private void Disambiguate(ArrayList sparsetree)
{
foreach (ParseTree ps in SParseTrees)
{
ListifyParseTree(ps);
PrepareInputArrOfDisambiguation();
double Score;
WordSenseDisambiguator wsd = new WordSenseDisambiguator();
DisambiguateRes = new MyWordInfo[size];
DisambiguateRes = wsd.MMG_Disambiguate(mwi2, out Score);
ScoresOfParseTree.Add(Score);
Senses.Add(DisambiguateRes);
}
checkParseTreeRepitition();
// AddArrStems(NewSParseTrees);
// fe problem fel disambiguation law ra7laha parsetrees fehom kaza pt l 1 sintence
//law fe gomla liha 3 parse trees el code da hayetsaraf ?
}
public void ConstructMapping()
{
string concept = "";
string word = "";
int ID = -1;
string senseNo = "";
string Sense = "";
string Pos = "";
WordOlogy WO = new WordOlogy();
ArrayList wordologyArr = new ArrayList();
int conceptcounter = 0;
LoadOntology();
FileStream allConceptsFile = new FileStream(_ontologyDirectoryPath + @"\AllConcepts.txt", FileMode.Open);
StreamReader allConceptsFileReader = new StreamReader(allConceptsFile);
string _wordologyDirectoryPath =
@"..\..\..\wordology\";
BinaryFormatter bf = new BinaryFormatter();
FileStream fs = new FileStream(
_wordologyDirectoryPath + "\\wordology.txt", FileMode.Create);
int indxWatcherconceptCounter = 0;
int NoMapLexConcepts = 0;
int CannotGetSenseExeption = 0;
int AllSensesMapped = 0;
while ((concept = allConceptsFileReader.ReadLine()) != null)
{
indxWatcherconceptCounter++;
string Conceptpath = _ontologyDirectoryPath + @"\" + concept[0] + @"\" + concept;
Concept C = (Concept)Onto[concept];
Property maplexProperty = C.FullProperties["ENGLISH1"];
List <MyWordInfo> maplexsenses = new List <MyWordInfo>();
MyWordInfo mwi = new MyWordInfo();
int NoOfSensesSucceeded = 0;
if (maplexProperty != null)
{
for (int i = 0; i < maplexProperty.Fillers.Count; i++)
{
string tmp = maplexProperty.Fillers[i].ScalarFiller;
char[] charr = new char[] { '-', '_' };
string[] splt = tmp.Split(charr);
//there r fillers with no type & a-bomb masalan
if (splt.Length > 1)
{
mwi = new MyWordInfo();
for (int k = 0; k < splt.Length - 2; k++)
{
mwi.Word += splt[k] + " ";
}
mwi.Word += splt[splt.Length - 2];
if (splt[splt.Length - 1].Length == 2)
{
if (splt[splt.Length - 1][0] == 'v')
{
mwi.Pos = Wnlib.PartsOfSpeech.Verb;
}
else if (splt[splt.Length - 1][0] == 'n')
{
mwi.Pos = Wnlib.PartsOfSpeech.Noun;
}
else if (splt[splt.Length - 1][0] == 'a')
{
mwi.Pos = Wnlib.PartsOfSpeech.Adj;
}
else if (splt[splt.Length - 1][0] == 'r')
{
mwi.Pos = Wnlib.PartsOfSpeech.Adv;
}
else
{
mwi.Pos = Wnlib.PartsOfSpeech.Unknown;
}
}
else
{
mwi.Pos = Wnlib.PartsOfSpeech.Unknown;
mwi.Word += " " + splt[splt.Length - 1];
}
if (i == 0 || (maplexsenses.Count > 0 && (mwi.Word != maplexsenses[maplexsenses.Count - 1].Word || mwi.Pos != maplexsenses[maplexsenses.Count - 1].Pos)))
{
maplexsenses.Add(mwi);
}
}
//ne loop 3al ontology kolaha
}
if (maplexsenses.Count > 0)
{
MyWordInfo[] maplexArray = new MyWordInfo[maplexsenses.Count];
for (int j = 0; j < maplexsenses.Count; j++)
{
maplexArray[j] = maplexsenses[j];
}
WordSenseDisambiguator wsd = new WordSenseDisambiguator();
MyWordInfo[] res = new MyWordInfo[maplexArray.Length];
res = wsd.Disambiguate(maplexArray);
int i = 0;
foreach (MyWordInfo wi in res)
