{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "#r \"nuget:Microsoft.ML,1.5.1\"\n",
    "#r \"nuget:Microsoft.ML.FastTree,0.17.1\""
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "using System;\n",
    "using System.Collections.Generic;\n",
    "using System.IO;\n",
    "using System.Linq;\n",
    "using Microsoft.ML;\n",
    "using Microsoft.ML.Data;"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "class ModelInput\n",
    "{\n",
    "    [ColumnName(\"Need\"), LoadColumn(0)]\n",
    "    public string Need { get; set; }\n",
    "\n",
    "\n",
    "    [ColumnName(\"Percentage\"), LoadColumn(1)]\n",
    "    public float Percentage { get; set; }\n",
    "\n",
    "\n",
    "    [ColumnName(\"Priority\"), LoadColumn(2)]\n",
    "    public string Priority { get; set; }\n",
    "}\n",
    "\n",
    "public class ModelOutput\n",
    "{\n",
    "    // ColumnName attribute is used to change the column name from\n",
    "    // its default value, which is the name of the field.\n",
    "    [ColumnName(\"PredictedLabel\")]\n",
    "    public String Prediction { get; set; }\n",
    "    public float[] Score { get; set; }\n",
    "}"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "IEstimator<ITransformer> BuildTrainingPipeline(MLContext context)\n",
    "{\n",
    "    // Data process configuration with pipeline data transformations \n",
    "    var dataProcessPipeline = context.Transforms.Conversion.MapValueToKey(\"Priority\", \"Priority\")\n",
    "                                .Append(context.Transforms.Categorical.OneHotEncoding(new[] { new InputOutputColumnPair(\"Need\", \"Need\") }))\n",
    "                                .Append(context.Transforms.Concatenate(\"Features\", new[] { \"Need\", \"Percentage\" }))\n",
    "                                .AppendCacheCheckpoint(context);\n",
    "    // Set the training algorithm \n",
    "    var trainer = context.MulticlassClassification.Trainers.OneVersusAll(context.BinaryClassification.Trainers.FastTree(labelColumnName: \"Priority\", featureColumnName: \"Features\"), labelColumnName: \"Priority\")\n",
    "                                .Append(context.Transforms.Conversion.MapKeyToValue(\"PredictedLabel\",  \"PredictedLabel\"));\n",
    "\n",
    "    var trainingPipeline = dataProcessPipeline.Append(trainer);\n",
    "\n",
    "    return trainingPipeline;\n",
    "}"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "ITransformer TrainModel(MLContext context, IDataView trainingDataView, IEstimator<ITransformer> trainingPipeline)\n",
    "{\n",
    "    Console.WriteLine(\"=============== Training  model ===============\");\n",
    "\n",
    "    var model = trainingPipeline.Fit(trainingDataView);\n",
    "\n",
    "    Console.WriteLine(\"=============== End of training process ===============\");\n",
    "    return model;\n",
    "}"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "void SaveModel(MLContext mlContext, ITransformer mlModel, string modelRelativePath, DataViewSchema modelInputSchema)\n",
    "{\n",
    "    var modelPath = Path.Combine(\"models\", modelRelativePath);\n",
    "    var path = Path.GetFullPath(modelPath);\n",
    "    // Save/persist the trained model to a .ZIP file\n",
    "    Console.WriteLine($\"=============== Saving the model  ===============\");\n",
    "    mlContext.Model.Save(mlModel, modelInputSchema, path);\n",
    "    Console.WriteLine(\"The model is saved to {0}\", path);\n",
    "}"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()\n",
    "{\n",
    "    var modelPath = Path.Combine(\"models\", \"Servo.zip\");\n",
    "    var MLNetModelPath = Path.GetFullPath(modelPath);\n",
    "\n",
    "    // Create new MLContext\n",
    "    MLContext mlContext = new MLContext();\n",
    "\n",
    "    // Load model & create prediction engine\n",
    "    var mlModel = mlContext.Model.Load(MLNetModelPath, out var modelInputSchema);\n",
    "    var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);\n",
    "\n",
    "    return predEngine;\n",
    "}"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "void Predictions(ModelInput[] inputs)\n",
    "{\n",
    "    foreach (var input in inputs)\n",
    "    {\n",
    "        var engine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);\n",
    "        var result = engine.Value.Predict(input);\n",
    "        var priority = false;\n",
    "        \n",
    "        switch (result.Prediction)\n",
    "        {\n",
    "            default:\n",
    "            case \"FALSE\":\n",
    "                priority = false;\n",
    "                break;\n",
    "            case \"TRUE\":\n",
    "                priority = true;\n",
    "                break;\n",
    "        }\n",
    "        \n",
    "        Console.WriteLine($\"Need: {input.Need}{Environment.NewLine}\" + \n",
    "        $\"Percentage: {input.Percentage}{Environment.NewLine}\" + \n",
    "        $\"Predicted Priority value {priority}{Environment.NewLine}\" + \n",
    "        $\"Predicted Priority scores: [{string.Join(\",\", result.Score)}]\");\n",
    "    }\n",
    "}"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "var sampleData = new ModelInput[]\n",
    "{\n",
    "    new ModelInput\n",
    "    {\n",
    "        Need = \"Social\",\n",
    "        Percentage = 27F,\n",
    "    },\n",
    "    new ModelInput\n",
    "    {\n",
    "        Need = \"Bladder\",\n",
    "        Percentage = 76F,\n",
    "    },\n",
    "    new ModelInput\n",
    "    {\n",
    "        Need = \"Energy\",\n",
    "        Percentage = 55F,\n",
    "    },\n",
    "    new ModelInput\n",
    "    {\n",
    "        Need = \"Social\",\n",
    "        Percentage = 20F,\n",
    "    },\n",
    "    new ModelInput\n",
    "    {\n",
    "        Need = \"Fun\",\n",
    "        Percentage = 14F,\n",
    "    }\n",
    "};"
   ],
   "outputs": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "source": [
    "Predictions(sampleData);"
   ],
   "outputs": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".NET (C#)",
   "language": "C#",
   "name": ".net-csharp"
  },
  "language_info": {
   "file_extension": ".cs",
   "mimetype": "text/x-csharp",
   "name": "C#",
   "pygments_lexer": "csharp",
   "version": "8.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}