void CWord::DeleteHomonym(CHomonym& hom)
{
if (!hom.IsDeleted()) {
++m_DeletedHomCount;
hom.Delete();
}
}
void CHomonym::ResetGenderFromHomonym(const CHomonym& h)
{
if (h.HasGrammem(gFeminine))
Grammems.Replace(gMasculine, gFeminine);
if (h.HasGrammem(gMasculine))
Grammems.Replace(gFeminine, gMasculine);
}
void CMultiWordCreator::AddQuoteMultiWord(CWordSequence& ws, const TArticleRef& article)
{
SWordHomonymNum wh;
Wtroka str;
CWord* pNewWord = GetWordForMultiWord(ws, str, wh);
if (pNewWord->m_SourceWords.Size() == 1 && pNewWord->HasOnlyUnknownPOS()) {
size_t firstId = pNewWord->IterHomonyms().GetID();
CHomonym& h = pNewWord->GetRusHomonym(firstId);
h.SetSourceWordSequence(&ws);
h.PutArticle(article);
wh.m_HomNum = firstId;
} else {
pNewWord->m_SourceWords.SetPair(ws.FirstWord(), ws.LastWord());
if (str.size() == 0)
str = pNewWord->m_txt;
TMorph::ToLower(str);
CHomonym* pNewHom = new CHomonym(TMorph::GetMainLanguage(), str);
pNewHom->SetSourceWordSequence(&ws);
pNewHom->PutArticle(article);
wh.m_HomNum = pNewWord->AddRusHomonym(pNewHom);
}
if (article.AuxDic().IsValid()) {
const article_t* pArt = GlobalDictsHolder->GetAuxArticle(article.AuxDic());
YASSERT(pArt != NULL);
AddFoundArticle(pArt->get_kw_type(), pArt->get_title(), wh);
} else {
YASSERT(!article.Gzt().Empty());
AddFoundArticle(article.Gzt().GetType(), article.Gzt().GetTitle(), wh);
}
m_wordSequences.push_back(&ws);
}
void CWord::UndeleteHomonym(CHomonym& hom)
{
if (hom.IsDeleted()) {
if (m_DeletedHomCount > 0)
--m_DeletedHomCount;
hom.Undelete();
}
}
void CWord::AddFio(yset<Wtroka>& fioStrings, bool bIndexed)
{
yset<Wtroka>::iterator it = fioStrings.begin();
for (; it != fioStrings.end(); it++) {
CHomonym* pNewHom = new CHomonym(TMorph::GetMainLanguage(), *it);
if (*it == g_strFIONonTerminal)
pNewHom->SetIsDictionary(true);
else
pNewHom->SetIsDictionary(bIndexed);
m_variant.push_back(pNewHom);
}
}
bool CWord::CanAddHomonym(const CHomonym& h) const
{
//make sure that abbreviations (e.g. PO, OAO, etc.) are uppercased
if (h.HasGrammem(gAbbreviation) && !m_bIgnoreUpperCase) {
if (!m_bUp)
return false;
for (size_t i = 0; i < m_txt.size(); ++i)
if (!::IsUpper(m_txt[i]))
return false;
}
if ((h.IsGeo() || h.IsName()) && h.IsMorphNoun() && !m_bUp && !m_bIgnoreUpperCase)
return false;
return true;
}
bool CLemWord::AddNextHomonym(const char* sPlmLine)
{
char buffer[CriticalGraphemLineLength*2];
assert (strlen(sPlmLine) < CriticalGraphemLineLength*2);
strcpy(buffer, sPlmLine);
char* strPlmLine = buffer;
assert (strPlmLine[0] == ' ');
assert (strPlmLine[1] == ' ');
strPlmLine += 2;
assert ( GetHomonymsCount() );
size_t iPlmLineLen = strlen(strPlmLine);
rtrim(strPlmLine,&iPlmLineLen);
size_t iFirstFieldLen = strcspn(strPlmLine," ");
size_t iSomeUnusefulDigitsLen = strspn(strPlmLine + iFirstFieldLen," -1234567890");
CHomonym* pHomonym = AddNewHomonym();
strPlmLine += iFirstFieldLen + iSomeUnusefulDigitsLen;
int MorphSectionOffset = ProcessGraphematicalDescriptors(strPlmLine);
if (!pHomonym->ProcessLemmaAndGrammems(strPlmLine+MorphSectionOffset)) return false;
{
const CHomonym* pFirstHom = GetHomonym(0);
pHomonym->m_bInOb = pFirstHom->m_bInOb;
