Ya hace predicciones y crea el ZIP

testML/DictionaryToObjectConverter.cs

@@ -16,9 +16,9 @@ namespace testML
 {
 	public static class DictionaryToObjectConverter
 	{
-		public static IEnumerable<object> Convert(List<Dictionary<string, object>> data, string toPredict, out Type classType, out DataViewSchema schema)
+		public static IEnumerable<object> Convert(List<Dictionary<string, object>> data, string toPredict, out Type classType, out Type classPredictionType, out DataViewSchema schema)
 		{
-			var schemaBuilder = new DataViewSchema.Builder();			
+			var schemaBuilder = new DataViewSchema.Builder();
 
 			var definition = new Dictionary<string, Type>();
 
@@ -32,12 +32,6 @@ namespace testML
 					if (sampleValue != null)
 					{
 						var keyType = sampleValue.GetType();
-
-						if (key == toPredict)
-						{
-							keyType = typeof(float);
-						}
-
 						definition.Add(key, keyType);
 
 						if (keyType == typeof(string))
@@ -101,42 +95,22 @@ namespace testML
 			var dllAssembly = compilerResults.CompiledAssembly;
 
 			classType = dllAssembly.GetType("DictionaryToObjectConverterNamespace." + converter.ClassName);
+			classPredictionType = dllAssembly.GetType("DictionaryToObjectConverterNamespace." + converter.ClassName + "Prediction");
 
 			Type listType = typeof(List<>);
 			Type genericType = listType.MakeGenericType(classType);
 
-			var result =(IList) Activator.CreateInstance(genericType) ;
+			var result = (IList)Activator.CreateInstance(genericType);
+
 
-			Dictionary<string, float> translate = new Dictionary<string, float>();
-			translate.Add(string.Empty, 0);
 			foreach (var inputData in data)
-			{
-				if (inputData.ContainsKey(toPredict) && inputData[toPredict] != null)
-				{
-					if (!translate.ContainsKey(inputData[toPredict] as string))
-					{
-						var max = translate.Values.Max()+1;
-						translate.Add(inputData[toPredict] as string, max);
-					}
-				}
-			}
-
-
-				foreach (var inputData in data)
 			{
 				var outputData = (IDictionaryToObjectConverter)Activator.CreateInstance(classType);
 				result.Add(outputData);
 
 				foreach (var key in inputData.Keys)
 				{
-					if (key == toPredict)
+					outputData[key] = inputData[key];
-					{
-						outputData[key] = translate[inputData[key] as string ?? string.Empty];
-					}
-					else
-					{
-						outputData[key] = inputData[key];
-					}
 				}
 			}
 

testML/DictionaryToObjectConverterClass.cs

@@ -28,8 +28,14 @@ namespace testML
         /// </summary>
         public virtual string TransformText()
         {
-            this.Write("\r\nusing System;\r\nusing System.Text;\r\nusing Microsoft.ML.Data;\r\n\r\nnamespace Dictio" +
+            
-                    "naryToObjectConverterNamespace\r\n{\r\n\tpublic class ");
+            #line 6 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+ var toPredictType = Definition[ToPredict]; 
+            
+            #line default
+            #line hidden
+            this.Write("using System;\r\nusing System.Text;\r\nusing Microsoft.ML.Data;\r\n\r\nnamespace Dictiona" +
+                    "ryToObjectConverterNamespace\r\n{\r\n\tpublic class ");
             
             #line 13 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
             this.Write(this.ToStringHelper.ToStringWithCulture(ClassName));
@@ -141,7 +147,77 @@ namespace testML
             
