private Net.Network.Network GetLoadedHotGymHierarchy()
{
Parameters p = NetworkTestHarness.GetParameters();
p = p.Union(NetworkTestHarness.GetNetworkDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
Net.Network.Network network = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer("2", p)
.Add(Anomaly.Create())
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer("3", p)
.Add(new SpatialPooler()))
.Connect("2", "3"))
.Add(Net.Network.Network.CreateRegion("r2")
.Add(Net.Network.Network.CreateLayer("1", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true)
.Add(new TemporalMemory())
.Add(new SpatialPooler())
.Add(FileSensor.Create(Net.Network.Sensor.FileSensor.Create, SensorParams.Create(
SensorParams.Keys.Path, "", ResourceLocator.Path("rec-center-hourly.csv"))))))
.Connect("r1", "r2");
return(network);
}
public void TestAutomaticClose()
{
Parameters p = NetworkTestHarness.GetParameters();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add((MultiEncoder)MultiEncoder.GetBuilder().Name("").Build())));
//.Close(); // Not necessary due to implicit call during start() or compute()
Region r1 = n.Lookup("r1");
r1.Start();
Assert.IsTrue(r1.IsClosed());
try
{
r1.Add(Net.Network.Network.CreateLayer <IInference>("5", p));
Assert.Fail();
}
catch (Exception e)
{
Assert.IsTrue(e.GetType().IsAssignableFrom(typeof(InvalidOperationException)));
Assert.AreEqual("Cannot add Layers when Region has already been closed.", e.Message);
}
}
public void TestAdd()
{
Parameters p = NetworkTestHarness.GetParameters();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add((MultiEncoder)MultiEncoder.GetBuilder().Name("").Build())));
Region r1 = n.Lookup("r1");
ILayer layer4 = r1.Lookup("4");
Assert.IsNotNull(layer4);
Assert.AreEqual("r1:4", layer4.GetName());
try
{
r1.Add(Net.Network.Network.CreateLayer <IInference>("4", p));
Assert.Fail();
}
catch (Exception e)
{
Assert.IsInstanceOfType(e, typeof(InvalidOperationException));
Assert.AreEqual("A Layer with the name: 4 has already been added to this Region.", e.Message);
}
}
public void TestEncoderPassesUpToTopLayer()
{
Parameters p = NetworkTestHarness.GetParameters();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
p.SetParameterByKey(Parameters.KEY.ANOMALY_KEY_MODE, Anomaly.Mode.PURE);
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("1", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true))
.Add(Net.Network.Network.CreateLayer <IInference>("2", p)
.Add(Anomaly.Create(p)))
.Add(Net.Network.Network.CreateLayer <IInference>("3", p)
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add(new SpatialPooler())
.Add((MultiEncoder)MultiEncoder.GetBuilder().Name("").Build())));
Region r1 = n.Lookup("r1");
r1.Connect("1", "2").Connect("2", "3").Connect("3", "4");
Assert.IsNotNull(r1.Lookup("1").GetEncoder());
}
private Net.Network.Network GetLoadedDayOfWeekNetwork()
{
Parameters p = NetworkTestHarness.GetParameters().Copy();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new XorshiftRandom(42));
Sensor <ObservableSensor <string[]> > sensor = Sensor <ObservableSensor <string[]> > .Create(
ObservableSensor <string[]> .Create, SensorParams.Create(SensorParams.Keys.Obs, new object[] { "name",
PublisherSupplier.GetBuilder()
.AddHeader("dayOfWeek")
.AddHeader("number")
.AddHeader("B").Build() }));
Net.Network.Network network = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer("1", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true)
.Add(Anomaly.Create())
.Add(new TemporalMemory())
.Add(new SpatialPooler())
.Add(sensor)));
return(network);
}
public void TestRegionSerialisation()
{
Net.Network.Network n = new Net.Network.Network("network", Parameters.Empty());
Region r = new Region("myRegion", n);
IPersistenceAPI pa = Persistence.Get(new SerialConfig(null, SerialConfig.SERIAL_TEST_DIR));
byte[] bytes = pa.Serializer().Serialize(r);
Region r2 = pa.Serializer().Deserialize <Region>(bytes);
