void TDzsKwBs::AddHits(const PBook& Book, const TStr& FNm){
PSOut SOut=TFOut::New(FNm);
TOLx Lx(SOut, TFSet()|oloFrcEoln|oloSigNum|oloTabSep);
int KwToSecIdVP=KwToSecIdVH.FFirstKeyId();
while (KwToSecIdVH.FNextKeyId(KwToSecIdVP)){
TChA QueryChA=KwToSecIdVH.GetKey(KwToSecIdVP);
for (int QueryChN=0; QueryChN<QueryChA.Len(); QueryChN++){
if ((' '<QueryChA[QueryChN])&&(QueryChA[QueryChN]<='~')){
if (QueryChA[QueryChN]!='*'){
if (!TCh::IsAlNum(QueryChA[QueryChN])){
QueryChA.PutCh(QueryChN, ' ');}
}
}
}
TStr QueryStr=QueryChA;
PBixRes BixRes=Book->Search(QueryStr);
if (BixRes->IsOk()){
Lx.PutUQStr(QueryStr); Lx.PutUQStr(TInt::GetStr(BixRes->GetHits()));
printf("%s - %d\n", QueryStr.CStr(), BixRes->GetHits());
int Hits=BixRes->GetHits(); // TInt::GetMn(BixRes->GetHits(), 10);
for (int HitN=0; HitN<Hits; HitN++){
TStr SecIdStr; TStr TitleStr; TStr SecStr; TStr CtxStr;
BixRes->GetHitInfo(
Book, HitN, -1, 100, SecIdStr, TitleStr, SecStr, CtxStr);
KwToSecIdVH[KwToSecIdVP].Add(SecIdStr);
Lx.PutUQStr(SecIdStr);
}
Lx.PutLn();
} else {
Lx.PutUQStr(QueryStr); Lx.PutUQStr("Error"); Lx.PutStr(BixRes->GetMsgStr());
printf("%s - Error - %s\n", QueryStr.CStr(), BixRes->GetMsgStr().CStr());
Lx.PutLn();
}
}
}
void TWebTxtBsTrmSrv::OnLn(const int& CltSockId, const TStr& Ln){
TChA InNotifyChA;
InNotifyChA+=TStr("[")+TInt::GetStr(CltSockId)+"] ";
//InNotifyChA+=TStr("[")+GetCltPeerNm(CltSockId)+"] ";
InNotifyChA+=TStr("[")+TSecTm::GetCurTm().GetStr()+"] ";
InNotifyChA+=Ln;
TNotify::OnNotify(WebTxtBs->GetNotify(), ntInfo, InNotifyChA);
PSIn SIn=TMIn::New(Ln);
TILx Lx(SIn, TFSet()|iloExcept);
TChA MsgChA; bool SendMsgToClt=true;
Lx.GetSym();
if (Lx.Sym==syIdStr){
try {
ParseAndExeCmLn(CltSockId, Lx, SendMsgToClt, MsgChA);
}
catch (PExcept Except){
MsgChA=TStr("Error: ")+Except->GetMsgStr()+"\r\n";
}
} else {
if (Lx.Sym!=syEof){
MsgChA="Error: No command keyword.\r\n";}
}
if (SendMsgToClt){
SendTxtToClt(CltSockId, MsgChA);
SendTxtToClt(CltSockId, ">");
}
}
void TYFSelBs::SaveTxt(
const PSOut& SOut, const PYBs& YBs, const PYDsBs& YDsBs){
PYWordDs NegWordDs=TYDmDs::GetNegWordDs(YNegDsType, YBs, YDsBs);
TOLx Lx(SOut, TFSet()|oloFrcEoln|oloSigNum);
for (int DocId=0; DocId<DocIdToWordIdEstVV.Len(); DocId++){
TIntFltKdV& WordIdEstKdV=DocIdToWordIdEstVV[DocId];
TFltIntKdV WordEstIdKdV(WordIdEstKdV.Len(), 0);
for (int WordIdN=0; WordIdN<WordIdEstKdV.Len(); WordIdN++){
int WordId=WordIdEstKdV[WordIdN].Key;
double WordEst=WordIdEstKdV[WordIdN].Dat;
WordEstIdKdV.Add(TFltIntKd(WordEst, WordId));
}
WordEstIdKdV.Sort(false);
Lx.PutVarStr("UrlStr", YBs->GetDocUrlStr(DocId));
Lx.PutVarInt("DocId", DocId);
PYWordDs WordDs=YDsBs->GetWordDs(DocId);
Lx.PutVar("WordIdEstKdV", true, true);
{for (int WordIdN=0; WordIdN<WordEstIdKdV.Len(); WordIdN++){
double WordEst=WordEstIdKdV[WordIdN].Key;
