public static DataFrame FastPredictOrTransform(ScikitPipeline pipe, DataFrame df, int conc = 1)
{
DataFrame res = null;
pipe.Predict(df, ref res, conc);
return(res);
}
private void ScikitAPI_SimpleTransform_Load(bool removeFirstTransform)
{
var methodName = System.Reflection.MethodBase.GetCurrentMethod().Name;
var output = FileHelper.GetOutputFile($"model{removeFirstTransform}.zip", methodName);
var inputs = new[] {
new ExampleA()
{
X = new float[] { 1, 10, 100 }
},
new ExampleA()
{
X = new float[] { 2, 3, 5 }
}
};
var inputs2 = new[] {
new ExampleA()
{
X = new float[] { -1, -10, -100 }
},
new ExampleA()
{
X = new float[] { -2, -3, -5 }
}
};
string expected = null;
using (var host = EnvHelper.NewTestEnvironment(conc: 1))
{
var data = host.CreateStreamingDataView(inputs);
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, host: host))
{
var predictor = pipe.Train(data);
Assert.IsTrue(predictor != null);
var data2 = host.CreateStreamingDataView(inputs2);
var predictions = pipe.Transform(data2);
var df = DataFrameIO.ReadView(predictions);
Assert.AreEqual(df.Shape, new Tuple <int, int>(2, 9));
var dfs = df.ToString();
var dfs2 = dfs.Replace("\n", ";");
expected = dfs2;
Assert.AreEqual(dfs2, "X.0,X.1,X.2,X.3,X.4,X.5,X.6,X.7,X.8;-1,-10,-100,1,10,100,100,1000,10000;-2,-3,-5,4,6,10,9,15,25");
pipe.Save(output, removeFirstTransform);
}
}
using (var host = EnvHelper.NewTestEnvironment(conc: 1))
{
var data2 = host.CreateStreamingDataView(inputs2);
using (var pipe2 = new ScikitPipeline(output, host))
{
var predictions = pipe2.Transform(data2);
var df = DataFrameIO.ReadView(predictions);
Assert.AreEqual(df.Shape, new Tuple <int, int>(2, 9));
var dfs = df.ToString();
var dfs2 = dfs.Replace("\n", ";");
Assert.AreEqual(expected, dfs2);
}
}
}
public void TestTagTrainOrScoreTransformCustomScorer()
{
var methodName = System.Reflection.MethodBase.GetCurrentMethod().Name;
var dataFilePath = FileHelper.GetTestFile("mc_iris.txt");
var outModelFilePath = FileHelper.GetOutputFile("outModelFilePath.zip", methodName);
var outData = FileHelper.GetOutputFile("outData1.txt", methodName);
using (var env = EnvHelper.NewTestEnvironment())
{
var loader = env.CreateLoader("Text{col=Label:R4:0 col=Slength:R4:1 col=Swidth:R4:2 col=Plength:R4:3 col=Pwidth:R4:4 header=-}",
new MultiFileSource(dataFilePath));
using (var pipe = new ScikitPipeline(new[] {
"Concat{col=Feature:Slength,Swidth}",
"TagTrainScore{tr=iova{p=ft{nl=10 iter=1}} lab=Label feat=Feature tag=model scorer=MultiClassClassifierScorer{ex=AA}}"
}, host: env))
{
pipe.Train(loader);
var pred = pipe.Predict(loader);
var df = DataFrameIO.ReadView(pred);
Assert.AreEqual(df.Shape, new Tuple <int, int>(150, 11));
var dfs = df.Head().ToString();
Assert.IsTrue(dfs.StartsWith("Label,Slength,Swidth,Plength,Pwidth,Feature.0,Feature.1,PredictedLabelAA,ScoreAA.0,ScoreAA.1,ScoreAA.2"));
}
}
}