Assert.AreEqual(r, r2);
}
public void TestAlgorithmRepetitionDetection()
{
Parameters p = NetworkTestHarness.GetParameters();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
// -- No overlap
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("2/3", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true)
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add(Sensor <FileSensor> .Create(FileSensor.Create, SensorParams.Create(
SensorParams.Keys.Path, "", ResourceLocator.Path(typeof(Resources), "days-of-week.Csv"))))
.Add(new SpatialPooler()))
.Connect("2/3", "4"));
Region r = n.Lookup("r1");
Assert.IsTrue(r.layersDistinct);
LayerMask flags = r.flagAccumulator;
flags ^= LayerMask.SpatialPooler;
flags ^= LayerMask.TemporalMemory;
flags ^= LayerMask.ClaClassifier;
Assert.AreEqual(LayerMask.None, flags);
Assert.AreEqual(r.Lookup("2/3").GetMask(), (LayerMask.TemporalMemory | LayerMask.ClaClassifier));
Assert.AreEqual(r.Lookup("4").GetMask(), LayerMask.SpatialPooler);
// -- Test overlap detection
n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("2/3", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true)
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add(Sensor <FileSensor> .Create(FileSensor.Create, SensorParams.Create(
SensorParams.Keys.Path, "", ResourceLocator.Path(typeof(Resources), "days-of-week.Csv"))))
.Add(new TemporalMemory())
.Add(new SpatialPooler()))
.Connect("2/3", "4"));
r = n.Lookup("r1");
Assert.IsFalse(r.layersDistinct);
Assert.AreEqual(r.Lookup("2/3").GetMask(), (LayerMask.TemporalMemory | LayerMask.ClaClassifier));
Assert.AreEqual(r.Lookup("4").GetMask(), (LayerMask.SpatialPooler | LayerMask.TemporalMemory));
}
public void TestHierarchicalNetwork()
{
Net.Network.Network network = GetLoadedHotGymHierarchy();
try
{
SerialConfig config = new SerialConfig("testSerializeHierarchy", SerialConfig.SERIAL_TEST_DIR);
IPersistenceAPI api = Persistence.Get(config);
api.Store(network);
}
catch (Exception e)
{
Console.WriteLine(e);
Assert.Fail();
}
}
public void TestIsLearn()
{
Parameters p = NetworkTestHarness.GetParameters().Copy();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.COLUMN_DIMENSIONS, new int[] { 30 });
p.SetParameterByKey(Parameters.KEY.SYN_PERM_INACTIVE_DEC, 0.1);
p.SetParameterByKey(Parameters.KEY.SYN_PERM_ACTIVE_INC, 0.1);
p.SetParameterByKey(Parameters.KEY.SYN_PERM_TRIM_THRESHOLD, 0.05);
p.SetParameterByKey(Parameters.KEY.SYN_PERM_CONNECTED, 0.4);
p.SetParameterByKey(Parameters.KEY.MAX_BOOST, 10.0);
p.SetParameterByKey(Parameters.KEY.DUTY_CYCLE_PERIOD, 7);
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
p.SetParameterByKey(Parameters.KEY.ANOMALY_KEY_MODE, Anomaly.Mode.PURE);
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("1", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true))
.Add(Net.Network.Network.CreateLayer <IInference>("2", p)
.Add(Anomaly.Create(p)))
.Add(Net.Network.Network.CreateLayer <IInference>("3", p)
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add(new SpatialPooler())
.Add((MultiEncoder)MultiEncoder.GetBuilder().Name("").Build()))
.Connect("1", "2")
.Connect("2", "3")
.Connect("3", "4"));
n.Lookup("r1").Close();
n.SetLearn(false);
Assert.IsFalse(n.IsLearn());
Region r1 = n.Lookup("r1");
Assert.IsFalse(n.IsLearn());
ILayer layer = r1.GetTail();
Assert.IsFalse(layer.SetIsLearn());
while (layer.GetNext() != null)
{
layer = layer.GetNext();
Assert.IsFalse(layer.SetIsLearn());
}
}
public void TestThreadedPublisher_SpatialPoolerNetwork()
{
Net.Network.Network network = CreateAndRunTestSpatialPoolerNetwork(0, 6);
ILayer l = network.Lookup("r1").Lookup("1");
Connections cn = l.GetConnections();
SerialConfig config = new SerialConfig("testThreadedPublisher_SpatialPoolerNetwork", SerialConfig.SERIAL_TEST_DIR);
IPersistenceAPI api = Persistence.Get(config);
byte[] bytes = api.Write(cn);
//Serialize above Connections for comparison with same run but unserialized below...
