void TVizMapContext::GetSelectDIdV(TIntV& DIdV) {
DIdV.Gen(SelPointV.Len(), 0);
const int SelPoints = SelPointV.Len();
for (int SelPointN = 0; SelPointN < SelPoints; SelPointN++) {
DIdV.Add(VizMapFrame->GetPoint(SelPointV[SelPointN])->GetDocId());
}
}
void TNmObjBs::GetNmObjDIdV(
const PBowDocBs& BowDocBs, TIntV& BowDIdV,
const TStr& NmObjStr1, const TStr& NmObjStr2) const {
// get first named-object-id
int NmObjId1=GetNmObjId(NmObjStr1);
TIntV NmObjDocIdV1; GetNmObjDocIdV(NmObjId1, NmObjDocIdV1);
NmObjDocIdV1.Sort();
// get second named-object-id
TIntV NmObjDocIdV2;
if (!NmObjStr2.Empty()){
int NmObjId2=GetNmObjId(NmObjStr2);
GetNmObjDocIdV(NmObjId2, NmObjDocIdV2);
NmObjDocIdV2.Sort();
}
// create joint doc-id-vector
TIntV NmObjDocIdV;
if (NmObjDocIdV2.Empty()){
NmObjDocIdV=NmObjDocIdV1;
} else {
NmObjDocIdV1.Intrs(NmObjDocIdV2, NmObjDocIdV);
}
// traverse named-object-documents to collect bow-document-ids
BowDIdV.Gen(NmObjDocIdV.Len(), 0);
for (int NmObjDocIdN=0; NmObjDocIdN<NmObjDocIdV.Len(); NmObjDocIdN++){
TStr DocNm=GetDocNm(NmObjDocIdV[NmObjDocIdN]);
int DId=BowDocBs->GetDId(DocNm);
if (DId!=-1){
BowDIdV.Add(DId);
}
}
}
void TBPGraph::GetNIdV(TIntV& NIdV) const {
NIdV.Gen(GetNodes(), 0);
for (int N=LeftH.FFirstKeyId(); LeftH.FNextKeyId(N); ) {
NIdV.Add(LeftH.GetKey(N)); }
for (int N=RightH.FFirstKeyId(); RightH.FNextKeyId(N); ) {
NIdV.Add(RightH.GetKey(N)); }
}
void TBlobBs::GetBlockLenV(const PFRnd& FBlobBs, TIntV& BlockLenV){
EAssert(FBlobBs->GetStr(BlockLenVNm.Len())==BlockLenVNm);
BlockLenV.Gen(FBlobBs->GetInt());
for (int BlockLenN=0; BlockLenN<BlockLenV.Len(); BlockLenN++){
BlockLenV[BlockLenN]=FBlobBs->GetInt();}
EAssert(FBlobBs->GetInt()==-1);
}
void TYInvIx::GetDocIdV(
const PYWordDs& WordDs, const int& MnDocFq, TIntV& DocIdV){
IAssert(MnDocFq>=0);
if (MnDocFq==0){
DocIdV=AllDocIdV;
} else {
TIntIntH DocIdFqH(100); int MxDocFq=0;
int WordIdN=WordDs->FFirstWordId(); int WordId; double WordFq;
while (WordDs->FNextWordId(WordIdN, WordId, WordFq)){
if (WordIdToFirstDocIdNH.IsKey(WordId)){
int DocIdN=FFirstDocId(WordId); int DocId;
while (FNextWordId(DocIdN, DocId)){
DocIdFqH.AddDat(DocId)+=int(WordFq);
MxDocFq=TInt::GetMx(MxDocFq, DocIdFqH.GetDat(DocId));
}
}
}
int NewMnDocFq=(MnDocFq<=MxDocFq) ? MnDocFq : MxDocFq-3;
DocIdV.Gen(DocIdFqH.Len(), 0);
int DocIdP=DocIdFqH.FFirstKeyId();
while (DocIdFqH.FNextKeyId(DocIdP)){
int DocId=DocIdFqH.GetKey(DocIdP);
int DocFq=DocIdFqH[DocIdP];
if (DocFq>=NewMnDocFq){DocIdV.Add(DocId);}
}
}
}
void TYFSelBs::GetBestWordIdV(
