ET_ReturnCode CFormBuilderNouns::eBuild()
{
ASSERT(m_pLexeme); // we assume base class ctor took care of this
ET_ReturnCode rc = H_NO_ERROR;
m_pEndings = new CNounEndings(m_pLexeme);
if (NULL == m_pEndings)
{
return H_ERROR_POINTER;
}
if (rc != H_NO_ERROR)
{
return rc;
}
ET_Animacy eAnimacy = m_pLexeme->eAnimacy();
ET_Gender eoGender = m_pLexeme->eGender();
CHasher gramIterator;
gramIterator.Initialize(eoGender, eAnimacy);
do
{
if ((L"мн." == m_pLexeme->sMainSymbol()) && (gramIterator.m_eNumber == NUM_SG))
{
continue;
}
if (NUM_PL == gramIterator.m_eNumber &&
(CASE_PART == gramIterator.m_eCase || CASE_LOC == gramIterator.m_eCase))
{
continue;
}
if (CASE_PART == gramIterator.m_eCase && !m_pLexeme->bSecondGenitive())
{
continue;
}
if (CASE_LOC == gramIterator.m_eCase && !m_pLexeme->bSecondLocative())
{
continue;
}
// Handle acc ending
ET_Case eEndingCase = gramIterator.m_eCase;
if (CASE_ACC == gramIterator.m_eCase)
{
rc = eHandleAccEnding (gramIterator.m_eNumber, eEndingCase);
if (rc != H_NO_ERROR)
{
return rc;
}
}
if (CASE_PART == gramIterator.m_eCase || CASE_LOC == gramIterator.m_eCase)
{
eEndingCase = CASE_DAT;
}
CEString sLemma (m_pLexeme->sGraphicStem());
if (m_pLexeme->bHasIrregularForms())
{
bool bSkipRegular = false;
rc = eCheckIrregularForms (gramIterator.m_eGender,
gramIterator.m_eAnimacy,
gramIterator.m_eCase,
eEndingCase,
gramIterator.m_eNumber,
bSkipRegular);
if (rc != H_NO_ERROR)
{
return rc;
}
if (bSkipRegular)
{
// Workaround for lack of "исх. форма иррег." mark in current source
if (GENDER_M == gramIterator.m_eGender && NUM_SG == gramIterator.m_eNumber && CASE_NOM == gramIterator.m_eCase)
{
m_bIrregularSourceForm = true;
}
continue;
}
}
rc = eHandleStemAugment (sLemma, gramIterator.m_eNumber, gramIterator.m_eCase);
if (rc != H_NO_ERROR)
{
return rc;
}
ET_StressLocation eStress = STRESS_LOCATION_UNDEFINED;
if (CASE_LOC == gramIterator.m_eCase)
{
eStress = STRESS_LOCATION_ENDING;
}
else
{
rc = eGetStressType (gramIterator.m_eNumber, eEndingCase, eStress);
if (rc != H_NO_ERROR)
{
return rc;
}
}
((CNounEndings *)m_pEndings)->eSelect(gramIterator.m_eNumber, eEndingCase, eStress);
int iNumEndings = m_pEndings->iCount();
if (iNumEndings < 1)
{
if (m_pLexeme->iType() != 0)
{
ASSERT(0);
ERROR_LOG(L"No endings");
}
continue;
}
CEString sSavedLemma (sLemma); // lemma can change, e.g. because of a fleetimg vowel
for (int iEnding = 0; iEnding < iNumEndings; ++iEnding, sLemma = sSavedLemma)
{
// Get ending and modify as necessary
CEString sEnding;
unsigned __int64 llEndingKey = -1;
rc = m_pEndings->eGetEnding(iEnding, sEnding, llEndingKey);
if (rc != H_NO_ERROR)
{
return rc;
}
if (8 == m_pLexeme->iType() && GENDER_N != m_pLexeme->eGender())
{
if (sLemma.bEndsWithOneOf (L"шжчщц"))
{
if (sEnding.bStartsWith (L"я"))
{
continue;
}
}
else
{
