public static bool GetReceiptTotalTopLevelOnly(BaiOcrLine ocr)
{
var newContent = RefineToMoney(ocr.Content);
var length = newContent.Length;
Regex moneyPattern = new Regex(@"[.\d,]+");
bool isMatched = (moneyPattern.Match(newContent).Value.Length == length);
if (((length > 4 && GetPercentNumber(newContent) > 70) ||
(length > 3 && GetPercentNumber(newContent) == 100)) && isMatched)
{
string result = string.Empty;
foreach (char c in newContent)
{
if (char.IsNumber(c) || char.IsDigit(c) || c == '.' || c == ',')
{
result += c;
int len = result.Length;
int index = (len - 2) - 1;
if (index > 1)
{
if (result[index] == '.')
{
Global.CurrentReciept.Amount = result;
return(true);
}
}
}
}
}
return(false);
}
public static void GetReceiptTINValue(BaiOcrLine ocr)
{
var newContent = RefineToTIN(ocr.Content);
var length = newContent.Length;
if (length > 5)
{
Global.CurrentReciept.Tax_Identification = newContent;
return;
}
else
{
int i = ocr.LineNo + 1;
var nextLine = Global.OcrLines.Where(o => o.LineNo == i).FirstOrDefault();
if (nextLine == null || nextLine.LineNo == 0)
{
Global.CurrentReciept.Tax_Identification = RefineToTIN(Global.OcrLines.OrderByDescending(o => o.WeightedAsVendorTINTitle).First().Content);
}
else
{
GetReceiptTINValue(nextLine);
}
};
}
public static void GetReceiptTotalValue(BaiOcrLine ocr)
{ //do 1st level
if (!GetReceiptTotalTopLevelOnly(ocr))
{
//check next
int i = ocr.LineNo + 1;
var nextLine = Global.OcrLines.Where(o => o.LineNo == i).FirstOrDefault();
if (nextLine == null || nextLine.LineNo == 0)
{
Global.CurrentReciept.Amount = RefineToMoney(Global.OcrLines.OrderByDescending(o => o.WeightedAsTotalTitle).First().Content);
return;
}
else
{
//do 2nd level
if (!GetReceiptTotalTopLevelOnly(nextLine))
{
GetReceiptTotalValue(nextLine);
}
}
//if not found dig more...
GetReceiptTotalValue(nextLine);
}
;
}
private void btnTest_Click(object sender, EventArgs e)
{
//TestCmd();
List <BaiOcrLine> testLines = new List <BaiOcrLine>();
BaiOcrLine ocr = new BaiOcrLine
{
Content = "Date:10/11/19 TIN:123456789 address: 154 bisalao st. Maligay Nov Q.C. Vendor: SEven eleven"
};
testLines.Add(ocr);
BaiOcrLine ocr1 = new BaiOcrLine
{
Content = "TOTAL:100 CHANGE:123 Cash: 500"
};
testLines.Add(ocr1);
List <BaiOcrLine> result = new List <BaiOcrLine>();
foreach (var thisOcr in testLines)
{
RocsTextService.RefineOcr(thisOcr, ref result);
}
int i = 0;
foreach (BaiOcrLine item in result)
{
i += 1;
item.LineNo = i;
Console.WriteLine(string.Format("{0}. {1}", item.LineNo.ToString(), item.Content));
}
}
public static void GetReceiptDateValue(BaiOcrLine ocr)
{
var newContent = RefineToDate(ocr.Content);
var year = DateTime.Now.Date.Year;
string strYear = year.ToString();
strYear = strYear.Substring(2, 2);
var length = newContent.Length;
if (length >= 8)
{
if (newContent.EndsWith(strYear))
{
Global.CurrentReciept.Date = newContent;
}
}
else
{
int i = ocr.LineNo + 1;
var nextLine = Global.OcrLines.Where(o => o.LineNo == i).FirstOrDefault();
if (nextLine == null || nextLine.LineNo == 0)
{
Global.CurrentReciept.Date = RefineToDate(Global.OcrLines.OrderByDescending(o => o.WeightedAsDateTitle).First().Content);
}
else
{
GetReceiptDateValue(nextLine);
}
};
}
//remove all elected first then do this policy
public static void AssertAsDetail(ref BaiOcrLine ocrLine)
{
var dim = ReceiptParts.Detail.ToString();
using (MyDBContext db = new MyDBContext())
{
var wfactors = db.WeightFactors.ToList();
var factors = wfactors.Where(f => f.Dimension == dim);
foreach (var f in factors)
{
if (ocrLine.Content.ToLower().Contains(f.keyWord.ToLower()))
