Fix lint errors and set read-only properties

Author: Planeshifter

lib/node_modules/@stdlib/ml/online-binary-classification/lib/loss/modified_huber.js

@@ -38,7 +38,7 @@ function modifiedHuber( weights, x, y, eta, lambda ) {
 		if ( p < -1.0 ) {
 			weights.add( x, ( eta * 4.0 * y ) );
 		} else if ( p < 1.0 ) {
-			weights.add( x, eta * ( y - p*y ) );
+			weights.add( x, eta * ( y - (p*y) ) );
 		}
 	}
 } // end FUNCTION modifiedHuber()

lib/node_modules/@stdlib/ml/online-binary-classification/lib/weight_vector.js

@@ -8,58 +8,27 @@
 
 var isPositiveInteger = require( '@stdlib/assert/is-positive-integer' );
 var isBoolean = require( '@stdlib/assert/is-boolean' );
+var setReadOnly = require( '@stdlib/utils/define-read-only-property' );
 var pow = require( '@stdlib/math/base/special/pow' );
 var dot = require( './dot.js' );
 
 
 // VARIABLES //
 
-var MIN_SCALE = 1e-11;
+var MIN_SCALE = 1.0e-11;
 
 
-// MAIN //
-
-/**
-* Creates a WeightVector.
-*
-* @constructor
-* @param {PositiveInteger} dim - number of feature weights (excluding bias/intercept term)
-* @param {boolean} intercept - boolean indicating whether a bias/intercept weight should be implicitly assumed
-*/
-function WeightVector( dim, intercept ) {
-	var i;
-	if ( !(this instanceof WeightVector) ) {
-		return new WeightVector( dim, intercept );
-	}
-	if ( !isPositiveInteger( dim ) ) {
-		throw new TypeError( 'invalid input argument. First argument `dim` must be a positive integer. Value: `' + dim + '`.' );
-	}
-	if ( !isBoolean( intercept ) ) {
-		throw new TypeError( 'invalid input argument. Second argument `intercept` must be a boolean primitive. Value: `' + intercept + '`.' );
-	}
-
-	this.scale = 1.0;
-	this.norm = 0.0;
-	this.intercept = intercept;
-	this.nWeights = dim + ( this.intercept ? 1 : 0 );
-
-	this._data = new Array( this.nWeights );
-
-	// Initialize weights to zero:
-	for ( i = 0; i < this.nWeights; i++ ) {
-		this._data[ i ] = 0.0;
-	}
-} // end FUNCTION WeightVector()
+// FUNCTIONS //
 
 /**
 * Scale elements of the weight vector by the supplied factor.
 *
-* @memberof WeightVector.prototype
-* @function scaleTo
+* @private
 * @param {number} factor - scaling factor
 * @throws {RangeError} `lambda` times `eta` must be large enough for the scaling weight to be nonnegative
 */
-WeightVector.prototype.scaleTo = function scaleTo( factor ) {
+function scaleTo( factor ) {
+	/* eslint-disable no-invalid-this */
 	var i;
 	if ( this.scale < MIN_SCALE ) {
 		// Scale vector to one:
@@ -76,23 +45,22 @@ WeightVector.prototype.scaleTo = function scaleTo( factor ) {
 	} else {
 		throw new RangeError( 'Scaling weight vector by nonpositive value, likely due to too large value of eta * lambda. Value: `' + factor + '`' );
 	}
-}; // end METHOD scaleTo()
-
+} // end FUNCTION scaleTo()
 
 /**
 * Adds vector `x` to the weight vector after scaling its elements.
 *
-* @memberof WeightVector.prototype
-* @function add
+* @private
 * @param {NumericArray} x - vector to add
 * @param {number} [xScale=1.0] - number to scale the elements of x with
 */
-WeightVector.prototype.add = function add( x, xScale ) {
+function add( x, xScale ) {
+	/* eslint-disable no-invalid-this */
 	var xscaled;
 	var inner;
 	var i;
 
-	inner = 0;
+	inner = 0.0;
 	if ( xScale === void 0 ) {
 		xScale = 1.0;
 	}
@@ -105,23 +73,22 @@ WeightVector.prototype.add = function add( x, xScale ) {
 	if ( this.intercept ) {
 		xscaled = 1.0 * xScale;
 		inner += this._data[ i ] * xscaled;
-		this._data[ i ] = this._data[ i ] + xscaled / this.scale;
+		this._data[ i ] = this._data[ i ] + ( xscaled / this.scale );
 	}
-	this.norm += ( dot( x, x ) + ( this.intercept ? 1.0 : 0.0  ) ) *
-		pow( xScale, 2 ) +
+	this.norm += ( ( dot( x, x ) + ( this.intercept ? 1.0 : 0.0 ) ) *
+		pow( xScale, 2 ) ) +
 		( 2.0 * this.scale * inner );
-}; // end METHOD add()
-
+} // end FUNCTION add()
 