int WordId=WordEstIdKdV[WordIdN].Dat;
TStr WordStr=YBs->GetWordStr(WordId);
double PosWordPrb=WordDs->GetWordPrb(WordId);
double NegWordPrb=NegWordDs->GetWordPrb(WordId);
Lx.PutQStr(WordStr);
Lx.PutTab(); Lx.PutFlt(WordEst);
Lx.PutIndent(1); Lx.PutSym(syLBracket);
Lx.PutFlt(PosWordPrb); Lx.PutFlt(NegWordPrb);
Lx.PutSym(syRBracket);
Lx.PutLn();
}}
Lx.PutSym(syRBracket); Lx.PutLn();
}
}
void TWebTxtBsSrv::PrepTplSet(const TStr& FNm, const TStr& DfFNm){
// get template input-stream (priority order: FNm, DfFNm, DfTplSetStr)
PSIn SIn;
bool FExists=false; SIn=TFIn::New(FNm, FExists);
if (!FExists){SIn=TFIn::New(DfFNm, FExists);}
if (!FExists){SIn=TStrIn::New(GetDfTplSetStr());}
TILx Lx(SIn, TFSet()|iloRetEoln|iloSigNum|iloUniStr);
Lx.GetStrToEolnAndCh('.');
do {
TStr SecNm=Lx.GetStr();
if (SecNm=="End"){Lx.GetStrToEolnAndCh('.');}
else if (SecNm=="Stop"){break;}
else if (SecNm=="Template"){
Lx.GetStr("ContentType"); Lx.GetSym(syEq); TStr ContTypeStr=Lx.GetStr();
Lx.GetStr("Category"); Lx.GetSym(syEq); TStr CatStr=Lx.GetStr();
Lx.GetEoln();
TStr Tpl=Lx.GetStrToEolnAndCh('.');
Tpl.ChangeStrAll(HostNmMacro, HostNm);
Tpl.ChangeStrAll(PortMacro, PortNStr);
AddTpl(ContTypeStr, CatStr, Tpl);
} else if (SecNm=="Variable"){
Lx.GetStr("ContentType"); Lx.GetSym(syEq); TStr ContTypeStr=Lx.GetStr();
TStr VarNm=Lx.GetStr(); Lx.GetSym(syEq); TStr VarVal=Lx.GetStr();
Lx.GetEoln();
AddVar(ContTypeStr, VarNm, VarVal);
Lx.GetStrToEolnAndCh('.');
} else if (SecNm=="Remark"){
Lx.GetStrToEolnAndCh('.');
} else {
Fail;
}
} while (!Lx.IsEof());
}
void TNmObjBs::LoadNrWordBs(const PSIn& SIn){
if (SIn.Empty()){return;}
TILx Lx(SIn, TFSet(iloCmtAlw, iloRetEoln, iloExcept));
// traverse lines
Lx.GetSym(syQStr, syEoln, syEof);
while (Lx.Sym!=syEof){
if (Lx.Sym==syQStr){
// get normalized word
TStr NrWordStr=Lx.Str;
// get inflected words
Lx.GetSym(syColon);
Lx.GetSym(syQStr, syEoln);
while (Lx.Sym!=syEoln){
// get inflected word
TStr WordStr=Lx.Str;
// test if inflected word already exists
if (WordStrToNrH.IsKey(WordStr)){
printf("Word already normalized (%s)", WordStr.CStr());}
// add inflected word and corresponding normalized word
WordStrToNrH.AddDat(WordStr, NrWordStr);
//printf("'%s' ->'%s'\n", WordStr.CStr(), NrWordStr.CStr());
Lx.GetSym(syQStr, syEoln);
}
Lx.GetSym(syQStr, syEoln, syEof);
} else
if (Lx.Sym==syEoln){
// empty line
Lx.GetSym(syQStr, syEoln, syEof);
} else {
Fail;
}
}
}
PExpVal TExp::LoadAndEvalExpL(
const TStr& ExpLStr, bool& Ok, TStr& MsgStr, const PExpEnv& ExpEnv){
// create final expression value
PExpVal ExpVal;
// transform exp. str. to input stream
PSIn SIn=TStrIn::New(ExpLStr);
// create lexical
TILx Lx(SIn, TFSet()|iloCmtAlw|iloCsSens|iloExcept);
TFSet Expect=TFSet()|sySemicolon|syEof;
// load & evaluate expression separated by semicolon
while (Lx.Sym!=syEof){
// create expression