Connections serializedConnections = api.Read <Connections>(bytes);
Net.Network.Network network2 = CreateAndRunTestSpatialPoolerNetwork(0, 6);
ILayer l2 = network2.Lookup("r1").Lookup("1");
Connections newCons = l2.GetConnections();
//Compare the two Connections (both serialized and regular runs) - should be equal
bool b = newCons.IsDeepEqual(serializedConnections);
DeepCompare(newCons, serializedConnections);
Assert.IsTrue(b);
}
public void TestThreadedPublisher_TemporalMemoryNetwork()
{
Net.Network.Network network = CreateAndRunTestTemporalMemoryNetwork();
ILayer l = network.Lookup("r1").Lookup("1");
Connections cn = l.GetConnections();
SerialConfig config = new SerialConfig("testThreadedPublisher_TemporalMemoryNetwork", SerialConfig.SERIAL_TEST_DIR);
IPersistenceAPI api = Persistence.Get(config);
byte[] bytes = api.Write(cn);
Connections serializedConnections = api.Read <Connections>(bytes);
Net.Network.Network network2 = CreateAndRunTestTemporalMemoryNetwork();
ILayer l2 = network2.Lookup("r1").Lookup("1");
Connections newCons = l2.GetConnections();
newCons.ShouldDeepEqual(serializedConnections);
bool b = newCons.IsDeepEqual(serializedConnections);
DeepCompare(newCons, serializedConnections);
Assert.IsTrue(b);
}
public void TestSearchAndListPreviousCheckPoint()
{
Parameters p = NetworkTestHarness.GetParameters();
Net.Network.Network network = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer("1", p)
.Add(Anomaly.Create())
.Add(new TemporalMemory())
.Add(new SpatialPooler())));
IPersistenceAPI pa = Persistence.Get(new SerialConfig(null, SerialConfig.SERIAL_TEST_DIR));
ArrayUtils.Range(0, 5).ToList().ForEach(i =>
((Persistence.PersistenceAccess)pa).GetCheckPointFunction <Net.Network.Network>(network)(network));
Dictionary <string, DateTime> checkPointFiles = pa.ListCheckPointFiles();
Assert.IsTrue(checkPointFiles.Count > 4);
Assert.AreEqual(checkPointFiles.ElementAt(checkPointFiles.Count - 2).Key,
pa.GetPreviousCheckPoint(checkPointFiles.ElementAt(checkPointFiles.Count - 1)));
}
private Net.Network.Network CreateAndRunTestTemporalMemoryNetwork()
{
Publisher manual = Publisher.GetBuilder()
.AddHeader("dayOfWeek")
.AddHeader("darr")
.AddHeader("B").Build();
Sensor <ObservableSensor <String[]> > sensor = Sensor <ObservableSensor <string[]> > .Create(
ObservableSensor <string[]> .Create, SensorParams.Create(SensorParams.Keys.Obs, new Object[] { "name", manual }));
Parameters p = GetParameters();
Map <String, Map <String, Object> > settings = NetworkTestHarness.SetupMap(
null, // map
20, // n
0, // w
0, // min
0, // max
0, // radius
0, // resolution
null, // periodic
null, // clip
true, // forced
"dayOfWeek", // fieldName
"darr", // fieldType (dense array as opposed to sparse array or "sarr")
"SDRPassThroughEncoder"); // encoderType
p.SetParameterByKey(Parameters.KEY.FIELD_ENCODING_MAP, settings);
Net.Network.Network network = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer("1", p)
.Add(new TemporalMemory())
.Add(sensor)));
network.Start();
network.Observe().Subscribe(i => { }, e => Console.WriteLine(e));
// new Subscriber<Inference>() {
// @Override public void onCompleted() { }
// @Override public void onError(Throwable e) { e.printStackTrace(); }
// @Override public void onNext(Inference i) { }
//});
int[] input1 = new int[] { 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0 };
int[] input2 = new int[] { 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0 };
int[] input3 = new int[] { 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0 };
int[] input4 = new int[] { 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0 };
int[] input5 = new int[] { 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0 };
int[] input6 = new int[] { 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1 };
int[] input7 = new int[] { 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0 };
int[][] inputs = { input1, input2, input3, input4, input5, input6, input7 };
// Run until TemporalMemory is "warmed up".
