/********************************************************
* CLASS METHODS
*********************************************************/
/// <summary>
/// Primary Analysis Thread:
/// </summary>
public static void Main(string[] args)
{
//Initialize:
var algorithmPath = "";
string mode = "RELEASE";
AlgorithmNodePacket job = null;
var algorithm = default(IAlgorithm);
var startTime = DateTime.Now;
Log.LogHandler = Composer.Instance.GetExportedValueByTypeName<ILogHandler>(Config.Get("log-handler", "CompositeLogHandler"));
#if DEBUG
mode = "DEBUG";
#endif
//Name thread for the profiler:
Thread.CurrentThread.Name = "Algorithm Analysis Thread";
Log.Trace("Engine.Main(): LEAN ALGORITHMIC TRADING ENGINE v" + Constants.Version + " Mode: " + mode);
Log.Trace("Engine.Main(): Started " + DateTime.Now.ToShortTimeString());
Log.Trace("Engine.Main(): Memory " + OS.ApplicationMemoryUsed + "Mb-App " + +OS.TotalPhysicalMemoryUsed + "Mb-Used " + OS.TotalPhysicalMemory + "Mb-Total");
//Import external libraries specific to physical server location (cloud/local)
try
{
// grab the right export based on configuration
Api = Composer.Instance.GetExportedValueByTypeName<IApi>(Config.Get("api-handler"));
Notify = Composer.Instance.GetExportedValueByTypeName<IMessagingHandler>(Config.Get("messaging-handler"));
JobQueue = Composer.Instance.GetExportedValueByTypeName<IJobQueueHandler>(Config.Get("job-queue-handler"));
}
catch (CompositionException compositionException)
{ Log.Error("Engine.Main(): Failed to load library: " + compositionException);
}
//Setup packeting, queue and controls system: These don't do much locally.
Api.Initialize();
Notify.Initialize();
JobQueue.Initialize();
//Start monitoring the backtest active status:
var statusPingThread = new Thread(StateCheck.Ping.Run);
statusPingThread.Start();
try
{
//Reset algo manager internal variables preparing for a new algorithm.
AlgorithmManager.ResetManager();
//Reset thread holders.
var initializeComplete = false;
Thread threadFeed = null;
Thread threadTransactions = null;
Thread threadResults = null;
Thread threadRealTime = null;
do
{
//-> Pull job from QuantConnect job queue, or, pull local build:
job = JobQueue.NextJob(out algorithmPath); // Blocking.
// if the job version doesn't match this instance version then we can't process it
// we also don't want to reprocess redelivered live jobs
if (job.Version != Constants.Version || (LiveMode && job.Redelivered))
{
Log.Error("Engine.Run(): Job Version: " + job.Version + " Deployed Version: " + Constants.Version);
//Tiny chance there was an uncontrolled collapse of a server, resulting in an old user task circulating.
//In this event kill the old algorithm and leave a message so the user can later review.
JobQueue.AcknowledgeJob(job);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError, _collapseMessage);
Notify.SetChannel(job.Channel);
Notify.RuntimeError(job.AlgorithmId, _collapseMessage);
job = null;
}
} while (job == null);
//-> Initialize messaging system
Notify.SetChannel(job.Channel);
//-> Create SetupHandler to configure internal algorithm state:
SetupHandler = GetSetupHandler(job.SetupEndpoint);
//-> Set the result handler type for this algorithm job, and launch the associated result thread.
ResultHandler = GetResultHandler(job);
threadResults = new Thread(ResultHandler.Run, 0) {Name = "Result Thread"};
threadResults.Start();
try
{
// Save algorithm to cache, load algorithm instance:
algorithm = SetupHandler.CreateAlgorithmInstance(algorithmPath);
//Initialize the internal state of algorithm and job: executes the algorithm.Initialize() method.
initializeComplete = SetupHandler.Setup(algorithm, out _brokerage, job);
//If there are any reasons it failed, pass these back to the IDE.
if (!initializeComplete || algorithm.ErrorMessages.Count > 0 || SetupHandler.Errors.Count > 0)
{
initializeComplete = false;
//Get all the error messages: internal in algorithm and external in setup handler.