             #line default
             #line hidden
-            this.Write("\r\n\t\t\t}\r\n\t\t\treturn null;\r\n\t\t}\r\n\t}\r\n}");
+            this.Write("\r\n\t\t\t}\r\n\t\t\treturn null;\r\n\t\t}\r\n\t}\r\n\r\n\tpublic class ");
+            
+            #line 54 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(ClassName));
+            
+            #line default
+            #line hidden
+            this.Write("Prediction: testML.IDictionaryToObjectConverter\r\n\t{\r\n\t\t[ColumnName(\"PredictedLabe" +
+                    "l\")]\t\r\n\t\tpublic ");
+            
+            #line 57 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(toPredictType.FullName));
+            
+            #line default
+            #line hidden
+            this.Write(" ");
+            
+            #line 57 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(ToPredict));
+            
+            #line default
+            #line hidden
+            this.Write(@" { get; set; }
+
+		public object this[string propertyName]
+		{
+			get { return GetValue(propertyName); }
+			set { SetValue(propertyName, value); }
+		}
+
+		public void SetValue(string propertyName, object value)
+		{
+			switch(propertyName)
+			{
+				case """);
+            
+            #line 69 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(ToPredict));
+            
+            #line default
+            #line hidden
+            this.Write("\":\t");
+            
+            #line 69 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(ToPredict));
+            
+            #line default
+            #line hidden
+            this.Write(" = (");
+            
+            #line 69 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(toPredictType.FullName));
+            
+            #line default
+            #line hidden
+            this.Write(")value;\tbreak;\r\n\t\t\t}\r\n\t\t}\r\n\r\n\t\tpublic object GetValue(string propertyName)\r\n\t\t{\r\n" +
+                    "\t\t\tswitch(propertyName)\r\n\t\t\t{\r\n\t\t\t\tcase \"");
+            
+            #line 77 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(ToPredict));
+            
+            #line default
+            #line hidden
+            this.Write("\":\treturn ");
+            
+            #line 77 "C:\Users\miguel.maldonado\Documents\Subversion\TestML\testML\DictionaryToObjectConverterClass.tt"
+            this.Write(this.ToStringHelper.ToStringWithCulture(ToPredict));
+            
+            #line default
+            #line hidden
+            this.Write(";\r\n\t\t\t}\r\n\t\t\treturn null;\r\n\t\t}\r\n\t}\r\n}");
             return this.GenerationEnvironment.ToString();
         }
     }

testML/DictionaryToObjectConverterClass.tt

@@ -3,7 +3,7 @@
 <#@ import namespace="System.Linq" #>
 <#@ import namespace="System.Text" #>
 <#@ import namespace="System.Collections.Generic" #>
-
+<# var toPredictType = Definition[ToPredict]; #>
 using System;
 using System.Text;
 using Microsoft.ML.Data;
@@ -50,4 +50,33 @@ namespace DictionaryToObjectConverterNamespace
 			return null;
 		}
 	}
+
+	public class <#= ClassName #>Prediction: testML.IDictionaryToObjectConverter
+	{
+		[ColumnName("PredictedLabel")]	
+		public <#= toPredictType.FullName #> <#= ToPredict #> { get; set; }
+
+		public object this[string propertyName]
+		{
+			get { return GetValue(propertyName); }
+			set { SetValue(propertyName, value); }
+		}
+
+		public void SetValue(string propertyName, object value)
+		{
+			switch(propertyName)
+			{
+				case "<#= ToPredict #>":	<#= ToPredict #> = (<#= toPredictType.FullName #>)value;	break;
+			}
+		}
+
+		public object GetValue(string propertyName)
+		{
+			switch(propertyName)
+			{
+				case "<#= ToPredict #>":	return <#= ToPredict #>;
+			}
+			return null;
+		}
+	}
 }

testML/Program.cs

@@ -2,6 +2,8 @@
 using Microsoft.ML.AutoML;
 using Microsoft.ML.Data;
 using Microsoft.ML.Trainers;
+using Microsoft.ML.Transforms.Text;
+using Microsoft.SqlServer.Server;
 using NPOI.XSSF.UserModel;
 using System;
 using System.Collections.Generic;
@@ -24,7 +26,8 @@ namespace testML
 		static void Main(string[] args)
 		{
 			XSSFWorkbook wb;
-			using (FileStream file = new FileStream(@"C:\Users\miguel.maldonado\Downloads\entrenar_IAMenos.xlsx", FileMode.Open, FileAccess.Read))
+			//using (FileStream file = new FileStream(@"C:\Users\miguel.maldonado\Downloads\entrenar_IAMenos.xlsx", FileMode.Open, FileAccess.Read))
+			using (FileStream file = new FileStream(@"C:\Users\miguel.maldonado\Downloads\entrenar_IA.xlsx", FileMode.Open, FileAccess.Read))
 			{
 				wb = new XSSFWorkbook(file);
 			}
@@ -40,7 +43,7 @@ namespace testML
 
 			for (var r = 1; r < sheet.LastRowNum - 1; r++)
 			{
-				if (r == 300) break;
+				//if (r == 50) break;
 				Console.WriteLine(string.Format("{0} / {1}", r, sheet.LastRowNum - 1));
 				var row = sheet.GetRow(r);
 