const int& DocId, const double& EstExp, const double& SumEstPrb,
const PYWordDs& IntrsWordDs, TIntV& BestWordIdV){
TIntFltKdV& WordIdEstKdV=DocIdToWordIdEstVV[DocId];
TFltIntKdV WordEstIdKdV(WordIdEstKdV.Len(), 0);
double MnWordEst=TFlt::Mx;
for (int WordIdN=0; WordIdN<WordIdEstKdV.Len(); WordIdN++){
int WordId=WordIdEstKdV[WordIdN].Key;
double WordEst=pow(WordIdEstKdV[WordIdN].Dat, EstExp);
if (IntrsWordDs->IsWordId(WordId)){
WordEstIdKdV.Add(TFltIntKd(WordEst, WordId));
MnWordEst=TFlt::GetMn(WordEst, MnWordEst);
}
}
double SumWordEst=0;
{for (int WordIdN=0; WordIdN<WordEstIdKdV.Len(); WordIdN++){
SumWordEst+=(WordEstIdKdV[WordIdN].Key-=MnWordEst);}}
WordEstIdKdV.Sort(false);
{BestWordIdV.Gen(WordEstIdKdV.Len(), 0);
SumWordEst*=SumEstPrb; int WordIdN=0;
while ((SumWordEst>=0)&&(WordIdN<WordEstIdKdV.Len())){
double WordEst=WordEstIdKdV[WordIdN].Key;
int WordId=WordEstIdKdV[WordIdN].Dat;
SumWordEst-=WordEst;
BestWordIdV.Add(WordId);
WordIdN++;
}}
}
///Generate sequence from Power law
void TAGMUtil::GenPLSeq(TIntV& SzSeq, const int& SeqLen, const double& Alpha, TRnd& Rnd, const int& Min, const int& Max) {
SzSeq.Gen(SeqLen, 0);
while (SzSeq.Len() < SeqLen) {
int Sz = (int) TMath::Round(Rnd.GetPowerDev(Alpha));
if (Sz >= Min && Sz <= Max) {
SzSeq.Add(Sz);
}
}
}
void TGraphCascade::TopologicalSort(TIntV& SortedNIdV) {
int Nodes = Graph.GetNodes();
SortedNIdV.Gen(Nodes, 0); // result
THash<TInt, TBool> Marks(Nodes); // nodeid -> mark map
THash<TInt,TBool> TempMarks(Nodes); // nodeid -> temp mark map
THash<TInt, TBool> Added(Nodes);
TIntV NIdV; Graph.GetNIdV(NIdV); // all node ids
// set marks
for (int NodeN = 0; NodeN < Nodes; NodeN++) {
int NodeId = NIdV[NodeN];
Marks.AddDat(NodeId, false);
TempMarks.AddDat(NodeId, false);
Added.AddDat(NodeId, false);
}
TSStack<TInt> Stack;
for (int NodeN = 0; NodeN < Nodes; NodeN++) {
int NodeId = NIdV[NodeN];
// select an unmarked node
if (!Marks.GetDat(NodeId)) {
Stack.Push(NodeId);
while (!Stack.Empty()) {
// visit TopNode
int TopNodeId = Stack.Top();
Marks.GetDat(TopNodeId) = true;
TempMarks.GetDat(TopNodeId) = true;
// add children, set their temp marks to true
TNGraph::TNodeI NI = Graph.GetNI(TopNodeId);
int Children = NI.GetOutDeg();
bool IsFinal = true;
for (int ChildN = 0; ChildN < Children; ChildN++) {
int ChildId = NI.GetOutNId(ChildN);
EAssertR(!TempMarks.GetDat(ChildId), "TGraphCascade::TopologicalSort: the graph is not a DAG!");
if (!Marks.GetDat(ChildId)) {
// unvisited node
IsFinal = false;
Stack.Push(ChildId);
}
}
if (IsFinal) {
// push TopNode to tail