int timeUntilStable = 602;
for (int j = 0; j < timeUntilStable; j++)
{
for (int i = 0; i < inputs.Length; i++)
{
manual.OnNext(Arrays.ToString(inputs[i]));
}
}
manual.OnComplete();
ILayer l = network.Lookup("r1").Lookup("1");
try
{
l.GetLayerThread().Wait();
Console.WriteLine(Arrays.ToString(SDR.AsCellIndices(l.GetConnections().GetActiveCells())));
}
catch (Exception e)
{
Assert.AreEqual(typeof(ThreadInterruptedException), e.GetType());
}
return(network);
}
public void Test2LayerAssemblyWithSensor()
{
Parameters p = NetworkTestHarness.GetParameters();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("2/3", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true)
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add(Sensor <FileSensor> .Create(FileSensor.Create, SensorParams.Create(
SensorParams.Keys.Path, "", ResourceLocator.Path(typeof(Resources), "days-of-week.Csv"))))
.Add(new SpatialPooler()))
.Connect("2/3", "4"));
int[][] inputs = new int[7][];
inputs[0] = new[] { 1, 1, 0, 0, 0, 0, 0, 1 };
inputs[1] = new[] { 1, 1, 1, 0, 0, 0, 0, 0 };
inputs[2] = new[] { 0, 1, 1, 1, 0, 0, 0, 0 };
inputs[3] = new[] { 0, 0, 1, 1, 1, 0, 0, 0 };
inputs[4] = new[] { 0, 0, 0, 1, 1, 1, 0, 0 };
inputs[5] = new[] { 0, 0, 0, 0, 1, 1, 1, 0 };
inputs[6] = new[] { 0, 0, 0, 0, 0, 1, 1, 1 };
Region r1 = n.Lookup("r1");
// Observe the top layer
r1.Lookup("4").Observe().Subscribe(
// next
i =>
{
Console.WriteLine("onNext() called (4)");
Assert.IsTrue(Arrays.AreEqual(inputs[idx0++], i.GetEncoding()));
},
//error
e => { Console.WriteLine(e); },
//completed
() => { Console.WriteLine("onCompleted() called (4)"); }
);
//r1.Lookup("4").Observe().Subscribe(new Subscriber<Inference>()
//{
// public void onCompleted() { }
// public void onError(Throwable e) { e.printStackTrace(); }
// public void onNext(Inference i)
// {
// Assert.IsTrue(Arrays.equals(inputs[idx0++], i.GetEncoding()));
// }
//});
// Observe the bottom layer
r1.Lookup("2/3").Observe().Subscribe(
// next
i =>
{
Console.WriteLine("onNext() called (2/3)");
Assert.IsTrue(Arrays.AreEqual(inputs[idx1++], i.GetEncoding()));
},
//error
e => { Console.WriteLine(e); },
//completed
() => { Console.WriteLine("onCompleted() called (2/3)"); }
);
//r1.Lookup("2/3").Observe().Subscribe(new Subscriber<Inference>() {
// public void onCompleted() { }
// public void onError(Throwable e) { e.printStackTrace(); }
// public void onNext(Inference i)
// {
// Assert.IsTrue(Arrays.equals(inputs[idx1++], i.GetEncoding()));
// }
//});
// Observe the Region output
r1.Observe().Subscribe(
// next
i =>
{
Console.WriteLine("onNext() called (out)");
Assert.IsTrue(Arrays.AreEqual(inputs[idx2++], i.GetEncoding()));
},
//error
e => { Console.WriteLine(e); },
//completed
() => { Console.WriteLine("onCompleted() called (out)"); }
);
//r1.Observe().Subscribe(new Subscriber<Inference>() {
// public void onCompleted() { }
// public void onError(Throwable e) { e.printStackTrace(); }
// public void onNext(Inference i)
// {
// Assert.IsTrue(Arrays.equals(inputs[idx2++], i.GetEncoding()));
// }
//});
r1.Start();
try
{
r1.Lookup("4").GetLayerThread().Wait();
}
catch (Exception e)
{
Console.WriteLine(e);
}
Assert.AreEqual(7, idx0);
Assert.AreEqual(7, idx1);
Assert.AreEqual(7, idx2);
}
public void TestMultiLayerAssemblyNoSensor()
{
Parameters p = NetworkTestHarness.GetParameters().Copy();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.COLUMN_DIMENSIONS, new int[] { 30 });
p.SetParameterByKey(Parameters.KEY.SYN_PERM_INACTIVE_DEC, 0.1);
p.SetParameterByKey(Parameters.KEY.SYN_PERM_ACTIVE_INC, 0.1);
p.SetParameterByKey(Parameters.KEY.SYN_PERM_TRIM_THRESHOLD, 0.05);
p.SetParameterByKey(Parameters.KEY.SYN_PERM_CONNECTED, 0.4);
p.SetParameterByKey(Parameters.KEY.MAX_BOOST, 10.0);