if (sEnding.bStartsWith (L"а"))
{
continue;
}
}
}
bool bHasFleetingVowel = false;
rc = eFleetingVowelCheck (gramIterator.m_eNumber,
eEndingCase,
gramIterator.m_eGender,
eStress,
SUBPARADIGM_NOUN,
sEnding,
sLemma);
if (rc != H_NO_ERROR)
{
continue;
}
vector<int> vecStress;
rc = eGetStressPositions (sLemma, sEnding, eStress, vecStress);
if (rc != H_NO_ERROR)
{
continue;
}
CWordForm * pWordForm = NULL;
rc = eCreateFormTemplate (gramIterator.m_eNumber, gramIterator.m_eCase, sLemma, pWordForm);
if (rc != H_NO_ERROR)
{
continue;
}
if (1 == vecStress.size() || m_pLexeme->bIsMultistressedCompound())
{
vector<int>::iterator itStressPos = vecStress.begin();
for (; itStressPos != vecStress.end(); ++itStressPos)
{
pWordForm->m_mapStress[*itStressPos] = STRESS_PRIMARY; // primary
rc = eHandleYoAlternation (eStress, *itStressPos, pWordForm->m_sLemma, sEnding);
if (rc != H_NO_ERROR)
{
continue;
}
pWordForm->m_sEnding = sEnding;
pWordForm->m_llEndingDataId = llEndingKey;
pWordForm->m_sWordForm = pWordForm->m_sLemma + sEnding;
}
m_pLexeme->AddWordForm (pWordForm);
}
else
{
vector<int>::iterator itStressPos = vecStress.begin();
for (; itStressPos != vecStress.end(); ++itStressPos)
{
if (itStressPos != vecStress.begin())
{
CWordForm * pwfVariant = NULL;
CloneWordForm (pWordForm, pwfVariant);
pwfVariant->m_mapStress.clear();
pWordForm = pwfVariant;
}
pWordForm->m_mapStress[*itStressPos] = STRESS_PRIMARY; // primary
rc = eHandleYoAlternation (eStress, *itStressPos, pWordForm->m_sLemma, sEnding);
if (rc != H_NO_ERROR)
{
continue;
}
pWordForm->m_sWordForm = pWordForm->m_sLemma + sEnding;
pWordForm->m_sEnding = sEnding;
pWordForm->m_llEndingDataId = llEndingKey;
m_pLexeme->AddWordForm (pWordForm);
}
}
} // for (int iEnding = 0; ... )
} while (gramIterator.bIncrement());
return H_NO_ERROR;
} // eBuildNounForms()
ET_ReturnCode CFormBuilderPronounAdj::eBuild()
{
ASSERT(m_pLexeme); // we assume base class ctor took care of this
ET_ReturnCode rc = H_NO_ERROR;
m_pEndings = new CAdjPronounEndings(m_pLexeme);
if (NULL == m_pEndings)
{
return H_ERROR_POINTER;
}
CHasher gramTmp;
gramTmp.Initialize (POS_PRONOUN_ADJ);
gramTmp.SetParadigm (SUBPARADIGM_PRONOUN_ADJ);
do
{
bool bSkipRegular = false;
if (m_pLexeme->bHasIrregularForms())
{
int iPronAdj = iGramHash (m_pLexeme->ePartOfSpeech(),
gramTmp.m_eCase, gramTmp.m_eNumber,
gramTmp.m_eGender, gramTmp.m_eAnimacy);
if (m_pLexeme->bHasIrregularForm(iPronAdj))
{
map<CWordForm *, bool> mapPronAdj;
rc = m_pLexeme->eGetIrregularForms(iPronAdj, mapPronAdj);
if (rc != H_NO_ERROR)
{
return rc;
}
map<CWordForm *, bool>::iterator it = mapPronAdj.begin();
for (; it != mapPronAdj.end(); ++it)
{
CWordForm * pWordForm = NULL;
rc = eCreateFormTemplate(gramTmp.m_eGender,
gramTmp.m_eNumber,
gramTmp.m_eCase,
gramTmp.m_eAnimacy,
L"",
-1,
pWordForm);
if (rc != H_NO_ERROR)
