
+    const isCode = content[0] === "`" && content[len - 1] === "`";
+    let pre_Markdown = "";
+    if (isCode) {
+      pre_Markdown = formatCode(node.innerHTML);
+    }
+    const res = isCode ? pre_Markdown : content;
+    return "```\n" + res + "\n```\n";
+  }
+}).addRule("getImage", {
+  filter: ["img"],
+  replacement(content, node) {
+    const src = node.getAttribute("data-src") || "";
+    return src ? `
+
+ ![](${src}) 
+
+` : "";
+  }
+}).addRule("lineBreaks", {
+  filter: "br",
+  replacement: () => "\n"
+}).addRule("img2Code", {
+  filter: ["figure"],
+  replacement(content, node) {
+    const res = figure2markdown(node.innerHTML);
+    return res || "";
+  }
+});
+
+// src/index.ts
+var getError = (code) => {
+  return {
+    code,
+    success: false,
+    msg: errObj[code]
+  };
+};
+async function transformHtml2Markdown(url) {
+  let json = await import_axios.default.request({
+    url,
+    method: "get",
+    timeout: 3e4,
+    transformResponse(res) {
+      return res;
+    }
+  }).then((res) => {
+    const $ = import_cheerio2.default.load(res["data"]);
+    let title = $("#activity-name").text();
+    title = title.trim() || "";
+    const author = Array.from(new Set($("#js_name").text().split("\n").map((item) => item.trim()).filter(Boolean))).join("\n");
+    const html = $("#js_content").html();
+    if (html && html.length > 0) {
+      let res2 = turndownService.turndown(html);
+      res2 = `## ${title} 
+ 
+## \u4F5C\u8005 ${author} 
+ 
+` + res2;
+      return {
+        success: true,
+        code: Status.Success,
+        data: {
+          title,
+          author,
+          content: res2
+        }
+      };
+    }
+    return getError(Status.Fail);
+  }).catch((err) => {
+    console.log(err);
+    return err;
+  });
+  return json;
+}
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  Status
+});
+//# sourceMappingURL=index.cjs.map

package/dist/index.cjs.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/index.ts", "../node_modules/tsup/assets/cjs_shims.js", "../src/error.ts", "../src/type.ts", "../src/turndownCode.ts", "../src/formatHtml.ts"],
+  "sourcesContent": ["import axios from 'axios'\nimport cheerio from 'cheerio'\nimport { errObj } from './error'\nimport { TurnDownResult, Status } from './type'\nimport { turndownService } from './turndownCode'\n\nconst getError = (code: number) => {\n    return {\n        code,\n        success: false,\n        msg: errObj[code],\n    }\n}\n\nexport { TurnDownResult, Status }\n\nexport default async function transformHtml2Markdown(\n    url: string\n): Promise<TurnDownResult> {\n    let json: TurnDownResult = await axios\n        .request({\n            url,\n            method: 'get',\n            timeout: 30000,\n            transformResponse(res) {\n                return res\n            },\n        })\n        .then((res) => {\n            const $ = cheerio.load(res['data'])\n\n            let title = $('#activity-name').text()\n\n            title = title.trim() || ''\n\n            const author = Array.from(\n                new Set(\n                    $('#js_name')\n                        .text()\n                        .split('\\n')\n                        .map((item) => item.trim())\n                        .filter(Boolean)\n                )\n            ).join('\\n')\n\n            const html = $('#js_content').html()\n\n            if (html && html.length > 0) {\n                let res = turndownService.turndown(html)\n\n                res = `## ${title} \\n \\n` + `## \u4F5C\u8005 ${author} \\n \\n` + res\n\n                return {\n                    success: true,\n                    code: Status.Success,\n                    data: {\n                        title,\n                        author,\n                        content: res,\n                    },\n                }\n            }\n\n            return getError(Status.Fail)\n        })\n        .catch((err) => {\n            console.log(err)\n            return err\n        })\n\n    return json\n}\n", "export const importMetaUrlShim =\n  typeof document === 'undefined'\n    ? new (require('u' + 'rl').URL)('file:' + __filename).href\n    : (document.currentScript && document.currentScript.src) ||\n      new URL('main.js', document.baseURI).href\n", "export const errObj: {\n    [key: number]: string\n} = {\n    '400': '\u5185\u5BB9\u89E3\u6790\u5931\u8D25',\n}\n", "export interface TurnDownResult {\n    success: boolean\n    code: number\n    data?: {\n        title?: string\n        author?: string\n        content?: string\n    }\n    msg?: string\n}\n\nexport const enum Status {\n    Success = 200,\n    Fail = 400,\n}\n", "/**\n *  html \u8F6C\u6362 markdown \u683C\u5F0F\n */\nimport turnDownService from 'turndown'\nimport { gfm } from 'turndown-plugin-gfm'\nimport { formatCode, figure2markdown } from './formatHtml'\n\nconst turndownService = new turnDownService({\n    codeBlockStyle: 'fenced',\n    hr: '',\n})\n\nturndownService.use(gfm)\n\n// \u81EA\u5B9A\u4E49\u914D\u7F6E\nturndownService\n    .addRule('pre2Code', {\n        filter: ['pre'],\n        replacement(content, node: any) {\n            const len = content.length\n            // \u5FAE\u4FE1\u6587\u7AE0\u83B7\u53D6\u5230\u7684 content\uFF0C \u4F1A\u51FA\u73B0\u9996\u5C3E\u90FD\u6709 '`'\n            const isCode = content[0] === '`' && content[len - 1] === '`'\n\n            let pre_Markdown = ''\n\n            if (isCode) {\n                pre_Markdown = formatCode(node.innerHTML)\n            }\n\n            const res = isCode ? pre_Markdown : content\n\n            return '```\\n' + res + '\\n```\\n'\n        },\n    })\n    .addRule('getImage', {\n        filter: ['img'],\n        replacement(content, node: any) {\n            const src = node.getAttribute('data-src') || ''\n\n            return src ? `\\n\\n ![](${src}) \\n\\n` : ''\n        },\n    })\n    .addRule('lineBreaks', {\n        filter: 'br',\n        replacement: () => '\\n',\n    })\n    .addRule('img2Code', {\n        filter: ['figure'],\n        replacement(content, node: any) {\n            const res = figure2markdown(node.innerHTML)\n            return res || ''\n        },\n    })\n\nexport { turndownService }\n", "import cheerio from 'cheerio'\n\n/**\n * \u5FAE\u4FE1\u4E0D\u540C\u4EE3\u7801\u98CE\u683C\n * 1. <code><span>code</span></code>\n * 2. <code><span><span>123</span><br></span></code>\n * turndown \u4E0D\u89E3\u6790 code \u4E0B\u7684 br \u6807\u7B7E\uFF0C\u9700\u8981\u4F7F\u7528\u6B63\u5219\u66FF\u6362 br \u6807\u7B7E\u4E3A \\n \u624D\u53EF\u4EE5\u7EE7\u7EED\u89E3\u6790\n * @param htmlStr\n * @returns\n */\n\nexport function formatCode(htmlStr: string) {\n    let code = htmlStr\n\n    code = code.replace(/<br>/gi, '\\n')\n\n    code = code.replace(/&nbsp;/gi, ' ')\n\n    code = code.replace(/&lt;/gi, '<')\n\n    code = code.replace(/&gt;/gi, '>')\n\n    code = code.replace(/&amp;/gi, '&')\n\n    code = code.replace(/&quot;/gi, '\"')\n\n    code = code.replace(/&apos;/gi, '\u2018')\n\n    code = code.replace(/&times;/gi, '*')\n\n    code = code.replace(/&divide;/gi, '%')\n\n    const $ = cheerio.load(code)\n\n    return $.text()\n}\n\n/**\n * \u89E3\u51B3\u5982\u4E0B\u683C\u5F0F\n * <figcaption><img><figcaption></figcaption></figcaption>\n * @param figureHTML\n * @returns\n */\nexport function figure2markdown(figureHTML: string) {\n    const imgRegex = /<img.*?data-src=['\"](.*?)['\"]/\n\n    const descRegex = /\\<figcaption .*?>(.+)<\\/figcaption>/\n\n    const imgArr = figureHTML.match(imgRegex)\n\n    const descArr = figureHTML.match(descRegex)\n\n    let imgUrl = ''\n\n    let desc = ''\n\n    if (Array.isArray(imgArr)) {\n        imgUrl = imgArr[1]\n    }\n\n    if (Array.isArray(descArr)) {\n        desc = descArr[1]\n    }\n\n    // img \u53EF\u80FD\u6CA1\u6709\u56FE\u7247\u8BF4\u660E\n    if (imgUrl) {\n        return `\\n\\n ![${desc}](${imgUrl}) \\n\\n`\n    }\n\n    return\n}\n"],
+  "mappings": ";;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;;;ACAO,IAAM,oBACX,OAAO,aAAa,cAChB,IAAK,SAAQ,QAAY,IAAK,UAAU,YAAY,OACnD,SAAS,iBAAiB,SAAS,cAAc,OAClD,IAAI,IAAI,WAAW,SAAS,SAAS;;;ADJ3C,mBAAkB;AAClB,sBAAoB;;;AEDb,IAAM,SAET;AAAA,EACA,OAAO;AAAA;;;ACQJ,IAAW;AAAX,UAAW,SAAX;AACH,+BAAU,OAAV;AACA,4BAAO,OAAP;AAAA,GAFc;;;ACRlB,sBAA4B;AAC5B,iCAAoB;;;ACJpB,qBAAoB;AAWb,oBAAoB,SAAiB;AACxC,MAAI,OAAO;AAEX,SAAO,KAAK,QAAQ,UAAU;AAE9B,SAAO,KAAK,QAAQ,YAAY;AAEhC,SAAO,KAAK,QAAQ,UAAU;AAE9B,SAAO,KAAK,QAAQ,UAAU;AAE9B,SAAO,KAAK,QAAQ,WAAW;AAE/B,SAAO,KAAK,QAAQ,YAAY;AAEhC,SAAO,KAAK,QAAQ,YAAY;AAEhC,SAAO,KAAK,QAAQ,aAAa;AAEjC,SAAO,KAAK,QAAQ,cAAc;AAElC,QAAM,IAAI,uBAAQ,KAAK;AAEvB,SAAO,EAAE;AAAA;AASN,yBAAyB,YAAoB;AAChD,QAAM,WAAW;AAEjB,QAAM,YAAY;AAElB,QAAM,SAAS,WAAW,MAAM;AAEhC,QAAM,UAAU,WAAW,MAAM;AAEjC,MAAI,SAAS;AAEb,MAAI,OAAO;AAEX,MAAI,MAAM,QAAQ,SAAS;AACvB,aAAS,OAAO;AAAA;AAGpB,MAAI,MAAM,QAAQ,UAAU;AACxB,WAAO,QAAQ;AAAA;AAInB,MAAI,QAAQ;AACR,WAAO;AAAA;AAAA,KAAU,SAAS;AAAA;AAAA;AAAA;AAG9B;AAAA;;;AD9DJ,IAAM,kBAAkB,IAAI,wBAAgB;AAAA,EACxC,gBAAgB;AAAA,EAChB,IAAI;AAAA;AAGR,gBAAgB,IAAI;AAGpB,gBACK,QAAQ,YAAY;AAAA,EACjB,QAAQ,CAAC;AAAA,EACT,YAAY,SAAS,MAAW;AAC5B,UAAM,MAAM,QAAQ;AAEpB,UAAM,SAAS,QAAQ,OAAO,OAAO,QAAQ,MAAM,OAAO;AAE1D,QAAI,eAAe;AAEnB,QAAI,QAAQ;AACR,qBAAe,WAAW,KAAK;AAAA;AAGnC,UAAM,MAAM,SAAS,eAAe;AAEpC,WAAO,UAAU,MAAM;AAAA;AAAA,GAG9B,QAAQ,YAAY;AAAA,EACjB,QAAQ,CAAC;AAAA,EACT,YAAY,SAAS,MAAW;AAC5B,UAAM,MAAM,KAAK,aAAa,eAAe;AAE7C,WAAO,MAAM;AAAA;AAAA,OAAY;AAAA;AAAA,IAAc;AAAA;AAAA,GAG9C,QAAQ,cAAc;AAAA,EACnB,QAAQ;AAAA,EACR,aAAa,MAAM;AAAA,GAEtB,QAAQ,YAAY;AAAA,EACjB,QAAQ,CAAC;AAAA,EACT,YAAY,SAAS,MAAW;AAC5B,UAAM,MAAM,gBAAgB,KAAK;AACjC,WAAO,OAAO;AAAA;AAAA;;;AJ5C1B,IAAM,WAAW,CAAC,SAAiB;AAC/B,SAAO;AAAA,IACH;AAAA,IACA,SAAS;AAAA,IACT,KAAK,OAAO;AAAA;AAAA;AAMpB,sCACI,KACuB;AACvB,MAAI,OAAuB,MAAM,qBAC5B,QAAQ;AAAA,IACL;AAAA,IACA,QAAQ;AAAA,IACR,SAAS;AAAA,IACT,kBAAkB,KAAK;AACnB,aAAO;AAAA;AAAA,KAGd,KAAK,CAAC,QAAQ;AACX,UAAM,IAAI,wBAAQ,KAAK,IAAI;AAE3B,QAAI,QAAQ,EAAE,kBAAkB;AAEhC,YAAQ,MAAM,UAAU;AAExB,UAAM,SAAS,MAAM,KACjB,IAAI,IACA,EAAE,YACG,OACA,MAAM,MACN,IAAI,CAAC,SAAS,KAAK,QACnB,OAAO,WAElB,KAAK;AAEP,UAAM,OAAO,EAAE,eAAe;AAE9B,QAAI,QAAQ,KAAK,SAAS,GAAG;AACzB,UAAI,OAAM,gBAAgB,SAAS;AAEnC,aAAM,MAAM;AAAA;AAAA,kBAAyB;AAAA;AAAA,IAAiB;AAEtD,aAAO;AAAA,QACH,SAAS;AAAA,QACT,MAAM,OAAO;AAAA,QACb,MAAM;AAAA,UACF;AAAA,UACA;AAAA,UACA,SAAS;AAAA;AAAA;AAAA;AAKrB,WAAO,SAAS,OAAO;AAAA,KAE1B,MAAM,CAAC,QAAQ;AACZ,YAAQ,IAAI;AACZ,WAAO;AAAA;AAGf,SAAO;AAAA;",
+  "names": []
+}

package/dist/index.d.ts

@@ -0,0 +1,18 @@
+interface TurnDownResult {
+    success: boolean;
+    code: number;
+    data?: {
+        title?: string;
+        author?: string;
+        content?: string;
+    };
+    msg?: string;
+}
+declare const enum Status {
+    Success = 200,
+    Fail = 400
+}
+
+declare function transformHtml2Markdown(url: string): Promise<TurnDownResult>;
+
+export { Status, TurnDownResult, transformHtml2Markdown as default };

package/dist/index.js

@@ -0,0 +1,149 @@
+// src/index.ts
+import axios from "axios";
+import cheerio2 from "cheerio";
+
+// src/error.ts
+var errObj = {
+  "400": "\u5185\u5BB9\u89E3\u6790\u5931\u8D25"
+};
+
+// src/type.ts
+var Status;
+(function(Status2) {
+  Status2[Status2["Success"] = 200] = "Success";
+  Status2[Status2["Fail"] = 400] = "Fail";
+})(Status || (Status = {}));
+
+// src/turndownCode.ts
+import turnDownService from "turndown";
+import { gfm } from "turndown-plugin-gfm";
+
+// src/formatHtml.ts
+import cheerio from "cheerio";
+function formatCode(htmlStr) {
+  let code = htmlStr;
+  code = code.replace(/<br>/gi, "\n");
+  code = code.replace(/&nbsp;/gi, " ");
+  code = code.replace(/&lt;/gi, "<");
+  code = code.replace(/&gt;/gi, ">");
+  code = code.replace(/&amp;/gi, "&");
+  code = code.replace(/&quot;/gi, '"');
+  code = code.replace(/&apos;/gi, "\u2018");
+  code = code.replace(/&times;/gi, "*");
+  code = code.replace(/&divide;/gi, "%");
+  const $ = cheerio.load(code);
+  return $.text();
+}
+function figure2markdown(figureHTML) {
+  const imgRegex = /<img.*?data-src=['"](.*?)['"]/;