{
ocrLine.WeightedAsProduct += f.Weight;
}
}
}
if (ocrLine.PercentNumber < 25)
{
ocrLine.WeightedAsAddress += 2;
}
if (ocrLine.Content.Contains(":"))
{
ocrLine.WeightedAsAddress -= 1;
}
}
public static void RefineOcr(BaiOcrLine ocr, ref List <BaiOcrLine> newLines)
{
//string totalContent = ocr.Content;
var splitAll = ocr.Content.Split(':').ToList();
HashSet <string> hashResult = new HashSet <string>();
for (int i = 0; i < splitAll.Count; i++)
{
string content = string.Empty;
content = splitAll[i];
Console.WriteLine(content);
if (string.IsNullOrEmpty(content))
{
continue;
}
content = content.Trim();
var head = content.Split().Last();
var first = content.Replace(head, string.Empty);
if (splitAll.Count == 1)
{
head = splitAll[0];
// first = splitAll[1];
hashResult.Add(head);
}
else if (splitAll.Count == 2)
{
hashResult.Add(splitAll[0]);
hashResult.Add(splitAll[1]);
}
else if (i == splitAll.Count - 1)
{
hashResult.Add(content);
}
else
{
if (!string.IsNullOrEmpty(first))
{
hashResult.Add(first);
}
hashResult.Add(head);
}
}
foreach (string l in hashResult)
{
BaiOcrLine ocr1 = new BaiOcrLine
{
Content = l
};
newLines.Add(ocr1);
Console.WriteLine(l);
}
}
public static double GetPercentLocation(BaiOcrLine ocr)
{
var totalCnt = Convert.ToDouble(Global.OcrLines.Count());
var i = Convert.ToDouble(ocr.LineNo);
double percent = i / totalCnt;
return(percent * 100);
}
public static void AssertAsDateTitle(ref BaiOcrLine ocrLine)
{
var dim = ReceiptParts.DateTitle.ToString();
using (MyDBContext db = new MyDBContext())
{
var wfactors = db.WeightFactors.ToList();
var factors = wfactors.Where(f => f.Dimension == dim);
foreach (var f in factors)
{
if (ocrLine.Content.ToLower().Contains(f.keyWord.ToLower()))
{
ocrLine.WeightedAsDateTitle += f.Weight;
}
}
//as observed of SI entry
if ((!string.IsNullOrEmpty(ocrLine.Content) &&
ocrLine.Content.Trim() == "SI"))
{
ocrLine.WeightedAsDateTitle += 3;
}
}
//do not for 11/11/11
if (ocrLine.PercentNumber == 0)
{
ocrLine.WeightedAsDateTitle += 1;
}
if (ocrLine.Content.Contains("/"))
{
ocrLine.WeightedAsDateTitle -= 1;
}
var y = DateTime.Now.Year.ToString();
if (ocrLine.Content.Contains(y))
{
ocrLine.WeightedAsDateTitle -= 2;
}
var y2 = y.Substring(2, 2);
if (ocrLine.Content.Contains(y2))
{
ocrLine.WeightedAsDateTitle -= 1;
}
//add weight by line
if (ocrLine.WeightedAsDateTitle > 2)
{
var maxLine = Global.OcrLines.Count;
var weightLine = maxLine - ocrLine.LineNo;
ocrLine.WeightedAsDateTitle += weightLine;
}
}
public static void AssertAsAddress(ref BaiOcrLine ocrLine)
{
var dim = ReceiptParts.Address.ToString();
using (MyDBContext db = new MyDBContext())
{
var wfactors = db.WeightFactors.ToList();
var factors = wfactors.Where(f => f.Dimension == dim);
foreach (var f in factors)
{
if (ocrLine.Content.ToLower().Contains(f.keyWord.ToLower()))
{
ocrLine.WeightedAsAddress += f.Weight;
}
}
}
//add weight by line
if (ocrLine.WeightedAsAddress > 2)
{
var maxLine = Global.OcrLines.Count;
var weightLine = maxLine - ocrLine.LineNo;
ocrLine.WeightedAsAddress += weightLine;
}
if (ocrLine.LineNo < 3)
{
ocrLine.WeightedAsAddress += 3;
}
if (ocrLine.LineNo == 1)
{
ocrLine.WeightedAsAddress += -3;
}
if (ocrLine.PercentNumber < 25)
{
ocrLine.WeightedAsAddress += 3;
}
var split = ocrLine.Content.Split();
if (split.Length == 1)
{
ocrLine.WeightedAsAddress += -5;
}
if (split.Length > 3)
{
ocrLine.WeightedAsAddress += 2;
}
}
// Retrieve the recognized text
private async Task GetTextAsync(
ComputerVisionClient computerVision, string operationLocation)
{
try
{