if (!Added.GetDat(TopNodeId)) {
SortedNIdV.Add(TopNodeId);
Added.GetDat(TopNodeId) = true;
}
TempMarks.GetDat(TopNodeId) = false;
Stack.Pop();
}
}
}
}
SortedNIdV.Reverse();
}
/////////////////////////////////////////////////
// SkyGrid-Entity
void TSkyGridEnt::GetDocIdV(const TSkyGridBs* SkyGridBs,
const uint64& MnTm, const uint64& MxTm, TIntV& DocIdV) const {
DocIdV.Gen(GetDocIds(), 0);
for (int DocN=0; DocN<GetDocIds(); DocN++){
int DocId=GetDocId(DocN);
PSkyGridDoc Doc=SkyGridBs->GetDoc(DocId);
uint64 DocTm=Doc->GetTm();
if (((MnTm==0)||(MnTm<=DocTm))&&((MxTm==0)||(DocTm<MxTm))){
DocIdV.Add(DocId);
}
}
}
void TNmObjBs::GetMergedNmObj(TIntV& NewNmObjIdV){
// matching constraints
int MnPfxLen=3; int MxSfxLen=2;
// create transformation vector
int NmObjs=NmObjWordStrVToDocIdVH.Len();
NewNmObjIdV.Gen(NmObjs); NewNmObjIdV.PutAll(-1);
// merging single words
// merging statistics
{int SingleWords=0; int ReducedSingleWords=0;
// collect single words according to prefix
TStrIntVH PfxStrToNmObjIdVH;
for (int NmObjId=0; NmObjId<NmObjs; NmObjId++){
if (NewNmObjIdV[NmObjId]!=-1){continue;}
const TStrV& WordStrV=NmObjWordStrVToDocIdVH.GetKey(NmObjId);
if (WordStrV.Len()==1){
TStr PfxStr=WordStrV[0].GetSubStr(0, 2);
PfxStrToNmObjIdVH.AddDat(PfxStr).Add(NmObjId);
SingleWords++;
}
}
// traverse word-groups with the same prefix
int Pfxs=PfxStrToNmObjIdVH.Len();
for (int PfxId=0; PfxId<Pfxs; PfxId++){
// get & traverse word-group
TIntV& NmObjIdV=PfxStrToNmObjIdVH[PfxId];
for (int NmObjIdN=0; NmObjIdN<NmObjIdV.Len(); NmObjIdN++){
int NmObjId=NmObjIdV[NmObjIdN];
if (NewNmObjIdV[NmObjId]!=-1){continue;}
NewNmObjIdV[NmObjId]=NmObjId;
const TStr& WordStr=NmObjWordStrVToDocIdVH.GetKey(NmObjId)[0];
int Fq=NmObjWordStrVToDocIdVH[NmObjId].Len();
TIntPrV FqNmObjIdPrV(NmObjIdV.Len(), 0);
FqNmObjIdPrV.Add(TIntPr(Fq, NmObjId));
// traverse rest of the word-group for matching words
for (int SubNmObjIdN=NmObjIdN+1; SubNmObjIdN<NmObjIdV.Len(); SubNmObjIdN++){
int SubNmObjId=NmObjIdV[SubNmObjIdN];
if (NewNmObjIdV[SubNmObjId]!=-1){continue;}
const TStr& SubWordStr=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[0];
// test matching
if (IsMatchPfx(WordStr, SubWordStr, MnPfxLen, MxSfxLen)){
NewNmObjIdV[SubNmObjId]=NmObjId;
int SubFq=NmObjWordStrVToDocIdVH[SubNmObjId].Len();
FqNmObjIdPrV.Add(TIntPr(SubFq, SubNmObjId));
//printf("%s -> %s\n", WordStr.CStr(), SubWordStr.CStr());
}
}
// increment number of equivalence word-groups
ReducedSingleWords++;
// collapse matching words into most frequent word