pHomonym->m_bOborot1 = pFirstHom->m_bOborot1;
pHomonym->m_bOborot2 = pFirstHom->m_bOborot2;
pHomonym->m_OborotNo = pFirstHom->m_OborotNo;
}
InitLevelSpecific(pHomonym);
return true;
}
void CWord::AddHyphenSurnameLemma(int iH, const THomonymGrammems& forms, const Wtroka& strLemma)
{
if (iH == -1) {
if (m_Homonyms.size() == 1) {
m_Homonyms[0]->SetLemma(strLemma);
m_Homonyms[0]->SetGrammems(forms);
m_Homonyms[0]->SetNameType(Surname);
return;
} else {
CHomonym* pH = new CHomonym(TMorph::GetMainLanguage(), strLemma);
pH->SetGrammems(forms);
pH->SetNameType(Surname);
AddRusHomonym(pH);
return;
}
} else {
CHomonym& h = GetRusHomonym(iH);
h.SetLemma(strLemma);
h.SetNameType(Surname);
}
}
bool CLemWord::ProcessPlmLineForTheFirstHomonym(const char* sPlmLine, MorphLanguageEnum langua, int& OborotNo)
{
char buffer[CriticalGraphemLineLength*2];
assert (strlen(sPlmLine) < CriticalGraphemLineLength*2);
strcpy(buffer, sPlmLine);
char* strPlmLine = buffer;
// откусываю признаки конца строки, если они по¤вились
size_t iPlmLineLen = strlen(strPlmLine);
rtrim(strPlmLine,&iPlmLineLen);
size_t iFirstFieldLen = strcspn(strPlmLine," ");
char WordBuffer[CriticalTokenLength+1];
strncpy(WordBuffer, strPlmLine, iFirstFieldLen);
WordBuffer[iFirstFieldLen] = '\0';
SetWordStr(WordBuffer, langua);
size_t iSomeUnusefulDigitsLen = strspn(strPlmLine + iFirstFieldLen," -1234567890");
// reading file position of an item from graline
if (sscanf(strPlmLine + iFirstFieldLen, "%i %i", &m_GraphematicalUnitOffset, &m_TokenLengthInFile) != 2)
return false;
strPlmLine += iFirstFieldLen + iSomeUnusefulDigitsLen;
int MorphSectionOffset = ProcessGraphematicalDescriptors(strPlmLine);
if( m_bSpace )
return true;
CHomonym* pHomonym = AddNewHomonym();
if (HasDes(OEXPR1))
OborotNo = ParseOborotNo(m_UnparsedGraphemDescriptorsStr);
if (OborotNo != -1)
{
pHomonym->m_OborotNo = OborotNo;
pHomonym->m_bInOb = true;
pHomonym->m_bOborot1 = HasDes(OEXPR1);
pHomonym->m_bOborot2 = HasDes(OEXPR2);
}
if (HasDes(OEXPR2))
OborotNo = -1;
if( MorphSectionOffset == strlen(strPlmLine) )
pHomonym->SetLemma(m_strWord);
else
if (!pHomonym->ProcessLemmaAndGrammems(strPlmLine+MorphSectionOffset))
return false;
InitLevelSpecific(pHomonym);
if ( (m_strWord == "\"")
|| (m_strWord == "'")
)
DelDes(OPun);
if (pHomonym->m_LemSign != '+')
m_bPredicted = true;
return true;
}
void CAnalyticFormBuilder::ChangeGrammemsAsAnalyticForm(CHomonym& H, const CHomonym& VerbHomonym)
{
THomonymGrammems old_grammems;
H.Grammems.Swap(old_grammems);
for (THomonymGrammems::TFormIter old = old_grammems.IterForms(); old.Ok(); ++old)
for (THomonymGrammems::TFormIter verb = VerbHomonym.Grammems.IterForms(); verb.Ok(); ++verb) {
Stroka strPos;
// auxiliary verb grammems
const TGramBitSet& VerbGrammems = *verb;
// meaningful part grammems
TGramBitSet MainGrammems = *old;
// final grammems to set
TGramBitSet ResultedGrammems;
if (MainGrammems.Has(gInfinitive)) {
ResultedGrammems = MainGrammems & ~TMorph::AllPOS();
if (VerbGrammems.Has(gImperative)) {
// analytical form for imperatives in singular number does not exist
if (VerbGrammems.Has(gSingular))
continue;
ResultedGrammems.Set(gImperative); // "будем же жить!"