@@ -102,126 +105,208 @@ namespace testML
 				tmpData.Add(rowData);
 			}
 
+
+			var columnToPredict = "DESCENDIENTE_S4i001";
+
+			var firstRow = tmpData[0] as IDictionary<string, object>;
+			foreach (var key in firstRow.Keys.ToArray())
+			{
+				var firstValue = (from x in tmpData where x.ContainsKey(key) && x[key] != null && !string.IsNullOrEmpty(x[key] as string) select x[key]).FirstOrDefault();
+				if (firstValue == null)
+				{
+					foreach (var item in tmpData)
+					{
+						if (item.ContainsKey(key))
+						{
+							item.Remove(key);
+						}
+					}
+				}
+			}
 			#endregion
 
 
 
 			MLContext mlContext = new MLContext();
 
-			var dataConverted = DictionaryToObjectConverter.Convert(tmpData, "DESCENDIENTE_S4i001", out Type classType, out DataViewSchema schema);
+
+
+			mlContext.Log += (_, e) =>
+			{
+				if (e.Kind == Microsoft.ML.Runtime.ChannelMessageKind.Trace && e.Source.EndsWith(" Cursor")) { return; }
+				if (e.Kind == Microsoft.ML.Runtime.ChannelMessageKind.Trace && e.Source.EndsWith(" CursorSplitter")) { return; }
+				if (e.Kind == Microsoft.ML.Runtime.ChannelMessageKind.Trace && e.Source.EndsWith(" Consolidate")) { return; }
+				if (e.Kind == Microsoft.ML.Runtime.ChannelMessageKind.Trace && e.Source.EndsWith(" Training")) { return; }
+
+				if (e.Kind == Microsoft.ML.Runtime.ChannelMessageKind.Trace && e.Source.Equals("RangeFilter; Checking parameters")) { return; }
+
+
+				//if (e.Source.Equals("AutoMLExperiment"))
+				//{
+				Console.WriteLine(e.RawMessage);
+				//}
+			};
+
+			var dataConverted = DictionaryToObjectConverter.Convert(tmpData, columnToPredict, out Type classType, out Type classPredictionType, out DataViewSchema schema);
+
+			tmpData = null; //Liberamos la memoria
+
 
 			var loadMethod = mlContext.Data.GetType().GetMethods().Where(x => x.Name == "LoadFromEnumerable" && x.IsGenericMethodDefinition).FirstOrDefault();
-
 			var loadMethodObj = loadMethod.MakeGenericMethod(classType);
 			var data = (IDataView)loadMethodObj.Invoke(mlContext.Data, new object[] { dataConverted, null });
 
-
-			//var data = mlContext.Data.LoadFromEnumerable(dataConverted, schema);
-
-			//var data = new DictionaryView<Expando>(tmpData, schema.ToSchema(), converter);
-
 			#region Cortamos los datos de entrenamiento en (Datos para entenar y Datos para hacer el test de precisión)
 
-			DataOperationsCatalog.TrainTestData dataSplit = mlContext.Data.TrainTestSplit(data, testFraction: 0.1);
+			DataOperationsCatalog.TrainTestData dataSplit = mlContext.Data.TrainTestSplit(data, testFraction: 0.5);
 			IDataView trainData = dataSplit.TrainSet;
 			IDataView testData = dataSplit.TestSet;
 
 			#endregion
 
+
 			#region Preparamos los datos de entrada y salida
 
-			//var trainer = mlContext.Regression.Trainers.Sdca(maximumNumberOfIterations: 100);
-			var trainer = mlContext.Regression.Trainers.OnlineGradientDescent(numberOfIterations: 100, learningRate: 0.01f);
 