{
string tmp = maplexProperty.Fillers[i].ScalarFiller;
char[] charr = new char[] { '-', '_' };
string[] splt = tmp.Split(charr);
if (splt.Length > 1 && splt[splt.Length - 1].Length == 2)
{
WO.SenseNo = splt[splt.Length - 1];
}
else
{
// "sense doesn't have POS";
}
Wnlib.PartOfSpeech p = Wnlib.PartOfSpeech.of((Wnlib.PartsOfSpeech)wi.Pos);
try
{
Wnlib.Index index = Wnlib.Index.lookup(wi.Word.ToLower(), p);
SynSet sense = new SynSet(index, res[i].Sense, null);
WO.Sense = sense.defn;
AllSensesMapped++;
NoOfSensesSucceeded++;
try
{
WO.Pos = p.name;
}
catch
{
WO.Pos = wi.Pos.ToString();
}
ID++;
WO.Word = wi.Word;
WO.ID = ID;
WO.Concept = concept;
WO.Word = word;
}
catch
{
};
if (NoOfSensesSucceeded == 0)
{
CannotGetSenseExeption++;
}
i++;
// bf.Serialize(fs, "\n" + WO);
wordologyArr.Add(WO);
}
conceptcounter++;
}
}
else
{
NoMapLexConcepts++;
//new part
Wnlib.Index index;
Wnlib.PartOfSpeech p;
Search se;
try
{
index = Wnlib.Index.lookup(concept.ToLower(), PartOfSpeech.of(PartsOfSpeech.Noun));
if (index != null)
{
WO.Pos = "noun";
Opt[] relatedness = WordsMatching.Relatedness.GetRelatedness(PartsOfSpeech.Noun);
foreach (Opt o in relatedness)
{
for (int senseNumber = 0; senseNumber < index.sense_cnt; senseNumber++)
{
se = new Search(concept, true, PartOfSpeech.of("noun"), o.sch, senseNumber);
SynSet sense = new SynSet(index, senseNumber, se);
WO.Concept = concept;
WO.Word = concept;
WO.Sense = sense.defn;
WO.ID = ID;
ID++;
NoOfSensesSucceeded++;
AllSensesMapped++;
//bf.Serialize(fs, "\n" + WO);
wordologyArr.Add(WO);
}
}
}
}
catch
{ }
try
{
index = Wnlib.Index.lookup(concept.ToLower(), PartOfSpeech.of(PartsOfSpeech.Verb));
if (index != null)
{
WO.Pos = "verb";
Opt[] relatedness = WordsMatching.Relatedness.GetRelatedness(PartsOfSpeech.Verb);
foreach (Opt o in relatedness)
{
for (int senseNumber = 0; senseNumber < index.sense_cnt; senseNumber++)
{
se = new Search(concept, true, PartOfSpeech.of("verb"), o.sch, senseNumber);
SynSet sense = new SynSet(index, senseNumber, se);
WO.Sense = sense.defn;
WO.Concept = concept;
WO.Word = concept;
WO.ID = ID;
ID++;
NoOfSensesSucceeded++;
AllSensesMapped++;
//bf.Serialize(fs, "\n" + WO);
wordologyArr.Add(WO);
}
}
}
}
catch
{ }
try
{
index = Wnlib.Index.lookup(concept.ToLower(), PartOfSpeech.of(PartsOfSpeech.Adj));
if (index != null)
{
WO.Pos = "adj";
Opt[] relatedness = WordsMatching.Relatedness.GetRelatedness(PartsOfSpeech.Adj);
foreach (Opt o in relatedness)
{
for (int senseNumber = 0; senseNumber < index.sense_cnt; senseNumber++)
{
se = new Search(concept, true, PartOfSpeech.of("adj"), o.sch, senseNumber);
SynSet sense = new SynSet(index, senseNumber, se);
WO.Sense = sense.defn;
WO.Concept = concept;
WO.Word = concept;
WO.ID = ID;
ID++;
NoOfSensesSucceeded++;
AllSensesMapped++;
//bf.Serialize(fs, "\n" + WO);
wordologyArr.Add(WO);
}
}
}
}
catch
{ }
try
{
index = Wnlib.Index.lookup(concept.ToLower(), PartOfSpeech.of(PartsOfSpeech.Adv));
if (index != null)
{
WO.Pos = "adv";
Opt[] relatedness = WordsMatching.Relatedness.GetRelatedness(PartsOfSpeech.Noun);
foreach (Opt o in relatedness)
{
for (int senseNumber = 0; senseNumber < index.sense_cnt; senseNumber++)
{
se = new Search(concept, true, PartOfSpeech.of("adv"), o.sch, senseNumber);