var errorMessage = String.Join(",", algorithm.ErrorMessages);
errorMessage += String.Join(",", SetupHandler.Errors);
ResultHandler.RuntimeError(errorMessage);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError);
}
}
catch (Exception err)
{
var runtimeMessage = "Algorithm.Initialize() Error: " + err.Message + " Stack Trace: " + err.StackTrace;
ResultHandler.RuntimeError(runtimeMessage, err.StackTrace);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError, runtimeMessage);
}
//-> Using the job + initialization: load the designated handlers:
if (initializeComplete)
{
//-> Reset the backtest stopwatch; we're now running the algorithm.
startTime = DateTime.Now;
//Set algorithm as locked; set it to live mode if we're trading live, and set it to locked for no further updates.
algorithm.SetAlgorithmId(job.AlgorithmId);
algorithm.SetLiveMode(LiveMode);
algorithm.SetLocked();
//Load the associated handlers for data, transaction and realtime events:
ResultHandler.SetAlgorithm(algorithm);
DataFeed = GetDataFeedHandler(algorithm, job);
TransactionHandler = GetTransactionHandler(algorithm, _brokerage, ResultHandler, job);
RealTimeHandler = GetRealTimeHandler(algorithm, _brokerage, DataFeed, ResultHandler, job);
//Set the error handlers for the brokerage asynchronous errors.
SetupHandler.SetupErrorHandler(ResultHandler, _brokerage);
//Send status to user the algorithm is now executing.
ResultHandler.SendStatusUpdate(job.AlgorithmId, AlgorithmStatus.Running);
//Launch the data, transaction and realtime handlers into dedicated threads
threadFeed = new Thread(DataFeed.Run) {Name = "DataFeed Thread"};
threadTransactions = new Thread(TransactionHandler.Run) {Name = "Transaction Thread"};
threadRealTime = new Thread(RealTimeHandler.Run) {Name = "RealTime Thread"};
//Launch the data feed, result sending, and transaction models/handlers in separate threads.
threadFeed.Start(); // Data feed pushing data packets into thread bridge;
threadTransactions.Start(); // Transaction modeller scanning new order requests
threadRealTime.Start(); // RealTime scan time for time based events:
// Result manager scanning message queue: (started earlier)
ResultHandler.DebugMessage(string.Format("Launching analysis for {0} with LEAN Engine v{1}", job.AlgorithmId, Constants.Version));
try
{
// Execute the Algorithm Code:
var complete = Isolator.ExecuteWithTimeLimit(SetupHandler.MaximumRuntime, AlgorithmManager.TimeLoopWithinLimits, () =>
{
try
{
//Run Algorithm Job:
// -> Using this Data Feed,
// -> Send Orders to this TransactionHandler,
// -> Send Results to ResultHandler.
AlgorithmManager.Run(job, algorithm, DataFeed, TransactionHandler, ResultHandler, SetupHandler, RealTimeHandler);
}
catch (Exception err)
{
//Debugging at this level is difficult, stack trace needed.
Log.Error("Engine.Run", err);
}
Log.Trace("Engine.Run(): Exiting Algorithm Manager");
}, job.UserPlan == UserPlan.Free ? 1024 : MaximumRamAllocation);
if (!complete)
{
Log.Error("Engine.Main(): Failed to complete in time: " + SetupHandler.MaximumRuntime.ToString("F"));
throw new Exception("Failed to complete algorithm within " + SetupHandler.MaximumRuntime.ToString("F") + " seconds. Please make it run faster.");
}
// Algorithm runtime error:
if (algorithm.RunTimeError != null)
{
throw algorithm.RunTimeError;
}
}
catch (Exception err)
{
//Error running the user algorithm: purge datafeed, send error messages, set algorithm status to failed.
Log.Error("Engine.Run(): Breaking out of parent try-catch: " + err.Message + " " + err.StackTrace);
if (DataFeed != null) DataFeed.Exit();
if (ResultHandler != null)
{
var message = "Runtime Error: " + err.Message;
Log.Trace("Engine.Run(): Sending runtime error to user...");
ResultHandler.LogMessage(message);
ResultHandler.RuntimeError(message, err.StackTrace);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError, message + " Stack Trace: " + err.StackTrace);
}
}
//Send result data back: this entire code block could be rewritten.
// todo: - Split up statistics class, its enormous.
// todo: - Make a dedicated Statistics.Benchmark class.
// todo: - Move all creation and transmission of statistics out of primary engine loop.
