private void CollectCoordinatedNounPhraseMentions(IParse nounPhrase, List <Mention> entities)
{
//System.err.println("collectCoordNp: "+np);
List <IParse> nounPhraseTokens = nounPhrase.Tokens;
bool inCoordinatedNounPhrase = false;
int lastNounPhraseTokenIndex = mHeadFinder.GetHeadIndex(nounPhrase);
for (int tokenIndex = lastNounPhraseTokenIndex - 1; tokenIndex >= 0; tokenIndex--)
{
IParse token = nounPhraseTokens[tokenIndex];
string tokenText = token.ToString();
if (tokenText == "and" || tokenText == "or")
{
if (lastNounPhraseTokenIndex != tokenIndex)
{
if (tokenIndex - 1 >= 0 && (nounPhraseTokens[tokenIndex - 1]).SyntacticType.StartsWith("NN"))
{
Util.Span nounPhraseSpan = new Util.Span((nounPhraseTokens[tokenIndex + 1]).Span.Start, (nounPhraseTokens[lastNounPhraseTokenIndex]).Span.End);
Mention nounPhraseSpanExtent = new Mention(nounPhraseSpan, nounPhraseSpan, token.EntityId, null, "CNP");
entities.Add(nounPhraseSpanExtent);
//System.err.println("adding extent for conjunction in: "+np+" preeceeded by "+((Parse) npTokens.get(ti-1)).getSyntacticType());
inCoordinatedNounPhrase = true;
}
else
{
break;
}
}
lastNounPhraseTokenIndex = tokenIndex - 1;
}
else if (inCoordinatedNounPhrase && tokenText.Equals(","))
{
if (lastNounPhraseTokenIndex != tokenIndex)
{
Util.Span nounPhraseSpan = new Util.Span((nounPhraseTokens[tokenIndex + 1]).Span.Start, (nounPhraseTokens[lastNounPhraseTokenIndex]).Span.End);
Mention nounPhraseSpanExtent = new Mention(nounPhraseSpan, nounPhraseSpan, token.EntityId, null, "CNP");
entities.Add(nounPhraseSpanExtent);
//System.err.println("adding extent for comma in: "+np);
}
lastNounPhraseTokenIndex = tokenIndex - 1;
}
else if (inCoordinatedNounPhrase && tokenIndex == 0 && lastNounPhraseTokenIndex >= 0)
{
Util.Span nounPhraseSpan = new Util.Span((nounPhraseTokens[tokenIndex]).Span.Start, (nounPhraseTokens[lastNounPhraseTokenIndex]).Span.End);
Mention nounPhraseSpanExtent = new Mention(nounPhraseSpan, nounPhraseSpan, token.EntityId, null, "CNP");
entities.Add(nounPhraseSpanExtent);
//System.err.println("adding extent for start coord in: "+np);
}
}
}
private void CollectCoordinatedNounPhraseMentions(IParse nounPhrase, List <Mention> entities)
{
List <IParse> nounPhraseTokens = nounPhrase.Tokens;
bool inCoordinatedNounPhrase = false;
int lastNounPhraseTokenIndex = mHeadFinder.GetHeadIndex(nounPhrase);
for (int tokenIndex = lastNounPhraseTokenIndex - 1; tokenIndex >= 0; tokenIndex--)
{
IParse token = nounPhraseTokens[tokenIndex];
string tokenText = token.ToString();
if (tokenText == "and" || tokenText == "or")
{
if (lastNounPhraseTokenIndex != tokenIndex)
{
if (tokenIndex - 1 >= 0 && PartsOfSpeech.IsNoun(nounPhraseTokens[tokenIndex - 1].SyntacticType))
{
var nounPhraseSpan = new Util.Span((nounPhraseTokens[tokenIndex + 1]).Span.Start, (nounPhraseTokens[lastNounPhraseTokenIndex]).Span.End);
var nounPhraseSpanExtent = new Mention(nounPhraseSpan, nounPhraseSpan, token.EntityId, null, "CNP");
entities.Add(nounPhraseSpanExtent);
inCoordinatedNounPhrase = true;
}
else
{
break;
}
}
lastNounPhraseTokenIndex = tokenIndex - 1;
}
else if (inCoordinatedNounPhrase && tokenText == PartsOfSpeech.Comma)
{
if (lastNounPhraseTokenIndex != tokenIndex)
{
var nounPhraseSpan = new Util.Span((nounPhraseTokens[tokenIndex + 1]).Span.Start, (nounPhraseTokens[lastNounPhraseTokenIndex]).Span.End);
var nounPhraseSpanExtent = new Mention(nounPhraseSpan, nounPhraseSpan, token.EntityId, null, "CNP");
entities.Add(nounPhraseSpanExtent);
}
lastNounPhraseTokenIndex = tokenIndex - 1;
}
else if (inCoordinatedNounPhrase && tokenIndex == 0 && lastNounPhraseTokenIndex >= 0)
{
var nounPhraseSpan = new Util.Span((nounPhraseTokens[tokenIndex]).Span.Start, (nounPhraseTokens[lastNounPhraseTokenIndex]).Span.End);
var nounPhraseSpanExtent = new Mention(nounPhraseSpan, nounPhraseSpan, token.EntityId, null, "CNP");