p.SetParameterByKey(Parameters.KEY.DUTY_CYCLE_PERIOD, 7);
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
p.SetParameterByKey(Parameters.KEY.ANOMALY_KEY_MODE, Anomaly.Mode.PURE);
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("1", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true))
.Add(Net.Network.Network.CreateLayer <IInference>("2", p)
.Add(Anomaly.Create(p)))
.Add(Net.Network.Network.CreateLayer <IInference>("3", p)
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add(new SpatialPooler())
.Add((MultiEncoder)MultiEncoder.GetBuilder().Name("").Build()))
.Connect("1", "2")
.Connect("2", "3")
.Connect("3", "4"));
Region r1 = n.Lookup("r1");
r1.Lookup("3").Using(r1.Lookup("4").GetConnections()); // How to share Connections object between Layers
//r1.Observe().Subscribe(new Subscriber<Inference>()
//{
// public void onCompleted() { }
// public void onError(Throwable e) { e.printStackTrace(); }
// public void onNext(Inference i)
// {
// // UNCOMMENT TO VIEW STABILIZATION OF PREDICTED FIELDS
// System.Out.println("Day: " + r1.GetInput() + " - predictive cells: " + i.GetPreviousPredictiveCells() +
// " - " + Arrays.toString(i.GetFeedForwardSparseActives()) + " - " +
// ((int)Math.Rint(((Number)i.GetClassification("dayOfWeek").GetMostProbableValue(1)).doubleValue())));
// }
//});
const int NUM_CYCLES = 400;
const int INPUT_GROUP_COUNT = 7; // Days of Week
Map <String, Object> multiInput = new Map <string, object>();
for (int i = 0; i < NUM_CYCLES; i++)
{
for (double j = 0; j < INPUT_GROUP_COUNT; j++)
{
multiInput.Add("dayOfWeek", j);
r1.Compute(multiInput);
}
r1.Reset();
}
r1.SetLearn(false);
r1.Reset();
// Test that we get proper output after prediction stabilization
//r1.Observe().Subscribe(new Subscriber<Inference>()
//{
// public void onCompleted() { }
// public void onError(Throwable e) { e.printStackTrace(); }
// public void onNext(Inference i)
// {
// int nextDay = ((int)Math.Rint(((Number)i.GetClassification("dayOfWeek").GetMostProbableValue(1)).doubleValue()));
// Assert.AreEqual(6, nextDay);
// }
//});
multiInput.Add("dayOfWeek", 5.0);
r1.Compute(multiInput);
}
public void TestHalt()
{
Parameters p = NetworkTestHarness.GetParameters();
p = p.Union(NetworkTestHarness.GetDayDemoTestEncoderParams());
p.SetParameterByKey(Parameters.KEY.RANDOM, new MersenneTwister(42));
p.SetParameterByKey(Parameters.KEY.ANOMALY_KEY_MODE, Anomaly.Mode.PURE);
Net.Network.Network n = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("1", p)
.AlterParameter(Parameters.KEY.AUTO_CLASSIFY, true))
.Add(Net.Network.Network.CreateLayer <IInference>("2", p)
.Add(Anomaly.Create(p)))
.Add(Net.Network.Network.CreateLayer <IInference>("3", p)
.Add(new TemporalMemory()))
.Add(Net.Network.Network.CreateLayer <IInference>("4", p)
.Add(Sensor <FileSensor> .Create(FileSensor.Create, SensorParams.Create(
SensorParams.Keys.Path, "", ResourceLocator.Path(typeof(Resources), "days-of-week.csv"))))
.Add(new SpatialPooler()))
.Connect("1", "2")
.Connect("2", "3")
.Connect("3", "4"));
Region r1 = n.Lookup("r1");
int seq = 0;
r1.Observe().Subscribe(
// next
i =>
{
if (seq == 2)
{
isHalted = true;
}
seq++;
},
//error
e => { Console.WriteLine(e); },
//completed
() => { Console.WriteLine("onCompleted() called"); }
);
//r1.Observe().Subscribe(new Subscriber<Inference>() {
// int seq = 0;
// public void onCompleted()
// {
// // System.Out.println("onCompleted() called");
// }
// public void onError(Throwable e) { e.printStackTrace(); }
// public void onNext(Inference i)
// {
// if (seq == 2)
// {
// isHalted = true;
// }
// seq++;
// // System.Out.println("output: " + i.GetSDR());
// }
//});
new Thread(() =>
{
while (!isHalted)
{
