static void Extract(bool recoveryEnabled)
{
List <int> errorSessionList = new List <int>();
int sessionsX = 0;
int global = 0;
var pdfReader = new PdfReader(_filePath);
int lastPage;
string format = "D" + pdfReader.NumberOfPages.ToString().Length;
int lastValidIdFromPreviousPage = -1;
for (int p = 0; p < pdfReader.NumberOfPages; p++)
{
Console.Clear();
Console.SetCursorPosition(0, 0);
Console.WriteLine("Processing page " + (p + 1).ToString(format) + " of " + pdfReader.NumberOfPages);
string textFromPage = "";
lastPage = p;
while (lastPage + 1 <= pdfReader.NumberOfPages && lastPage + 1 != pdfReader.NumberOfPages)
{
lastPage++;
textFromPage = PdfTextExtractor.GetTextFromPage(pdfReader, lastPage + 1);
if (textFromPage.Contains("Sess"))
{
break;
}
}
string segments = PdfTextExtractor.GetTextFromPage(pdfReader, p + 1);
List <string> sessionList = new List <string>();
while (segments.Length > 0)
{
int i1 = segments.IndexOf("Sess");
//text goes over more than 2 pages but we already got the session info
if (i1 == -1)
{
break;
}
int i2 = segments.IndexOf("Sess", i1 + "Session".Length);
if (i2 != -1)
{
string[] kArr = segments.Substring(i2, 20).Split(' ');
if (!kArr[1].Contains("#"))
{
string temp = segments.Substring(i1, i2 - i1);
sessionList.Add(temp);
segments = segments.Substring(i2, segments.Length - i2);
}
else
{
i2 = segments.IndexOf("Sess", i2 + "Session".Length);
if (i2 != -1)
{
string temp = segments.Substring(i1, i2 - i1);
sessionList.Add(temp);
segments = segments.Substring(i2, segments.Length - i2);
}
else
{
string temp = segments.Substring(i1, segments.Length - i1);
sessionList.Add(temp);
segments = "";
}
}
}
else
{
string temp = segments.Substring(i1, segments.Length - i1);
sessionList.Add(temp);
segments = "";
}
}
List <Info> infoList = new List <Info>();
int sessions = 0;
var tes = new MyLocationTextExtractionStrategy();
var ex = PdfTextExtractor.GetTextFromPage(pdfReader, p + 1, tes);
for (int chunk = 0; chunk < tes.myPoints.Count; chunk++)
{
if (tes.myPoints[chunk].Text.Contains("Sess"))
{
if (chunk + 1 < tes.myPoints.Count && tes.myPoints[chunk + 1].Text.Contains("#"))
{
continue;
}
Info info;
info.sessionId = "NA";
info.rect = tes.myPoints[chunk].Rect;
//get the Session number
Regex regex = new Regex(@"^[0-9]+$");
string[] tokens = sessionList[sessions].Split(' ');
for (int t = 0; t < tokens.Length; t++)
{
if (tokens[t].Contains("Sess"))
{
/**** do a regex to any number of digits****/
if (t + 1 < tokens.Length && regex.IsMatch(tokens[t + 1].Trim()))
{
info.sessionId = tokens[t + 1].Trim();
}
break;
}
}
//info.Print();
infoList.Add(info);
sessions++;
}
}
if (infoList.Count == 0)
{
Console.WriteLine("Fatal: PDF resolution did not allow to recover word rendering locations");
return;
}
/*** Recovering session numbers lost ***/
if (recoveryEnabled)
{
int indexFound = -1;
int s = 0;
for (; s < infoList.Count; s++)
{
if (indexFound != -1)
{
if (!infoList[s].sessionId.Contains("NA"))
{
if (int.Parse(infoList[indexFound].sessionId) - int.Parse(infoList[s].sessionId) == indexFound - s)
{
break;
}
}
}
if (!infoList[s].sessionId.Contains("NA"))
{
indexFound = s;
if (lastValidIdFromPreviousPage != -1)
{
if (Math.Abs(lastValidIdFromPreviousPage - int.Parse(infoList[s].sessionId)) == s + 1)
{
break;
}
}
}
}
if (s == infoList.Count && indexFound != s - 1)
