using HtmlAgilityPack;

using System;

using System.IO;

using System.Linq;

using System.Net;

​

namespace WebScraping

{

    public class Program

    {

        public static void Main(string[] args)

        {

            var url = "https://www.politie.nl/nieuws/2019/november/15/09-operatie-alfa---grote-vangst-op-kamp-verdachten-langer-vast.html";

            var web = new HtmlWeb();

            var htmlDoc = web.Load(url);

            // nullable operator doesnt work here  var content = doc.DocumentNode.Descendants(0).FirstOrDefault(n => n.HasClass("content-blocks"))?.InnerText;

            var content = htmlDoc.DocumentNode.Descendants(0)

                            .FirstOrDefault(n => n.HasClass("content-blocks")).InnerText;

​

            // Find all H2's / Strong tags in the document

            // Might use these to wrap tags around the "innerText" of this document

            var headings = htmlDoc.DocumentNode.SelectNodes("//h2");

            var subheadings = htmlDoc.DocumentNode.SelectNodes("//strong");

            WriteCollection(headings, "Headings");

            WriteCollection(subheadings, "SubHeadings");

            // Two known ways to retreive image.

            var image = htmlDoc.DocumentNode.Descendants(0).FirstOrDefault(n => n.HasClass("nieuwsafbeelding")) ?? htmlDoc.DocumentNode.Descendants(0).FirstOrDefault(n => n.HasClass("opsporing-image"));

​

            if (image != null)

            {

                var imageSrc = image.SelectNodes("//img").FirstOrDefault(n => n.Attributes["src"].Value.Contains("/binaries")).GetAttributeValue("src", null).ToString();

                Console.WriteLine("Image found: https://politie.nl " +imageSrc);

                Console.WriteLine("-----------------");

            }

            Console.WriteLine("-----------------");

            Console.WriteLine("...::[CONTENT]::..");

            Console.WriteLine(content.Trim());

        }

        public static void WriteCollection (HtmlNodeCollection nodeCollection, string name) 

        {

            if (nodeCollection != null)

            {

                Console.WriteLine(String.Format("..::FOUND {0} [{1}] on this page::..", nodeCollection.Count, name.ToUpper()));

                Console.WriteLine("-------------------------------------");

​

                foreach (var node in nodeCollection)

                {

                    Console.WriteLine(String.Format("> {0}", node.InnerText));

                }

                return;

           }

            Console.WriteLine("-----------------");

            Console.WriteLine("..::[{0}]::..", name);

            Console.WriteLine("> none found");

            Console.WriteLine("-----------------");

        }

    }

}

​