 /**
 * Calculates the inner product of the weights and supplied vector `x`.
 *
-* @memberof WeightVector.prototype
-* @function innerProduct
+* @private
 * @param {NumericArray} x - input vector
 * @returns {number} inner product
 */
-WeightVector.prototype.innerProduct = function innerProduct( x ) {
+function innerProduct( x ) {
+	/* eslint-disable no-invalid-this */
 	var ret = 0;
 	var i;
 	for ( i = 0; i < x.length; i++ ) {
@@ -130,7 +97,72 @@ WeightVector.prototype.innerProduct = function innerProduct( x ) {
 	ret += this.intercept ? this._data[ i ] : 0.0;
 	ret *= this.scale;
 	return ret;
-}; // end METHOD innerProduct()
+} // end FUNCTION innerProduct()
+
+
+// MAIN //
+
+/**
+* Creates a WeightVector.
+*
+* @constructor
+* @param {PositiveInteger} dim - number of feature weights (excluding bias/intercept term)
+* @param {boolean} intercept - boolean indicating whether a bias/intercept weight should be implicitly assumed
+*/
+function WeightVector( dim, intercept ) {
+	var i;
+	if ( !(this instanceof WeightVector) ) {
+		return new WeightVector( dim, intercept );
+	}
+	if ( !isPositiveInteger( dim ) ) {
+		throw new TypeError( 'invalid input argument. First argument `dim` must be a positive integer. Value: `' + dim + '`.' );
+	}
+	if ( !isBoolean( intercept ) ) {
+		throw new TypeError( 'invalid input argument. Second argument `intercept` must be a boolean primitive. Value: `' + intercept + '`.' );
+	}
+
+	this.scale = 1.0;
+	this.norm = 0.0;
+	this.intercept = intercept;
+	this.nWeights = dim + ( this.intercept ? 1 : 0 );
+
+	this._data = new Array( this.nWeights );
+
+	// Initialize weights to zero:
+	for ( i = 0; i < this.nWeights; i++ ) {
+		this._data[ i ] = 0.0;
+	}
+} // end FUNCTION WeightVector()
+
+/**
+* Scale elements of the weight vector by the supplied factor.
+*
+* @memberof WeightVector.prototype
+* @function scaleTo
+* @param {number} factor - scaling factor
+* @throws {RangeError} `lambda` times `eta` must be large enough for the scaling weight to be nonnegative
+*/
+setReadOnly( WeightVector.prototype, 'scaleTo', scaleTo );
+
+/**
+* Adds vector `x` to the weight vector after scaling its elements.
+*
+* @memberof WeightVector.prototype
+* @function add
+* @param {NumericArray} x - vector to add
+* @param {number} [xScale=1.0] - number to scale the elements of x with
+*/
+setReadOnly( WeightVector.prototype, 'add', add );
+
+/**
+* Calculates the inner product of the weights and supplied vector `x`.
+*
+* @memberof WeightVector.prototype
+* @function innerProduct
+* @param {NumericArray} x - input vector
+* @returns {number} inner product
+*/
+setReadOnly( WeightVector.prototype, 'innerProduct', innerProduct );
 
 
 // EXPORTS //

lib/node_modules/@stdlib/ml/online-binary-classification/test/test.learner.js

@@ -249,6 +249,7 @@ tape( 'the `predict` method calculates prediction probabilities when `type` is s
 
 	x = [ 2.0, 2.0 ];
 	actual = model.predict( x, 'probability' );
+
 	// Add intercept...
 	x.push( 1.0 );
 	wx = 0;

lib/node_modules/@stdlib/ml/online-binary-classification/test/test.weight_vector.js

@@ -111,6 +111,7 @@ tape( 'the `innerProduct` method calculates the inner product of the weights and
 });
 
 tape( 'the `add` method adds a vector `x` to the weights', function test( t ) {
+	/* eslint-disable no-underscore-dangle */
 	var expected;
 	var weights;
 
@@ -166,7 +167,7 @@ tape( 'the `scaleTo` method scales the weights to one if the internal scale is s
 	weights.scaleTo( 1.0 );
 
 	t.equal( weights.scale, 1.0, 'scale is equal to 1.0' );
-	t.deepEqual( weights._data, [ 2e-13, 2e-13, 2e-13 ], '_data array is equal to [2e-13,2e-13,2e-13]' );
+	t.deepEqual( weights._data, [ 2e-13, 2e-13, 2e-13 ], '_data array is equal to [2e-13,2e-13,2e-13]' ); // eslint-disable-line no-underscore-dangle
 	t.end();
 });
 

lib/node_modules/@stdlib/ml/online-sgd-regression/lib/weight_vector.js

@@ -8,6 +8,7 @@
 
 var isPositiveInteger = require( '@stdlib/assert/is-positive-integer' );
 var isBoolean = require( '@stdlib/assert/is-boolean' );
+var setReadOnly = require( '@stdlib/utils/define-read-only-property' );
 var pow = require( '@stdlib/math/base/special/pow' );
 var dot = require( './dot.js' );
 