if (FqNmObjIdPrV.Len()>1){
FqNmObjIdPrV.Sort(false);
int MainNmObjId=FqNmObjIdPrV[0].Val2;
NewNmObjIdV[MainNmObjId]=MainNmObjId;
TStr MainWordStr=NmObjWordStrVToDocIdVH.GetKey(MainNmObjId)[0];
//printf("[%s:", MainWordStr.CStr());
for (int FqNmObjIdPrN=1; FqNmObjIdPrN<FqNmObjIdPrV.Len(); FqNmObjIdPrN++){
int SubNmObjId=FqNmObjIdPrV[FqNmObjIdPrN].Val2;
NewNmObjIdV[SubNmObjId]=MainNmObjId;
//TStr& SubWordStr=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[0];
//printf(" %s", SubWordStr.CStr());
}
//printf("]\n");
}
}
}
// print statistics
//printf("SingleWords:%d ReducedSingleWords:%d\n",
// SingleWords, ReducedSingleWords);
}
// merging double words
// merging statistics
{int DoubleWords=0; int ReducedDoubleWords=0;
// collect double words according to prefix
TStrIntVH PfxStrToNmObjIdVH;
for (int NmObjId=0; NmObjId<NmObjs; NmObjId++){
if (NewNmObjIdV[NmObjId]!=-1){continue;}
const TStrV& WordStrV=NmObjWordStrVToDocIdVH.GetKey(NmObjId);
if (WordStrV.Len()==2){
TStr PfxStr=WordStrV[0].GetSubStr(0, 2)+WordStrV[1].GetSubStr(0, 2);
PfxStrToNmObjIdVH.AddDat(PfxStr).Add(NmObjId);
DoubleWords++;
}
}
// traverse word-groups with the same prefix
int Pfxs=PfxStrToNmObjIdVH.Len();
for (int PfxId=0; PfxId<Pfxs; PfxId++){
// get & traverse word-group
TIntV& NmObjIdV=PfxStrToNmObjIdVH[PfxId];
for (int NmObjIdN=0; NmObjIdN<NmObjIdV.Len(); NmObjIdN++){
int NmObjId=NmObjIdV[NmObjIdN];
if (NewNmObjIdV[NmObjId]!=-1){continue;}
NewNmObjIdV[NmObjId]=NmObjId;
const TStr& WordStr1=NmObjWordStrVToDocIdVH.GetKey(NmObjId)[0];
const TStr& WordStr2=NmObjWordStrVToDocIdVH.GetKey(NmObjId)[1];
int Fq=NmObjWordStrVToDocIdVH[NmObjId].Len();
TIntPrV FqNmObjIdPrV(NmObjIdV.Len(), 0);
FqNmObjIdPrV.Add(TIntPr(Fq, NmObjId));
// traverse rest of the word-group for matching words
for (int SubNmObjIdN=NmObjIdN+1; SubNmObjIdN<NmObjIdV.Len(); SubNmObjIdN++){
int SubNmObjId=NmObjIdV[SubNmObjIdN];
if (NewNmObjIdV[SubNmObjId]!=-1){continue;}
const TStr& SubWordStr1=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[0];
const TStr& SubWordStr2=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[1];
// test matching
if (IsMatchPfx(WordStr1, SubWordStr1, MnPfxLen, MxSfxLen+1)&&
IsMatchPfx(WordStr2, SubWordStr2, MnPfxLen, MxSfxLen+1)){
NewNmObjIdV[SubNmObjId]=NmObjId;
int SubFq=NmObjWordStrVToDocIdVH[SubNmObjId].Len();
FqNmObjIdPrV.Add(TIntPr(SubFq, SubNmObjId));
//printf("%s_%s -> %s_%s\n",
// WordStr1.CStr(), WordStr2.CStr(),
// SubWordStr1.CStr(), SubWordStr2.CStr());
}
}
// increment number of equivalence word-groups
