public static DocumentJSON.Document CreateJSON(DocumentCore docCore, int version)
{
DocumentJSON.Document docJSON = new DocumentJSON.Document();
string now = DateTime.Now.ToString("yyyymmdd");
docJSON.date = now;
docJSON.version = version;
//Create a list of paragraphs
List <DocumentJSON.Paragraph> parJSONList = new List <DocumentJSON.Paragraph>();
List <DocumentJSON.Image> imageJSONList = new List <DocumentJSON.Image>();
List <DocumentJSON.Cell> cellJSONList = new List <DocumentJSON.Cell>();
//loops through all sections
int parId = 0;
int imgId = 0;
bool imgLoop = false;
bool parLoop = false;
int tableId = 0;
foreach (Section sec in docCore.GetChildElements(false, ElementType.Section))
{
//loop through tables
foreach (Table t in sec.GetChildElements(false, ElementType.Table))
{
foreach (TableRow r in t.GetChildElements(false, ElementType.TableRow))
{
foreach (TableCell c in r.GetChildElements(false, ElementType.TableCell))
{
DocumentJSON.Cell cell = new DocumentJSON.Cell();
cell.hash = CalculateHash(Encoding.UTF8.GetBytes(c.Content.ToString()));
cell.content = c.Content.ToString();
cell.version = 1;
cell.date = now;
cell.status = "n";
cellJSONList.Add(cell);
}
}
docJSON.lastTableId++;
}
docJSON.cells = cellJSONList.ToArray();
//gets all text on each section
foreach (Paragraph par in sec.GetChildElements(false, ElementType.Paragraph))
{
//gets images
foreach (Picture p in par.GetChildElements(false, ElementType.Picture))
{
imgLoop = true;
MemoryStream stream = new MemoryStream();
stream = p.ImageData.GetStream();
BinaryReader binaryReader = new BinaryReader(stream);
Byte[] data = binaryReader.ReadBytes((int)stream.Length);
string hash = Document.CalculateHash(data);
if (imageJSONList.Where(x => x.hash.Equals(hash)).Count() == 0)
{
DocumentJSON.Image image = new DocumentJSON.Image();
image.hash = hash;
image.date = now;
image.numberOfRepetition = 0;
image.id = imgId++;
image.status = "n";
image.version = version;
imageJSONList.Add(image);
}
else
{
DocumentJSON.Image image = imageJSONList.Where(x => x.hash.Equals(hash)).First();
image.numberOfRepetition++;
}
}
//guarantees that only paragraphs with actual text are saved
if (!par.Content.ToString().Equals("\r\n") && !par.Content.ToString().Equals(""))
{
parLoop = true;
DocumentJSON.Paragraph parJSON = new DocumentJSON.Paragraph();
parJSON.content = par.Content.ToString().Replace("trial", "");
if (parJSON.content.Contains("Created by the version of Document .Net 3.3.3.27!"))
