SummerSchoolKA-Data-Preprocessing

using System;
using System.IO;
using System.Linq;
using System.Collections.Generic;
using System.Threading;
using System.Threading.Tasks;

namespace sska
{
	class Tuple
	{
		public int Day { get; set;}

		public String WeekDay { get; set; }

		public int WeekDayIndex { get; set;}

		public int Hour { get; set;}

		public int Minute { get; set; }

		private Lazy<DateTime> dateTime;
		internal DateTime DateTime { get { return this.dateTime.Value; } }

		public String Instance { get; set; }

		public int ResponseTime { get; set; }

		public double? ResponseTimeAverageDay { get; set; }

		public int? MaxResponseTimeDay { get; set; }

		public int? MinResponseTimeDay { get; set; }

		internal Tuple() {
			this.dateTime = new Lazy<DateTime>(() => {
				try {
					// year, month, day, hour, minute, and second
					return new DateTime(1, 1, 1 + this.Day, this.Hour, this.Minute, 0);
				} catch {
					return DateTime.Now;
				}
			});
		}

		public int InstanceId { get; set; }

		public int TimeStampRelative { get; set; }

		const string ToStringFormat =
			"{0},{1},{2},{3},{4},{5},{6},{7},{8},{9},{10},{11}";

		public override string ToString ()
		{
			return String.Format(ToStringFormat,
            	Day, WeekDay, WeekDayIndex, Hour, Minute, Instance, ResponseTime,
                ResponseTimeAverageDay, MaxResponseTimeDay, MinResponseTimeDay,
                InstanceId, TimeStampRelative);
		}
	}

	class MainClass
	{
		const string CSVHeader = "day,weekday,weekdayIndex,hour,minute,instance,responseTime," +
			"responseTimeAverageDay,maxResponseTimeDay,minResponseTimeDay," +
			"instanceId,timestampRelative\n{0}";

		public static void Main (string[] args)
		{
			if (args.Length == 0) {
				Console.WriteLine("Please specify two arguments. The first argument " +
                  "must be the absolute path to the source file and " +
                  "the other might be the output file. If you omit the " +
                  "2nd argument, all output will be written to console.");
				Environment.Exit(1);
			}
			var data = File.ReadAllLines(args[0]);

			var tuples = data.Skip(1).Where(l => !String.IsNullOrEmpty(l)).Select(l => {
				var vals = l.Split (';').Select(s => s.Replace("\"", "").Trim()).ToArray();
				// the CSV-header looks like timestamp,responsetime,championServiceInstance
				// after processing, it looks like this:
				// day, weekday, weekdayIndex, hour, minute, instanceName, respTime

				var tupleDate = default(DateTime);
				DateTime.TryParse(vals[0], out tupleDate);

				return new Tuple {
					Day = tupleDate.Day,
					WeekDay = tupleDate.DayOfWeek.ToString(),
					WeekDayIndex = (int)tupleDate.DayOfWeek,
					Hour = tupleDate.Hour,
					Minute = tupleDate.Minute,
					Instance = vals[1],
					ResponseTime = int.Parse(vals[2])
				};
			}).ToList();

			// very first thing is creating instanceIds and relative timeStamps
			int autoId = 0;
			var dictWithIds = tuples.GroupBy(t => t.Instance).Select(g => g.Key)
				.ToDictionary(kv => kv, kv => Interlocked.Increment(ref autoId));
			tuples.ForEach(t => t.InstanceId = dictWithIds[t.Instance]);

			// now the relative timestamps; exact to the minute
			// weekday * 1440 + hour * 60 + minute
			tuples.ForEach(t => t.TimeStampRelative =
            	(t.WeekDayIndex * 1440 + t.Hour * 60 + t.Minute) / 10);

			// first thing is to calculate the m/m response times each day for each server
			// (MaxResponseTimeDay, MinResponseTimeDay)
			var dayGroups = tuples.GroupBy (t => t.Day);

			foreach (var dayGroup in dayGroups) {
				// now we calc m/m & avg for each server
				var serverGroups = dayGroup.GroupBy (t => t.Instance);

				foreach (var serverGroup in serverGroups) {
					// now each group corresponds to one server..
					var foo = serverGroup.Select (s => s.ResponseTime).ToList ();
					var max = foo.Max ();
					var min = foo.Min ();
					var avg = foo.Average ();

					// now apply to each tuple
					serverGroup.ToList ().ForEach (s => {
						s.MaxResponseTimeDay = max;
						s.MinResponseTimeDay = min;
						s.ResponseTimeAverageDay = avg;
					}
					);
				}
			}

			// this is what we write to the console (or file if specified)
			var output = String.Format(CSVHeader,
               String.Join("\n", tuples.Select(t => t.ToString())));

			// export or print..
			if (args.Length > 1) {
				File.WriteAllText(args[1], output);
			} else {
				Console.Write(output);
			}
		}
	}
}