PExp Exp; Ok=true; MsgStr="Ok";
try {
Lx.GetSym(ExpExpect);
Exp=LoadTxtExp(Lx, Expect);
}
catch (PExcept Except){
Ok=false; MsgStr=Except->GetMsgStr();
Exp=PExp(new TExp(etUndef));
}
// evaluate expression
if (Ok){
ExpVal=Exp->Eval(Ok, MsgStr, ExpEnv);
//printf("%s\n", ExpVal->GetStr().CStr());
if (!Ok){
return NULL;}
} else {
return NULL;
}
}
return ExpVal;
}
void TCpDoc::SaveAcmTechNewsToCpd(
const TStr& InFPath, const TStr& OutCpdFNm, const int& MxDocs){
// create output file
PSOut SOut=TFOut::New(OutCpdFNm);
// processing xml files
TFFile FFile(TStrV()+InFPath, TStrV()+".Html"+".Htm", "", true);
TStr FNm; int Docs=0;
while (FFile.Next(FNm)){
printf("Processing file '%s'\r", FNm.CStr());
PSIn SIn=TFIn::New(FNm);
THtmlLx Lx(SIn);
while (Lx.GetSym()!=hsyEof){
//printf("%d\r", Docs);
if ((MxDocs!=-1)&&(Docs>MxDocs)){break;}
Lx.MoveToBTagOrEof("<SPAN>");
if (Lx.GetArg("CLASS")!="title"){continue;}
Lx.MoveToBTagOrEof("<A>");
TStr TitleStr=Lx.GetStrToETag("<A>", false); TitleStr="";
Lx.MoveToETagOrEof("<SPAN>");
Lx.MoveToBTagOrEof("<P>");
TStr ParStr=Lx.GetStrToETag("<P>", false);
if (!ParStr.Empty()){
Docs++;
PCpDoc CpDoc=TCpDoc::New(TInt::GetStr(Docs), TitleStr, ParStr);
CpDoc->Save(*SOut);
}
}
}
printf("\n");
}
/*!
Compute and return the interaction matrix \f$ L \f$ for the whole
features or a part of them.
\param select : Selection of a subset of the possible features.
- To compute the interaction matrix for all the features use
vpBasicFeature::FEATURE_ALL. In that case the dimension of the interaction
matrix is \f$ [number of features \times 6] \f$
- To compute the interaction matrix for only one of the component
feature you have to say which one you want to take into
account. If it is the first one set select to
vpBasicFeature::FEATURE_LINE[0], if it is the second one set
select to vpBasicFeature::FEATURE_LINE[1], and so on. In that case
the returned interaction matrix is \f$ [1 \times 6] \f$ dimension.
- To compute the interaction matrix for only two of the component
features you have to say which ones you want to take into
account. If it is the first one and the second one set select to
vpBasicFeature::FEATURE_LINE[0] | vpBasicFeature::FEATURE_LINE[1]. In
that case the returned interaction matrix is \f$ [2 \times 6] \f$
dimension.
\return The interaction matrix computed from the features.
The code below shows how to compute the interaction matrix associated to the
first visual feature.
\code
// Creation of the current feature s
vpGenericFeature s(3);
s.set_s(0, 0, 0);
// Here you have to compute the interaction matrix L for all the three features
s.setInteractionMatrix(L);
vpMatrix L_x = s.interaction( vpBasicFeature::FEATURE_LINE[0] );
\endcode
The code below shows how to compute the interaction matrix associated to two
visual features over three.