-			//var pipeline = mlContext.Transforms.CopyColumns(outputColumnName: "Label", inputColumnName: "DESCENDIENTE_S4i001");
+			var columnNameAndTypes = new Dictionary<string, Type>();
-			//IEstimator<ITransformer> pipe = (IEstimator<ITransformer>)pipeline;
+			foreach (var item in (from x in firstRow.Keys
-
+								  select new { Key = x, Type = (from y in dataConverted.Cast<IDictionaryToObjectConverter>() where y.GetValue(x) != null select y.GetValue(x).GetType()).FirstOrDefault() })
-			//pipe = pipe.Append(mlContext.Transforms.Text.NormalizeText("Label"));
+				)
-			//pipe = pipe.Append(mlContext.Transforms.Text.FeaturizeText("Label"));
-			var firstRow = tmpData[0] as IDictionary<string, object>;
-
-			var columnInference = new ColumnInformation()
 			{
-				LabelColumnName = "DESCENDIENTE_S4i001"
+				columnNameAndTypes.Add(item.Key, item.Type);
-			};
-
-			foreach (var key in firstRow.Keys)
-			{
-				if (key == columnInference.LabelColumnName)
-				{
-					continue;
-				}
-
-				if (key.Contains("_S4i") || key.Contains("_SNP"))
-				{
-					columnInference.CategoricalColumnNames.Add(key);
-				}
 			}
 
-			mlContext.Log += (_, e) => {
+			var pipeline = ProcessData(mlContext, columnToPredict, columnNameAndTypes);
-				if (e.Source.Equals("AutoMLExperiment"))
+			var trainingPipeline = BuildAndTrainModel(mlContext, trainData, pipeline, classType, classPredictionType);
-				{
-					Console.WriteLine(e.RawMessage);
-				}
-			};
 
+			Console.WriteLine("Training...");
+			var _trainedModel = trainingPipeline.Fit(trainData);
 
-			SweepablePipeline pipeline = mlContext.Auto().Featurizer(data, columnInference)
+			mlContext.Model.Save(_trainedModel, data.Schema, columnToPredict + ".zip");
-				.Append(mlContext.Auto().Regression(labelColumnName: columnInference.LabelColumnName));
 
-			AutoMLExperiment experiment = mlContext.Auto().CreateExperiment();
+			var createPredictionEngineMethod = mlContext.Model.GetType().GetMethods().Where(x => x.Name == "CreatePredictionEngine" && x.IsGenericMethodDefinition).FirstOrDefault();
+			var createPredictionEngineMethodObj = createPredictionEngineMethod.MakeGenericMethod(classType, classPredictionType);
+			var _predEngine = createPredictionEngineMethodObj.Invoke(mlContext.Model, new object[] { _trainedModel, null, null, null });
 
-			experiment
+			//Test
-				.SetPipeline(pipeline)
+			var predictMethod = _predEngine.GetType().GetMethods().Where(x => x.Name == "Predict" && x.GetParameters().Length == 1 && x.GetParameters()[0].ParameterType == classType).FirstOrDefault();
-				.SetRegressionMetric(RegressionMetric.RSquared, labelColumn: columnInference.LabelColumnName)
-				.SetTrainingTimeInSeconds(10)
-				.SetDataset(trainData);
 
-			var result = experiment.Run();
+			var ok = 0;
-
+			var fail = 0;
-			#endregion
+			foreach (var item in dataConverted.Cast<IDictionaryToObjectConverter>())
-
-			
-			//Entrenamos el modelo
-			//ITransformer model = pipe.Fit(trainData);
-
-			#region Hacemos un test para medir el % de error
-
-			// Use trained model to make inferences on test data
-			IDataView testDataPredictions = result.Model.Transform(testData);
-
-			// Extract model metrics and get RSquared
-			RegressionMetrics trainedModelMetrics = mlContext.Regression.Evaluate(testDataPredictions, labelColumnName: columnInference.LabelColumnName);
-			double rSquared = trainedModelMetrics.RSquared;
-
-			Console.WriteLine("ModelMetrics: {0}", rSquared);
-
-			#endregion
-
-
-			#region Ponemos a prueba haciendo algunas predicciones
-
-			var predictionFunction = mlContext.Model.CreatePredictionEngine<Data, DataPrediction>(result.Model);
-
-			for (var c = 0; c < 25; c++)
 			{
-				var test = CreateRandomData();
+				var expected = item.GetValue(columnToPredict);
-				var expected = test.IntegerNumber;
+				if (expected == null || string.IsNullOrEmpty(expected as string)) { continue; }
-				test.IntegerNumber = 0;
 
