Bet.Extensions.ML.Sentiment

Bet.Extensions.ML.Sentiment

The second letter in the Hebrew alphabet is the ב bet/beit. Its meaning is "house". In the ancient pictographic Hebrew it was a symbol resembling a tent on a landscape.

Note: Pre-release packages are distributed via feedz.io.

Summary

This goal of this repo is to create ML.NET self contained ML.NET nuget package that can be used with production DotNetCore applications for Sentiment prediction.

Give a Star! ⭐

If you like or are using this project to learn or start your solution, please give it a star. Thanks!

Install

    dotnet add package Bet.Extensions.ML.Sentiment

Usage

For complete examples please refer to sample projects:

1. Bet.AspNetCore.Sample - AspNetCore Web App with spam prediction models.
2. Bet.Hosting.Sample - DotNetCore Worker based scheduled job for generating ML.NET Models.

    // 1. register
    services.AddSentimentModelBuilder();

    services.AddTimedHostedService<ModelBuilderHostedService>(options =>
    {
        options.Interval = TimeSpan.FromMinutes(30);
        options.FailMode = FailMode.LogAndRetry;
        options.RetryInterval = TimeSpan.FromSeconds(30);
    });

    // 2. build model

    public class ModelBuilderHostedService : TimedHostedService
    {
        private readonly IEnumerable<IModelBuilderService> _modelBuilders;

        public ModelBuilderHostedService(
            IEnumerable<IModelBuilderService> modelBuilders,
            IOptionsMonitor<TimedHostedServiceOptions> options,
            IEnumerable<ITimedHostedLifeCycleHook> lifeCycleHooks,
            ILogger<ITimedHostedService> logger) : base(options, lifeCycleHooks, logger)
        {
            TaskToExecuteAsync = (token) => RunModelGenertorsAsync(token);
            _modelBuilders = modelBuilders ?? throw new ArgumentNullException(nameof(modelBuilders));
        }

        public async Task RunModelGenertorsAsync(CancellationToken cancellationToken)
        {
            foreach (var modelBuilder in _modelBuilders)
            {
                try
                {
                    await modelBuilder.TrainModelAsync(cancellationToken);

                    await modelBuilder.ClassifyTestAsync(cancellationToken);

                    await modelBuilder.SaveModelAsync(cancellationToken);
                }
                catch (Exception ex)
                {
                    Logger.LogError("{modelBuilder} failed with exception: {message}", modelBuilder.GetType(), ex.Message);
                }
            }
        }
    }

    // 3. predict
    [Route("api/[controller]")]
    [ApiController]
    public class PredictionController : ControllerBase
    {
        private readonly IModelPredictionEngine<SentimentIssue, SentimentPrediction> _sentimentModel;

        public PredictionController(
            IModelPredictionEngine<SentimentIssue, SentimentPrediction> sentimentModel)
        {
            _sentimentModel = sentimentModel ?? throw new ArgumentNullException(nameof(sentimentModel));
        }

        [HttpPost()]
        public ActionResult<SentimentPrediction> GetSentiment(SentimentIssue input)
        {
            return _sentimentModel.Predict(input);
        }
    }