try { Thread.Sleep(1); } catch (Exception e) { Console.WriteLine(e); }
}
r1.Halt();
}).Start();
// (new Thread()
// {
// public void run()
// {
// while (!isHalted)
// {
// try { Thread.Sleep(1); } catch (Exception e) { e.printStackTrace(); }
// }
// r1.Halt();
// }
//}).Start();
r1.Start();
try
{
r1.Lookup("4").GetLayerThread().Wait();
}
catch (Exception e)
{
Console.WriteLine(e);
}
}
private Net.Network.Network CreateAndRunTestSpatialPoolerNetwork(int start, int runTo)
{
Publisher manual = Publisher.GetBuilder()
.AddHeader("dayOfWeek")
.AddHeader("darr")
.AddHeader("B")
.Build();
Sensor <ObservableSensor <string[]> > sensor = Sensor <ObservableSensor <String[]> > .Create(
ObservableSensor <string[]> .Create, SensorParams.Create(SensorParams.Keys.Obs, new Object[] { "name", manual }));
Parameters p = NetworkTestHarness.GetParameters().Copy();
p.SetParameterByKey(Parameters.KEY.RANDOM, new XorshiftRandom(42));
Map <String, Map <String, Object> > settings = NetworkTestHarness.SetupMap(
null, // map
8, // n
0, // w
0, // min
0, // max
0, // radius
0, // resolution
null, // periodic
null, // clip
true, // forced
"dayOfWeek", // fieldName
"darr", // fieldType (dense array as opposed to sparse array or "sarr")
"SDRPassThroughEncoder"); // encoderType
p.SetParameterByKey(Parameters.KEY.FIELD_ENCODING_MAP, settings);
Net.Network.Network network = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer("1", p)
.Add(new SpatialPooler())
.Add(sensor)));
network.Start();
int[][] inputs = new int[7][];
inputs[0] = new int[] { 1, 1, 0, 0, 0, 0, 0, 1 };
inputs[1] = new int[] { 1, 1, 1, 0, 0, 0, 0, 0 };
inputs[2] = new int[] { 0, 1, 1, 1, 0, 0, 0, 0 };
inputs[3] = new int[] { 0, 0, 1, 1, 1, 0, 0, 0 };
inputs[4] = new int[] { 0, 0, 0, 1, 1, 1, 0, 0 };
inputs[5] = new int[] { 0, 0, 0, 0, 1, 1, 1, 0 };
inputs[6] = new int[] { 0, 0, 0, 0, 0, 1, 1, 1 };
int[] expected0 = new int[] { 0, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0 };
int[] expected1 = new int[] { 0, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1 };
int[] expected2 = new int[] { 1, 1, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1 };
int[] expected3 = new int[] { 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0 };
int[] expected4 = new int[] { 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0 };
int[] expected5 = new int[] { 1, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0 };
int[] expected6 = new int[] { 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1 };
int[][] expecteds = new int[][] { expected0, expected1, expected2, expected3, expected4, expected5, expected6 };
//TestObserver<Inference> tester;
int test = 0;
network.Observe().Subscribe(
spatialPoolerOutput =>
{
Console.WriteLine(test + " E: " + Arrays.ToString(expecteds[test]) + " -- "
+ "A: " + Arrays.ToString(spatialPoolerOutput.GetSdr()));
Assert.IsTrue(Arrays.AreEqual(expecteds[test], spatialPoolerOutput.GetSdr()));
test++;
},
e =>
{
Console.WriteLine(e);
},
() =>
{
});
// Now push some fake data through so that "onNext" is called above
for (int i = start; i <= runTo; i++)
{
manual.OnNext(Arrays.ToString(inputs[i]));
}
manual.OnComplete();
try
{
network.Lookup("r1").Lookup("1").GetLayerThread().Wait();
}
catch (Exception e) { Console.WriteLine(e); }
//checkObserver(tester);
return(network);
}
public void TestSerializedUnStartedNetworkRuns()
{
const int NUM_CYCLES = 600;
const int INPUT_GROUP_COUNT = 7; // Days of Week
Net.Network.Network network = GetLoadedDayOfWeekNetwork();
SerialConfig config = new SerialConfig("testSerializedUnStartedNetworkRuns", SerialConfig.SERIAL_TEST_DIR);
IPersistenceAPI api = Persistence.Get(config);
api.Store(network);
//Serialize above Connections for comparison with same run but unserialized below...
