use xgboost regressor

Signed-off-by: Huamin Chen <[email protected]>

pkg/model/estimator/local/lr.go

@@ -112,16 +112,37 @@ func (weights ModelWeights) predict(usageMetricNames []string, usageMetricValues
 		basePower += coeffMap[systemMetaDataFeatureValues[index]].Weight
 	}
 	var powers []float64
-	for _, vals := range usageMetricValues {
-		power := basePower
-		for index, coeff := range numericalWeights {
-			if coeff.Weight == 0 {
-				continue
+	switch weights.RegressorType {
+	case types.LinearRegressor:
+		for _, vals := range usageMetricValues {
+			power := basePower
+			for index, coeff := range numericalWeights {
+				if coeff.Weight == 0 {
+					continue
+				}
+				normalizedX := vals[index] / coeff.Scale
+				power += coeff.Weight * normalizedX
 			}
-			normalizedX := vals[index] / coeff.Scale
-			power += coeff.Weight * normalizedX
+			powers = append(powers, power)
 		}
-		powers = append(powers, power)
+	case types.XGBoostRegressor:
+		for _, vals := range usageMetricValues {
+			data := make([]float32, len(vals))
+			for index, coeff := range numericalWeights {
+				if coeff.Weight == 0 {
+					continue
+				}
+				data[index] = float32(vals[index] / coeff.Scale)
+			}
+			power, err := weights.XGBoostModel.PredictFromData(data)
+			if err != nil {
+				klog.Errorf("XGBoostModel.PredictFromData failed: %v", err)
+				return []float64{}
+			}
+			powers = append(powers, power[0])
+		}
+	default:
+		klog.Errorf("RegressorType %v is not supported", weights.RegressorType)
 	}
 	return powers
 }

pkg/model/estimator/local/xgboost_model_weight.go

@@ -125,27 +125,3 @@ func (m *XGBoostModelWeight) PredictFromData(data []float32) ([]float64, error)
 	C.free(unsafe.Pointer(output))
 	return output_array, nil
 }
-
-func (m XGBoostModelWeight) predict(usageMetricNames []string, usageMetricValues [][]float64, systemMetaDataFeatureNames, systemMetaDataFeatureValues []string) []float64 {
-	output := make([]float64, len(usageMetricValues))
-	// FIXME: how to use systmeMetaDataFeatureNames and systemMetaDataFeatureValues?
-	for i, usageMetricValue := range usageMetricValues {
-		if m.num_features != len(usageMetricValue) {
-			panic("xgboost model features and usageMetricValues length not equal")
-		}
-		// TODO we should really make a float32 array since xgboost only support float32
-		data := make([]float32, len(usageMetricValue))
-		for i, v := range usageMetricValue {
-			data[i] = float32(v)
-		}
-		val, err := m.PredictFromData(data)
-		if err != nil {
-			panic(err)
-		}
-		if len(val) != 1 {
-			panic("xgboost model predict length not equal to 1")
-		}
-		output[i] = val[0]
-	}
-	return output
-}