+  const descRegex = /\<figcaption .*?>(.+)<\/figcaption>/;
+  const imgArr = figureHTML.match(imgRegex);
+  const descArr = figureHTML.match(descRegex);
+  let imgUrl = "";
+  let desc = "";
+  if (Array.isArray(imgArr)) {
+    imgUrl = imgArr[1];
+  }
+  if (Array.isArray(descArr)) {
+    desc = descArr[1];
+  }
+  if (imgUrl) {
+    return `
+
+ ![${desc}](${imgUrl}) 
+
+`;
+  }
+  return;
+}
+
+// src/turndownCode.ts
+var turndownService = new turnDownService({
+  codeBlockStyle: "fenced",
+  hr: ""
+});
+turndownService.use(gfm);
+turndownService.addRule("pre2Code", {
+  filter: ["pre"],
+  replacement(content, node) {
+    const len = content.length;
+    const isCode = content[0] === "`" && content[len - 1] === "`";
+    let pre_Markdown = "";
+    if (isCode) {
+      pre_Markdown = formatCode(node.innerHTML);
+    }
+    const res = isCode ? pre_Markdown : content;
+    return "```\n" + res + "\n```\n";
+  }
+}).addRule("getImage", {
+  filter: ["img"],
+  replacement(content, node) {
+    const src = node.getAttribute("data-src") || "";
+    return src ? `
+
+ ![](${src}) 
+
+` : "";
+  }
+}).addRule("lineBreaks", {
+  filter: "br",
+  replacement: () => "\n"
+}).addRule("img2Code", {
+  filter: ["figure"],
+  replacement(content, node) {
+    const res = figure2markdown(node.innerHTML);
+    return res || "";
+  }
+});
+
+// src/index.ts
+var getError = (code) => {
+  return {
+    code,
+    success: false,
+    msg: errObj[code]
+  };
+};
+async function transformHtml2Markdown(url) {
+  let json = await axios.request({
+    url,
+    method: "get",
+    timeout: 3e4,
+    transformResponse(res) {
+      return res;
+    }
+  }).then((res) => {
+    const $ = cheerio2.load(res["data"]);
+    let title = $("#activity-name").text();
+    title = title.trim() || "";
+    const author = Array.from(new Set($("#js_name").text().split("\n").map((item) => item.trim()).filter(Boolean))).join("\n");
+    const html = $("#js_content").html();
+    if (html && html.length > 0) {
+      let res2 = turndownService.turndown(html);
+      res2 = `## ${title} 
+ 
+## \u4F5C\u8005 ${author} 
+ 
+` + res2;
+      return {
+        success: true,
+        code: Status.Success,
+        data: {
+          title,
+          author,
+          content: res2
+        }
+      };
+    }
+    return getError(Status.Fail);
+  }).catch((err) => {
+    console.log(err);
+    return err;
+  });
+  return json;
+}
+export {
+  Status,
+  transformHtml2Markdown as default
+};
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": ["../src/index.ts", "../src/error.ts", "../src/type.ts", "../src/turndownCode.ts", "../src/formatHtml.ts"],
