/// <summary>
/// Tests the prediction using the current algorithm on the supplied data.
/// </summary>
/// <param name="scenarioId">
/// The scenario id.
/// </param>
/// <returns>
/// The predictions for the specified feature ID.
/// </returns>
public TestPredictions Test(string scenarioId, string featureId, Contents contents)
{
using (var dbContext = new OpenAIEntities1())
{
var scenario = dbContext.Scenarios.Find(scenarioId);
if (scenario == null)
{
return(null);
}
ScenarioTrainings trainings;
if (!trainingByScenario.TryGetValue(scenarioId, out trainings))
{
trainings = this.Train(scenarioId);
}
var table = this.ParseCsv(contents.Data);
TrainerHelper training;
if (!trainings.TrainingByFeatureId.TryGetValue(featureId, out training))
{
return(null);
}
var predictions = trainer.Decide(training, table, featureId);
return(predictions);
}
}
/// <summary>
/// The train.
/// </summary>
/// <param name="scenarioId">
/// The scenario id.
/// </param>
/// <returns>
/// The <see cref="ScenarioTrainings"/>.
/// </returns>
public ScenarioTrainings Train(string scenarioId)
{
using (var dbContext = new OpenAIEntities1())
{
Scenario scenario = dbContext.Scenarios.Find(scenarioId);
if (scenario == null)
{
return(null);
}
DataTable table = this.ParseCsv(scenario.Contents);
// Get hypothesis for each feature / output
var trainings = new ScenarioTrainings
{
ScenarioId = scenarioId,
TrainingByFeatureId = new Dictionary <string, TrainerHelper>()
};
trainingByScenario[scenarioId] = trainings;
foreach (DataColumn column in table.Columns)
{
TrainerHelper thing = trainer.Train(table, column.ColumnName);
trainings.TrainingByFeatureId[column.ColumnName] = thing;
}
return(trainings);
}
}
/// <summary>
/// The create scenario.
/// </summary>
/// <param name="scenario">
/// The scenario.
/// </param>
public void CreateScenario(Models.Scenario scenario)
{
using (var dbContext = new OpenAIEntities1())
{
Scenario entity = dbContext.Scenarios.Find(scenario.Id);
if (entity == null)
{
entity = new Scenario();
entity.ScenarioId = scenario.Id;
dbContext.Scenarios.Add(entity);
}
entity.Name = scenario.Name;
entity.Contents = scenario.Contents;
// Add the features
DataTable table = this.ParseCsv(scenario.Contents);
int position = 1;
foreach (DataColumn column in table.Columns)
{
if (entity.Features.Any(f => f.FeatureId == column.ColumnName))
{
continue;
}
var feature = new Feature();
feature.FeatureId = column.ColumnName;
feature.ScenarioId = scenario.Id;
feature.Position = position++;
entity.Features.Add(feature);
}
dbContext.SaveChanges();
}
}
/// <summary>
/// The create inputs.
/// </summary>
/// <param name="scenarioId">
/// The scenario id.
/// </param>
/// <param name="inputSetId">
/// The input set id.
/// </param>
/// <param name="inputs">
/// The inputs.
/// </param>
public PredictionSet CreateInputs(string scenarioId, string inputSetId, List <Models.Input> inputs)
{
if (inputs == null)
{
return(new PredictionSet());
}
using (var dbContext = new OpenAIEntities1())
{
InputSet inputSet =
dbContext.InputSets.FirstOrDefault(i => i.ScenarioId == scenarioId && i.InputSetId == inputSetId);
if (inputSet == null)
{
inputSet = new InputSet();
inputSet.ScenarioId = scenarioId;
inputSet.InputSetId = inputSetId;
dbContext.InputSets.Add(inputSet);
}
Dictionary <string, Input> existing =
dbContext.Inputs.Where(i => i.ScenarioId == scenarioId && i.InputSetId == inputSetId)
.ToDictionary(i => i.FeatureId);
foreach (Models.Input input in inputs)
{
Input entity;
existing.TryGetValue(input.InputId, out entity);
if (entity == null)
{
entity = new Input();
dbContext.Inputs.Add(entity);
entity.ScenarioId = scenarioId;
entity.InputSetId = inputSetId;
entity.FeatureId = input.InputId;
}
entity.Value = input.Value;
}
dbContext.SaveChanges();
}
return(this.Predict(scenarioId, inputSetId));
}
public void CompleteInputSet(string scenarioId, string inputSetId)
{
using (var dbContext = new OpenAIEntities1())
{
Scenario entity = dbContext.Scenarios.Find(scenarioId);
if (entity == null)
{
throw new Exception(string.Format("Unable to find scenario for {0}", scenarioId));
}
List <Feature> features =
dbContext.Features.Where(f => f.ScenarioId == scenarioId).OrderBy(f => f.Position).ToList();
Dictionary <string, Input> inputsById =
dbContext.Inputs.Where(i => i.ScenarioId == scenarioId && i.InputSetId == inputSetId)
.ToDictionary(i => i.FeatureId);
var sortedInputs = new List <Input>();
foreach (Feature feature in features)
{
Input input;
if (inputsById.TryGetValue(feature.FeatureId, out input))
{
sortedInputs.Add(input);
}
else
{
sortedInputs.Add(new Input {
FeatureId = feature.FeatureId
});
}
}
var newData = string.Join(",", sortedInputs.Select(i => i.Value ?? "0"));
entity.Contents = entity.Contents + Environment.NewLine + newData;
dbContext.SaveChanges();
this.Train(scenarioId);
}
}
/// <summary>
/// The predict.
/// </summary>
/// <param name="scenarioId">
/// The scenario id.
/// </param>
/// <param name="inputSetId">
/// The input set id.
/// </param>
/// <returns>
/// The <see cref="PredictionSet"/>.
/// </returns>
public PredictionSet Predict(string scenarioId, string inputSetId)
{
var predictions = new PredictionSet {
Predictions = new List <Prediction>()
};
ScenarioTrainings trainings;
if (!trainingByScenario.TryGetValue(scenarioId, out trainings))
{
trainings = this.Train(scenarioId);
}
if (trainings == null)
{
return(predictions);
}
using (var dbContext = new OpenAIEntities1())
{
List <Feature> features =
dbContext.Features.Where(f => f.ScenarioId == scenarioId).OrderBy(f => f.Position).ToList();
Dictionary <string, Input> inputsById =
dbContext.Inputs.Where(i => i.ScenarioId == scenarioId && i.InputSetId == inputSetId)
.ToDictionary(i => i.FeatureId);
var sortedInputs = new List <Input>();
foreach (Feature feature in features)
{
Input input;
if (inputsById.TryGetValue(feature.FeatureId, out input))
{
sortedInputs.Add(input);
}
else
{
sortedInputs.Add(new Input {
FeatureId = feature.FeatureId
});
}
}
List <string> unset =
features.Select(f => f.FeatureId).Except(inputsById.Values.Select(i => i.FeatureId)).ToList();
foreach (string inputId in unset)
{
TrainerHelper training;
if (!trainings.TrainingByFeatureId.TryGetValue(inputId, out training))
{
continue;
}
var inputsMinusFeature = sortedInputs.Where(i => i.FeatureId != inputId).Select(i => i.Value).ToArray();
KeyValuePair <string, double> valueAndConfidence = trainer.Decide(training, inputsMinusFeature, inputId);
var prediction = new Prediction
{
InputId = inputId,
Confidence = valueAndConfidence.Value,
Value = valueAndConfidence.Key
};
predictions.Predictions.Add(prediction);
}
}
return(predictions);
}