SynSet sense = new SynSet(index, senseNumber, se);
WO.Sense = sense.defn;
WO.Concept = concept;
WO.Word = concept;
WO.ID = ID;
ID++;
NoOfSensesSucceeded++;
AllSensesMapped++;
//bf.Serialize(fs, "\n" + WO);
wordologyArr.Add(WO);
}
}
}
}
catch
{ }
if (NoOfSensesSucceeded != 0)
{
conceptcounter++;
}
}
}//end while
allConceptsFileReader.Close();
allConceptsFile.Close();
bf.Serialize(fs, wordologyArr);
fs.Close();
MessageBox.Show("no map-lex concepts number = " + NoMapLexConcepts.ToString());
MessageBox.Show("can't getsense pos number = " + CannotGetSenseExeption.ToString());
MessageBox.Show(conceptcounter.ToString());
}
private void FillSenses(ParseTree parsetree, ParseNode node, ref ArrayList wordinfoArr, ref int j)
{
if (node.Children != null)
{
for (int i = 0; i < node.Children.Count; i++)
{
ParseNode pn = (ParseNode)node.Children[i];
FillSenses(parsetree, pn, ref wordinfoArr, ref j);
}
}
else
{
SentenceParser dummysp = new SentenceParser();
string str = node.Goal;
if (str == "PPN")
{
WordSenseDisambiguator wsd = new WordSenseDisambiguator();
double score;
double num;
if (double.TryParse(node.Text, out num))
{
MyWordInfo[] ret = wsd.MMG_Disambiguate(new MyWordInfo[] { new MyWordInfo("Number", PartsOfSpeech.Noun) }, out score);
PartOfSpeech pos = PartOfSpeech.of(PartsOfSpeech.Noun);
Index index = Wnlib.Index.lookup(ret[0].Word.ToLower(), pos);
SynSet sense = new SynSet(index, ret[0].Sense, null);
node.Sense = sense.defn;
node.SenseNo = ret[0].Sense;
}
else
{
MyWordInfo[] ret = wsd.MMG_Disambiguate(new MyWordInfo[] { new MyWordInfo("proper_name", PartsOfSpeech.Noun) }, out score);
PartOfSpeech pos = PartOfSpeech.of(PartsOfSpeech.Noun);
Index index = Wnlib.Index.lookup(ret[0].Word.ToLower(), pos);
SynSet sense = new SynSet(index, ret[0].Sense, null);
node.Sense = sense.defn;
node.SenseNo = ret[0].Sense;
}
}
else
if (str == "N" || str.Contains("NPP") || (str == "VING") || str.Contains("PPJ") || (str == "VPSP") || (str == "BE1") || (str == "BE2") || (str == "V") || str.Contains("CPADJ") || str.Contains("ADJ") || str.Contains("PADV") || str.Contains("ADV") || str == "VINF")
{
///ta3deelat 5/7/////////
string [] spltstr = DisambRes[j].ToString().Split(':');
VerbSense VS = new VerbSense();
string[] spltstr2 = new string[10];
string[] spltstr3 = new string[10];
if (node.Senses != null && node.Goal.Contains("V"))
{
VS = (VerbSense)node.Senses[0];
spltstr2 = VS.Sense.ToString().Split('#');
spltstr3 = spltstr[0].Split(' ');
ArrayList results = dummysp.GetINFOfVerb(spltstr2[0]);
if (results.Count > 0)
{
spltstr2[0] = (string)results[0];
}
if (spltstr3[0] == spltstr2[0])
{
if (NodesSenses.Count > 0)
{
node.Sense = (string)NodesSenses[0];
node.SenseNo = (int)SensesNos[0];
SensesNos.RemoveAt(0);
NodesSenses.RemoveAt(0);
j++;
}
}
}
if (str == "N")
{
node.Sense = (string)NodesSenses[0];
node.SenseNo = (int)SensesNos[0];
SensesNos.RemoveAt(0);
NodesSenses.RemoveAt(0);
j++;
}
else
{
string NodeWord = SyntacticAnalyzer.SentenceParser.GetWordString(parsetree, node);
spltstr3 = spltstr[0].Split(' ');
string nodeWord = NodeWord.ToLower();
if (node.Goal == "V" || node.Goal == "BE1" || node.Goal == "VINF" || node.Goal == "VPSP" || node.Goal == "VING")
{
