KHInsider downloader

khdownload.cs

using System;
using System.Collections.Generic;
using System.Diagnostics;
using System.IO;
using System.Net;
using System.Text.RegularExpressions;
using System.Threading.Tasks;
using Newtonsoft.Json;

namespace KHDownload
{
    internal static class Program
    {
        private static readonly Regex RegexJs = new Regex(@"<script>(eval[\s\S]*?)<\/script>\s+<h2>(.*?)</h2>");
        private static readonly Regex RegexParams = new Regex(@"}\('(.+)',(\d+),(\d+),'([^']+)'\.split\('\|'\)");
        private static readonly Regex RegexTracks = new Regex(@"mediaPath=\\'(.*?)\\',extension=\\'\\',tracks=\[(.*?),\],trackCount=");

        private static void Main()
        {
            while (true) {
                Console.Write("URL: ");
                string url = Console.ReadLine();
                DoDownload(url);

                Console.WriteLine("Finished downloading.");
                Console.WriteLine();
            }
        }

        private static void DoDownload(string url)
        {
            string albumSource = new WebClient().DownloadString(url);
            var matchJs = RegexJs.Match(albumSource);
            string js = matchJs.Groups[1].Value.Trim().TrimEnd(';').TrimEnd();
            string albumName = matchJs.Groups[2].Value;

            var args = RegexParams.Match(js);  // could pass RegexParams immediately
            string obfCode = args.Groups[1].Value;
            int keyspace = int.Parse(args.Groups[2].Value);
            int stringsLen = int.Parse(args.Groups[3].Value);
            string[] strings = args.Groups[4].Value.Split('|');
            Debug.Assert(keyspace == 62);
            Debug.Assert(stringsLen == strings.Length);

            var unpackedJs = Unpack(obfCode, strings);

            var matchTracks = RegexTracks.Match(unpackedJs);
            string mediaPath = matchTracks.Groups[1].Value;
            string tracksJson = "[" + matchTracks.Groups[2].Value + "]";

            var tracks = JsonConvert.DeserializeObject<TrackModel[]>(tracksJson);
            Console.WriteLine($"Found {tracks.Length} tracks");

            Directory.CreateDirectory(albumName);
            Parallel.ForEach(tracks, tr =>
            {
                string fileName = tr.Track + ". " + WebUtility.HtmlDecode(tr.Name);
                if (!fileName.EndsWith(".mp3")) fileName += ".mp3";

                string filePath = Path.Combine(albumName, fileName);
                if (File.Exists(filePath)) {
                    Console.WriteLine($"Skipping {fileName} because it already exists!");
                } else {
                    new WebClient().DownloadFile(mediaPath + tr.File, filePath);
                    Console.WriteLine($"Downloaded {fileName}");
                }
            });
        }

        private static string Unpack(string codeParam, string[] strings)
        {
            var dic = new Dictionary<string, string>();

            for (int i = strings.Length - 1; i >= 0; i--) {
                string key = Base62(i);
                dic[key] = !string.IsNullOrEmpty(strings[i]) ? strings[i] : key;
            }

            return Regex.Replace(codeParam, @"\b\w+\b", match => dic[match.Value], RegexOptions.ECMAScript);

            string Base62(int c)
            {
                const string keyspace = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ";

                string recursive = c < keyspace.Length ? string.Empty : Base62(c / keyspace.Length);
                return recursive + keyspace[c % keyspace.Length];
            }
        }

        private class TrackModel
        {
            public string Track, Name, Length, File;
        }
    }
}