/// <summary>
/// Used to update the retrained model. It first checks to ensure that the job completed successfuly.
/// </summary>
/// <param name="jobId"></param>
/// <returns></returns>
public async Task <Boolean> UpdateModel(string jobId)
{
BatchScoreStatus status;
bool isUdpated = false;
try
{
using (HttpClient client = new HttpClient())
{
// Check the job
string jobLocation = _mlretrainmodelurl + "/" + jobId + "?api-version=2.0";
client.DefaultRequestHeaders.Authorization = new AuthenticationHeaderValue("Bearer", _mlretrainerkey);
HttpResponseMessage response = await client.GetAsync(jobLocation);
if (!response.IsSuccessStatusCode)
{
return(false);
}
status = await response.Content.ReadAsAsync <BatchScoreStatus>();
}
AzureBlobDataReference res = status.Results["output2"];
isUdpated = await UpdateRetrainedModel(res.BaseLocation, res.RelativeLocation, res.SasBlobToken, _storageConnectionString);
//isUdpated = await UpdateRetrainedModel2(res.BaseLocation, res.RelativeLocation, res.SasBlobToken, _storageConnectionString);
return(isUdpated);
}
catch
{
return(false);
}
}
/// <summary>
/// Used to setup the job retraining. You nust receive the jobID from this call and submit to the start method. This method only gets retraining
/// ready but it MUST be started with the returned jobID after the sucessful completion of this call
/// This will submit the default SQL Query pulled from the ML Storage container and use it unless it was nodified with other LIB calls
/// </summary>
/// <returns>This is a jobID used to then start the retraining job</returns>
public async Task <string> QueueRetrainingAsync(TRAINING_DATA_SOURCE source, Dictionary <string, string> globalOptions)
{
string jobId = null;
_currentModelTraining = retrainerPrefix + DateTime.Now.ToString("yyyy'-'MM'-'dd'T'HH':'mm':'ss");
BatchExecutionRequest request = null;
AzureBlobDataReference adr = null;
if (source == TRAINING_DATA_SOURCE.DATA_UPLOAD && _trainingBlob == null)
{
throw new Exception("Training set not uploaded referenced");
}
else if (source == TRAINING_DATA_SOURCE.DATA_UPLOAD && _trainingBlob != null)
{
adr = new AzureBlobDataReference()
{
ConnectionString = _storageConnectionString,
RelativeLocation = _trainingBlob.Uri.LocalPath
};
}
//Now do outputs
var outputs = new Dictionary <string, AzureBlobDataReference>()
{
{
"output2",
new AzureBlobDataReference()
{
ConnectionString = _storageConnectionString,
RelativeLocation = string.Format("/{0}/{1}.ilearner", _mlstoragecontainer, _currentModelTraining)
}
},
{
"output1",
new AzureBlobDataReference()
{
ConnectionString = _storageConnectionString,
RelativeLocation = string.Format("/{0}/{1}.csv", _mlstoragecontainer, _currentModelTraining)
}
},
};
//Now setup request object
request = new BatchExecutionRequest()
{
Outputs = outputs
};
if (globalOptions != null)
{
request.GlobalParameters = globalOptions;
}
else
{
request.GlobalParameters = new Dictionary <string, string>();
}
if (adr != null)
{
request.Input = adr;
}
using (HttpClient client = new HttpClient())
{
client.DefaultRequestHeaders.Authorization = new AuthenticationHeaderValue("Bearer", _mlretrainerkey);
// WARNING: The 'await' statement below can result in a deadlock if you are calling this code from the UI thread of an ASP.Net application.
// One way to address this would be to call ConfigureAwait(false) so that the execution does not attempt to resume on the original context.
// For instance, replace code such as:
// result = await DoSomeTask()
// with the following:
// result = await DoSomeTask().ConfigureAwait(false)
// submit the job
string uploadJobURL = _mlretrainmodelurl + "?api-version=2.0";
HttpResponseMessage response = await client.PostAsJsonAsync(uploadJobURL, request);
if (!response.IsSuccessStatusCode)
{
return(null);
}
jobId = await response.Content.ReadAsAsync <string>(); //Used to reference the job for start and monitoring of completion
}
return(jobId);
}