} else
ResultedGrammems |= VerbGrammems & NSpike::AllTimes; // "я стал пилить" или "стану писать"
ResultedGrammems |= VerbGrammems & NSpike::AllPersons;
ResultedGrammems |= VerbGrammems & NSpike::AllNumbers;
ResultedGrammems |= VerbGrammems & NSpike::AllGenders;
//copy all POS grammems from verb to main
ResultedGrammems |= VerbGrammems & TMorph::AllPOS();
H.PutAuxArticle(SDictIndex(TEMPLATE_DICT, VerbHomonym.GetAuxArticleIndex(TEMPLATE_DICT)));
strPos = "Г";
} else if (TMorph::IsShortParticiple(MainGrammems)) {
// "*будем же взяты!"
if (VerbGrammems.Has(gImperative))
continue;
ResultedGrammems = MainGrammems & ~TMorph::AllPOS();
// remove any time grammems from participle
ResultedGrammems &= ~NSpike::AllTimes;
ResultedGrammems |= VerbGrammems & NSpike::AllPersons;
ResultedGrammems |= VerbGrammems & NSpike::AllTimes;
if (VerbGrammems.Has(gImperative)) // ??? the same check second time, always false?
ResultedGrammems.Set(gImperative);
strPos = "ПРИЧАСТИЕ";
ResultedGrammems |= TGramBitSet(gParticiple, gShort);
} else if (TMorph::IsShortAdjective(MainGrammems)) {
if (VerbGrammems.Has(gImperative))
continue; // будем cчитать, что "будем же красивы!" - это плохо!
// на самом деле, просто не хочется вводить ее кучу кодов.
ResultedGrammems = VerbGrammems;
ResultedGrammems |= MainGrammems & (NSpike::AllNumbers | NSpike::AllGenders | TGramBitSet(gAnimated, gInanimated));
ResultedGrammems &= ~TMorph::AllPOS();
if (ResultedGrammems.Has(gActive))
ResultedGrammems &= ~TGramBitSet(gActive);
ResultedGrammems |= TGramBitSet(gAdjective, gShort);
strPos = "П";
} else if (MainGrammems.Has(gPraedic)) // "мне было больно"
{
ResultedGrammems = VerbGrammems;
ResultedGrammems |= NSpike::AllCases & MainGrammems; //copied from PronounPredk code (commented below) - preserve cases if any
ResultedGrammems &= ~TMorph::AllPOS();
if (ResultedGrammems.Has(gActive))
ResultedGrammems.Reset(gActive);
strPos = "ПРЕДК";
ResultedGrammems |= MainGrammems & TMorph::AllPOS();
} else if (MainGrammems.Has(gComparative)) // он был больше тебя
{
ResultedGrammems = (VerbGrammems & ~TMorph::AllPOS()) | TGramBitSet(gComparative);
if (ResultedGrammems.Has(gActive))
ResultedGrammems.Reset(gActive);
strPos = "П";
ResultedGrammems |= MainGrammems & TMorph::AllPOS();
} else if (TMorph::IsFullAdjective(MainGrammems))
// resolve disambiguity of homonyms, because analytical forms with full adjectives do not exist.