Net.Network.Network serializedNetwork = api.Load();
Assert.AreEqual(serializedNetwork, network);
DeepCompare(network, serializedNetwork);
int cellsPerCol = (int)serializedNetwork.GetParameters().GetParameterByKey(Parameters.KEY.CELLS_PER_COLUMN);
serializedNetwork.Observe().Subscribe(
inf => { dayOfWeekPrintout(inf, cellsPerCol); },
e => { Console.WriteLine(e); },
() => { });
// new Observer<Inference>() {
// @Override public void onCompleted() { }
// @Override public void onError(Throwable e) { e.printStackTrace(); }
// @Override
// public void onNext(Inference inf)
// {
// /** see {@link #createDayOfWeekInferencePrintout()} */
// dayOfWeekPrintout.apply(inf, cellsPerCol);
// }
//});
Publisher pub = serializedNetwork.GetPublisher();
serializedNetwork.Start();
int cycleCount = 0;
for (; cycleCount < NUM_CYCLES; cycleCount++)
{
for (double j = 0; j < INPUT_GROUP_COUNT; j++)
{
pub.OnNext("" + j);
}
serializedNetwork.Reset();
if (cycleCount == 284)
{
break;
}
}
pub.OnComplete();
try
{
Region r1 = serializedNetwork.Lookup("r1");
r1.Lookup("1").GetLayerThread().Wait();
}
catch (Exception e)
{
Console.WriteLine(e);
Assert.Fail();
}
}
public void TestCopy()
{
Map <string, NamedTuple> classifierInput = new Map <string, NamedTuple>();
NamedTuple classifiers = new NamedTuple(new[] { "one", "two" }, 1, 2);
object layerInput = new object();
int[] sdr = new[] { 20 };
int[] encoding = new int[40];
int[] activeColumns = new int[25];
int[] sparseActives = new int[2];
HashSet <Cell> activeCells = new HashSet <Cell>(); activeCells.Add(new Cell(new Column(4, 0), 1));
HashSet <Cell> previousPrediction = new HashSet <Cell>(); previousPrediction.Add(new Cell(new Column(4, 0), 2));
HashSet <Cell> currentPrediction = new HashSet <Cell>(); currentPrediction.Add(new Cell(new Column(4, 0), 3));
Classification <object> classification = new Classification <object>();
double anomalyScore = 0.48d;
object customObject = new Net.Network.Network("MI Network", NetworkTestHarness.GetNetworkDemoTestEncoderParams());
ManualInput mi = new ManualInput()
.SetClassifierInput(classifierInput)
.SetLayerInput(layerInput)
.SetSdr(sdr)
.SetEncoding(encoding)
.SetFeedForwardActiveColumns(activeColumns)
.SetFeedForwardSparseActives(sparseActives)
.SetPredictiveCells(previousPrediction)
.SetPredictiveCells(currentPrediction) // last prediction internally becomes previous
.SetActiveCells(activeCells)
.SetClassifiers(classifiers)
.StoreClassification("foo", classification)
.SetAnomalyScore(anomalyScore)
.SetCustomObject(customObject);
ManualInput copy = mi.Copy();
Assert.IsTrue(copy.GetClassifierInput().DictEquals(classifierInput));
Assert.IsFalse(copy.GetClassifierInput() == classifierInput);
Assert.IsTrue(copy.GetLayerInput() == layerInput);
Assert.IsTrue(Arrays.AreEqual(copy.GetSdr(), sdr));
Assert.IsFalse(copy.GetSdr() == sdr);
Assert.IsTrue(Arrays.AreEqual(copy.GetEncoding(), encoding));
Assert.IsFalse(copy.GetEncoding() == encoding);
Assert.IsTrue(Arrays.AreEqual(copy.GetFeedForwardActiveColumns(), activeColumns));
Assert.IsFalse(copy.GetFeedForwardActiveColumns() == activeColumns);
Assert.IsTrue(Arrays.AreEqual(copy.GetFeedForwardSparseActives(), sparseActives));