ReducedDoubleWords++;
// collapse matching words into most frequent word
if (FqNmObjIdPrV.Len()>1){
FqNmObjIdPrV.Sort(false);
int MainNmObjId=FqNmObjIdPrV[0].Val2;
NewNmObjIdV[MainNmObjId]=MainNmObjId;
TStr MainWordStr1=NmObjWordStrVToDocIdVH.GetKey(MainNmObjId)[0];
TStr MainWordStr2=NmObjWordStrVToDocIdVH.GetKey(MainNmObjId)[1];
//printf("[%s_%s:", MainWordStr1.CStr(), MainWordStr2.CStr());
for (int FqNmObjIdPrN=1; FqNmObjIdPrN<FqNmObjIdPrV.Len(); FqNmObjIdPrN++){
int SubNmObjId=FqNmObjIdPrV[FqNmObjIdPrN].Val2;
NewNmObjIdV[SubNmObjId]=MainNmObjId;
//TStr& SubWordStr1=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[0];
//TStr& SubWordStr2=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[1];
//printf(" %s_%s", SubWordStr1.CStr(), SubWordStr2.CStr());
}
//printf("]\n");
}
}
}
// print statistics
//printf("DoubleWords:%d ReducedDoubleWords:%d\n",
// DoubleWords, ReducedDoubleWords);
}
// merging triples to doubles
// ... (prefix, first-name, last-name) to (first-name, last-name)
// merging statistics
{int TripleWords=0; int ReducedTripleWords=0;
// collect single words according to prefix
TStrIntVH PfxStrToNmObjIdVH;
for (int NmObjId=0; NmObjId<NmObjs; NmObjId++){
if (NewNmObjIdV[NmObjId]!=-1){continue;}
const TStrV& WordStrV=NmObjWordStrVToDocIdVH.GetKey(NmObjId);
if (WordStrV.Len()==3){
TripleWords++;
TStrV DbWordStrV(2, 0);
DbWordStrV.Add(WordStrV[1]); DbWordStrV.Add(WordStrV[2]);
int DbNmObjId=NmObjWordStrVToDocIdVH.GetKeyId(DbWordStrV);
if (DbNmObjId!=-1){
ReducedTripleWords++;
int NewDbNmObjId=NewNmObjIdV[DbNmObjId];
NewNmObjIdV[NmObjId]=NewDbNmObjId;
//TStr NmObjStr=GetNmObjStr(NmObjId);
//TStr DbNmObjStr=GetNmObjStr(DbNmObjId);
//TStr NewDbNmObjStr=GetNmObjStr(NewDbNmObjId);
//printf("%s -> %s -> %s\n",
// NmObjStr.CStr(), DbNmObjStr.CStr(), NewDbNmObjStr.CStr());
}
}
}
//printf("TripleWords:%d ReducedTripleWords:%d\n",
// TripleWords, ReducedTripleWords);
}
// merging triple words
// merging statistics
{int TripleWords=0; int ReducedTripleWords=0;
// collect triple words according to prefix
TStrIntVH PfxStrToNmObjIdVH;
for (int NmObjId=0; NmObjId<NmObjs; NmObjId++){
if (NewNmObjIdV[NmObjId]!=-1){continue;}
const TStrV& WordStrV=NmObjWordStrVToDocIdVH.GetKey(NmObjId);
if (WordStrV.Len()==3){
TStr PfxStr=WordStrV[0].GetSubStr(0, 2)+WordStrV[1].GetSubStr(0, 2)+WordStrV[2].GetSubStr(0, 2);
PfxStrToNmObjIdVH.AddDat(PfxStr).Add(NmObjId);
TripleWords++;
}
}
// traverse word-groups with the same prefix
int Pfxs=PfxStrToNmObjIdVH.Len();
for (int PfxId=0; PfxId<Pfxs; PfxId++){