public void TestScikitAPI_DelegateEnvironment()
{
var inputs = new[] {
new ExampleA()
{
X = new float[] { 1, 10, 100 }
},
new ExampleA()
{
X = new float[] { 2, 3, 5 }
}
};
var inputs2 = new[] {
new ExampleA()
{
X = new float[] { -1, -10, -100 }
},
new ExampleA()
{
X = new float[] { -2, -3, -5 }
}
};
var stdout = new List <string>();
var stderr = new List <string>();
ILogWriter logout = new LogWriter((string s) =>
{
stdout.Add(s);
});
ILogWriter logerr = new LogWriter((string s) =>
{
stderr.Add(s);
});
using (var host = new DelegateEnvironment(conc: 1, outWriter: logout, errWriter: logerr, verbose: 3))
using (var ch = host.Start("Train Pipeline"))
{
ComponentHelper.AddStandardComponents(host);
ch.Info(MessageSensitivity.All, "Polynomial");
var data = host.CreateStreamingDataView(inputs);
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, host: host))
{
var predictor = pipe.Train(data);
if (predictor == null)
{
throw new Exception("Predictor is null");
}
}
}
if (stdout.Count == 0)
{
throw new Exception("stdout is empty.");
}
if (stderr.Count != 0)
{
throw new Exception($"stderr not empty\n{string.Join("\n", stderr)}");
}
}
public void TestScikitAPI_SimplePredictor()
{
var inputs = new[] {
new ExampleA()
{
X = new float[] { 1, 10, 100 }
},
new ExampleA()
{
X = new float[] { 2, 3, 5 }
},
new ExampleA()
{
X = new float[] { 2, 4, 5 }
},
new ExampleA()
{
X = new float[] { 2, 4, 7 }
},
};
var inputs2 = new[] {
new ExampleA()
{
X = new float[] { -1, -10, -100 }
},
new ExampleA()
{
X = new float[] { -2, -3, -5 }
},
new ExampleA()
{
X = new float[] { 3, 4, 5 }
},
new ExampleA()
{
X = new float[] { 3, 4, 7 }
},
};
/*using (*/
var host = EnvHelper.NewTestEnvironment(conc: 1);
{
var data = DataViewConstructionUtils.CreateFromEnumerable(host, inputs);
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, "km{k=2}", host))
{
var predictor = pipe.Train(data, feature: "X");
Assert.IsTrue(predictor != null);
var data2 = new StreamingDataFrame(DataViewConstructionUtils.CreateFromEnumerable(host, inputs2));
var predictions = pipe.Predict(data2);
var df = DataFrameIO.ReadView(predictions);
Assert.AreEqual(df.Shape, new Tuple <int, int>(4, 12));
var dfs = df.ToString();
var dfs2 = dfs.Replace("\n", ";");
Assert.IsTrue(dfs2.StartsWith("X.0,X.1,X.2,X.3,X.4,X.5,X.6,X.7,X.8,PredictedLabel,Score.0,Score.1;-1,-10,-100,1,10,100,100,1000,10000"));
}
}
}
public void Train()
{
using (var env = new ConsoleEnvironment())
{
var df = DataFrameIO.ReadCsv(_dataset, sep: ',',
dtypes: new ColumnType[] { NumberType.R4 });
var concat = "Concat{col=Features:F0,F1,F2,F3,F4,F5,F6,F7,F8,F9}";
var pipe = new ScikitPipeline(new[] { concat }, "ftr{iter=10}");
pipe.Train(df, "Features", "Label");
_pipeline = pipe;
}
}
public void TestScikitAPI_TrainingDiabete()
{
var diab = FileHelper.GetTestFile("diabete.csv");
var cols = Enumerable.Range(0, 10).Select(c => NumberType.R4).ToArray();