{
continue;
}
pWordForm->m_sWordForm = (*it).first->m_sWordForm;
pWordForm->m_mapStress = (*it).first->m_mapStress;
m_pLexeme->AddWordForm(pWordForm);
bSkipRegular = !(*it).second; // i.e. no regular variant
}
}
} // if (m_pLexeme->m_bHasIrregularForms)
if (bSkipRegular)
{
continue;
}
ET_StressLocation eStressType = STRESS_LOCATION_UNDEFINED;
switch (m_pLexeme->eAccentType1())
{
case AT_A:
{
eStressType = STRESS_LOCATION_STEM;
break;
}
case AT_B:
{
eStressType = STRESS_LOCATION_ENDING;
break;
}
case AT_F:
{
if (NUM_SG == gramTmp.m_eNumber)
{
eStressType = STRESS_LOCATION_ENDING;
}
else if (CASE_NOM == gramTmp.m_eCase)
{
eStressType = STRESS_LOCATION_STEM;
}
else if (CASE_ACC == gramTmp.m_eCase && ANIM_NO == gramTmp.m_eAnimacy)
{
eStressType = STRESS_LOCATION_STEM;
}
else
{
eStressType = STRESS_LOCATION_ENDING;
}
break;
}
default:
{
ASSERT(0);
ERROR_LOG (L"Illegal accent type.");
return H_ERROR_UNEXPECTED;
}
} // switch (m_pLexeme->eAccentType1())
rc = ((CAdjPronounEndings *)m_pEndings)->eSelect(gramTmp.m_eGender, gramTmp.m_eNumber, gramTmp.m_eCase, gramTmp.m_eAnimacy);
int iNumEndings = m_pEndings->iCount();
if (iNumEndings < 1)
{
ERROR_LOG (L"No endings");
continue;
}
for (int iEnding = 0; iEnding < iNumEndings; ++iEnding)
{
CEString sEnding;
unsigned __int64 llEndingKey = -1;
rc = m_pEndings->eGetEnding(iEnding, sEnding, llEndingKey);
if (rc != H_NO_ERROR)
{
return rc;
}
if (m_pLexeme->bFleetingVowel() &&
(gramTmp.m_eCase != CASE_NOM || gramTmp.m_eGender != GENDER_M) &&
(gramTmp.m_eCase != CASE_ACC || gramTmp.m_eGender != GENDER_M ||
gramTmp.m_eAnimacy != ANIM_NO))
{
rc = eFleetingVowelCheck (gramTmp.m_eNumber, gramTmp.m_eCase,
gramTmp.m_eGender, eStressType,
SUBPARADIGM_LONG_ADJ, sEnding, m_sLemma);
if (rc != H_NO_ERROR)
{
return rc;
}
}
CWordForm * pWordForm = NULL;
rc = eCreateFormTemplate(gramTmp.m_eGender,
gramTmp.m_eNumber,
gramTmp.m_eCase,
gramTmp.m_eAnimacy,
sEnding,
llEndingKey,
pWordForm);
if (rc != H_NO_ERROR)
{
return rc;
}
vector<int> vecStress;
rc = eGetStressPositions (sEnding, eStressType, vecStress);
if (rc != H_NO_ERROR)
{
return rc;
}
if (1 == vecStress.size() || m_pLexeme->bIsMultistressedCompound())
{
vector<int>::iterator itStressPos = vecStress.begin();
for (; itStressPos != vecStress.end(); ++itStressPos)
{
pWordForm->m_mapStress[*itStressPos] = STRESS_PRIMARY;
}
m_pLexeme->AddWordForm (pWordForm);
}
else
{
vector<int>::iterator itStressPos = vecStress.begin();
for (; itStressPos != vecStress.end(); ++itStressPos)
{
if (itStressPos != vecStress.begin())
{
CWordForm * pWfVariant = NULL;
CloneWordForm (pWordForm, pWfVariant);
pWordForm = pWfVariant;
}
pWordForm->m_mapStress[*itStressPos] = STRESS_PRIMARY;
m_pLexeme->AddWordForm (pWordForm);
}
}
}
} while (gramTmp.bIncrement());
return H_NO_ERROR;
} // eBuild (...)