qlhb/Ropin.IOT.MLService/Controllers/WeatherForecastController.cs

using InfluxData.Net.Common.Enums;
using Microsoft.AspNetCore.Mvc;
using Microsoft.Extensions.Logging;
using Microsoft.ML;
using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using InfluxData.Net.Common.Enums;
using InfluxData.Net.InfluxDb;
using Microsoft.ML.Data;
using Microsoft.ML.Transforms;

namespace Ropin.IOT.MLService.Controllers
{
    [ApiController]
    [Route("[controller]")]
    public class WeatherForecastController : ControllerBase
    {
        private static readonly string[] Summaries = new[]
        {
            "Freezing", "Bracing", "Chilly", "Cool", "Mild", "Warm", "Balmy", "Hot", "Sweltering", "Scorching"
        };

        private readonly ILogger<WeatherForecastController> _logger;

        public WeatherForecastController(ILogger<WeatherForecastController> logger)
        {
            _logger = logger;
        }

        [HttpGet]
        public IEnumerable<WeatherForecast> Get()
        {
            var rng = new Random();
            return Enumerable.Range(1, 5).Select(index => new WeatherForecast
            {
                Date = DateTime.Now.AddDays(index),
                TemperatureC = rng.Next(-20, 55),
                Summary = Summaries[rng.Next(Summaries.Length)]
            })
            .ToArray();
        }
        [HttpPatch]
        public async void PatchAI()
        {
            // Create a new context for ML operations
            var mlContext = new MLContext();

            // Sample data points in-memory collection
            IDataView dataPoints = mlContext.Data.LoadFromEnumerable(new[]
            {
            new DataPoint() { Feature = 1, Label = 2 },
            new DataPoint() { Feature = 2, Label = 4 },
            new DataPoint() { Feature = 3, Label = 6 },
            new DataPoint() { Feature = 4, Label = 8 },
            new DataPoint() { Feature = 5, Label = 10 }
        });

            // Define the pipeline
            var pipeline = mlContext.Transforms.Concatenate("Features", nameof(DataPoint.Feature))
                .Append(mlContext.Regression.Trainers.Sdca());

            // Train the model
            var trainedModel = pipeline.Fit(dataPoints);

            // Create a prediction engine to make predictions on individual data points
            var predEngine = mlContext.Model.CreatePredictionEngine<DataPoint, Prediction>(trainedModel);

            // Make a prediction
            var prediction = predEngine.Predict(new DataPoint() { Feature = 6 });
            Console.WriteLine($"Predicted value: {prediction.PredictedValue}");

        }

        public class DataPoint
        {
            [LoadColumn(0)]
            public float Feature { get; set; }

            [LoadColumn(1)]
            public float Label { get; set; }
        }
        public class DeviceStatusPoint
        {
            [LoadColumn(0)]
            public DateTime TimeStamp { get; set; }

            [LoadColumn(1)]
            public float Temperature { get; set; }

            [LoadColumn(2)]
            public float Pressure { get; set; }

            [LoadColumn(3)]
            public bool IsFaulty { get; set; }


            [LoadColumn(4)]
            public float Label { get; set; }

        }

        public class Prediction
        {
            [ColumnName("Score")]
            public float PredictedValue { get; set; }
        }


        [HttpPost]
        public async void GetAI()
        {
            await GetData();
           
        }
        private async Task GetData()
        {

            //// 查询最近 10 条 CPU 使用率数据
            //    var query1 = "SELECT * FROM fanyidev ORDER BY time DESC LIMIT 10";
            //    var results = await client.QueryMultiSeriesAsync("fanyidb", query1);

            //    if (results != null && results.Count > 0)
            //    {

            //    }
            //     if (results != null && results.Count > 0)
            //    {
            //        // 处理查询结果
            //        foreach (var series in results)
            //        {
            //            Console.WriteLine($"查询结果集: {series.Name}");
            //            foreach (var point in series.Entries)
            //            {
            //                // 根据字段名获取值
            //                var time = point.GetTimeAsDateTime();
            //        var host = point.GetTagAsString("host");
            //        var value = point.GetFieldAsDouble("value");
            //        Console.WriteLine($"时间: {time}, 主机: {host}, CPU 使用率: {value}");
            //            }
            //}
            //    }





            //传入查询命令，支持多条
            var queries = new[]
            {

                " SELECT * FROM fanyidev WHERE time> now() -  1h "
            };

            var dbName = "fanyidb";