Assert.IsFalse(copy.GetFeedForwardSparseActives() == sparseActives);
Assert.IsTrue(copy.GetPredictiveCells().SetEquals(currentPrediction));
Assert.IsFalse(copy.GetPredictiveCells() == currentPrediction);
Assert.IsTrue(copy.GetActiveCells().SetEquals(activeCells));
Assert.IsFalse(copy.GetActiveCells() == activeCells);
Assert.IsTrue(copy.GetPreviousPredictiveCells().SetEquals(previousPrediction));
Assert.IsFalse(copy.GetPreviousPredictiveCells() == previousPrediction);
Assert.IsTrue(copy.GetClassifiers().Equals(classifiers));
Assert.IsFalse(copy.GetClassifiers() == classifiers);
Assert.IsTrue(copy.GetClassification("foo").Equals(classification));
Assert.AreEqual(copy.GetAnomalyScore(), anomalyScore, 0.0); // zero deviation
Assert.AreEqual(copy.GetCustomObject(), customObject);
}
public void TestThreadedPublisher()
{
Publisher manual = Publisher.GetBuilder()
.AddHeader("dayOfWeek")
.AddHeader("darr")
.AddHeader("B").Build();
Sensor <ObservableSensor <string[]> > sensor = Sensor <ObservableSensor <string[]> > .Create(
ObservableSensor <string[]> .Create, SensorParams.Create(SensorParams.Keys.Obs, new Object[] { "name", manual }));
Parameters p = GetParameters();
Map <String, Map <String, Object> > settings = NetworkTestHarness.SetupMap(
null, // map
20, // n
0, // w
0, // min
0, // max
0, // radius
0, // resolution
null, // periodic
null, // clip
true, // forced
"dayOfWeek", // fieldName
"darr", // fieldType (dense array as opposed to sparse array or "sarr")
"SDRPassThroughEncoder"); // encoderType
p.SetParameterByKey(Parameters.KEY.FIELD_ENCODING_MAP, settings);
Net.Network.Network network = Net.Network.Network.Create("test network", p)
.Add(Net.Network.Network.CreateRegion("r1")
.Add(Net.Network.Network.CreateLayer <IInference>("1", p)
.Add(new TemporalMemory())
.Add(sensor)));
network.Start();
network.Observe().Subscribe(Observer.Create <IInference>(
output =>
{
},
Console.WriteLine,
() => { }
));
int[] input1 = new int[] { 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0 };
int[] input2 = new int[] { 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0 };
int[] input3 = new int[] { 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0 };
int[] input4 = new int[] { 0, 0, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0 };
int[] input5 = new int[] { 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0 };
int[] input6 = new int[] { 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1 };
int[] input7 = new int[] { 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0 };
int[][] inputs = { input1, input2, input3, input4, input5, input6, input7 };
// Now push some warm up data through so that "onNext" is called above
int timeUntilStable = 602;
for (int j = 0; j < timeUntilStable; j++)
{
for (int i = 0; i < inputs.Length; i++)
{
manual.OnNext(Arrays.ToString(inputs[i]));
}
}
manual.OnComplete();
ILayer l = network.Lookup("r1").Lookup("1");
try
{
l.GetLayerThread().Wait();
ComputeCycle cc = l.GetInference().GetComputeCycle();
TEST_AGGREGATION[TM_NAPI] = SDR.AsCellIndices(cc.activeCells);
}
catch (Exception e)
{
Debug.WriteLine(e.ToString());
Assert.IsInstanceOfType(e, typeof(ThreadAbortException));
}
}
