github.com/jingcheng-WU/gonum@v0.9.1-0.20210323123734-f1a2a11a8f7b/optimize/functions/validate.go (about) 1 // Copyright ©2015 The Gonum Authors. All rights reserved. 2 // Use of this source code is governed by a BSD-style 3 // license that can be found in the LICENSE file. 4 5 package functions 6 7 import ( 8 "math" 9 "testing" 10 11 "github.com/jingcheng-WU/gonum/diff/fd" 12 "github.com/jingcheng-WU/gonum/floats" 13 ) 14 15 // function represents an objective function. 16 type function interface { 17 Func(x []float64) float64 18 } 19 20 type gradient interface { 21 Grad(grad, x []float64) []float64 22 } 23 24 // minimumer is an objective function that can also provide information about 25 // its minima. 26 type minimumer interface { 27 function 28 29 // Minima returns _known_ minima of the function. 30 Minima() []Minimum 31 } 32 33 // Minimum represents information about an optimal location of a function. 34 type Minimum struct { 35 // X is the location of the minimum. X may not be nil. 36 X []float64 37 // F is the value of the objective function at X. 38 F float64 39 // Global indicates if the location is a global minimum. 40 Global bool 41 } 42 43 type funcTest struct { 44 X []float64 45 46 // F is the expected function value at X. 47 F float64 48 // Gradient is the expected gradient at X. If nil, it is not evaluated. 49 Gradient []float64 50 } 51 52 // TODO(vladimir-ch): Decide and implement an exported testing function: 53 // func Test(f Function, ??? ) ??? { 54 // } 55 56 const ( 57 defaultTol = 1e-12 58 defaultGradTol = 1e-9 59 defaultFDGradTol = 1e-5 60 ) 61 62 // testFunction checks that the function can evaluate itself (and its gradient) 63 // correctly. 64 func testFunction(f function, ftests []funcTest, t *testing.T) { 65 // Make a copy of tests because we may append to the slice. 66 tests := make([]funcTest, len(ftests)) 67 copy(tests, ftests) 68 69 // Get information about the function. 70 fMinima, isMinimumer := f.(minimumer) 71 fGradient, isGradient := f.(gradient) 72 73 // If the function is a Minimumer, append its minima to the tests. 74 if isMinimumer { 75 for _, minimum := range fMinima.Minima() { 76 // Allocate gradient only if the function can evaluate it. 77 var grad []float64 78 if isGradient { 79 grad = make([]float64, len(minimum.X)) 80 } 81 tests = append(tests, funcTest{ 82 X: minimum.X, 83 F: minimum.F, 84 Gradient: grad, 85 }) 86 } 87 } 88 89 for i, test := range tests { 90 F := f.Func(test.X) 91 92 // Check that the function value is as expected. 93 if math.Abs(F-test.F) > defaultTol { 94 t.Errorf("Test #%d: function value given by Func is incorrect. Want: %v, Got: %v", 95 i, test.F, F) 96 } 97 98 if test.Gradient == nil { 99 continue 100 } 101 102 // Evaluate the finite difference gradient. 103 fdGrad := fd.Gradient(nil, f.Func, test.X, &fd.Settings{ 104 Formula: fd.Central, 105 Step: 1e-6, 106 }) 107 108 // Check that the finite difference and expected gradients match. 109 if !floats.EqualApprox(fdGrad, test.Gradient, defaultFDGradTol) { 110 dist := floats.Distance(fdGrad, test.Gradient, math.Inf(1)) 111 t.Errorf("Test #%d: numerical and expected gradients do not match. |fdGrad - WantGrad|_∞ = %v", 112 i, dist) 113 } 114 115 // If the function is a Gradient, check that it computes the gradient correctly. 116 if isGradient { 117 grad := make([]float64, len(test.Gradient)) 118 fGradient.Grad(grad, test.X) 119 120 if !floats.EqualApprox(grad, test.Gradient, defaultGradTol) { 121 dist := floats.Distance(grad, test.Gradient, math.Inf(1)) 122 t.Errorf("Test #%d: gradient given by Grad is incorrect. |grad - WantGrad|_∞ = %v", 123 i, dist) 124 } 125 } 126 } 127 }