Migrate from tsc 2.3.x to tsc 2.4.x (#38)

* migrate to typescript 2.4.x

* shorten travis log by supressing npm run err statements

* Merge master into tsc

Author: dsmilkov

.travis.yml

@@ -3,5 +3,5 @@ node_js: "8"
 install:
   - npm install
 script:
-  - npm run build
-  - npm run lint
+  - npm run build --silent
+  - npm run lint --silent

package.json

@@ -25,7 +25,7 @@
     "tsify": "~3.0.1",
     "tslint": "~5.6.0",
     "typedoc": "~0.7.2",
-    "typescript": "2.3.4",
+    "typescript": "2.4.2",
     "watchify": "~3.9.0"
   },
   "scripts": {

src/math/activation_functions.ts

@@ -23,13 +23,13 @@ export interface ActivationFunction {
 }
 
 export class TanHFunc implements ActivationFunction {
-  output(math: NDArrayMath, x: NDArray) {
+  output<T extends NDArray>(math: NDArrayMath, x: T) {
     return math.scope(() => {
       return math.tanh(x);
     });
   }
 
-  der(math: NDArrayMath, x: NDArray, y: NDArray) {
+  der<T extends NDArray>(math: NDArrayMath, x: T, y: T) {
     return math.scope(() => {
       const ySquared = math.elementWiseMul(y, y);
       // 1 - y^2.
@@ -39,27 +39,27 @@ export class TanHFunc implements ActivationFunction {
 }
 
 export class ReLUFunc implements ActivationFunction {
-  output(math: NDArrayMath, x: NDArray) {
+  output<T extends NDArray>(math: NDArrayMath, x: T) {
     return math.scope(() => {
       return math.relu(x);
     });
   }
 
-  der(math: NDArrayMath, x: NDArray, y: NDArray) {
+  der<T extends NDArray>(math: NDArrayMath, x: T, y: T) {
     return math.scope(() => {
       return math.step(x);
     });
   }
 }
 
 export class SigmoidFunc implements ActivationFunction {
-  output(math: NDArrayMath, x: NDArray) {
+  output<T extends NDArray>(math: NDArrayMath, x: T) {
     return math.scope(() => {
       return math.sigmoid(x);
     });
   }
 
-  der(math: NDArrayMath, x: NDArray, y: NDArray) {
+  der<T extends NDArray>(math: NDArrayMath, x: T, y: T) {
     return math.scope(() => {
       // y * (1 - y) = y - y^2
       const ySquared = math.elementWiseMul(y, y);
@@ -69,13 +69,13 @@ export class SigmoidFunc implements ActivationFunction {
 }
 
 export class SquareFunc implements ActivationFunction {
-  output(math: NDArrayMath, x: NDArray) {
+  output<T extends NDArray>(math: NDArrayMath, x: T) {
     return math.scope(() => {
       return math.elementWiseMul(x, x);
     });
   }
 
-  der(math: NDArrayMath, x: NDArray, y: NDArray) {
+  der<T extends NDArray>(math: NDArrayMath, x: T, y: T) {
     return math.scope(() => {
       // dy/dx = 2*x.
       return math.scalarTimesArray(Scalar.TWO, x);

src/math/cost_functions.ts

@@ -28,7 +28,7 @@ export interface ElementWiseCostFunction {
 export class SquareCostFunc implements ElementWiseCostFunction {
   private halfOne = Scalar.new(0.5);
 
-  cost(math: NDArrayMath, x1: NDArray, x2: NDArray): NDArray {
+  cost<T extends NDArray>(math: NDArrayMath, x1: T, x2: T): T {
     const diff = math.sub(x1, x2);
     const diffSquared = math.elementWiseMul(diff, diff);
     const result = math.scalarTimesArray(this.halfOne, diffSquared);
@@ -39,7 +39,7 @@ export class SquareCostFunc implements ElementWiseCostFunction {
     return result;
   }
 