// Retrieve the URI where the recognized text will be
// stored from the Operation-Location header
string operationId = operationLocation.Substring(
operationLocation.Length - numberOfCharsInOperationId);
Console.WriteLine("\nCalling GetReadOperationResultAsync()");
ReadOperationResult result =
await computerVision.GetReadOperationResultAsync(operationId);
// Wait for the operation to complete
int i = 0;
int maxRetries = 10;
while ((result.Status == TextOperationStatusCodes.Running ||
result.Status == TextOperationStatusCodes.NotStarted) && i++ < maxRetries)
{
Console.WriteLine(
"Server status: {0}, waiting {1} seconds...", result.Status, i);
await Task.Delay(1000);
result = await computerVision.GetReadOperationResultAsync(operationId);
}
// Display the results
Console.WriteLine();
var recResults = result.RecognitionResults;
int lineNo = 0;
foreach (TextRecognitionResult recResult in recResults)
{
foreach (Line line in recResult.Lines)
{
lineNo += 1;
Console.WriteLine(line.Text);
BaiOcrLine ocr = new BaiOcrLine
{
LineNo = lineNo,
Content = line.Text
};
RawList.Add(ocr);
}
}
Console.WriteLine();
Global.ProcessStatus = ProcessStatus.Ready.ToString();
OnReadDone?.Invoke(this, EventArgs.Empty);
}
catch (Exception err)
{
throw err;
}
}
public static void AssertAsTotalTitle(ref BaiOcrLine ocrLine)
{
var dim = ReceiptParts.PriceTitle.ToString();
using (MyDBContext db = new MyDBContext())
{
var wfactors = db.WeightFactors.ToList();
var factors = wfactors.Where(f => f.Dimension == dim);
foreach (var f in factors)
{
if (ocrLine.Content.ToLower().Contains(f.keyWord.ToLower()))
{
ocrLine.WeightedAsTotalTitle += f.Weight;
}
}
}
}
public static void AssertAsChangeTitle(ref BaiOcrLine ocrLine)
{
var dim = ReceiptParts.ChangeTitle.ToString();
using (MyDBContext db = new MyDBContext())
{
var wfactors = db.WeightFactors.ToList();
var factors = wfactors.Where(f => f.Dimension == dim);
foreach (var f in factors)
{
if (ocrLine.Content.ToLower().Contains(f.keyWord.ToLower()))
{
ocrLine.WeightedAsChangeTitle += f.Weight;
}
}
}
if (ocrLine.GetPercentLocation > 40 && ocrLine.GetPercentLocation < 70)
{
ocrLine.WeightedAsTenderTitle += 2;
}
}
public static void DigValue(BaiOcrLine ocr, ref double?value)
{
var newContent = RefineToMoney(ocr.Content);
var length = newContent.Length;
Regex moneyPattern = new Regex(@"[.\d,]+");
bool isMatched = (moneyPattern.Match(newContent).Value.Length == length);
if (((length > 4 && GetPercentNumber(newContent) > 70) ||
(length > 3 && GetPercentNumber(newContent) == 100)) && isMatched)
{
try
{
value = Convert.ToDouble(newContent);
return;
}
catch (Exception)
{
//throw;
//continue search
}
}
int i = ocr.LineNo + 1;
var nextLine = Global.OcrLines.Where(o => o.LineNo == i).FirstOrDefault();
if (nextLine == null || nextLine.LineNo == 0)
{
value = null;
return;
}
else
{
DigValue(nextLine, ref value);
}
}
public static void AssertAsVendorName(ref BaiOcrLine ocrLine)
{
var dim = ReceiptParts.VendorName.ToString();
using (MyDBContext db = new MyDBContext())
{
var wfactors = db.WeightFactors.ToList();
var factors = wfactors.Where(f => f.Dimension == dim);
foreach (var f in factors)
{
if (ocrLine.Content.ToLower().Contains(f.keyWord.ToLower()))
{
ocrLine.WeightedAsVendorName += f.Weight;
}
}
}
if (ocrLine.LineNo < 3)
{
ocrLine.WeightedAsVendorName += 3;
}
if (ocrLine.PercentNumber < 10)
{
ocrLine.WeightedAsVendorName += 3;
}
//add weight by line
if (ocrLine.WeightedAsVendorName > 2)
{
var maxLine = Global.OcrLines.Count;
var weightLine = maxLine - ocrLine.LineNo;
ocrLine.WeightedAsVendorName += weightLine;
}
}
