Adds broken links support

README.md

@@ -9,5 +9,20 @@ Get Broken images and screenshot of page with broken images from a list of page
 4. (HT_USER and HT_PASS are optional, use only if your site is password protected).
 5. npm start
 
+## Instructions (Broken Links)
+1. npm install.
+2. create an .env file with content like [this](env.example)
+3. SAME_SITE_LINKS is a comma separated list of urls from the original domain (without the domain or starting slash)
+4. (HT_USER and HT_PASS are optional, use only if your site is password protected).
+5. (NEW_DOMAIN is required and is the new domain to check if those urls exist)
+6. npm run broken-links
+
+## Instruction (Featured Image)
+1. npm install.
+2. create an .env file with content like [this](env.example)
+3. SAME_SITE_LINKS is a comma separated list of urls to test
+4. (HT_USER and HT_PASS are optional, use only if your site is password protected).
+5. npm run featured-image
+
 ### To Do
 * Code Refactoring

broken-links.js

@@ -0,0 +1,65 @@
+require('dotenv').config()
+const puppeteer = require('puppeteer');
+var fs = require('fs');
+
+async function asyncForEach(array, callback) {
+  for (let index = 0; index < array.length; index++) {
+    await callback(array[index], index, array)
+  }
+}
+
+async function getImages(url , index=0){
+  let username = process.env.HT_USER;
+  let password = process.env.HT_PASS;
+  let browser = await puppeteer.launch();
+  let page = await browser.newPage();
+  // let old_url = process.env.OLD_DOMAIN.replace(/\/$/, '') + '/' +url;
+  let new_url = process.env.NEW_DOMAIN.replace(/\/$/, '') + '/' +url;
+  try {
+    if (username && password) {
+      await page.authenticate({username, password});
+    }
+
+    await page.goto(new_url, {
+      waitUntil: 'networkidle2',
+      timeout: 0
+    });
+    let images_links = await page.$$eval('body.error404', image => image);
+    if (images_links.length > 0) {
+      // await page.screenshot({path: `page-${index}.png`, fullPage: true });
+      console.log(new_url);
+
+      fs.appendFile('broken-links.txt', new_url+'\n', function (err) {
+        if (err) throw err;
+        console.log('Saved!');
+      });
+    }
+    await browser.close();
+  }
+  catch (e){
+    console.error(e);
+  }
+}
+
+(async() => {
+  fs.writeFile('broken-links.txt', '', function (err) {
+    if (err) throw err;
+    console.log('Replaced!');
+  });
+
+  const same_site_links = process.env.SAME_SITE_LINKS.split(',');
+  try {
+
+  await asyncForEach(same_site_links, async (link, index) => {
+    try {
+      await getImages(link, index);
+    }
+    catch (e){
+      console.error(e);
+    }
+  });
+  }
+  catch (e){
+    console.error(e);
+  }
+})();

env.example

@@ -1,3 +1,4 @@
 HT_USER=your_user
 HT_PASS=your_password
 SAME_SITE_LINKS=https://your_link.com,https://your_link_2.com
+NEW_DOMAIN=https://google.com

package.json

@@ -6,7 +6,8 @@
   "scripts": {
     "test": "echo \"Error: no test specified\" && exit 1",
     "start": "node index.js",
-    "featured-image": "node featured-images.js"
+    "featured-image": "node featured-images.js",
+    "broken-links": "node broken-links.js"
   },
   "repository": {
     "type": "git",