void CMultiWordCreator::NormalizeMultiWordHomonym(CWord* pW, CHomonym* pH)
{
CWordSequence* pWS = pH->GetSourceWordSequence();
pWS->ClearLemmas();
Wtroka lemma_text;
if (pWS->HasAuxArticle()) {
const article_t* pArt = GlobalDictsHolder->GetAuxArticle(pWS->GetAuxArticleIndex());
if (pArt->has_lemma())
lemma_text = pArt->get_lemma();
} else if (pWS->HasGztArticle()) {
const TGztArticle& gzt_article = pWS->GetGztArticle();
const NGzt::TMessage* lemma_info = gzt_article.GetLemmaInfo();
if (lemma_info != NULL)
lemma_text = gzt_article.GetLemmaText(*lemma_info);
}
if (!lemma_text.empty()) {
//если есть лемма, указанная в статье
Wtroka capitalizedLemma = CNormalization::GetCapitalizedLemma(*pW, lemma_text);
pWS->AddLemma(SWordSequenceLemma(lemma_text, capitalizedLemma));
} else {
if (pWS->Size() == 1) {
// for single-word sequences - just take a lemma of corresponding word
pWS->AddLemma(SWordSequenceLemma(m_Words[pWS->GetMainWord()].GetLemma()), true);
} else for (int i = pWS->FirstWord(); i <= pWS->LastWord(); ++i) {
// not normalized now, will be normalized lazily on request
pWS->AddLemma(SWordSequenceLemma(m_Words[i].GetText()), false);
}
}
}
void CSitFactInterpretation::FillConstFieldValue(const fact_field_reference_t& fact_field, yvector<CFactFields>& newFacts, const SWordHomonymNum& valValue)
{
const CWord& w = m_Words.GetWord(valValue);
switch (fact_field.m_Field_type) {
case TextField: {
CTextWS ws;
ws.SetArtificialPair(w.GetSourcePair());
ws.AddLemma(SWordSequenceLemma(fact_field.m_StringValue));
FillWSFactField(fact_field, ws, newFacts);
break;
} case BoolField: {
CBoolWS ws(fact_field.m_bBoolValue);
ws.SetPair(w.GetSourcePair());
FillWSFactField(fact_field, ws, newFacts);
break;
} default:
break;
}
}
void CQuotesFinder::CreateTextField(const Stroka& TextFieldName, CSentence* pSent, const SWordHomonymNum& WH,
CFactFields& factFields, const Wtroka& Postfix)
{
CNormalization InterpNorm(pSent->m_Words);
const CHomonym& h = pSent->m_Words[WH];
const CFioWordSequence* pFioWS = dynamic_cast<const CFioWordSequence*>(h.GetSourceWordSequence());
if (pFioWS != NULL) {
CFioWS newFioWS(*pFioWS);
newFioWS.SetMainWord(WH);
newFioWS.AddLemma(SWordSequenceLemma (InterpNorm.GetArtificialLemma(WH)));
CTextWS newTextWS;
newTextWS.SetPair(*pFioWS);
newTextWS.SetMainWord(WH);
newTextWS.ResetLemmas(newFioWS.GetLemmas(), true);
if (!Postfix.empty())
newTextWS.AddLemma(SWordSequenceLemma(Postfix));
factFields.AddValue(TextFieldName, newTextWS);
} else {
const CFactsWS* pFactWS = dynamic_cast<const CFactsWS*>(h.GetSourceWordSequence());
if (pFactWS) {
CTextWS newTextWS;
newTextWS.SetPair(*pFactWS);
newTextWS.SetMainWord(WH);
newTextWS.ResetLemmas(pFactWS->GetLemmas(), true);
if (!Postfix.empty())
newTextWS.AddLemma(SWordSequenceLemma(Postfix));
factFields.AddValue(TextFieldName, newTextWS);
} else {
const CTextWS* pTextWS = dynamic_cast<const CTextWS*>(h.GetSourceWordSequence());
if (pTextWS) {
CTextWS newTextWS = *pTextWS;
if (!Postfix.empty())
newTextWS.AddLemma(SWordSequenceLemma(Postfix));
factFields.AddValue(TextFieldName, newTextWS);
} else {
if (h.GetSourceWordSequence() == 0 || h.GetSourceWordSequence()->Size() == 1) {
CTextWS newTextWS;
newTextWS.SetPair(pSent->m_Words.GetWord(WH).GetSourcePair());
newTextWS.SetMainWord(WH);
newTextWS.AddLemma(SWordSequenceLemma(h.GetLemma()));