{
indexFound = -1;
}
if (indexFound != -1)
{
int val = int.Parse(infoList[indexFound].sessionId);
for (int tt = 0; tt < infoList.Count; tt++)
{
Info info = infoList[tt];
if (info.sessionId.Contains("NA") && tt < indexFound)
{
int newVal = val - Math.Abs(indexFound - tt);
info.sessionId = (newVal).ToString();
infoList[tt] = info;
errorSessionList.Add(newVal);
}
else if (info.sessionId.Contains("NA") && tt > indexFound)
{
int newVal = val + Math.Abs(indexFound - tt);
info.sessionId = (newVal).ToString();
infoList[tt] = info;
errorSessionList.Add(newVal);
}
else if (!info.sessionId.Contains("NA"))
{
if (lastValidIdFromPreviousPage != -1)
{
int iVal = lastValidIdFromPreviousPage + tt + 1;
if (int.Parse(infoList[tt].sessionId) != iVal)
{
//Correcting OCR error
info.sessionId = iVal.ToString();
infoList[tt] = info;
errorSessionList.Add(iVal);
}
}
else
{
int iVal = int.Parse(infoList[indexFound].sessionId) - indexFound + tt;
if (int.Parse(infoList[tt].sessionId) != iVal)
{
//Correcting OCR error
info.sessionId = iVal.ToString();
infoList[tt] = info;
errorSessionList.Add(iVal);
}
}
}
}
lastValidIdFromPreviousPage = int.Parse(infoList[infoList.Count - 1].sessionId);
}
else //try again in case the whole page failed
{
if (lastValidIdFromPreviousPage != -1)
{
int ss = lastValidIdFromPreviousPage + 1;
for (int tt = 0; tt < infoList.Count; tt++)
{
Info info = infoList[tt];
if (info.sessionId.Contains("NA"))
{
info.sessionId = (ss + tt).ToString();
infoList[tt] = info;
}
}
if (infoList.Count - 1 >= 0)
{
lastValidIdFromPreviousPage = int.Parse(infoList[infoList.Count - 1].sessionId);
}
}
else
{
if (!int.TryParse(infoList[infoList.Count - 1].sessionId, out lastValidIdFromPreviousPage))
{
Console.WriteLine("Fatal: Could not recover from this error on page " + (p + 1));
lastValidIdFromPreviousPage = -1;
}
}
}
//double check
foreach (Info info in infoList)
{
if (info.sessionId.Contains("NA"))
{
Console.WriteLine("Could not recover session Id on page " + (p + 1));
sessionsX++;
}
}
}
else
{
//just count
foreach (Info info in infoList)
{
if (info.sessionId.Contains("NA"))
{
sessionsX++;
}
}
}
//Single-page template
Document doc1 = new Document();
PdfCopy copy1;
PdfImportedPage importedPage1;
using (FileStream fs1 = new FileStream(outputDir + "template1.pdf", FileMode.Create))
{
copy1 = new PdfCopy(doc1, fs1);
doc1.Open();
importedPage1 = copy1.GetImportedPage(pdfReader, p + 1);
copy1.AddPage(importedPage1);
doc1.Close();
}
//Multi-page template
int pTemp;
Document docN = new Document();
PdfCopy copyN = null;
PdfImportedPage importedPageN = null;
using (FileStream fs = new FileStream(outputDir + "templateN.pdf", FileMode.Create))
{
copyN = new PdfCopy(docN, fs);
docN.Open();
pTemp = p;
while (pTemp <= lastPage && pTemp != pdfReader.NumberOfPages)
{
importedPageN = copyN.GetImportedPage(pdfReader, pTemp + 1);
copyN.AddPage(importedPageN);
pTemp++;
}
docN.Close();
}
//retrieve the info from for the last page where the next session is
Info infoEnd;
infoEnd.sessionId = "Last";
Rectangle r = pdfReader.GetPageSize(lastPage + 1);
infoEnd.rect = new Rectangle(0, r.Top, 0, r.Bottom); ///swapping top
tes = new MyLocationTextExtractionStrategy();
ex = PdfTextExtractor.GetTextFromPage(pdfReader, lastPage + 1, tes);
for (int chunk = 0; chunk < tes.myPoints.Count; chunk++)