{
parJSON.content = parJSON.content.Replace("Created by the version of Document .Net 3.3.3.27!\r\nThe version sometimes inserts \"\" into random places.\r\nGet the full version of Document .Net.\r\n", "");
}
parJSON.hash = CalculateHash(Encoding.UTF8.GetBytes(parJSON.content));
parJSON.id = parId++;
parJSON.version = version;
parJSON.date = now;
parJSON.status = "n";
parJSON.sentence = null;
if (!parJSON.content.Equals(""))
{
parJSONList.Add(parJSON);
}
}
}
}
docJSON.images = imageJSONList.ToArray();
if (parLoop)
{
docJSON.lastParId = --parId;
}
if (imgLoop)
{
docJSON.lastImageId = --imgId;
}
docJSON.paragraphs = parJSONList.ToArray();
return(docJSON);
}
public static DocumentJSON.Document CompareParagraphs(DocumentJSON.Document oldDocument, DocumentJSON.Document newDocument)
{
//DocumentJSON.Document refDoc = getJSONObject(docId, version - 1);
string now = DateTime.Now.ToString("yyyymmdd");
List <DocumentJSON.Paragraph> newParList = new List <DocumentJSON.Paragraph>(newDocument.paragraphs);
List <DocumentJSON.Paragraph> oldParList = new List <DocumentJSON.Paragraph>(oldDocument.paragraphs);
newDocument.lastParId = oldDocument.lastParId;
foreach (DocumentJSON.Paragraph par in newParList)
{
//Checks for pre-existent paragraphs
if (oldParList.Where(x => x.hash.Equals(par.hash)).Count() != 0)
{
if (!oldParList.Where(x => x.hash.Equals(par.hash)).First().status.Equals("d"))
{
par.status = "o";
}
else
{
par.status = "d";
}
DocumentJSON.Paragraph refPar = (DocumentJSON.Paragraph)oldParList.Where(x => x.hash == par.hash).First();
par.version = refPar.version;
par.date = refPar.date;
par.id = refPar.id;
oldParList.Remove(refPar);
}
//Checks for new and modified paragraph
else
{
//Checks for modified paragraph
string[] sentences = par.content.Split('.');
List <DocumentJSON.Sentence> sentenceList = new List <DocumentJSON.Sentence>();
//creates a similarityCounter
List <KeyVal <int, int> > similarityList = new List <KeyVal <int, int> >();
//creates a reference paragraph list
List <DocumentJSON.Paragraph> refPars = new List <DocumentJSON.Paragraph>();
//sets version and date for the paragraph
par.version = newDocument.version;
par.date = now;
//loops through the sentences verifying the content on the paragraph
foreach (string s in sentences)
{
refPars = (List <DocumentJSON.Paragraph>)oldParList.Where(x => x.content.Contains(s.Replace(".", ""))).ToList();
if (refPars.Count > 0)
{
foreach (DocumentJSON.Paragraph refPar in refPars)
{
if (similarityList.Where(x => x.Id == refPar.id).Count() > 0)
{
similarityList.Where(x => x.Id == refPar.id).First().Text += 1;
}
else
{
similarityList.Add(new KeyVal <int, int>(refPar.id, 1));
}
}
}
else
{
DocumentJSON.Sentence stc = new DocumentJSON.Sentence()
{
content = s, status = "m"
};
sentenceList.Add(stc);
}
}
KeyVal <int, int> mostSimilarPar = new KeyVal <int, int>(-1, -1);
foreach (KeyVal <int, int> key in similarityList)
{
if (key.Text > mostSimilarPar.Text)
{
mostSimilarPar = key;
}
}
//if the content has more than 50% of similarity the paragraph is considered as modified
if (mostSimilarPar.Text / ((double)sentences.Count()) > paragraphSensivity)
{
DocumentJSON.Paragraph oldPar = oldParList.Where(x => x.id == mostSimilarPar.Id).First();
string[] oldSentences = oldPar.content.Split('.');
foreach (string s in oldSentences)
{
int i = 0;
foreach (string c in sentences)
{
sentences[i] = c.Trim();
i++;
}
if (!sentences.Contains(s.Trim()))
{
sentenceList.Add(new DocumentJSON.Sentence()
{
content = s, status = "d"
});
}
}
par.status = "m";
par.id = mostSimilarPar.Id;
par.sentence = sentenceList.ToArray();
if (refPars.Where(x => x.id == mostSimilarPar.Id).Count() > 0)
{
oldParList.Remove(refPars.Where(x => x.id == mostSimilarPar.Id).First());
}
}
//if the paragraph has less or equal to 50% of similarity, it is considered as new
else
{
par.status = "n";
//gets a new id for the paragraph
par.id = ++newDocument.lastParId;
//refDocParList.Remove(refPar.First());
}
}
}
//Paragraphs not found are added as deleted on the document.
foreach (DocumentJSON.Paragraph par in oldParList)
{
if (!par.status.Equals("d"))
{
par.status = "d";
par.version = newDocument.version;
par.date = DateTime.Now.ToString("yyyymmdd");
newParList.Add(par);
}
}
newDocument.paragraphs = newParList.ToArray();
return(newDocument);
}