@@ -141,7 +142,7 @@ function WeightVector( dim, intercept ) {
 * @param {number} factor - scaling factor
 * @throws {RangeError} `lambda` times `eta` must be large enough for the scaling weight to be nonnegative
 */
-WeightVector.prototype.scaleTo = scaleTo;
+setReadOnly( WeightVector.prototype, 'scaleTo', scaleTo );
 
 /**
 * Adds vector `x` to the weight vector after scaling its elements.
@@ -151,7 +152,7 @@ WeightVector.prototype.scaleTo = scaleTo;
 * @param {NumericArray} x - vector to add
 * @param {number} [xScale=1.0] - number to scale the elements of x with
 */
-WeightVector.prototype.add = add;
+setReadOnly( WeightVector.prototype, 'add', add );
 
 /**
 * Calculates the inner product of the weights and supplied vector `x`.
@@ -161,7 +162,7 @@ WeightVector.prototype.add = add;
 * @param {NumericArray} x - input vector
 * @returns {number} inner product
 */
-WeightVector.prototype.innerProduct = innerProduct;
+setReadOnly( WeightVector.prototype, 'innerProduct', innerProduct );
 
 
 // EXPORTS //

lib/node_modules/@stdlib/ml/online-sgd-regression/test/test.loss.epsilon_insensitive.js

@@ -27,7 +27,7 @@ tape( 'the weights are unchanged for errors smaller than epsilon in magnitude (n
 	lambda = 0.0;
 
 	weights = new WeightVector( 3, false );
-	weights.add( [1.0,2.0,3.0] );
+	weights.add( [ 1.0, 2.0, 3.0 ] );
 	epsilon = 0.1;
 	eta = 0.02;
 
@@ -37,12 +37,13 @@ tape( 'the weights are unchanged for errors smaller than epsilon in magnitude (n
 	expected = [ 1.0, 2.0, 3.0 ];
 
 	epsilonInsensitiveLoss( weights, y, x, eta, lambda, epsilon );
-	t.deepEqual( weights._data, expected, 'weights are correctly updated' );
+	t.deepEqual( weights._data, expected, 'weights are correctly updated' ); // eslint-disable-line no-underscore-dangle
 
 	t.end();
 });
 
 tape( 'the sub-gradient of the linear loss times the learning rate is added to the weights for absolute errors greater or equal than epsilon (no regularization)', function test( t ) {
+	/* eslint-disable no-underscore-dangle */
 	var expected;
 	var weights;
 	var epsilon;
@@ -54,7 +55,7 @@ tape( 'the sub-gradient of the linear loss times the learning rate is added to t
 	lambda = 0.0;
 
 	weights = new WeightVector( 3, false );
-	weights.add( [1.0,2.0,3.0] );
+	weights.add( [ 1.0, 2.0, 3.0 ] );
 	epsilon = 0.1;
 	eta = 0.02;
 

lib/node_modules/@stdlib/ml/online-sgd-regression/test/test.loss.huber.js

@@ -40,11 +40,12 @@ tape( 'the sub-gradient of the squared-error loss times the learning rate is add
 
 	huberLoss( weights, x, y, eta, lambda, epsilon );
 
-	t.deepEqual( weights._data, expected, 'weights are correctly updated' );
+	t.deepEqual( weights._data, expected, 'weights are correctly updated' ); // eslint-disable-line no-underscore-dangle
 	t.end();
 });
 
 tape( 'the sub-gradient of the linear loss times the learning rate is added to the weights for absolute errors greater or equal than epsilon (no regularization)', function test( t ) {
+	/* eslint-disable no-underscore-dangle */
 	var expected;
 	var weights;
 	var epsilon;

lib/node_modules/@stdlib/ml/online-sgd-regression/test/test.loss.squared_error.js

@@ -27,7 +27,7 @@ tape( 'the sub-gradient of the squared-error loss times the learning rate is add
 	lambda = 0.0;
 
 	weights = new WeightVector( 3, false );
-	weights.add( [1.0,2.0,3.0] );
+	weights.add( [ 1.0, 2.0, 3.0 ] );
 	epsilon = 0.1;
 	eta = 0.02;
 
@@ -39,6 +39,6 @@ tape( 'the sub-gradient of the squared-error loss times the learning rate is add
 
 	squaredErrorLoss( weights, x, y, eta, lambda, epsilon );
 
-	t.deepEqual( weights._data, expected, 'weights are correctly updated' );
+	t.deepEqual( weights._data, expected, 'weights are correctly updated' ); // eslint-disable-line no-underscore-dangle
 	t.end();
 });