-  der(math: NDArrayMath, x1: NDArray, x2: NDArray): NDArray {
+  der<T extends NDArray>(math: NDArrayMath, x1: T, x2: T): T {
     return math.sub(x1, x2);
   }
 

src/math/math_cpu.ts

@@ -69,7 +69,7 @@ export class NDArrayMathCPU extends NDArrayMath {
     const outputShape =
         concat3d_util.computeConcat3DOutputShape(x1.shape, x2.shape, axis);
 
-    const values = NDArray.zeros<Array3D>(outputShape);
+    const values = Array3D.zeros(outputShape);
 
     for (let i = 0; i < outputShape[0]; i++) {
       for (let j = 0; j < outputShape[1]; j++) {

src/math/math_cpu_test.ts

@@ -245,8 +245,8 @@ describe('NDArrayMathCPU matMul', () => {
   });
 
   it('A x B^t shapes do not match', () => {
-    const a = NDArray.zeros<Array2D>([2, 3]);
-    const b = NDArray.zeros<Array2D>([3, 2]);
+    const a = Array2D.zeros([2, 3]);
+    const b = Array2D.zeros([3, 2]);
     const f = () => {
       math.matMul(
           a, b, MatrixOrientation.REGULAR, MatrixOrientation.TRANSPOSED);
@@ -255,8 +255,8 @@ describe('NDArrayMathCPU matMul', () => {
   });
 
   it('A^t x B shapes do not match', () => {
-    const a = NDArray.zeros<Array2D>([2, 3]);
-    const b = NDArray.zeros<Array2D>([3, 2]);
+    const a = Array2D.zeros([2, 3]);
+    const b = Array2D.zeros([3, 2]);
     const f = () => {
       math.matMul(
           a, b, MatrixOrientation.TRANSPOSED, MatrixOrientation.REGULAR);
@@ -265,8 +265,8 @@ describe('NDArrayMathCPU matMul', () => {
   });
 
   it('A^t x B^t shapes do not match', () => {
-    const a = NDArray.zeros<Array2D>([3, 2]);
-    const b = NDArray.zeros<Array2D>([3, 2]);
+    const a = Array2D.zeros([3, 2]);
+    const b = Array2D.zeros([3, 2]);
     const f = () => {
       math.matMul(
           a, b, MatrixOrientation.TRANSPOSED, MatrixOrientation.TRANSPOSED);

src/math/ndarray.ts

@@ -99,9 +99,9 @@ export class NDArray {
   }
 
   /** Creates a ndarray of zeros with the specified shape. */
-  static zeros<T extends NDArray>(shape: number[]): T {
+  static zeros(shape: number[]): NDArray {
     const values = new Float32Array(util.sizeFromShape(shape));
-    return NDArray.make<T>(shape, {values});
+    return NDArray.make(shape, {values});
   }
 
   /** Creates a ndarray of zeros with the same shape as the specified ndarray.
@@ -390,7 +390,7 @@ export class Array1D extends NDArray {
   }
 
   static zeros(shape: [number]): Array1D {
-    return NDArray.zeros<Array1D>(shape);
+    return NDArray.zeros(shape) as Array1D;
   }
 }
 
@@ -441,7 +441,7 @@ export class Array2D extends NDArray {
   }
 
   static zeros(shape: [number, number]): Array2D {
-    return NDArray.zeros<Array2D>(shape);
+    return NDArray.zeros(shape) as Array2D;
   }
 }
 
@@ -496,7 +496,7 @@ export class Array3D extends NDArray {
   }
 
   static zeros(shape: [number, number, number]): Array3D {
-    return NDArray.zeros<Array3D>(shape);
+    return NDArray.zeros(shape) as Array3D;
   }
 }
 
@@ -559,7 +559,7 @@ export class Array4D extends NDArray {
   }
 
   static zeros(shape: [number, number, number, number]): Array4D {
-    return NDArray.zeros<Array4D>(shape);
+    return NDArray.zeros(shape) as Array4D;
   }
 }