var colsName = string.Join(',', Enumerable.Range(0, 10).Select(c => $"F{c}"));
var df = DataFrameIO.ReadCsv(diab, sep: ',', dtypes: cols);
var pipe = new ScikitPipeline(new string[] { $"Concat{{col=Features:{colsName}}}" }, "ols");
pipe.Train(df, "Features", "Label");
DataFrame pred = null;
pipe.Predict(df, ref pred);
Assert.AreEqual(pred.Shape, new ShapeType(83, 13));
}
public void TestScikitAPI_TrainingWithIris()
{
var iris = FileHelper.GetTestFile("iris.txt");
var df = DataFrameIO.ReadCsv(iris, sep: '\t');
df.AddColumn("LabelI", df["Label"].AsType(NumberType.R4));
var pipe = new ScikitPipeline(new[] { $"Concat{{col=Features:{df.Columns[1]},{df.Columns[2]}}}" }, "mlr");
pipe.Train(df, "Features", "LabelI");
DataFrame pred = null;
pipe.Predict(df, ref pred);
Assert.AreEqual(pred.Shape, new ShapeType(150, 9));
}
public void TestLoadModelFromNimbusML()
{
var iris = FileHelper.GetTestFile("model_iris.zip");
using (var env = EnvHelper.NewTestEnvironment())
{
try
{
using (var pipe2 = new ScikitPipeline(iris, env))
{
}
}
catch (Exception e)
{
Assert.IsTrue(e.ToString().Contains("because the model is too old"));
}
}
}
public void TestScikitAPI_DelegateEnvironmentVerbose0()
{
var inputs = new[] {
new ExampleA()
{
X = new float[] { 1, 10, 100 }
},
new ExampleA()
{
X = new float[] { 2, 3, 5 }
}
};
var inputs2 = new[] {
new ExampleA()
{
X = new float[] { -1, -10, -100 }
},
new ExampleA()
{
X = new float[] { -2, -3, -5 }
}
};
var stdout = new List <string>();
var stderr = new List <string>();
ILogWriter logout = new LogWriter(s => stdout.Add(s));
ILogWriter logerr = new LogWriter(s => stderr.Add(s));
/*using (*/
var host = new DelegateEnvironment(seed: 0, outWriter: logout, errWriter: logerr, verbose: 0);
{
ComponentHelper.AddStandardComponents(host);
var data = DataViewConstructionUtils.CreateFromEnumerable(host, inputs);
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, "km{k=2}", host: host))
{
var predictor = pipe.Train(data, feature: "X");
Assert.IsTrue(predictor != null);
}
}
Assert.AreEqual(stdout.Count, 0);
Assert.AreEqual(stderr.Count, 0);
}
public void TestScikitAPI_MKL_TrainingDiabete()
{
try
{
var diab = FileHelper.GetTestFile("diabete.csv");
var cols = Enumerable.Range(0, 10).Select(c => NumberDataViewType.Single).ToArray();
var colsName = string.Join(',', Enumerable.Range(0, 10).Select(c => $"F{c}"));
var df = DataFrameIO.ReadCsv(diab, sep: ',', dtypes: cols);
var pipe = new ScikitPipeline(new string[] { $"Concat{{col=Features:{colsName}}}" }, "ols");
pipe.Train(df, "Features", "Label");
DataFrame pred = null;
pipe.Predict(df, ref pred);
Assert.AreEqual(pred.Shape, new ShapeType(83, 13));
}
catch (DllNotFoundException e)
{
var os = Environment.OSVersion;
if (os.Platform == PlatformID.Unix)
{
Console.WriteLine("FAIL(1): TestScikitAPI_MKL due to {0}", e.ToString());
return;
}
else
{
Console.WriteLine("FAIL(1): TestScikitAPI_MKL, OS={0}", os.ToString());