\code
// Creation of the current feature s
vpGenericFeature s(3);
s.set_s(0, 0, 0);
// Here you have to compute the interaction matrix L
s.setInteractionMatrix(L);
vpMatrix L_x = s.interaction( vpBasicFeature::FEATURE_LINE[0]|vpBasicFeature::FEATURE_LINE[1] );
\endcode
*/
vpMatrix
vpGenericFeature::interaction(const unsigned int select)
{
if (L.getRows() == 0)
{
std::cout << "interaction matrix " << L << std::endl ;
vpERROR_TRACE("Interaction has not been initialized");
std::cout << "A possible reason (may be) is that you have set" << std::endl ;
std::cout << "the interaction matrix for s and compute a control " << std::endl ;
std::cout << "with Ls=s* (default) or vice versa" << std::endl ;
throw(vpFeatureException(vpFeatureException::notInitializedError,
"size mismatch between s* dimension "
"and feature dimension"));
}
vpMatrix Ls ;
Ls.resize(0,6) ;
for (unsigned int i=0 ; i < dim_s ; i++)
if (FEATURE_LINE[i] & select )
{
vpMatrix Lx(1,6) ; Lx = 0;
for (int j=0 ; j < 6 ; j++)
Lx[0][j] = L[i][j] ;
Ls = vpMatrix::stackMatrices(Ls,Lx) ;
}
return Ls ;
}
Color Phong(const MatteMaterial &m, float NdotL, float RdotV) {
Color Lx(1,1,1);
Color I(0,0,0);
I.r = m.Ax.r*m.Ka*m.Dx.r + m.att*Lx.r*(m.Kd*m.Dx.r*(NdotL) + m.Ks*m.Sx.r*pow(RdotV, m.n));
I.g = m.Ax.g*m.Ka*m.Dx.g + m.att*Lx.g*(m.Kd*m.Dx.g*(NdotL) + m.Ks*m.Sx.g*pow(RdotV, m.n));
I.b = m.Ax.b*m.Ka*m.Dx.b + m.att*Lx.b*(m.Kd*m.Dx.b*(NdotL) + m.Ks*m.Sx.b*pow(RdotV, m.n));
return I;
}
// for growing
void base_bin_rel::move_from (const base_bin_rel & other)
{
POMAGMA_DEBUG("Copying base_bin_rel");
size_t min_item_dim = min(item_dim(), other.item_dim());
size_t min_word_dim = min(word_dim(), other.word_dim());
m_support.move_from(other.m_support);
if (_symmetric()) {
POMAGMA_ASSERT(other._symmetric(), "symmetry mismatch");
for (size_t i = 1; i <= min_item_dim; ++i) {
memcpy(Lx(i), other.Lx(i), sizeof(Word) * min_word_dim);
}
} else {
POMAGMA_ASSERT(not other._symmetric(), "symmetry mismatch");
for (size_t i = 1; i <= min_item_dim; ++i) {
memcpy(Lx(i), other.Lx(i), sizeof(Word) * min_word_dim);
memcpy(Rx(i), other.Rx(i), sizeof(Word) * min_word_dim);
}
}
}
/////////////////////////////////////////////////
// Dzs-Keyword-Base
TDzsKwBs::TDzsKwBs(const TStr& FNm):
KwToSecIdVH(10000){
PSIn SIn=TFIn::New(FNm);
TILx Lx(SIn, TFSet()|iloTabSep);
Lx.GetSym(syStr, syEoln, syEof);
while (Lx.Sym!=syEof){
if (Lx.Sym==syStr){
KwToSecIdVH.AddDat(Lx.Str);
while (Lx.GetSym()!=syEoln){}
}
Lx.GetSym(syStr, syEoln, syEof);
}
}
PExp TExp::LoadTxt(
const PSIn& SIn, bool& Ok, TStr& MsgStr, const TFSet& Expect){
TILx Lx(SIn, TFSet()|iloCmtAlw|iloCsSens|iloExcept);
PExp Exp; Ok=true; MsgStr="Ok";
try {
Lx.GetSym(ExpExpect);
Exp=LoadTxtExp(Lx, Expect);
}
catch (PExcept Except){
Ok=false; MsgStr=Except->GetMsgStr();
Exp=PExp(new TExp(etUndef));
}
return Exp;
}
/////////////////////////////////////////////////
// SqlDm-Program
PSqlDmProg TSqlDmProg::LoadTxt(PSIn& SIn){
// create object
PSqlDmProg Prog=TSqlDmProg::New();
// parse object
TSqlDmLx Lx(SIn);
Lx.GetSym(TFSet(TSqlDmStat::StatExpect)|dsyEof);
while (Lx.Sym!=dsyEof){
PSqlDmStat Stat=TSqlDmStat::LoadTxt(Lx, TFSet(dsySemicolon, dsyEof));
Prog->StatV.Add(Stat);
Lx.GetSym(TFSet(TSqlDmStat::StatExpect)|dsyEof);