// get & traverse word-group
TIntV& NmObjIdV=PfxStrToNmObjIdVH[PfxId];
for (int NmObjIdN=0; NmObjIdN<NmObjIdV.Len(); NmObjIdN++){
int NmObjId=NmObjIdV[NmObjIdN];
if (NewNmObjIdV[NmObjId]!=-1){continue;}
NewNmObjIdV[NmObjId]=NmObjId;
const TStr& WordStr1=NmObjWordStrVToDocIdVH.GetKey(NmObjId)[0];
const TStr& WordStr2=NmObjWordStrVToDocIdVH.GetKey(NmObjId)[1];
const TStr& WordStr3=NmObjWordStrVToDocIdVH.GetKey(NmObjId)[2];
int Fq=NmObjWordStrVToDocIdVH[NmObjId].Len();
TIntPrV FqNmObjIdPrV(NmObjIdV.Len(), 0);
FqNmObjIdPrV.Add(TIntPr(Fq, NmObjId));
// traverse rest of the word-group for matching words
for (int SubNmObjIdN=NmObjIdN+1; SubNmObjIdN<NmObjIdV.Len(); SubNmObjIdN++){
int SubNmObjId=NmObjIdV[SubNmObjIdN];
if (NewNmObjIdV[SubNmObjId]!=-1){continue;}
const TStr& SubWordStr1=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[0];
const TStr& SubWordStr2=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[1];
const TStr& SubWordStr3=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[2];
// test matching
if (IsMatchPfx(WordStr1, SubWordStr1, MnPfxLen, MxSfxLen+1)&&
IsMatchPfx(WordStr2, SubWordStr2, MnPfxLen, MxSfxLen+1)&&
IsMatchPfx(WordStr3, SubWordStr3, MnPfxLen, MxSfxLen+1)){
NewNmObjIdV[SubNmObjId]=NmObjId;
int SubFq=NmObjWordStrVToDocIdVH[SubNmObjId].Len();
FqNmObjIdPrV.Add(TIntPr(SubFq, SubNmObjId));
//printf("%s_%s_%s -> %s_%s_%s\n",
// WordStr1.CStr(), WordStr2.CStr(), WordStr3.CStr(),
// SubWordStr1.CStr(), SubWordStr2.CStr(), SubWordStr3.CStr());
}
}
// increment number of equivalence word-groups
ReducedTripleWords++;
// collapse matching words into most frequent word
if (FqNmObjIdPrV.Len()>1){
FqNmObjIdPrV.Sort(false);
int MainNmObjId=FqNmObjIdPrV[0].Val2;
NewNmObjIdV[MainNmObjId]=MainNmObjId;
TStr MainWordStr1=NmObjWordStrVToDocIdVH.GetKey(MainNmObjId)[0];
TStr MainWordStr2=NmObjWordStrVToDocIdVH.GetKey(MainNmObjId)[1];
TStr MainWordStr3=NmObjWordStrVToDocIdVH.GetKey(MainNmObjId)[2];
//printf("[%s_%s_%s:", MainWordStr1.CStr(), MainWordStr2.CStr(), MainWordStr3.CStr());
for (int FqNmObjIdPrN=1; FqNmObjIdPrN<FqNmObjIdPrV.Len(); FqNmObjIdPrN++){
int SubNmObjId=FqNmObjIdPrV[FqNmObjIdPrN].Val2;
NewNmObjIdV[SubNmObjId]=MainNmObjId;
//TStr& SubWordStr1=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[0];
//TStr& SubWordStr2=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[1];
//TStr& SubWordStr3=NmObjWordStrVToDocIdVH.GetKey(SubNmObjId)[2];
//printf(" %s_%s_%s", SubWordStr1.CStr(), SubWordStr2.CStr(), SubWordStr3.CStr());
}
//printf("]\n");
}
}
}