ArrayList results = dummysp.GetINFOfVerb(nodeWord);
if (results.Count > 0)
{
nodeWord = (string)results[0];
}
}
string dummy = spltstr3[0];
dummy = dummy.Remove(dummy.Length - 1);
dummy = dummy + "ies";
string dummy2 = nodeWord + "ing";
string dummy3 = dummy2 + "s";
string dummy4 = nodeWord.Remove(nodeWord.Length - 1);
dummy4 = dummy4 + "ing";
if (spltstr3[0] == nodeWord || spltstr3[0] + 's' == nodeWord || spltstr3[0] + "es" == nodeWord || dummy == nodeWord || spltstr3[0] == dummy2 || spltstr3[0] == dummy3 || spltstr3[0] == dummy4 || spltstr3[0] == dummy4 + 's')
{
if (NodesSenses.Count > 0)
{
node.Sense = (string)NodesSenses[0];
node.SenseNo = (int)SensesNos[0];
SensesNos.RemoveAt(0);
NodesSenses.RemoveAt(0);
j++;
}
}
}
}
}
}
private void MapConceptsWithMapLex(string concept, Property maplexProperty)
{
MyWordInfo mwi;
WordOlogy WO = new WordOlogy();
List <MyWordInfo> maplexsenses = new List <MyWordInfo>();
int NoOfSensesSucceeded = 0;
for (int i = 0; i < maplexProperty.Fillers.Count; i++)
{
string tmp = maplexProperty.Fillers[i].ScalarFiller;
char[] charr = new char[] { '-', '_' };
string[] splt = tmp.Split(charr);
//there r fillers with no type & a-bomb masalan
if (splt.Length > 1)
{
mwi = new MyWordInfo();
for (int k = 0; k < splt.Length - 2; k++)
{
mwi.Word += splt[k] + " ";
}
mwi.Word += splt[splt.Length - 2];
if (splt[splt.Length - 1].Length == 2)
{
if (splt[splt.Length - 1][0] == 'v')
{
mwi.Pos = Wnlib.PartsOfSpeech.Verb;
}
else if (splt[splt.Length - 1][0] == 'n')
{
mwi.Pos = Wnlib.PartsOfSpeech.Noun;
}
else if (splt[splt.Length - 1][0] == 'a')
{
mwi.Pos = Wnlib.PartsOfSpeech.Adj;
}
else if (splt[splt.Length - 1][0] == 'r')
{
mwi.Pos = Wnlib.PartsOfSpeech.Adv;
}
else
{
mwi.Pos = Wnlib.PartsOfSpeech.Unknown;
}
}
else
{
mwi.Pos = Wnlib.PartsOfSpeech.Unknown;
mwi.Word += " " + splt[splt.Length - 1];
}
if (i == 0 || (maplexsenses.Count > 0 && (mwi.Word != maplexsenses[maplexsenses.Count - 1].Word || mwi.Pos != maplexsenses[maplexsenses.Count - 1].Pos)))
{
maplexsenses.Add(mwi);
}
}
//ne loop 3al ontology kolaha
}
if (maplexsenses.Count > 0)
{
MyWordInfo[] maplexArray = new MyWordInfo[maplexsenses.Count];
for (int j = 0; j < maplexsenses.Count; j++)
{
maplexArray[j] = maplexsenses[j];
}
WordSenseDisambiguator wsd = new WordSenseDisambiguator();
MyWordInfo[] res = new MyWordInfo[maplexArray.Length];
res = wsd.Disambiguate(maplexArray);
int i = 0;
foreach (MyWordInfo wi in res)
{
string tmp = maplexProperty.Fillers[i].ScalarFiller;
char[] charr = new char[] { '-', '_' };
string[] splt = tmp.Split(charr);
if (splt.Length > 1 && splt[splt.Length - 1].Length == 2)
{
WO.SenseNo = splt[splt.Length - 1];
}
else
{
// "sense doesn't have POS";
}
Wnlib.PartOfSpeech p = Wnlib.PartOfSpeech.of((Wnlib.PartsOfSpeech)wi.Pos);
try
{
Wnlib.Index index = Wnlib.Index.lookup(wi.Word.ToLower(), p);
SynSet sense = new SynSet(index, res[i].Sense, null);
WO.Sense = sense.defn;
// AllSensesMapped++;
NoOfSensesSucceeded++;
try
{
WO.Pos = p.name;
}
catch
{
WO.Pos = wi.Pos.ToString();
}
ID++;
WO.Word = wi.Word;
WO.ID = ID;
WO.Concept = concept;
wordologyArr.Add(WO);
}
catch
{
};
if (NoOfSensesSucceeded == 0)
{
CannotGetSenseExeption++;
}
i++;
//bf.Serialize(fs, "\n" + WO);
}
conceptcounter++;
}
}