entities.Add(nounPhraseSpanExtent);
}
}
}
public MentionContext(Util.Span span, Util.Span headSpan, int entityId, IParse parse, string extentType, string nameType, int mentionIndex, int mentionsInSentence, int mentionIndexInDocument, int sentenceIndex, IHeadFinder headFinder) : base(span, headSpan, entityId, parse, extentType, nameType, headFinder)
{
mNounLocation = mentionIndex;
mMaxNounLocation = mentionsInSentence;
mNounNumber = mentionIndexInDocument;
mSentenceNumber = sentenceIndex;
mIndexSpan = parse.Span;
mPreviousToken = parse.PreviousToken;
mNextToken = parse.NextToken;
mHead = headFinder.GetLastHead(parse);
List <IParse> headTokens = mHead.Tokens;
Tokens = headTokens.ToArray();
mBasalNextToken = mHead.NextToken;
//System.err.println("MentionContext.init: "+ent+" "+ent.getEntityId()+" head="+head);
//mNonDescriptorStart = 0;
InitializeHeads(headFinder.GetHeadIndex(mHead));
mGender = Similarity.GenderEnum.Unknown;
mGenderProbability = 0d;
mNumber = Similarity.NumberEnum.Unknown;
mNumberProbability = 0d;
}
public MentionContext(Util.Span span, Util.Span headSpan, int entityId, IParse parse, string extentType, string nameType,
int mentionIndex, int mentionsInSentence, int mentionIndexInDocument, int sentenceIndex, IHeadFinder headFinder) :
base(span, headSpan, entityId, parse, extentType, nameType, headFinder)
{
NounPhraseSentenceIndex = mentionIndex;
MaxNounPhraseSentenceIndex = mentionsInSentence;
NounPhraseDocumentIndex = mentionIndexInDocument;
SentenceNumber = sentenceIndex;
IndexSpan = parse.Span;
PreviousToken = parse.PreviousToken;
NextToken = parse.NextToken;
Head = headFinder.GetLastHead(parse);
List <IParse> headTokens = Head.Tokens;
Tokens = headTokens.ToArray();
NextTokenBasal = Head.NextToken;
//mNonDescriptorStart = 0;
InitializeHeads(headFinder.GetHeadIndex(Head));
mGender = Similarity.GenderEnum.Unknown;
GenderProbability = 0d;
_number = Similarity.NumberEnum.Unknown;
NumberProbability = 0d;
}
public MentionContext(Util.Span span, Util.Span headSpan, int entityId, IParse parse, string extentType, string nameType, int mentionIndex, int mentionsInSentence, int mentionIndexInDocument, int sentenceIndex, IHeadFinder headFinder)
: base(span, headSpan, entityId, parse, extentType, nameType, headFinder)
{
mNounLocation = mentionIndex;
mMaxNounLocation = mentionsInSentence;
mNounNumber = mentionIndexInDocument;
mSentenceNumber = sentenceIndex;
mIndexSpan = parse.Span;
mPreviousToken = parse.PreviousToken;
mNextToken = parse.NextToken;
mHead = headFinder.GetLastHead(parse);
List<IParse> headTokens = mHead.Tokens;
Tokens = headTokens.ToArray();
mBasalNextToken = mHead.NextToken;
//System.err.println("MentionContext.init: "+ent+" "+ent.getEntityId()+" head="+head);
//mNonDescriptorStart = 0;
InitializeHeads(headFinder.GetHeadIndex(mHead));
mGender = Similarity.GenderEnum.Unknown;
mGenderProbability = 0d;
mNumber = Similarity.NumberEnum.Unknown;
mNumberProbability = 0d;
}
public MentionContext(Util.Span span, Util.Span headSpan, int entityId, IParse parse, string extentType, string nameType,
int mentionIndex, int mentionsInSentence, int mentionIndexInDocument, int sentenceIndex, IHeadFinder headFinder):
base(span, headSpan, entityId, parse, extentType, nameType, headFinder)
{
NounPhraseSentenceIndex = mentionIndex;
MaxNounPhraseSentenceIndex = mentionsInSentence;
NounPhraseDocumentIndex = mentionIndexInDocument;
SentenceNumber = sentenceIndex;
IndexSpan = parse.Span;
PreviousToken = parse.PreviousToken;
NextToken = parse.NextToken;
Head = headFinder.GetLastHead(parse);
List<IParse> headTokens = Head.Tokens;
Tokens = headTokens.ToArray();
NextTokenBasal = Head.NextToken;
//mNonDescriptorStart = 0;
InitializeHeads(headFinder.GetHeadIndex(Head));
mGender = Similarity.GenderEnum.Unknown;
GenderProbability = 0d;
_number = Similarity.NumberEnum.Unknown;
NumberProbability = 0d;
}