// print statistics
//printf("TripleWords:%d ReducedTripleWords:%d\n",
// TripleWords, ReducedTripleWords);
}
}
void TNmObjBs::FilterCandToNmObjIdV(
const TStrV& CandWordStrV, TIntV& NmObjIdV, const bool& DumpP){
// prepare candidate traversal
TVec<TStrV> NmObjIdWordStrVV;
int CandWordStrN=0; int CandWordStrs=CandWordStrV.Len();
while (CandWordStrN<CandWordStrs){
// get candidate
TStr WordStr=CandWordStrV[CandWordStrN];
//printf("%s ", WordStr.CStr());
// simple filters
if (WordStr.Len()<=1){CandWordStrN++; continue;}
if (WordStr==ParagraphTagStr){CandWordStrN++; continue;}
if (WordStr==BreakTagStr){CandWordStrN++; continue;}
if (WordStr==EofTagStr){CandWordStrN++; break;}
if (IsNumStr(WordStr)){CandWordStrN++; continue;}
TStr UcWordStr=ChDef->GetUcStr(WordStr);
//if (SwSet->IsIn(UcWordStr, true)){
// CandWordStrN++; continue;}
if ((WordStr==UcWordStr)&&((WordStr.Len()>4)&&(!IsNmObjAttr(WordStr, noaAcronym)))){
CandWordStrN++; continue;}
// unperiod
if (IsNmObjAttr(WordStr, noaUnperiod)&&(CandWordStrV[CandWordStrN+1]==PeriodTagStr)){
CandWordStrN+=1;
}
// period
if (WordStr==PeriodTagStr){
CandWordStrN++; WordStr=CandWordStrV[CandWordStrN];
if (IsTagStr(WordStr)){continue;}
if (IsNmObjAttr(WordStr, noaDefined)){
continue;
} else
if ((CandWordStrN>1)&&(IsNmObjAttr(CandWordStrV[CandWordStrN-2], noaUnperiod))){
continue;
} else {
TStr NextWordStr=CandWordStrV[CandWordStrN+1];
if (IsFirstCapWordStr(NextWordStr)||IsNmObjAttr(NextWordStr, noaAsCapitalized)){
continue;
} else
if (!IsNmObj(WordStr)){
CandWordStrN++; continue;
}
}
}
// if (WordStr=="British"){
// printf("");}
// ignore
if (IsNmObjAttr(WordStr, noaIgnore)){
CandWordStrN++; continue;
}
// collect named-object words
TStrV WordStrV;
forever {
WordStrV.Add(WordStr);
CandWordStrN++; WordStr=CandWordStrV[CandWordStrN];
if (IsTagStr(WordStr)){break;}
if (WordStr.Len()<=1){break;}
if (IsNmObjAttr(WordStr, noaIgnore)){CandWordStrN++; break;}
if (IsNmObjAttr(WordStr, noaStandalone)){break;}
if (IsNmObjAttr(WordStrV, noaStandalone)){break;}
}
// get normalized version of named-object
TStrV NrWordStrV; GetNrNmObjStrV(WordStrV, NrWordStrV);
// simple filters
if (IsNmObjAttr(NrWordStrV, noaIgnore)){continue;}
if (IsNmObjAttr(NrWordStrV, noaFirstName)){continue;}
if (NrWordStrV.Len()>5){
while (NrWordStrV.Len()>2){NrWordStrV.Del(0);}}
if (NrWordStrV.Len()==1){
TStr UcWordStr=ChDef->GetUcStr(NrWordStrV[0]);
if (SwSet->IsIn(UcWordStr, true)){continue;}
}
// add named object
NmObjIdWordStrVV.Add(NrWordStrV);
}
// merge similar words