// todo: - Statistics.Generate(algorithm, resulthandler, transactionhandler);
try
{
var charts = new Dictionary<string, Chart>(ResultHandler.Charts);
var orders = new Dictionary<int, Order>(algorithm.Transactions.Orders);
var holdings = new Dictionary<string, Holding>();
var statistics = new Dictionary<string, string>();
var banner = new Dictionary<string, string>();
try
{
//Generates error when things don't exist (no charting logged, runtime errors in main algo execution)
const string strategyEquityKey = "Strategy Equity";
const string equityKey = "Equity";
const string dailyPerformanceKey = "Daily Performance";
// make sure we've taken samples for these series before just blindly requesting them
if (charts.ContainsKey(strategyEquityKey) &&
charts[strategyEquityKey].Series.ContainsKey(equityKey) &&
charts[strategyEquityKey].Series.ContainsKey(dailyPerformanceKey))
{
var equity = charts[strategyEquityKey].Series[equityKey].Values;
var performance = charts[strategyEquityKey].Series[dailyPerformanceKey].Values;
var profitLoss =
new SortedDictionary<DateTime, decimal>(algorithm.Transactions.TransactionRecord);
statistics = Statistics.Statistics.Generate(equity, profitLoss, performance,
SetupHandler.StartingPortfolioValue, algorithm.Portfolio.TotalFees, 252);
}
}
catch (Exception err)
{
Log.Error("Algorithm.Node.Engine(): Error generating statistics packet: " + err.Message);
}
//Diagnostics Completed, Send Result Packet:
var totalSeconds = (DateTime.Now - startTime).TotalSeconds;
ResultHandler.DebugMessage(string.Format("Algorithm Id:({0}) completed in {1} seconds at {2}k data points per second. Processing total of {3} data points.",
job.AlgorithmId, totalSeconds.ToString("F2"), ((AlgorithmManager.DataPoints / (double)1000) / totalSeconds).ToString("F0"), AlgorithmManager.DataPoints.ToString("N0")));
ResultHandler.SendFinalResult(job, orders, algorithm.Transactions.TransactionRecord, holdings, statistics, banner);
}
catch (Exception err)
{
Log.Error("Engine.Main(): Error sending analysis result: " + err.Message + " ST >> " + err.StackTrace);
}
//Before we return, send terminate commands to close up the threads
TransactionHandler.Exit();
DataFeed.Exit();
RealTimeHandler.Exit();
}
//Close result handler:
ResultHandler.Exit();
StateCheck.Ping.Exit();
//Wait for the threads to complete:
var ts = Stopwatch.StartNew();
while ((ResultHandler.IsActive || (TransactionHandler != null && TransactionHandler.IsActive) || (DataFeed != null && DataFeed.IsActive)) && ts.ElapsedMilliseconds < 30 * 1000)
{
Thread.Sleep(100); Log.Trace("Waiting for threads to exit...");
}
//Terminate threads still in active state.
if (threadFeed != null && threadFeed.IsAlive) threadFeed.Abort();
if (threadTransactions != null && threadTransactions.IsAlive) threadTransactions.Abort();
if (threadResults != null && threadResults.IsAlive) threadResults.Abort();
if (statusPingThread != null && statusPingThread.IsAlive) statusPingThread.Abort();
if (_brokerage != null)
{
_brokerage.Disconnect();
}
if (SetupHandler != null)
{
SetupHandler.Dispose();
}
Log.Trace("Engine.Main(): Analysis Completed and Results Posted.");
}
catch (Exception err)
{
Log.Error("Engine.Main(): Error running algorithm: " + err.Message + " >> " + err.StackTrace);
}
finally
{
//No matter what for live mode; make sure we've set algorithm status in the API for "not running" conditions:
if (LiveMode && AlgorithmManager.State != AlgorithmStatus.Running && AlgorithmManager.State != AlgorithmStatus.RuntimeError)
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmManager.State);
//Delete the message from the job queue:
JobQueue.AcknowledgeJob(job);
Log.Trace("Engine.Main(): Packet removed from queue: " + job.AlgorithmId);
//Attempt to clean up ram usage:
GC.Collect();
}
//Final disposals.
