Add unit tests to check that the neural net can approximate x=y and y=sin(x)

Also rename the unit tests file since all these tests actually belong
to the neural net, not the tic tac toe.

Author: xanderdunn

src/test/scala/neuralNetTests.scala

@@ -0,0 +1,87 @@
+import org.scalatest._
+import collection.mutable.Stack
+import neuralNet._
+import neuralNet.NeuralNetUtilities._
+
+abstract class UnitSpec extends FlatSpec with Matchers with
+  OptionValues with Inside with Inspectors
+
+class ExampleSpec extends FlatSpec with Matchers {
+
+  "A NeuralNet" should "correctly convert a state and action into a featureVector" in {
+    var featureVetor = neuralNetFeatureVectorForStateAction(List("X", "", "", "", "", "", "" , "", ""), 2)
+    featureVetor should equal (Array(1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.0))
+    featureVetor = neuralNetFeatureVectorForStateAction(List("X", "", "", "O", "", "O", "" , "", ""), 9)
+    featureVetor should equal (Array(1.0, 0.0, 0.0, -1.0, 0.0, -1.0, 0.0, 0.0, 0.0, 9.0))
+  }
+
+  it should "be able to learn sin(x)" in {
+    val neuralNet = new NeuralNet(1, 20, 0.05, 1.0)
+    var i = 0
+    while (i < 100000) { // Train
+      val x = scala.util.Random.nextDouble()
+      val y = scala.math.sin(x)
+      neuralNet.train(Array(x), y)
+      i += 1
+    }
+    i = 0
+    while (i < 1000) { // Test it works to a degree
+      val x = scala.util.Random.nextDouble()
+      val y = scala.math.sin(x)
+      val result = neuralNet.feedForward(Array(x))
+      var withinRange = false
+      if (result < y + 0.1 && result > y - 0.1) {
+        withinRange = true
+      }
+      else {
+        println(s"x = ${x}")
+        println(s"result = ${result}")
+      }
+      withinRange should equal (true)
+      i += 1
+    }
+    while (i < 1000) { // Negative test to check that the test itself isn't broken
+      val x = scala.util.Random.nextDouble()
+      val y = scala.math.sin(x)
+      val result = neuralNet.feedForward(Array(x))
+      var withinRange = true
+      if (result < y + 0.01 && result > y - 0.01) {
+        withinRange = false
+      }
+      else {
+        println(s"x = ${x}")
+        println(s"result = ${result}")
+      }
+      withinRange should equal (false)
+      i += 1
+    }
+  }
+
+  it should "be able to learn x=y" in {
+    val neuralNet = new NeuralNet(1, 10, 0.1, 1.0)
+    var i = 0
+    while (i < 100000) { // Train
+      val x = scala.util.Random.nextDouble()
+      //println(s"input = ${x}")
+      val result = neuralNet.feedForward(Array(x))
+      //println(s"result = ${result}")
+      neuralNet.train(Array(x), x)
+      i += 1
+    }
+    i = 0
+    while (i < 1000) {
+      val x = scala.util.Random.nextDouble()
+      val result = neuralNet.feedForward(Array(x))
+      var withinRange = false
+      if (result < x + 0.1 && result > x - 0.1) {
+        withinRange = true
+      }
+      else {
+        println(s"x = ${x}")
+        println(s"result = ${result}")
+      }
+      withinRange should equal (true)
+      i += 1
+    }
+  }
+}