+  "sourcesContent": ["import axios from 'axios'\nimport cheerio from 'cheerio'\nimport { errObj } from './error'\nimport { TurnDownResult, Status } from './type'\nimport { turndownService } from './turndownCode'\n\nconst getError = (code: number) => {\n    return {\n        code,\n        success: false,\n        msg: errObj[code],\n    }\n}\n\nexport { TurnDownResult, Status }\n\nexport default async function transformHtml2Markdown(\n    url: string\n): Promise<TurnDownResult> {\n    let json: TurnDownResult = await axios\n        .request({\n            url,\n            method: 'get',\n            timeout: 30000,\n            transformResponse(res) {\n                return res\n            },\n        })\n        .then((res) => {\n            const $ = cheerio.load(res['data'])\n\n            let title = $('#activity-name').text()\n\n            title = title.trim() || ''\n\n            const author = Array.from(\n                new Set(\n                    $('#js_name')\n                        .text()\n                        .split('\\n')\n                        .map((item) => item.trim())\n                        .filter(Boolean)\n                )\n            ).join('\\n')\n\n            const html = $('#js_content').html()\n\n            if (html && html.length > 0) {\n                let res = turndownService.turndown(html)\n\n                res = `## ${title} \\n \\n` + `## \u4F5C\u8005 ${author} \\n \\n` + res\n\n                return {\n                    success: true,\n                    code: Status.Success,\n                    data: {\n                        title,\n                        author,\n                        content: res,\n                    },\n                }\n            }\n\n            return getError(Status.Fail)\n        })\n        .catch((err) => {\n            console.log(err)\n            return err\n        })\n\n    return json\n}\n", "export const errObj: {\n    [key: number]: string\n} = {\n    '400': '\u5185\u5BB9\u89E3\u6790\u5931\u8D25',\n}\n", "export interface TurnDownResult {\n    success: boolean\n    code: number\n    data?: {\n        title?: string\n        author?: string\n        content?: string\n    }\n    msg?: string\n}\n\nexport const enum Status {\n    Success = 200,\n    Fail = 400,\n}\n", "/**\n *  html \u8F6C\u6362 markdown \u683C\u5F0F\n */\nimport turnDownService from 'turndown'\nimport { gfm } from 'turndown-plugin-gfm'\nimport { formatCode, figure2markdown } from './formatHtml'\n\nconst turndownService = new turnDownService({\n    codeBlockStyle: 'fenced',\n    hr: '',\n})\n\nturndownService.use(gfm)\n\n// \u81EA\u5B9A\u4E49\u914D\u7F6E\nturndownService\n    .addRule('pre2Code', {\n        filter: ['pre'],\n        replacement(content, node: any) {\n            const len = content.length\n            // \u5FAE\u4FE1\u6587\u7AE0\u83B7\u53D6\u5230\u7684 content\uFF0C \u4F1A\u51FA\u73B0\u9996\u5C3E\u90FD\u6709 '`'\n            const isCode = content[0] === '`' && content[len - 1] === '`'\n\n            let pre_Markdown = ''\n\n            if (isCode) {\n                pre_Markdown = formatCode(node.innerHTML)\n            }\n\n            const res = isCode ? pre_Markdown : content\n\n            return '```\\n' + res + '\\n```\\n'\n        },\n    })\n    .addRule('getImage', {\n        