Api.Dispose();
// Make the console window pause so we can read log output before exiting and killing the application completely
if (IsLocal)
{
Log.Trace("Engine.Main(): Analysis Complete. Press any key to continue.");
Console.Read();
}
Log.LogHandler.Dispose();
}
/********************************************************
* CLASS METHODS
*********************************************************/
/// <summary>
/// Primary Analysis Thread:
/// </summary>
public static void Main(string[] args)
{
//Initialize:
var algorithmPath = "";
string mode = "RELEASE";
AlgorithmNodePacket job = null;
var algorithm = default(IAlgorithm);
var startTime = DateTime.Now;
Log.LogHandler = Composer.Instance.GetExportedValueByTypeName <ILogHandler>(Config.Get("log-handler", "CompositeLogHandler"));
#if DEBUG
mode = "DEBUG";
#endif
//Name thread for the profiler:
Thread.CurrentThread.Name = "Algorithm Analysis Thread";
Log.Trace("Engine.Main(): LEAN ALGORITHMIC TRADING ENGINE v" + Constants.Version + " Mode: " + mode);
Log.Trace("Engine.Main(): Started " + DateTime.Now.ToShortTimeString());
Log.Trace("Engine.Main(): Memory " + OS.ApplicationMemoryUsed + "Mb-App " + +OS.TotalPhysicalMemoryUsed + "Mb-Used " + OS.TotalPhysicalMemory + "Mb-Total");
//Import external libraries specific to physical server location (cloud/local)
try
{
// grab the right export based on configuration
Api = Composer.Instance.GetExportedValueByTypeName <IApi>(Config.Get("api-handler"));
Notify = Composer.Instance.GetExportedValueByTypeName <IMessagingHandler>(Config.Get("messaging-handler"));
JobQueue = Composer.Instance.GetExportedValueByTypeName <IJobQueueHandler>(Config.Get("job-queue-handler"));
}
catch (CompositionException compositionException)
{ Log.Error("Engine.Main(): Failed to load library: " + compositionException); }
//Setup packeting, queue and controls system: These don't do much locally.
Api.Initialize();
Notify.Initialize();
JobQueue.Initialize();
//Start monitoring the backtest active status:
var statusPingThread = new Thread(StateCheck.Ping.Run);
statusPingThread.Start();
try
{
//Reset algo manager internal variables preparing for a new algorithm.
AlgorithmManager.ResetManager();
//Reset thread holders.
var initializeComplete = false;
Thread threadFeed = null;
Thread threadTransactions = null;
Thread threadResults = null;
Thread threadRealTime = null;
do
{
//-> Pull job from QuantConnect job queue, or, pull local build:
job = JobQueue.NextJob(out algorithmPath); // Blocking.
// if the job version doesn't match this instance version then we can't process it
// we also don't want to reprocess redelivered live jobs
if (job.Version != Constants.Version || (LiveMode && job.Redelivered))
{
Log.Error("Engine.Run(): Job Version: " + job.Version + " Deployed Version: " + Constants.Version);
//Tiny chance there was an uncontrolled collapse of a server, resulting in an old user task circulating.
//In this event kill the old algorithm and leave a message so the user can later review.
JobQueue.AcknowledgeJob(job);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError, _collapseMessage);
Notify.SetChannel(job.Channel);
Notify.RuntimeError(job.AlgorithmId, _collapseMessage);
job = null;
}
} while (job == null);
//-> Initialize messaging system
Notify.SetChannel(job.Channel);
//-> Create SetupHandler to configure internal algorithm state:
SetupHandler = GetSetupHandler(job.SetupEndpoint);
//-> Set the result handler type for this algorithm job, and launch the associated result thread.
ResultHandler = GetResultHandler(job);
threadResults = new Thread(ResultHandler.Run, 0)
{
Name = "Result Thread"
};
threadResults.Start();
try
{
// Save algorithm to cache, load algorithm instance:
algorithm = SetupHandler.CreateAlgorithmInstance(algorithmPath);
//Initialize the internal state of algorithm and job: executes the algorithm.Initialize() method.
initializeComplete = SetupHandler.Setup(algorithm, out _brokerage, job);
//If there are any reasons it failed, pass these back to the IDE.
if (!initializeComplete || algorithm.ErrorMessages.Count > 0 || SetupHandler.Errors.Count > 0)
{
initializeComplete = false;
//Get all the error messages: internal in algorithm and external in setup handler.
var errorMessage = String.Join(",", algorithm.ErrorMessages);
errorMessage += String.Join(",", SetupHandler.Errors);
ResultHandler.RuntimeError(errorMessage);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError);
}
}
catch (Exception err)
{
var runtimeMessage = "Algorithm.Initialize() Error: " + err.Message + " Stack Trace: " + err.StackTrace;
ResultHandler.RuntimeError(runtimeMessage, err.StackTrace);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError, runtimeMessage);
}
//-> Using the job + initialization: load the designated handlers:
if (initializeComplete)
{
//-> Reset the backtest stopwatch; we're now running the algorithm.