-				var p = predictionFunction.Predict(test);
+				item.SetValue(columnToPredict, null);
+
+				var prediction = predictMethod.Invoke(_predEngine, new object[] { item }) as IDictionaryToObjectConverter;
+
+				var predicted = prediction.GetValue(columnToPredict);
+
+				if (expected is string a && predicted is string b)
+				{
+					Console.Write(item.GetValue("DESCENDIENTE") ?? string.Empty);
+					Console.Write(": ");
+
+					Console.Write(string.Format("Expected: {0}\t\tPredicted: {1}", a, b));
+
+					if (string.Equals(a, b))
+					{
+						ok++;
+						Console.WriteLine("\tOk");
+					}
+					else
+					{
+						fail++;
+						Console.WriteLine("\tERROR!!!");
+					}
+
+				}
 
-				Console.WriteLine("Found: {0:#,##0.00}\tExpected: {1:#,##0.00}\t\tDiff: {2:#,##0.00}", p.IntegerNumber, expected, expected - p.IntegerNumber);
 			}
 
+			Console.WriteLine(string.Format("Ok: {0}, Fail: {1}, Percent: {2}%", ok, fail, (((double)ok / (double)(ok + fail)) * 100.0).ToString("##0.0000")));
+
+
 			#endregion
-			
+
+			/*
+						//Entrenamos el modelo
+						//ITransformer model = pipe.Fit(trainData);
+
+						#region Hacemos un test para medir el % de error
+
+						// Use trained model to make inferences on test data
+						IDataView testDataPredictions = result.Model.Transform(testData);
+
+						// Extract model metrics and get RSquared
+						RegressionMetrics trainedModelMetrics = mlContext.Regression.Evaluate(testDataPredictions, labelColumnName: columnInference.LabelColumnName);
+						double rSquared = trainedModelMetrics.RSquared;
+
+						Console.WriteLine("ModelMetrics: {0}", rSquared);
+
+						#endregion
+
+
+						#region Ponemos a prueba haciendo algunas predicciones
+
+						var predictionFunction = mlContext.Model.CreatePredictionEngine<Data, DataPrediction>(result.Model);
+
+						for (var c = 0; c < 25; c++)
+						{
+							var test = CreateRandomData();
+							var expected = test.IntegerNumber;
+							test.IntegerNumber = 0;
+
+							var p = predictionFunction.Predict(test);
+
+							Console.WriteLine("Found: {0:#,##0.00}\tExpected: {1:#,##0.00}\t\tDiff: {2:#,##0.00}", p.IntegerNumber, expected, expected - p.IntegerNumber);
+						}
+
+						#endregion
+			*/
 			Console.WriteLine();
 			Console.WriteLine("Press enter to Exit");
 			Console.ReadLine();
 
 		}
 
+
+		private static IEstimator<ITransformer> ProcessData(MLContext mlContext, string predictColumnName, Dictionary<string, Type> columnNames)
+		{
+			IEstimator<ITransformer> pipeline = mlContext.Transforms.Conversion.MapValueToKey(inputColumnName: predictColumnName, outputColumnName: "Label");
+
+
+			var featured = new List<string>();
+
+			foreach (var key in columnNames.Keys)
+			{
+				if (key == predictColumnName) { continue; }
+				if (key.StartsWith("DESCENDIENTE_")) { continue; }
+
+				var type = columnNames[key];
+				if (type == typeof(string))
+				{
+					pipeline = pipeline.Append(mlContext.Transforms.Text.FeaturizeText(inputColumnName: key, outputColumnName: key));
+					featured.Add(key);
+				}
+			}
+
+			pipeline = pipeline.Append(mlContext.Transforms.Concatenate("Features", featured.ToArray()));
+
+			return pipeline;
+		}
+
+		public static IEstimator<ITransformer> BuildAndTrainModel(MLContext mlContext, IDataView trainingDataView, IEstimator<ITransformer> pipeline, Type modelType, Type prodelPredictionType)
+		{
+			var trainingPipeline = pipeline.Append(mlContext.MulticlassClassification.Trainers.SdcaMaximumEntropy("Label", "Features",  maximumNumberOfIterations: 1000))
+				.Append(mlContext.Transforms.Conversion.MapKeyToValue("PredictedLabel"));
+
+			return trainingPipeline;
+		}
+
+
 		private static string FixColumnName(string columnName)
 		{
 			var result = new StringBuilder(columnName.Length);