for (int NmObjN=0; NmObjN<NmObjIdWordStrVV.Len(); NmObjN++){
TStrV& WordStrV=NmObjIdWordStrVV[NmObjN];
if (WordStrV.Len()==1){
// merge single words
for (int SubNmObjN=0; SubNmObjN<NmObjIdWordStrVV.Len(); SubNmObjN++){
TStrV& SubWordStrV=NmObjIdWordStrVV[SubNmObjN];
if (SubWordStrV.Len()==1){
if (WordStrV[0]!=SubWordStrV[0]){
if (IsMatchPfx(WordStrV[0], SubWordStrV[0], 3, 4)){
// normalize to shorter string
if (WordStrV[0].Len()<SubWordStrV[0].Len()){SubWordStrV=WordStrV;}
else {WordStrV=SubWordStrV;}
}
}
}
}
} else
if (WordStrV.Len()>=2){
TStr LastNm=WordStrV.Last();
for (int SubNmObjN=0; SubNmObjN<NmObjIdWordStrVV.Len(); SubNmObjN++){
TStrV& SubWordStrV=NmObjIdWordStrVV[SubNmObjN];
if (SubWordStrV.Len()==1){
// merge last-name with [first-name,last-name] pairs
TStr SubLastNm=SubWordStrV[0];
if (LastNm!=SubLastNm){
if (IsMatchPfx(LastNm, SubLastNm, 3, 4)){
if (LastNm.Len()<SubLastNm.Len()){SubWordStrV=WordStrV;}
else {WordStrV=SubWordStrV;}
}
}
} else
if (false&&(SubWordStrV.Len()==2)){
// merge [first-name,last-name] with [first-name,last-name] pairs
if ((WordStrV[0]!=SubWordStrV[0])||(WordStrV[1]!=SubWordStrV[1])){
if ((IsMatchPfx(WordStrV[0], SubWordStrV[0], 3, 4))&&
(IsMatchPfx(WordStrV[1], SubWordStrV[1], 3, 4))){
// normalize to shorter string (first word)
if (WordStrV[0].Len()<SubWordStrV[0].Len()){
SubWordStrV[0]=WordStrV[0];}
else {WordStrV[0]=SubWordStrV[0];}
// normalize to shorter string (second word)
if (WordStrV[1].Len()<SubWordStrV[1].Len()){
SubWordStrV[1]=WordStrV[1];}
else {WordStrV[1]=SubWordStrV[1];}
}
}
}
}
}
}
// get named-objects-ids
NmObjIdV.Gen(NmObjIdWordStrVV.Len(), 0);
{for (int NmObjN=0; NmObjN<NmObjIdWordStrVV.Len(); NmObjN++){
TStrV& NmObjWordStrV=NmObjIdWordStrVV[NmObjN];
int NmObjId=GetNmObjId(NmObjWordStrV, true);
NmObjIdV.Add(NmObjId);
}}
// dump
if (DumpP){
printf("Named-Objects: ");
for (int NmObjN=0; NmObjN<NmObjIdV.Len(); NmObjN++){
int NmObjId=NmObjIdV[NmObjN];
TStr NmObjStr=GetNmObjStr(NmObjId);
printf("%s ", NmObjStr.CStr());
}
printf("\n");
}
}
void TUNGraph::GetNIdV(TIntV& NIdV) const {
NIdV.Gen(GetNodes(), 0);
for (int N=NodeH.FFirstKeyId(); NodeH.FNextKeyId(N); ) {
NIdV.Add(NodeH.GetKey(N)); }
}
void TNEGraph::GetEIdV(TIntV& EIdV) const {
EIdV.Gen(GetEdges(), 0);
for (int E=EdgeH.FFirstKeyId(); EdgeH.FNextKeyId(E); ) {
EIdV.Add(EdgeH.GetKey(E));
}
}
void TMultimodalGraphImplB::GetNIdV(TIntV& NIdV) const {
NIdV.Gen(GetNodes(), 0);
for (int N=NodeToModeMapping.FFirstKeyId(); NodeToModeMapping.FNextKeyId(N); ) {
NIdV.Add(NodeToModeMapping.GetKey(N));
}
}
