// METHOD ALAIN: KORTER, EVENTUEEL UITZOEKEN... WERKT NOG NIET ivm indexes
/*
* // Splitted string at index position
*
* string temp = str;
* for (int i = 0; i < indexList.Count; i++)
* {
* temp = temp.Insert(i + indexList[i], " ");
* Debug.WriteLine(temp);
* }
* string[] splitted = temp.Trim(new char[] { ' ' }).Split(' ');
*
* // Found date . First 4 lenght string on the end and add '*' before and at the end
* int p = splitted.Length - 1;
* bool DateFound = false;
* while (p >= 0 && !DateFound)
* {
* Debug.WriteLine(splitted[p]);
* if (splitted[p].Length == 4)
* {
* result.IndexDateStart = indexList[p - 1];
* result.IndexDateEnd = indexList[p] - 1;
* splitted[p] = $"*{splitted[p]}*";
* }
* p--;
* }
*
* // Concate all to one string
* string temp2 = String.Concat(splitted);
*
* // DOT*DATE*CLUTTER => DOT,DATE,CLUTTER
* string[] Splitted2 = temp2.Split('*');
* result.Dot = Splitted2[0];
* result.Date = Splitted2[1];
* string Clutter = Splitted2[2];
*
* return result;
* }
*/
public ResultB SegmentB(SegmentInput B, double devW, double devSpace)
{
//Remove clutter from end of string of B, we presume only one segment of clutter and with a bigger space
ResultB result = new ResultB(); //final result for B
SegmentResult segmentRes = Segment(B, devW); //Segment B --> list of indexes and spaces
int nSpace = segmentRes.SpaceList.Count;
int nSegm = segmentRes.SegmentIndexList.Count;
string str = segmentRes.FilteredString;
int nStr = segmentRes.FilteredString.Length;
int indexDotEnd = 0;
if (nStr > 0) //At least 1 char in string
{
indexDotEnd = nStr - 1; //We assume last segment is not clutter or have no previous space to compare to
if (nSpace > 1) //At least 2 spaces present--> compare last space to 1st space
{
double spaceRef = segmentRes.SpaceList[0]; //take first space as reference
if (segmentRes.SpaceList[nSpace - 1] > devSpace * spaceRef) //last space larger than allowed deviation (= devSpace * spaceRef)
{
//last segment is clutter
indexDotEnd = segmentRes.SegmentIndexList[nSegm - 1] - 1;//index of last segment - 1 = end of previous segm
}
}
result.Dot = str.Substring(0, indexDotEnd + 1);
}
else //if nStr=0
{
indexDotEnd = 0;
result.Dot = "";
}
result.IndexDotStart = 0;//We presume dot always starts at 0 (no clutter in beginning)
result.IndexDotEnd = indexDotEnd;
return(result);
}
public ResultA SegmentA(SegmentInput A, double devW)
{
//Segment A in Dot + Date
ResultA result = new ResultA(); //final result for A
SegmentResult segmentRes = Segment(A, devW); //Segment A --> list of indexes and spaces
int n = segmentRes.FilteredString.Length;
int indexStartDate = 0;
int indexEndDate = 0;
string str = segmentRes.FilteredString;
string date = "";
string dot = "";
List <int> indexList = new List <int>();
indexList = segmentRes.SegmentIndexList;
// METHOD ADRIAAN
//We have segmented first, now we want to check if the last segment is indeed the date timestamp
//We work from right to left!
if (n > 0 && indexList.Count > 1) //At least 2 segments
{
int i = (indexList.Count - 1); //Start from the last segment (ex. count = 12 --> last index = 11)
int j = n; //end of last segment
while (i > 0) //At least 2 segments
{
if ((j - indexList[i]) > 3) //timestamp = 4 karakters, length of segment
{
indexStartDate = indexList[i];
if (i != (indexList.Count - 1)) //if not the last segment
{
indexEndDate = indexList[i + 1] - 1; //previous index of segment -1, is end of current segment
}
else //if last segment
{
indexEndDate = n - 1;
}
break;
}
j = indexList[i]; // length of next segment = previous index - start index next segment
i--; //Work from right to left
}
}
if (n > 3 && indexList.Count == 1) //Only 1 segment and string at least 4 char --> we assume last 4 characters are timestamp
{
indexEndDate = n - 1;
indexStartDate = n - 4;
}
if (indexStartDate > 0 && indexEndDate >= (indexStartDate + 4))
{
date = str.Substring(indexStartDate, ((indexEndDate + 1) - indexStartDate));
dot = str.Substring(0, indexStartDate);
}
else
{
date = "0000";
indexEndDate = n;
dot = str;
}
result.Date = date;
result.Dot = dot;
result.IndexDateEnd = indexEndDate;
result.IndexDateStart = indexStartDate;
return(result);
}
public SegmentResult Segment(SegmentInput input, double devW)
{
//Segment into list of Segment indexes and spaces
SegmentResult result = new SegmentResult();
PreFilterResult filtered = PreFilter(input);//Filter rubbish end of string
input.OcrString = filtered.OcrFiltered;
ViDiReadResultList lists = GenerateList(input); //Use filtered string to generate lists
double w = 0;
double xCurr = 0;
double xPrev = 0;
int nChar = lists.Read.Length;
double dev = devW; //consider space when difference between two positions is > dev * characterwidth
List <double> x_list = new List <double>(); //list for width of characters
x_list = lists.XPosList;
List <double> w_list = new List <double>(); //list for width of characters
w_list = lists.WidthList;
List <double> segm_x_list = new List <double>(); //list for x-positions of start of new segment
List <int> segm_i_list = new List <int>(); //list for character indexes of start of new segment
List <double> space_list = new List <double>(); //list for size of spaces of segment
if (nChar > 0 && x_list.Count == w_list.Count && x_list.Count == nChar) //Lists have same number of elements(should be the case)
{
w = w_list.Max(); //retain widest character as reference for calculating space
//First segment starts at first character - index 0 for X-position
xCurr = x_list[0];
segm_x_list.Add(x_list[0]);
segm_i_list.Add(0);
//Loop through the remaining characters --> list of index positions with start of segments
xPrev = xCurr;
int i = 1; //start at 1, we handled 0 above
while (i < nChar)
{
xCurr = x_list[i];
if ((xCurr - xPrev) > (dev * w)) //consider space when difference between two positions is > dev * characterwidth -->New segment
{
segm_x_list.Add(xCurr);
segm_i_list.Add(i);
space_list.Add(xCurr - xPrev);
}
xPrev = xCurr;
i++;
}
}
else
{
//
}
result.Nsegments = segm_i_list.Count; //Get number of segments
result.SegmentIndexList = segm_i_list;
result.SpaceList = space_list;
result.FilteredString = lists.Read;
return(result);
}