throw e;
}
}
catch (NotSupportedException e)
{
var os = Environment.OSVersion;
if (os.Platform == PlatformID.Unix)
{
Console.WriteLine("FAIL(2): TestScikitAPI_MKL due to {0}", e.ToString());
return;
}
else
{
Console.WriteLine("FAIL(2): TestScikitAPI_MKL, OS={0}", os.ToString());
throw e;
}
}
}
public void TestScikitAPI_SimpleTransform()
{
var inputs = new[] {
new ExampleA()
{
X = new float[] { 1, 10, 100 }
},
new ExampleA()
{
X = new float[] { 2, 3, 5 }
}
};
var inputs2 = new[] {
new ExampleA()
{
X = new float[] { -1, -10, -100 }
},
new ExampleA()
{
X = new float[] { -2, -3, -5 }
}
};
/*using (*/
var host = EnvHelper.NewTestEnvironment(conc: 1);
{
var data = DataViewConstructionUtils.CreateFromEnumerable(host, inputs);
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, host: host))
{
var predictor = pipe.Train(data);
Assert.IsTrue(predictor != null);
var data2 = DataViewConstructionUtils.CreateFromEnumerable(host, inputs2);
var predictions = pipe.Transform(data2);
var df = DataFrameIO.ReadView(predictions);
Assert.AreEqual(df.Shape, new Tuple <int, int>(2, 9));
var dfs = df.ToString();
var dfs2 = dfs.Replace("\n", ";");
Assert.AreEqual(dfs2, "X.0,X.1,X.2,X.3,X.4,X.5,X.6,X.7,X.8;-1,-10,-100,1,10,100,100,1000,10000;-2,-3,-5,4,6,10,9,15,25");
}
}
}
public void TestTreePathInnerAPI()
{
using (var env = EnvHelper.NewTestEnvironment(conc: 1))
{
var methodName = System.Reflection.MethodBase.GetCurrentMethod().Name;
var iris = FileHelper.GetTestFile("iris.txt");
var df = DataFrameIO.ReadCsv(iris, sep: '\t', dtypes: new ColumnType[] { NumberType.R4 });
using (var pipe = new ScikitPipeline(new[] { "Concat{col=Feature:Sepal_length,Sepal_width}",
"TreeFeat{tr=ft{iter=2} lab=Label feat=Feature}" }))
{
pipe.Train(df);
var scorer = pipe.Predict(df);
var dfout = DataFrameIO.ReadView(scorer);
Assert.AreEqual(dfout.Shape, new Tuple <int, int>(150, 31));
var outfile = FileHelper.GetOutputFile("iris_path.txt", methodName);
dfout.ToCsv(outfile);
Assert.IsTrue(File.Exists(outfile));
}
}
}
public void TestOnnx_TrainingWithIris()
{
var methodName = System.Reflection.MethodBase.GetCurrentMethod().Name;
// direct call
var iris = FileHelper.GetTestFile("iris.txt");
var df = DataFrameIO.ReadCsv(iris, sep: '\t');
df.AddColumn("LabelI", df["Label"].AsType(NumberDataViewType.Single));
var pipe = new ScikitPipeline(new[] { $"Concat{{col=Features:{df.Columns[1]},{df.Columns[2]}}}" }, "mlr");
pipe.Train(df, "Features", "LabelI");
DataFrame pred = null;
pipe.Predict(df, ref pred);
// Onnx Save
var output = FileHelper.GetOutputFile("model.onnx", methodName);
var model = pipe.ToOnnx();
model.Save(output);
Assert.IsTrue(File.Exists(output));
// Onnx save no concat.
output = FileHelper.GetOutputFile("model_vector.onnx", methodName);
model = pipe.ToOnnx(1);
model.Save(output);
Assert.IsTrue(File.Exists(output));
// Onnx Load Not implemented yet.
