public IHttpActionResult UpdateDocumentType(int fileId)
{
HttpContent requestContent = Request.Content;
System.Diagnostics.Debug.WriteLine(requestContent);
string jsonContent = requestContent.ReadAsStringAsync().Result;
//System.Diagnostics.Debug.WriteLine(jsonContent.GetType());
//string text = JsonConvert.DeserializeObject<String>(jsonContent);
dynamic jsonText = JsonConvert.DeserializeObject(jsonContent);
string type = JsonConvert.SerializeObject(jsonText.type);
int val = Int32.Parse(type);
/*
* string text = "";
* for(int i = 0; i < 40; i++)
* {
* text += jsonText.text[i].Text;
* text += jsonText.text[i].Coords;
* }
*/
//System.Diagnostics.Debug.WriteLine(type);
//string text = jsonText.text;
OCRDatabaseEntities db = new OCRDatabaseEntities();
Document doc = db.Documents.Find(fileId);
if (doc == null)
{
return(NotFound());
}
doc.DocumentType = type;
System.Diagnostics.Debug.WriteLine("Type: " + type);
db.SaveChanges();
string dataFilePath = System.Web.HttpContext.Current.Server.MapPath("~/Data/data_train.csv");
string processedText = TextPreprocessorService.parseJSONText(db.Documents.Find(fileId).DocumentText);
//System.Diagnostics.Debug.WriteLine("Nakon JSON parse-a:" + processedText);
processedText = TextPreprocessorService.ProcessText(ref processedText);
FileIO.CSVWrite(processedText, val, dataFilePath);
fileService.UnlockDocument(fileId);
return(Ok());
//
//Document doc = db.Documents.Find(fileId);
//if (doc.DocumentText == null)
//{
// OCRService ocr = new OCRService();
// if (!ocr.RecognizeText(fileId))
// {
// //Debug.WriteLine("cao1");
// return NotFound();
// }
//}
////Debug.WriteLine("cao");
//return Ok(doc.DocumentText);
}
public Dictionary <int, double> PredictByText(string input)
{
// STEP 4: Read the data
string dataFilePath = System.Web.HttpContext.Current.Server.MapPath("~/Data/data_train.csv");
var dataTable = DataAccess.DataTable.New.ReadCsv(dataFilePath);
List <string> x = dataTable.Rows.Select(row => row["Text"]).ToList();
double[] y = dataTable.Rows.Select(row => double.Parse(row["Type"])).ToArray();
var vocabulary = x.SelectMany(GetWords).Distinct().OrderBy(word => word).ToList();
Console.WriteLine("Creating problem");
var problemBuilder = new DataPreprocess.TextClassificationProblemBuilder();
var problem = problemBuilder.CreateProblem(x, y, vocabulary.ToList());
// // If you want you can save this problem with :
// //ProblemHelper.WriteProblem(@"C:\Users\", problem);
// // And then load it again using:
// //var problem2 = ProblemHelper.ReadProblem(@"D:\MACHINE_LEARNING\SVM\Tutorial\sunnyData.problem");
System.Diagnostics.Debug.WriteLine("Creating model");
const int C = 1;
var model = new C_SVC(problem, KernelHelper.LinearKernel(), C, 100, true);
var accuracy = model.GetCrossValidationAccuracy(10);
System.Diagnostics.Debug.WriteLine(new string('=', 50));
System.Diagnostics.Debug.WriteLine("Accuracy of the model is {0:P}", accuracy);
model.Export(string.Format(@"model_{0}_accuracy.model", accuracy));
System.Diagnostics.Debug.WriteLine(new string('=', 50));
System.Diagnostics.Debug.WriteLine("The model is trained. \r\nEnter a sentence to make a prediction.");
System.Diagnostics.Debug.WriteLine(new string('=', 50));
_predictionDictionary = new Dictionary <int, string> {
{ 1, "ID" }, { 2, "Documents" }, { 3, "Forme" }
};
int numOFWords = 0;
string processedText = TextPreprocessorService.parseJSONText(input);
processedText = TextPreprocessorService.ProcessText(ref processedText);
Dictionary <int, double> dict = new Dictionary <int, double>()
{
{ 1, 0 }, { 2, 0 }, { 3, 0 }
};
if (processedText.Equals(""))
{
return(dict);
}
var newX = TextClassificationProblemBuilder.CreateNode(processedText, vocabulary);
var predictedY = model.Predict(newX);
System.Diagnostics.Debug.WriteLine(predictedY);
dict = model.PredictProbabilities(newX);
System.Diagnostics.Debug.WriteLine("Prob(1): " + dict[1]);
System.Diagnostics.Debug.WriteLine("Prob(2): " + dict[2]);
System.Diagnostics.Debug.WriteLine("Prob(3): " + dict[3]);
System.Diagnostics.Debug.WriteLine("The prediction is {0} value is {1} ", _predictionDictionary[(int)predictedY], predictedY);
return(dict);
}
public string ProcessText(string input)
{
return(TextPreprocessorService.ProcessText(ref input));
}
public string classify(int id)
{
OCRDatabaseEntities db = new OCRDatabaseEntities();
Document doc = db.Documents.Find(id);
string text = doc.DocumentText;
if ((text == null))
{
text = RecognizeDocText(id);
text = TextPreprocessorService.parseJSONText(text);
text = TextPreprocessorService.ProcessText(ref text);
if (text == null)
{
return(null);
}
}
Dictionary <int, double> dict = PredictByText(text);
System.Diagnostics.Debug.WriteLine("ByText");
System.Diagnostics.Debug.WriteLine(dict[1].ToString());
System.Diagnostics.Debug.WriteLine(dict[2].ToString());
System.Diagnostics.Debug.WriteLine(dict[3].ToString());
int predictionNumOFWords = PredictByNumOfWords();
System.Diagnostics.Debug.WriteLine(predictionNumOFWords);
if (predictionNumOFWords == 2)
{
dict[2] += 0.2;
}
System.Diagnostics.Debug.WriteLine("ByNumOfWords");
System.Diagnostics.Debug.WriteLine(dict[1].ToString());
System.Diagnostics.Debug.WriteLine(dict[2].ToString());
System.Diagnostics.Debug.WriteLine(dict[3].ToString());
System.Diagnostics.Debug.WriteLine(OCRService.FaceFlag);
int faceFlag = OCRService.FaceFlag;
System.Diagnostics.Debug.WriteLine("FaceFlag");
System.Diagnostics.Debug.WriteLine(faceFlag);
if (faceFlag == 1)
{
dict[1] += 0.75;
}
System.Diagnostics.Debug.WriteLine("ByFace");
System.Diagnostics.Debug.WriteLine(dict[1].ToString());
System.Diagnostics.Debug.WriteLine(dict[2].ToString());
System.Diagnostics.Debug.WriteLine(dict[3].ToString());
double maxPerc = dict.Values.Max();
if (maxPerc <= 0.1)
{
return("0");
}
string retVal = dict.FirstOrDefault(x => x.Value == maxPerc).Key.ToString();
string type = "-" + retVal;
doc.DocumentType = type;
System.Diagnostics.Debug.WriteLine(doc.DocumentType);
db.SaveChanges();
return(type);
}