            InfluxDbClient influxDbClient = new InfluxDbClient("http://10.126.126.131:8085/", "admin", "123456", InfluxDbVersion.Latest);
            //从指定库中查询数据

            var response = await influxDbClient.Client.QueryAsync(queries, dbName);

            if (response.Any())
            {
                var series = response.ToList();
                foreach (var value in series[0].Values)
                {
                    //var TimeStamp = DateTime.Parse((string)value[0]);
                    var Pressure = Convert.ToDouble(value[2]);
                }
                //var dataPoints = series[0].Values.Select(value =>
                //    new DeviceStatusDataPoint
                //    {
                //        //TimeStamp = DateTime.Parse((string)value[0]),
                //        //Temperature = Convert.ToDouble(value[1]),
                //        Pressure = Convert.ToDouble(value[2])//,
                //        //IsFaulty = (bool)value[3]
                //    }).ToList();

                var dataPoints = new List<DeviceStatusPoint>();
                var v1 = new DeviceStatusPoint
                {
                    TimeStamp = DateTime.Now.AddMilliseconds(-1),
                    Temperature = 70,
                    Pressure = 150,
                    IsFaulty = true,
                    Label = 10
                };
                var v2 = new DeviceStatusPoint
                {
                    TimeStamp = DateTime.Now,
                    Temperature = 60,
                    Pressure = 140,
                    IsFaulty = false,
                    Label = 20
                };
                dataPoints.Add(v1);
                dataPoints.Add(v2);

                // 使用 ML.NET 训练模型
                var mlContext = new MLContext();

                // 将数据加载到 IDataView
                IDataView dataView = mlContext.Data.LoadFromEnumerable(dataPoints);



                //// Define the pipeline
                //var pipeline1 = mlContext.Transforms.Concatenate("Features", nameof(DeviceStatusPoint.Temperature), nameof(DeviceStatusPoint.Pressure))
                //    .Append(mlContext.Regression.Trainers.Sdca());

                //// Train the model
                //var trainedModel = pipeline1.Fit(dataView);

                //// Create a prediction engine to make predictions on individual data points
                //var predEngine = mlContext.Model.CreatePredictionEngine<DeviceStatusPoint, FaultPrediction>(trainedModel);

                //// Make a prediction
                //var prediction1 = predEngine.Predict(new DeviceStatusPoint { Temperature = 75, Pressure = 150 });
                //Console.WriteLine($"Predicted value: {prediction1.PredictedLabel}");



                // 分割数据集为训练集和测试集
                var trainTestSplit = mlContext.Data.TrainTestSplit(dataView, testFraction: 0.2);
                var trainingData = trainTestSplit.TrainSet;
                var testData = trainTestSplit.TestSet;

                // 定义管道
                var pipeline = mlContext.Transforms.Concatenate("Features", nameof(DeviceStatusPoint.Temperature), nameof(DeviceStatusPoint.Pressure))
                    .Append(mlContext.BinaryClassification.Trainers.SdcaLogisticRegression(labelColumnName: nameof(DeviceStatusPoint.IsFaulty)));

                var progressHandler = new TrainingProgressHandler();

                //Console.WriteLine("Starting model training...");
                //var cvResults = mlContext.MulticlassClassification.CrossValidate(trainingData, pipeline, numberOfFolds: 5, progressHandler: progressHandler);

                //Console.WriteLine("Training completed.");

                // 训练模型
                var model = pipeline.Fit(trainingData);

                // 创建预测引擎
                var predictionEngine = mlContext.Model.CreatePredictionEngine<DeviceStatusPoint, FaultPrediction>(model);

                // 预测新样本
                var sampleData = new DeviceStatusPoint { Temperature = 75, Pressure = 150 };
                var prediction = predictionEngine.Predict(sampleData);
                Console.WriteLine($"Predicted IsFaulty: {prediction.PredictedLabel}, Probability: {prediction.Probability}");

                // 评估模型
                var predictions = model.Transform(testData);
                var metrics = mlContext.BinaryClassification.Evaluate(predictions);
                Console.WriteLine($"Accuracy: {metrics.Accuracy}");
            }
            else
            {
                Console.WriteLine("Failed to retrieve data from InfluxDB.");
            }


        }
    }

    public class TrainingProgressHandler : IProgress<TrainCatalogBase.CrossValidationResult<MulticlassClassificationMetrics>>
    {
        public void Report(TrainCatalogBase.CrossValidationResult<MulticlassClassificationMetrics> value)
        {
            //Console.WriteLine(value: $"Fold: {value.Model.Properties.TrainingTime}");
        }
    }
}