/*
* var restored = new ScikitPipeline(output);
* DataFrame pred2 = null;
* restored.Predict(df, ref pred2);
* pred.AssertAlmostEqual(pred2);
*/
}
public void TestScikitAPI_SimplePredictor_FastValueMapper()
{
var inputs = new[] {
new ExampleA()
{
X = new float[] { 1, 10, 100 }
},
new ExampleA()
{
X = new float[] { 2, 3, 5 }
},
new ExampleA()
{
X = new float[] { 2, 4, 5 }
},
new ExampleA()
{
X = new float[] { 2, 4, 7 }
},
};
var inputs2 = new[] {
new ExampleA()
{
X = new float[] { -1, -10, -100 }
},
new ExampleA()
{
X = new float[] { -2, -3, -5 }
},
new ExampleA()
{
X = new float[] { 3, 4, 5 }
},
new ExampleA()
{
X = new float[] { 3, 4, 7 }
},
};
DataFrame df1, df2, df3;
using (var host = EnvHelper.NewTestEnvironment(conc: 1))
{
var data = host.CreateStreamingDataView(inputs);
var data2 = host.CreateStreamingDataView(inputs2);
df1 = DataFrameIO.ReadView(data, env: host, keepVectors: true);
df2 = DataFrameIO.ReadView(data2, env: host, keepVectors: true);
df3 = DataFrameIO.ReadView(data2, env: host, keepVectors: true);
}
using (var host = EnvHelper.NewTestEnvironment(conc: 1))
{
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, "km{k=2}", host))
{
DataFrame pred = null, pred2 = null;
var predictor = pipe.Train(df1, feature: "X");
Assert.IsTrue(predictor != null);
pipe.Predict(df2, ref pred);
Assert.AreEqual(pred.Shape, new Tuple <int, int>(4, 3));
var dfs = pred.ToString();
var dfs2 = dfs.Replace("\n", ";");
if (!dfs2.StartsWith("X.0,X.1,X.2,X.3,X.4,X.5,X.6,X.7,X.8,PredictedLabel,Score.0,Score.1;-1,-10,-100,1,10,100,100,1000,10000"))
{
throw new Exception($"Wrong starts\n{dfs2}");
}
pipe.Predict(df3, ref pred2);
pred.AssertAlmostEqual(pred2);
}
}
}
public void Read(string name)
{
_pipeline = new ScikitPipeline(name);
}
private void ScikitAPI_SimplePredictor_Load(bool removeFirstTransform)
{
var methodName = System.Reflection.MethodBase.GetCurrentMethod().Name;
var output = FileHelper.GetOutputFile($"model{removeFirstTransform}.zip", methodName);
var inputs = new[] {
new ExampleA()
{
X = new float[] { 1, 10, 100 }
},
new ExampleA()
{
X = new float[] { 2, 3, 5 }
},
new ExampleA()
{
X = new float[] { 2, 4, 5 }
},
new ExampleA()
{
X = new float[] { 2, 4, 7 }
},
};
var inputs2 = new[] {
new ExampleA()
{
X = new float[] { -1, -10, -100 }
},
new ExampleA()
{
X = new float[] { -2, -3, -5 }
},
new ExampleA()
{
X = new float[] { 3, 4, 5 }
},
new ExampleA()
{
X = new float[] { 3, 4, 7 }
},
};
string expected = null;
using (var host = EnvHelper.NewTestEnvironment(conc: 1))
{
var data = host.CreateStreamingDataView(inputs);
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, "km{k=2}", host))
{
var predictor = pipe.Train(data, feature: "X");
Assert.IsTrue(predictor != null);
var data2 = host.CreateStreamingDataView(inputs2);
var predictions = pipe.Predict(data2);
var df = DataFrameIO.ReadView(predictions);
Assert.AreEqual(df.Shape, new Tuple <int, int>(4, 12));
var dfs = df.ToString();
var dfs2 = dfs.Replace("\n", ";");
if (!dfs2.StartsWith("X.0,X.1,X.2,X.3,X.4,X.5,X.6,X.7,X.8,PredictedLabel,Score.0,Score.1;-1,-10,-100,1,10,100,100,1000,10000"))
{