filter: ['img'],\n        replacement(content, node: any) {\n            const src = node.getAttribute('data-src') || ''\n\n            return src ? `\\n\\n ![](${src}) \\n\\n` : ''\n        },\n    })\n    .addRule('lineBreaks', {\n        filter: 'br',\n        replacement: () => '\\n',\n    })\n    .addRule('img2Code', {\n        filter: ['figure'],\n        replacement(content, node: any) {\n            const res = figure2markdown(node.innerHTML)\n            return res || ''\n        },\n    })\n\nexport { turndownService }\n", "import cheerio from 'cheerio'\n\n/**\n * \u5FAE\u4FE1\u4E0D\u540C\u4EE3\u7801\u98CE\u683C\n * 1. <code><span>code</span></code>\n * 2. <code><span><span>123</span><br></span></code>\n * turndown \u4E0D\u89E3\u6790 code \u4E0B\u7684 br \u6807\u7B7E\uFF0C\u9700\u8981\u4F7F\u7528\u6B63\u5219\u66FF\u6362 br \u6807\u7B7E\u4E3A \\n \u624D\u53EF\u4EE5\u7EE7\u7EED\u89E3\u6790\n * @param htmlStr\n * @returns\n */\n\nexport function formatCode(htmlStr: string) {\n    let code = htmlStr\n\n    code = code.replace(/<br>/gi, '\\n')\n\n    code = code.replace(/&nbsp;/gi, ' ')\n\n    code = code.replace(/&lt;/gi, '<')\n\n    code = code.replace(/&gt;/gi, '>')\n\n    code = code.replace(/&amp;/gi, '&')\n\n    code = code.replace(/&quot;/gi, '\"')\n\n    code = code.replace(/&apos;/gi, '\u2018')\n\n    code = code.replace(/&times;/gi, '*')\n\n    code = code.replace(/&divide;/gi, '%')\n\n    const $ = cheerio.load(code)\n\n    return $.text()\n}\n\n/**\n * \u89E3\u51B3\u5982\u4E0B\u683C\u5F0F\n * <figcaption><img><figcaption></figcaption></figcaption>\n * @param figureHTML\n * @returns\n */\nexport function figure2markdown(figureHTML: string) {\n    const imgRegex = /<img.*?data-src=['\"](.*?)['\"]/\n\n    const descRegex = /\\<figcaption .*?>(.+)<\\/figcaption>/\n\n    const imgArr = figureHTML.match(imgRegex)\n\n    const descArr = figureHTML.match(descRegex)\n\n    let imgUrl = ''\n\n    let desc = ''\n\n    if (Array.isArray(imgArr)) {\n        imgUrl = imgArr[1]\n    }\n\n    if (Array.isArray(descArr)) {\n        desc = descArr[1]\n    }\n\n    // img \u53EF\u80FD\u6CA1\u6709\u56FE\u7247\u8BF4\u660E\n    if (imgUrl) {\n        return `\\n\\n ![${desc}](${imgUrl}) \\n\\n`\n    }\n\n    return\n}\n"],
+  "mappings": ";AAAA;AACA;;;ACDO,IAAM,SAET;AAAA,EACA,OAAO;AAAA;;;ACQJ,IAAW;AAAX,UAAW,SAAX;AACH,+BAAU,OAAV;AACA,4BAAO,OAAP;AAAA,GAFc;;;ACRlB;AACA;;;ACJA;AAWO,oBAAoB,SAAiB;AACxC,MAAI,OAAO;AAEX,SAAO,KAAK,QAAQ,UAAU;AAE9B,SAAO,KAAK,QAAQ,YAAY;AAEhC,SAAO,KAAK,QAAQ,UAAU;AAE9B,SAAO,KAAK,QAAQ,UAAU;AAE9B,SAAO,KAAK,QAAQ,WAAW;AAE/B,SAAO,KAAK,QAAQ,YAAY;AAEhC,SAAO,KAAK,QAAQ,YAAY;AAEhC,SAAO,KAAK,QAAQ,aAAa;AAEjC,SAAO,KAAK,QAAQ,cAAc;AAElC,QAAM,IAAI,QAAQ,KAAK;AAEvB,SAAO,EAAE;AAAA;AASN,yBAAyB,YAAoB;AAChD,QAAM,WAAW;AAEjB,QAAM,YAAY;AAElB,QAAM,SAAS,WAAW,MAAM;AAEhC,QAAM,UAAU,WAAW,MAAM;AAEjC,MAAI,SAAS;AAEb,MAAI,OAAO;AAEX,MAAI,MAAM,QAAQ,SAAS;AACvB,aAAS,OAAO;AAAA;AAGpB,MAAI,MAAM,QAAQ,UAAU;AACxB,WAAO,QAAQ;AAAA;AAInB,MAAI,QAAQ;AACR,WAAO;AAAA;AAAA,KAAU,SAAS;AAAA;AAAA;AAAA;AAG9B;AAAA;;;AD9DJ,IAAM,kBAAkB,IAAI,gBAAgB;AAAA,EACxC,gBAAgB;AAAA,EAChB,IAAI;AAAA;AAGR,gBAAgB,IAAI;AAGpB,gBACK,QAAQ,YAAY;AAAA,EACjB,QAAQ,CAAC;AAAA,EACT,YAAY,SAAS,MAAW;AAC5B,UAAM,MAAM,QAAQ;AAEpB,UAAM,SAAS,QAAQ,OAAO,OAAO,QAAQ,MAAM,OAAO;AAE1D,QAAI,eAAe;AAEnB,QAAI,QAAQ;AACR,qBAAe,WAAW,KAAK;AAAA;AAGnC,UAAM,MAAM,SAAS,eAAe;AAEpC,WAAO,UAAU,MAAM;AAAA;AAAA,GAG9B,QAAQ,YAAY;AAAA,EACjB,QAAQ,CAAC;AAAA,EACT,YAAY,SAAS,MAAW;AAC5B,UAAM,MAAM,KAAK,aAAa,eAAe;AAE7C,WAAO,MAAM;AAAA;AAAA,OAAY;AAAA;AAAA,IAAc;AAAA;AAAA,GAG9C,QAAQ,cAAc;AAAA,EACnB,QAAQ;AAAA,EACR,aAAa,MAAM;AAAA,GAEtB,QAAQ,YAAY;AAAA,EACjB,QAAQ,CAAC;AAAA,EACT,YAAY,SAAS,MAAW;AAC5B,UAAM,MAAM,gBAAgB,KAAK;AACjC,WAAO,OAAO;AAAA;AAAA;;;AH5C1B,IAAM,WAAW,CAAC,SAAiB;AAC/B,SAAO;AAAA,IACH;AAAA,IACA,SAAS;AAAA,IACT,KAAK,OAAO;AAAA;AAAA;AAMpB,sCACI,KACuB;AACvB,MAAI,OAAuB,MAAM,MAC5B,QAAQ;AAAA,IACL;AAAA,IACA,QAAQ;AAAA,IACR,SAAS;AAAA,IACT,kBAAkB,KAAK;AACnB,aAAO;AAAA;AAAA,KAGd,KAAK,CAAC,QAAQ;AACX,UAAM,IAAI,SAAQ,KAAK,IAAI;AAE3B,QAAI,QAAQ,EAAE,kBAAkB;AAEhC,YAAQ,MAAM,UAAU;AAExB,UAAM,SAAS,MAAM,KACjB,IAAI,IACA,EAAE,YACG,OACA,MAAM,MACN,IAAI,CAAC,SAAS,KAAK,QACnB,OAAO,WAElB,KAAK;AAEP,UAAM,OAAO,EAAE,eAAe;AAE9B,QAAI,QAAQ,KAAK,SAAS,GAAG;AACzB,UAAI,OAAM,gBAAgB,SAAS;AAEnC,aAAM,MAAM;AAAA;AAAA,kBAAyB;AAAA;AAAA,IAAiB;AAEtD,aAAO;AAAA,QACH,SAAS;AAAA,QACT,MAAM,OAAO;AAAA,QACb,MAAM;AAAA,UACF;AAAA,UACA;AAAA,UACA,SAAS;AAAA;AAAA;AAAA;AAKrB,WAAO,SAAS,OAAO;AAAA,KAE1B,MAAM,CAAC,QAAQ;AACZ,YAAQ,IAAI;AACZ,WAAO;AAAA;AAGf,SAAO;AAAA;",
+  "names": []
+}

package/package.json

@@ -0,0 +1,66 @@
+{
+  "name": "@aiello/wechat-to-markdown",
+  "version": "1.2.3",
+  "description": "解析微信文章 URL 为 markdown",
+  "author": "Aiello Chan<aiello.chan@gmail.com>",
+  "keywords": [
+    "wechat to markdown",
+    "markdown"
+  ],
+  "main": "dist/index.cjs",
+  "module": "dist/index.js",
+  "types": "dist/index.d.ts",
+  "files": [
+    "dist"
+  ],
+  "type": "module",
+  "license": "MIT",
+  "devDependencies": {
+    "@ryan-liu/eslint-config-jsx-config": "^1.0.2",
+    "@types/fs-extra": "^9.0.13",
+    "@types/turndown": "^5.0.1",
+    "cross-spawn": "^7.0.3",
+    "eslint": "^7.32.0",
+    "husky": "^7.0.2",
+    "inquirer": "^8.2.0",
+    "lint-staged": "^11.1.2",
+    "prettier": "^2.4.1",
+    "tsup": "^5.4.0",
+    "typescript": "^4.4.3",
+    "vitest": "^1.3.1"
+  },
+  "scripts": {
+    "dev": "yarn build --watch",
+    "build": "tsup src/index.ts --dts --format cjs,esm",
+    "lint": "lint-staged",
+    "pub": "node tools/publish.js",
+    "test": "vitest"
+  },
+  "husky": {
+    "hooks": {
+      "pre-commit": "lint-staged"
+    }
+  },
+  "lint-staged": {
+    "*.{ts,js}": [
+      "eslint --fix --quiet"
+    ],
+    "*.json": [
+      "prettier --write"
+    ]
+  },
+  "tsup": {
+    "splitting": false,
+    "sourcemap": true,
+    "clean": true,
+    "external": [
+      "html2markdown"
+    ]
+  },
+  "dependencies": {
+    "axios": "^0.22.0",
+    "cheerio": "^1.0.0-rc.10",
+    "turndown": "^7.1.1",
+    "turndown-plugin-gfm": "^1.0.2"
+  }
+}