if (!Postfix.empty())
newTextWS.AddLemma(SWordSequenceLemma(Postfix));
factFields.AddValue(TextFieldName, newTextWS);
}
}
}
}
}
bool CQuotesFinder::CreateQuoteValue(const Wtroka& quoteStr, CFactFields& factFields) const
{
DECLARE_STATIC_RUS_WORD(kChto, "что ");
DECLARE_STATIC_RUS_WORD(kChtoby, "чтобы ");
Wtroka qstr = quoteStr;
CTextWS QuoteWS;
if (qstr.has_prefix(kChto))
qstr.erase(0,4);
if (qstr.has_prefix(kChtoby))
qstr.erase(0,6);
if (qstr.empty())
return false;
qstr = StripString(qstr);
if (qstr.size() > 1 && qstr[0] =='"' && qstr.back() =='"')
qstr = qstr.substr(1, qstr.size() - 2);
qstr = StripString(qstr);
QuoteWS.AddLemma(SWordSequenceLemma(qstr));
QuoteWS.SetArtificialPair(CWordsPair(0,0));
factFields.AddValue(CFactFields::QuoteValue, QuoteWS);
return true;
}
void CSentence::AddFIOWS(const CFIOOccurence& fioOccurence, const SFullFIO& fio, int iSimilarOccurencesCount)
{
TIntrusivePtr<CFioWordSequence> fioWS(new CFioWordSequence(fio));
*(CFIOOccurence*)(fioWS.Get()) = fioOccurence;
fioWS->PutWSType(FioWS);
if (fio.m_Genders.any()) {
THomonymGrammems gram = fioWS->GetGrammems();
gram.Replace(NSpike::AllGenders, fio.m_Genders);
fioWS->SetGrammems(gram);
}
fioWS->m_iSimilarOccurencesCount = iSimilarOccurencesCount;
bool isManualFio = true;
SWordHomonymNum wh = fioOccurence.m_NameMembers[Surname];
if (wh.IsValid())
if (m_Words.GetWord(wh).IsMultiWord())
isManualFio = false;
if (!fio.m_bFoundSurname &&
fioOccurence.m_NameMembers[Surname].IsValid() &&
!(fioOccurence.m_NameMembers[FirstName].IsValid() ||
fioOccurence.m_NameMembers[InitialName].IsValid()) &&
isManualFio) {
CNameFinder nameFinder(m_Words);
//если не смогли среди предсказанных фамилий
//найти совпадающую с фамилией из fio, то вываливаемся
if (!nameFinder.PredictSingleSurname(*fioWS, fio))
return;
}
fioWS->ClearLemmas();
if (!fio.m_strSurname.empty()) {
Wtroka capLemma;
if (fioOccurence.m_NameMembers[Surname].IsValid()) {
const CWord& w = m_Words.GetWord(fioOccurence.m_NameMembers[Surname]);
capLemma = GetCapitalizedLemma(w, -1, fio.m_strSurname);
} else {
capLemma = fio.m_strSurname;
NStr::ToFirstUpper(capLemma);
}
fioWS->AddLemma(SWordSequenceLemma(fio.m_strSurname, capLemma));
}
if (!fio.m_strName.empty()) {
Wtroka capLemma;
if (fioOccurence.m_NameMembers[FirstName].IsValid()) {
const CWord& w = m_Words.GetWord(fioOccurence.m_NameMembers[FirstName]);
capLemma = GetCapitalizedLemma(w, -1, fio.m_strName);
} else if (fioOccurence.m_NameMembers[InitialName].IsValid()) {
const CWord& w = m_Words.GetWord(fioOccurence.m_NameMembers[InitialName]);
capLemma = GetCapitalizedLemma(w, -1, fio.m_strName);
} else {
capLemma = fio.m_strName;
TMorph::ToTitle(capLemma);
}
fioWS->AddLemma(SWordSequenceLemma(fio.m_strName, capLemma));
}
if (!fio.m_strPatronomyc.empty()) {
Wtroka capLemma;
if (fioOccurence.m_NameMembers[Patronomyc].IsValid()) {
const CWord& w = m_Words.GetWord(fioOccurence.m_NameMembers[Patronomyc]);
capLemma = GetCapitalizedLemma(w, -1, fio.m_strPatronomyc);
} else if (fioOccurence.m_NameMembers[InitialPatronomyc].IsValid()) {
const CWord& w = m_Words.GetWord(fioOccurence.m_NameMembers[InitialPatronomyc]);
capLemma = GetCapitalizedLemma(w, -1, fio.m_strPatronomyc);
} else {
capLemma = fio.m_strPatronomyc;
TMorph::ToTitle(capLemma);
}
fioWS->AddLemma(SWordSequenceLemma(fio.m_strPatronomyc, capLemma));
}
TakeFioWS(fioWS);
}