startTime = DateTime.Now;
//Set algorithm as locked; set it to live mode if we're trading live, and set it to locked for no further updates.
algorithm.SetAlgorithmId(job.AlgorithmId);
algorithm.SetLiveMode(LiveMode);
algorithm.SetLocked();
//Load the associated handlers for data, transaction and realtime events:
ResultHandler.SetAlgorithm(algorithm);
DataFeed = GetDataFeedHandler(algorithm, job);
TransactionHandler = GetTransactionHandler(algorithm, _brokerage, ResultHandler, job);
RealTimeHandler = GetRealTimeHandler(algorithm, _brokerage, DataFeed, ResultHandler, job);
//Set the error handlers for the brokerage asynchronous errors.
SetupHandler.SetupErrorHandler(ResultHandler, _brokerage);
//Send status to user the algorithm is now executing.
ResultHandler.SendStatusUpdate(job.AlgorithmId, AlgorithmStatus.Running);
//Launch the data, transaction and realtime handlers into dedicated threads
threadFeed = new Thread(DataFeed.Run)
{
Name = "DataFeed Thread"
};
threadTransactions = new Thread(TransactionHandler.Run)
{
Name = "Transaction Thread"
};
threadRealTime = new Thread(RealTimeHandler.Run)
{
Name = "RealTime Thread"
};
//Launch the data feed, result sending, and transaction models/handlers in separate threads.
threadFeed.Start(); // Data feed pushing data packets into thread bridge;
threadTransactions.Start(); // Transaction modeller scanning new order requests
threadRealTime.Start(); // RealTime scan time for time based events:
// Result manager scanning message queue: (started earlier)
ResultHandler.DebugMessage(string.Format("Launching analysis for {0} with LEAN Engine v{1}", job.AlgorithmId, Constants.Version));
try
{
// Execute the Algorithm Code:
var complete = Isolator.ExecuteWithTimeLimit(SetupHandler.MaximumRuntime, AlgorithmManager.TimeLoopWithinLimits, () =>
{
try
{
//Run Algorithm Job:
// -> Using this Data Feed,
// -> Send Orders to this TransactionHandler,
// -> Send Results to ResultHandler.
AlgorithmManager.Run(job, algorithm, DataFeed, TransactionHandler, ResultHandler, SetupHandler, RealTimeHandler);
}
catch (Exception err)
{
//Debugging at this level is difficult, stack trace needed.
Log.Error("Engine.Run", err);
}
Log.Trace("Engine.Run(): Exiting Algorithm Manager");
}, job.UserPlan == UserPlan.Free ? 1024 : MaximumRamAllocation);
if (!complete)
{
Log.Error("Engine.Main(): Failed to complete in time: " + SetupHandler.MaximumRuntime.ToString("F"));
throw new Exception("Failed to complete algorithm within " + SetupHandler.MaximumRuntime.ToString("F") + " seconds. Please make it run faster.");
}
// Algorithm runtime error:
if (algorithm.RunTimeError != null)
{
throw algorithm.RunTimeError;
}
}
catch (Exception err)
{
//Error running the user algorithm: purge datafeed, send error messages, set algorithm status to failed.
Log.Error("Engine.Run(): Breaking out of parent try-catch: " + err.Message + " " + err.StackTrace);
if (DataFeed != null)
{
DataFeed.Exit();
}
if (ResultHandler != null)
{
var message = "Runtime Error: " + err.Message;
Log.Trace("Engine.Run(): Sending runtime error to user...");
ResultHandler.LogMessage(message);
ResultHandler.RuntimeError(message, err.StackTrace);
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmStatus.RuntimeError, message + " Stack Trace: " + err.StackTrace);
}
}
//Send result data back: this entire code block could be rewritten.
// todo: - Split up statistics class, its enormous.
// todo: - Make a dedicated Statistics.Benchmark class.
// todo: - Move all creation and transmission of statistics out of primary engine loop.