continue;
// "стал писать" "стану писать" "стать писать" - совершенный вид
if (VerbHomonym.Lemma == kStat)
ResultedGrammems.Reset(gImperfect).Set(gPerfect);
// if the auxiliary verb was an infinitive then it is all an infinitive
// "быть лучше" или "должно быть принесено"
if (VerbHomonym.HasGrammem(gInfinitive)) {
ResultedGrammems &= ~TMorph::AllPOS();
ResultedGrammems.Set(gInfinitive);
strPos = "ИНФИНИТИВ";
} else if (VerbHomonym.HasGrammem(gGerund)) // "будучи лишней"
{
ResultedGrammems &= ~TMorph::AllPOS();
ResultedGrammems.Set(gGerund);
strPos = "ДЕЕПРИЧАСТИЕ";
}
if (strPos.empty())
continue;
/* do some corrections (code taken from RusGramTab.ProcessPOSAndGrammems) */
if (ResultedGrammems.HasAll(NSpike::AllCases | TGramBitSet(gAdjPronoun)))
ResultedGrammems |= NSpike::AllGenders | NSpike::AllNumbers;
if (ResultedGrammems.Has(gMasFem))
ResultedGrammems |= TGramBitSet(gMasculine, gFeminine);
if (!ResultedGrammems.Has(gPraedic) && ResultedGrammems.HasAll(NSpike::AllCases) && !ResultedGrammems.Has(gSingular))
ResultedGrammems |= TGramBitSet(gSingular, gPlural);
H.Grammems.AddForm(ResultedGrammems);
}
}
bool CMAPost::SetFioFormat (const CFIOFormat* Format, CLineIter it)
{
vector<SmallHomonymsVec> Hypots;
SmallWordsVec FioWords;
Hypots.resize(Format->m_Items.size());
int CountOfVariants = 1;
for (long ItemNo = 0; ItemNo < Format->m_Items.size() && it != m_Words.end(); ItemNo++, it=NextNotSpace(it))
{
FioWords.Add(it);
CPostLemWord& W = *it;
for (int HomNo=0; HomNo < W.GetHomonymsCount(); HomNo++)
{
CHomonym* pH = W.GetHomonym(HomNo);
// иначе "“.≈. ќ“ ј«ј“№—я" будет ‘»ќ
if (IsPartOfNonSingleOborot(pH)) return false;
if (IsPartFio(*this, Format->m_Items[ItemNo], W, pH))
Hypots[ItemNo].Add(pH);
};
if ( Hypots[ItemNo].empty() ) return false;
CountOfVariants *= Hypots[ItemNo].size();
};
if (FioWords.size() != Format->m_Items.size()) return false; // не достроилось
SmallHomonymsVec V; // текущий вариант
vector<SmallHomonymsVec> Variants;
Variants.reserve(CountOfVariants);
V.m_ItemsCount = Hypots.size();
GetCommonVariants(Hypots, V, Variants, 0);
if (Format->m_GleicheCase)
for (long VarNo=0; VarNo < Variants.size(); VarNo++)
{
QWORD Grammems = rAllCases | rAllNumbers;
for (long i=0; i < Variants[VarNo].size(); i++)
{
Grammems &= Variants[VarNo].m_Items[i]->m_iGrammems;
};
if ((Grammems & rAllCases) == 0 || (Grammems & rAllNumbers) == 0 )
{
Variants.erase(Variants.begin()+VarNo);
VarNo--;
};
};
if (Variants.empty()) return false;
for (size_t i=0; i <FioWords.size(); i++)
{
CPostLemWord& W = *FioWords[i];
CHomonym* pH = Variants[0].m_Items[i];
W.SetHomonymsDel(true);
pH->m_bDelete = false;
pH->DeleteOborotMarks(); // удал¤ем однословные оборотыб (многословных там быть не может)
W.DeleteMarkedHomonymsBeforeClauses();
if (W.HasDes(OSentEnd))
{
/*
если ‘»ќ содержало конец предложени¤, а после ‘»ќ нет ни одной пометы
конца предложени¤, тогда надо поставить SENT_END на последнюю строчку ‘»ќ.
*/
W.DelDes(OSentEnd);
if ( W.m_strWord!="." )
FioWords.back()->AddDes(OSentEnd);
}
}
// ставим графем. пометы
FioWords[0]->AddDes(OFAM1);
FioWords.back()->AddDes(OFAM2);
return true;
};