release/1.0.0: Add Go templates for the LinearSVC estimator

readme.md

@@ -84,7 +84,7 @@ This table gives an overview over all supported combinations of estimators, prog
  <td></td>
  <td>×</td>
  <td>✓</td>
- <td></td>
+ <td>✓</td>
  <td>×</td>
  <td>✓</td>
  <td>✓</td>

sklearn_porter/estimator/LinearSVC/__init__.py

@@ -31,6 +31,7 @@ class LinearSVC(EstimatorBase, EstimatorApiABC):
  },
  enum.Language.GO: {
  enum.Template.ATTACHED: {enum.Method.PREDICT},
+ enum.Template.EXPORTED: {enum.Method.PREDICT},
  },
  enum.Language.JAVA: {
  enum.Template.ATTACHED: {enum.Method.PREDICT},

sklearn_porter/estimator/LinearSVC/templates/go/attached.class.jinja2

@@ -14,8 +14,8 @@ import (
 )
 
 type {{ class_name }} struct {
- coeffs []float64
- inters float64
+ Coeffs []float64
+ Inters float64
 }
 {% else %}
 import (
@@ -29,8 +29,8 @@ import (
 )
 
 type {{ class_name }} struct {
- coeffs [][]float64
- inters []float64
+ Coeffs [][]float64
+ Inters []float64
 }
 {% endif %}
 
@@ -43,10 +43,10 @@ type Response struct {
 {% if is_binary %}
 func (svc {{ class_name }}) Predict(features []float64) int {
  var prob float64
- for i := 0; i < len(svc.coeffs); i++ {
- prob = prob + svc.coeffs[i] * features[i]
+ for i := 0; i < len(svc.Coeffs); i++ {
+ prob = prob + svc.Coeffs[i] * features[i]
  }
- if (prob + svc.inters) > 0 {
+ if (prob + svc.Inters) > 0 {
  return 1
  }
  return 0
@@ -55,14 +55,14 @@ func (svc {{ class_name }}) Predict(features []float64) int {
 func (svc {{ class_name }}) Predict(features []float64) int {
  classIdx := 0
  classVal := math.Inf(-1)
- outerCount, innerCount := len(svc.coeffs), len(svc.coeffs[0])
+ outerCount, innerCount := len(svc.Coeffs), len(svc.Coeffs[0])
  for i := 0; i < outerCount; i++ {
  var prob float64
  for j := 0; j < innerCount; j++ {
- prob = prob + svc.coeffs[i][j] * features[j]
+ prob = prob + svc.Coeffs[i][j] * features[j]
  }
- if prob + svc.inters[i] > classVal {
- classVal = prob + svc.inters[i]
+ if prob + svc.Inters[i] > classVal {
+ classVal = prob + svc.Inters[i]
  classIdx = i
  }
  }

sklearn_porter/estimator/LinearSVC/templates/go/exported.class.jinja2

@@ -0,0 +1,109 @@
+{% extends 'base.exported.class' %}
+
+{% block content %}
+package main
+
+{% if is_binary %}
+import (
+ {% if is_test or to_json %}
+ "encoding/json"
+ {% endif %}
+ "fmt"
+ "io/ioutil"
+ "os"
+ "strconv"
+)
+
+type {{ class_name }} struct {
+ Coeffs []float64 `json:"coeffs"`
+ Inters float64 `json:"inters"`
+}
+{% else %}
+import (
+ {% if is_test or to_json %}
+ "encoding/json"
+ {% endif %}
+ "fmt"
+ "io/ioutil"
+ "math"
+ "os"
+ "strconv"
+)
+
+type {{ class_name }} struct {
+ Coeffs [][]float64 `json:"coeffs"`
+ Inters []float64 `json:"inters"`
+}
+{% endif %}
+
+{% if is_test or to_json %}
+type Response struct {
+ Predict int `json:"predict"`
+}
+{% endif %}
+
+{% if is_binary %}
+func (svc {{ class_name }}) Predict(features []float64) int {
+ var prob float64
+ for i := 0; i < len(svc.Coeffs); i++ {
+ prob = prob + svc.Coeffs[i] * features[i]
+ }
+ if (prob + svc.Inters) > 0 {
+ return 1
+ }
+ return 0
+}
+{% else %}
+func (svc {{ class_name }}) Predict(features []float64) int {
+ classIdx := 0
+ classVal := math.Inf(-1)
+ outerCount, innerCount := len(svc.Coeffs), len(svc.Coeffs[0])
+ for i := 0; i < outerCount; i++ {
+ var prob float64
+ for j := 0; j < innerCount; j++ {
+ prob = prob + svc.Coeffs[i][j] * features[j]
+ }
+ if prob + svc.Inters[i] > classVal {
+ classVal = prob + svc.Inters[i]
+ classIdx = i
+ }
+ }
+ return classIdx
+}
+{% endif %}
+
+func main() {
+
+ // Features:
+ var features []float64
+ for _, arg := range os.Args[2:] {
+ if n, err := strconv.ParseFloat(arg, 64); err == nil {
+ features = append(features, n)
+ }
+ }
+
+ // Model data:
+ jsonFile, err := os.Open(os.Args[1])
+ if err != nil {
+ fmt.Println(err)
+ }
+ defer jsonFile.Close()
+ byteValue, _ := ioutil.ReadAll(jsonFile)
+
+ // Estimator:
+ var clf {{ class_name }}
+ json.Unmarshal(byteValue, &clf)
+
+ {% if is_test or to_json %}
+ // Get JSON:
+ prediction := clf.Predict(features)
+ res, _ := json.Marshal(&Response{Predict: prediction})
+ fmt.Println(string(res))
+ {% else %}
+ // Get class prediction:
+ prediction := clf.Predict(features)
+ fmt.Printf("Predicted class: #%d\n", prediction)
+ {% endif %}
+
+}
+{% endblock %}