// todo: - Statistics.Generate(algorithm, resulthandler, transactionhandler);
try
{
var charts = new Dictionary <string, Chart>(ResultHandler.Charts);
var orders = new Dictionary <int, Order>(algorithm.Transactions.Orders);
var holdings = new Dictionary <string, Holding>();
var statistics = new Dictionary <string, string>();
var banner = new Dictionary <string, string>();
try
{
//Generates error when things don't exist (no charting logged, runtime errors in main algo execution)
const string strategyEquityKey = "Strategy Equity";
const string equityKey = "Equity";
const string dailyPerformanceKey = "Daily Performance";
// make sure we've taken samples for these series before just blindly requesting them
if (charts.ContainsKey(strategyEquityKey) &&
charts[strategyEquityKey].Series.ContainsKey(equityKey) &&
charts[strategyEquityKey].Series.ContainsKey(dailyPerformanceKey))
{
var equity = charts[strategyEquityKey].Series[equityKey].Values;
var performance = charts[strategyEquityKey].Series[dailyPerformanceKey].Values;
var profitLoss =
new SortedDictionary <DateTime, decimal>(algorithm.Transactions.TransactionRecord);
statistics = Statistics.Statistics.Generate(equity, profitLoss, performance,
SetupHandler.StartingPortfolioValue, algorithm.Portfolio.TotalFees, 252);
}
}
catch (Exception err)
{
Log.Error("Algorithm.Node.Engine(): Error generating statistics packet: " + err.Message);
}
//Diagnostics Completed, Send Result Packet:
var totalSeconds = (DateTime.Now - startTime).TotalSeconds;
ResultHandler.DebugMessage(string.Format("Algorithm Id:({0}) completed in {1} seconds at {2}k data points per second. Processing total of {3} data points.",
job.AlgorithmId, totalSeconds.ToString("F2"), ((AlgorithmManager.DataPoints / (double)1000) / totalSeconds).ToString("F0"), AlgorithmManager.DataPoints.ToString("N0")));
ResultHandler.SendFinalResult(job, orders, algorithm.Transactions.TransactionRecord, holdings, statistics, banner);
}
catch (Exception err)
{
Log.Error("Engine.Main(): Error sending analysis result: " + err.Message + " ST >> " + err.StackTrace);
}
//Before we return, send terminate commands to close up the threads
TransactionHandler.Exit();
DataFeed.Exit();
RealTimeHandler.Exit();
}
//Close result handler:
ResultHandler.Exit();
StateCheck.Ping.Exit();
//Wait for the threads to complete:
var ts = Stopwatch.StartNew();
while ((ResultHandler.IsActive || (TransactionHandler != null && TransactionHandler.IsActive) || (DataFeed != null && DataFeed.IsActive)) && ts.ElapsedMilliseconds < 30 * 1000)
{
Thread.Sleep(100); Log.Trace("Waiting for threads to exit...");
}
//Terminate threads still in active state.
if (threadFeed != null && threadFeed.IsAlive)
{
threadFeed.Abort();
}
if (threadTransactions != null && threadTransactions.IsAlive)
{
threadTransactions.Abort();
}
if (threadResults != null && threadResults.IsAlive)
{
threadResults.Abort();
}
if (statusPingThread != null && statusPingThread.IsAlive)
{
statusPingThread.Abort();
}
if (_brokerage != null)
{
_brokerage.Disconnect();
}
if (SetupHandler != null)
{
SetupHandler.Dispose();
}
Log.Trace("Engine.Main(): Analysis Completed and Results Posted.");
}
catch (Exception err)
{
Log.Error("Engine.Main(): Error running algorithm: " + err.Message + " >> " + err.StackTrace);
}
finally
{
//No matter what for live mode; make sure we've set algorithm status in the API for "not running" conditions:
if (LiveMode && AlgorithmManager.State != AlgorithmStatus.Running && AlgorithmManager.State != AlgorithmStatus.RuntimeError)
{
Api.SetAlgorithmStatus(job.AlgorithmId, AlgorithmManager.State);
}
//Delete the message from the job queue:
JobQueue.AcknowledgeJob(job);
Log.Trace("Engine.Main(): Packet removed from queue: " + job.AlgorithmId);
//Attempt to clean up ram usage:
GC.Collect();
}
//Final disposals.
Api.Dispose();
// Make the console window pause so we can read log output before exiting and killing the application completely
if (IsLocal)
{
Log.Trace("Engine.Main(): Analysis Complete. Press any key to continue.");
Console.Read();
}
Log.LogHandler.Dispose();
}
