Merge pull request #8 from Yushgoel/dataFrameCompatibility

Create dataframe_go.go

base/dataframe_go.go

@@ -0,0 +1,69 @@
+package base
+
+import (
+	"fmt"
+	"reflect"
+	"strconv"
+
+	dataframe "github.com/rocketlaunchr/dataframe-go"
+
+	"github.com/sjwhitworth/golearn/base"
+)
+
+// ConvertDataFrameToInstances converts a DataFrame-go dataframe object to Golearn Fixed Data Grid. Allows for compabitibility between dataframe and golearn's ML models.
+// df is the dataframe Object. classAttrIndex is the index of the class Attribute in the data.i
+func ConvertDataFrameToInstances(df *dataframe.DataFrame, classAttrIndex int) base.FixedDataGrid {
+
+	// Creating Attributes based on Datafraem
+	names := df.Names()
+	attrs := make([]base.Attribute, len(names))
+
+	newInst := base.NewDenseInstances()
+
+	for i := range names {
+		col := df.Series[i]
+		if reflect.TypeOf(col.Value(0)).Kind() == reflect.String {
+			attrs[i] = new(base.CategoricalAttribute)
+			attrs[i].SetName(names[i])
+		} else {
+			attrs[i] = base.NewFloatAttribute(names[i])
+		}
+	}
+
+	// Add the attributes
+	newSpecs := make([]base.AttributeSpec, len(attrs))
+	for i, a := range attrs {
+
+		newSpecs[i] = newInst.AddAttribute(a)
+	}
+	// Adding the class attribute
+	newInst.AddClassAttribute(attrs[classAttrIndex])
+
+	// Allocate space
+	nRows := df.NRows()
+	newInst.Extend(df.NRows())
+
+	// Write the data based on DataType
+	for i := 0; i < nRows; i++ {
+		for j := range names {
+			col := df.Series[j]
+
+			var val string
+			switch v := col.Value(i).(type) {
+			case string:
+				val = v
+			case int64:
+				val = strconv.FormatInt(v, 10)
+			case float64:
+				val = fmt.Sprintf("%f", v)
+			case float32:
+				val = fmt.Sprintf("%f", v)
+			}
+
+			newInst.Set(newSpecs[j], i, newSpecs[j].GetAttribute().GetSysValFromString(val))
+		}
+
+	}
+
+	return newInst
+}