{
if (tes.myPoints[chunk].Text.Contains("Sess"))
{
infoEnd.rect = tes.myPoints[chunk].Rect;
break;
}
}
//go through all info
for (int info = 0; info < infoList.Count; info++)
{
string outputFilename = infoList[info].sessionId.Trim();
if (File.Exists(outputDir + outputFilename + ".pdf"))
{
outputFilename += "-" + (global + 1).ToString("D4");
}
global++;
outputFilename += ".pdf";
PdfStamper stamper;
PdfDictionary page;
if (info + 1 < infoList.Count)
{
var pdfReader2 = new PdfReader(outputDir + "template1.pdf");
using (FileStream fs = new FileStream(outputDir + outputFilename, FileMode.Create))
{
stamper = new PdfStamper(pdfReader2, fs);
page = pdfReader2.GetPageN(1);
page.Put(PdfName.CROPBOX, new PdfArray(new float[] { 0, infoList[info + 1].rect.Top + gapV, importedPage1.Width, infoList[info].rect.Top + gapV }));
stamper.MarkUsed(page);
Dictionary <String, String> meta = pdfReader2.Info;
meta.Add("Keywords", infoList[info].sessionId);
stamper.MoreInfo = meta;
stamper.Close();
pdfReader2.Close();
}
}
else
{
PdfReader pdfReader2;
//last page
if (p + 1 == pdfReader.NumberOfPages)
{
pdfReader2 = new PdfReader(outputDir + "template1.pdf");
stamper = new PdfStamper(pdfReader2, new FileStream(outputDir + outputFilename, FileMode.Create));
page = pdfReader2.GetPageN(1);
page.Put(PdfName.CROPBOX, new PdfArray(new float[] { 0, 0, importedPageN.Width, infoList[info].rect.Top + gapV }));
stamper.MarkUsed(page);
Dictionary <String, String> meta2 = pdfReader2.Info;
meta2.Add("Keywords", infoList[info].sessionId);
stamper.MoreInfo = meta2;
stamper.Close();
pdfReader2.Close();
break;
}
else //last sesion
{
pdfReader2 = new PdfReader(outputDir + "templateN.pdf");
stamper = new PdfStamper(pdfReader2, new FileStream(outputDir + outputFilename, FileMode.Create));
page = pdfReader2.GetPageN(1);
page.Put(PdfName.CROPBOX, new PdfArray(new float[] { 0, 0, importedPageN.Width, infoList[info].rect.Top + gapV }));
stamper.MarkUsed(page);
//session with more than 2 pages
if (lastPage - p > 1)
{
int newP = 1;
int max = lastPage - p;
while (newP <= max)
{
page = pdfReader2.GetPageN(newP + 1);
if (newP != max)
{
page.Put(PdfName.CROPBOX, new PdfArray(new float[] { 0, 0, importedPageN.Width, importedPageN.Height }));
}
else
{
page.Put(PdfName.CROPBOX, new PdfArray(new float[] { 0, infoEnd.rect.Top + gapV, importedPageN.Width, importedPageN.Height }));
}
stamper.MarkUsed(page);
newP++;
}
p = lastPage - 1;
}
else
{
page = pdfReader2.GetPageN(2);
page.Put(PdfName.CROPBOX, new PdfArray(new float[] { 0, infoEnd.rect.Top + gapV, importedPageN.Width, importedPageN.Height }));
stamper.MarkUsed(page);
}
}
Dictionary <String, String> meta = pdfReader2.Info;
meta.Add("Keywords", infoList[info].sessionId);
stamper.MoreInfo = meta;
stamper.Close();
pdfReader2.Close();
}
}
}
Console.WriteLine("Sessions extracted: " + global);
if (recoveryEnabled)
{
Console.WriteLine("Session ID recovery was attempted. Percentage of undetected Session IDs after OCR reconstruction was " + (((float)sessionsX / (float)global) * 100).ToString("0.00") + "%");
}
else
{
Console.WriteLine("No Session ID recovery was attempted. Percentage of undetected Session IDs with no In/Out Digits was " + (((float)sessionsX / (float)global) * 100).ToString("0.00") + "%");
}
Console.WriteLine();
}
