Experimentally scraping data from HardX.

package.json

@@ -1,30 +1,41 @@
 {
-  "name": "traxxx",
+    "name": "traxxx",
-  "version": "1.0.0",
+    "version": "1.0.0",
-  "description": "All the latest porn releases in one place",
+    "description": "All the latest porn releases in one place",
-  "main": "src/app.js",
+    "main": "src/app.js",
-  "scripts": {
+    "scripts": {
-    "test": "echo \"Error: no test specified\" && exit 1"
+        "start": "node src/app.js",
-  },
+        "eslint": "eslint src/",
-  "repository": {
+        "eslint-watch": "esw --watch src/"
-    "type": "git",
+    },
-    "url": "https://gitea.unknown.name/niels/traxxx.git"
+    "repository": {
-  },
+        "type": "git",
-  "keywords": [
+        "url": "https://gitea.unknown.name/niels/traxxx.git"
-    "porn",
+    },
-    "releases",
+    "keywords": [
-    "updates",
+        "porn",
-    "nsfw"
+        "releases",
-  ],
+        "updates",
-  "author": "Niels Simenon",
+        "nsfw"
-  "license": "ISC",
+    ],
-  "devDependencies": {
+    "author": "Niels Simenon",
-    "@babel/cli": "^7.2.3",
+    "license": "ISC",
-    "@babel/core": "^7.3.4",
+    "devDependencies": {
-    "@babel/preset-env": "^7.3.4",
+        "@babel/cli": "^7.2.3",
-    "babel-preset-airbnb": "^3.2.0"
+        "@babel/core": "^7.3.4",
-  },
+        "@babel/preset-env": "^7.3.4",
-  "dependencies": {
+        "babel-eslint": "^10.0.1",
-    "config": "^3.0.1"
+        "babel-preset-airbnb": "^3.2.0",
-  }
+        "eslint": "^5.15.0",
+        "eslint-config-airbnb-base": "^13.1.0",
+        "eslint-plugin-import": "^2.16.0",
+        "eslint-watch": "^4.0.2"
+    },
+    "dependencies": {
+        "bhttp": "^1.2.4",
+        "cheerio": "^1.0.0-rc.2",
+        "config": "^3.0.1",
+        "date-fns": "^1.30.1",
+        "terminal-kit": "^1.27.0"
+    }
 }

src/app.js

@@ -0,0 +1,45 @@
+'use strict';
+
+const bhttp = require('bhttp');
+const cheerio = require('cheerio');
+const { parse, format } = require('date-fns');
+
+async function init() {
+    const baseUrl = 'https://www.hardx.com';
+    const res = await bhttp.get(`${baseUrl}/en/videos`, {});
+    const $ = cheerio.load(res.body.toString(), { normalizeWhitespace: true });
+    const scenesElements = $('.sceneInfo').toArray();
+
+    const scenes = scenesElements.map((element) => {
+        const sceneLinkElement = $(element).find('.sceneTitle a');
+        const url = `${baseUrl}${sceneLinkElement.attr('href')}`;
+        const title = sceneLinkElement.attr('title');
+
+        const [likes, dislikes] = $(element).find('.value')
+            .toArray()
+            .map(value => Number($(value).text()));
+        const stars = Math.floor(((likes * 5 + dislikes) / (likes + dislikes)) * 100) / 100;
+
+        const actors = $(element).find('.sceneActors a')
+            .map((actorIndex, actorElement) => $(actorElement).attr('title'))
+            .toArray();
+
+        const date = parse($(element).find('.sceneDate').text(), 'MM-DD-YYYY');
+
+        return {
+            url,
+            title,
+            actors,
+            date,
+            rating: {
+                likes,
+                dislikes,
+                stars,
+            },
+        };
+    });
+
+    console.log(scenes);
+}
+
+init();