throw new Exception($"Wrong starts\n{dfs2}");
}
expected = dfs2;
pipe.Save(output, removeFirstTransform);
}
}
using (var host = EnvHelper.NewTestEnvironment(conc: 1))
{
var data2 = host.CreateStreamingDataView(inputs2);
using (var pipe2 = new ScikitPipeline(output, host))
{
var predictions = pipe2.Predict(data2);
var df = DataFrameIO.ReadView(predictions);
Assert.AreEqual(df.Shape, new Tuple <int, int>(4, 12));
var dfs = df.ToString();
var dfs2 = dfs.Replace("\n", ";");
Assert.AreEqual(expected, dfs2);
}
}
}
public void TestTimeSeriesFloatRegression()
{
var inputs = new[] {
new InputOutput()
{
X = 5f, time = 0f, one = 1f
},
new InputOutput()
{
X = 7f, time = 1f, one = 1f
},
new InputOutput()
{
X = 9f, time = 2f, one = 1f
},
new InputOutput()
{
X = 11f, time = 3f, one = 1f
},
new InputOutput()
{
X = 5f, time = 0f, one = 1f
},
new InputOutput()
{
X = 7f, time = 1f, one = 1f
},
new InputOutput()
{
X = 9f, time = 2f, one = 1f
},
new InputOutput()
{
X = 11f, time = 3f, one = 1f
},
new InputOutput()
{
X = 5f, time = 0f, one = 1f
},
new InputOutput()
{
X = 7f, time = 1f, one = 1f
},
new InputOutput()
{
X = 9f, time = 2f, one = 1f
},
new InputOutput()
{
X = 11f, time = 3f, one = 1f
},
};
using (var host = EnvHelper.NewTestEnvironment())
{
var data = host.CreateStreamingDataView(inputs);
using (var pipe = new ScikitPipeline(new[] { "concat{col=xt:time,one}" }, "sasdcar{iter=50}", host))
{
pipe.Train(data, feature: "xt", label: "X");
var view = pipe.Predict(data);
var df = DataFrameIO.ReadView(view).Head(4).Copy();
df["diff"] = df["Score"] - df["X"];
var exp = DataFrameIO.ReadStr("null\n0\n0\n0\n0");
df["diff"].AssertAlmostEqual(exp["null"].AsType(NumberType.R4), precision: 1e-1);
}
}
}
/// <summary>
/// Runs a simple test.
/// </summary>
public static void TestScikitAPI()
{
var inputs = new[] {
new ExampleVector()
{
X = new float[] { 1, 10, 100 }
},
new ExampleVector()
{
X = new float[] { 2, 3, 5 }
},
new ExampleVector()
{
X = new float[] { 2, 4, 5 }
},
new ExampleVector()
{
X = new float[] { 2, 4, 7 }
},
};
var inputs2 = new[] {
new ExampleVector()
{
X = new float[] { -1, -10, -100 }
},
new ExampleVector()
{
X = new float[] { -2, -3, -5 }
},
new ExampleVector()
{
X = new float[] { 3, 4, 5 }
},
new ExampleVector()
{
X = new float[] { 3, 4, 7 }
},
};
using (var host = new ConsoleEnvironment(conc: 1))
{
ComponentHelper.AddStandardComponents(host);
var data = host.CreateStreamingDataView(inputs);
using (var pipe = new ScikitPipeline(new[] { "poly{col=X}" }, "km{k=2}", host))
{
var predictor = pipe.Train(data, feature: "X");
if (predictor == null)
{
throw new Exception("Test failed: no predictor.");
}
var data2 = host.CreateStreamingDataView(inputs2);
var predictions = pipe.Predict(data2);
var df = DataFrameIO.ReadView(predictions);
if (df.Shape.Item1 != 4 || df.Shape.Item2 != 12)
{
throw new Exception("Test failed: prediction failed.");
}
var dfs = df.ToString();
var dfs2 = dfs.Replace("\n", ";");
if (!dfs2.StartsWith("X.0,X.1,X.2,X.3,X.4,X.5,X.6,X.7,X.8,PredictedLabel,Score.0,Score.1;-1,-10,-100,1,10,100,100,1000,10000"))
{
throw new Exception("Test failed: prediction failed (header).");
}
}
}
}