}
// return object
return Prog;
}
void TYInvIx::SaveTxt(const PSOut& SOut, const PYBs& YBs){
TOLx Lx(SOut, TFSet()|oloFrcEoln|oloSigNum);
int WordIdToFirstDocIdNP=WordIdToFirstDocIdNH.FFirstKeyId();
while (WordIdToFirstDocIdNH.FNextKeyId(WordIdToFirstDocIdNP)){
int WordId=WordIdToFirstDocIdNH.GetKey(WordIdToFirstDocIdNP);
TStr WordStr;
if (YBs.Empty()){WordStr=TInt::GetStr(WordId);}
else {WordStr=YBs->GetWordStr(WordId);}
Lx.PutStr(WordStr); Lx.PutSym(syColon); Lx.PutSym(syLBracket);
int DocIdN=FFirstDocId(WordId); int DocId;
while (FNextWordId(DocIdN, DocId)){
Lx.PutInt(DocId);}
Lx.PutSym(syRBracket); Lx.PutLn();
}
}
PTb TTb::LoadTxtTab(const TStr& FNm, const PTb& Tb){
// open file
PSIn SIn=TFIn::New(FNm);
TILx Lx(SIn, TFSet()|iloTabSep);
// read variable names
Lx.GetStr();
bool IsTupNm=(Lx.UcStr=="NAME")||(Lx.UcStr=="ID");
if (IsTupNm){Lx.GetSym(syTab); Lx.GetStr();}
do {
PTbVarType VarType=PTbVarType(new TTbVarType());
PTbVar Var=PTbVar(new TGTbVar(Lx.Str, VarType));
Tb->AddVar(Var);
if (Lx.GetSym(syTab, syEoln, syEof)==syTab){
Lx.GetSym(syStr);}
} while (Lx.Sym==syStr);
// read tuples
if (Lx.Sym!=syEof){
Lx.GetSym(syStr, syEof);}
while (Lx.Sym!=syEof){
int TupN;
if (IsTupNm){
TupN=Tb->AddTup(Lx.GetSymStr());
Lx.GetSym(syTab); Lx.GetStr();}
else {TupN=Tb->AddTup();}
for (int VarN=0; VarN<Tb->GetVars(); VarN++){
if (VarN>0){Lx.GetSym(syTab); Lx.GetStr();}
if (Lx.Str=='?'){Tb->PutVal(TupN, VarN, TTbVal(tvtUnknw));}
else if (Lx.Str=='*'){Tb->PutVal(TupN, VarN, TTbVal(tvtUnimp));}
else if (Lx.Str=='/'){Tb->PutVal(TupN, VarN, TTbVal(tvtUnapp));}
else {
double Flt;
if (TStr(Lx.Str).IsFlt(Flt)){
Tb->PutFlt(TupN, VarN, Flt);
} else {
Tb->PutDsc(TupN, VarN, Lx.Str);
}
}
}
//printf("Tups: %d\r", Tb->GetTups());
Lx.GetSym(syEoln, syEof);
while (Lx.Sym==syEoln){Lx.GetSym(syStr, syEoln, syEof);}
}
Tb->DefVarTypes();
return Tb;
}
void TNmObjBs::LoadNmObjTypeBs(const PSIn& SIn){
if (SIn.Empty()){return;}
TILx Lx(SIn, TFSet(iloCmtAlw, iloRetEoln, iloExcept));
// traverse lines
Lx.GetSym(syQStr, syIdStr, syEoln, syEof);
while (Lx.Sym!=syEof){
if ((Lx.Sym==syQStr)||(Lx.Sym==syIdStr)){
TVec<TStrV> NmObjWordStrVV;
TB32Set NmObjAttrSet;
while ((Lx.Sym==syQStr)||(Lx.Sym==syIdStr)){
if (Lx.Sym==syQStr){
// named-object word-string
TStr WordStrVStr=Lx.Str;
TStrV WordStrV; WordStrVStr.SplitOnWs(WordStrV);
NmObjWordStrVV.Add(WordStrV);
} else
if (Lx.Sym==syIdStr){
// named-object attribute
TNmObjAttr NmObjAttr=TNmObjBs::GetNmObjTypeFromStr(Lx.Str);
NmObjAttrSet.Incl(NmObjAttr);
} else {
Fail;
}
Lx.GetSym(syQStr, syIdStr, syEoln, syEof);
}
// assign 'defined' attribute if 'not ignore'
if (!NmObjAttrSet.In(noaIgnore)){
NmObjAttrSet.Incl(noaDefined);}
// assign attribute-sets to word-vectors
for (int NmObjN=0; NmObjN<NmObjWordStrVV.Len(); NmObjN++){
WordStrVToNmObjAttrSetH.AddDat(NmObjWordStrVV[NmObjN])|=NmObjAttrSet;
}
// assign aliases
{for (int NmObjN=1; NmObjN<NmObjWordStrVV.Len(); NmObjN++){
NmObjWordStrVToNrH.AddDat(NmObjWordStrVV[NmObjN], NmObjWordStrVV[0]);
}}
// get eoln
} else
if (Lx.Sym==syEoln){
// empty line
Lx.GetSym(syQStr, syEoln, syEof);
} else {
Fail;
}
}
}
TWixExp::TWixExp(const TStr& _ExpStr):
ExpStr(_ExpStr), Ok(true), ErrMsg(), ExpItem(),
FactExpect(), TermExpect(), ExpExpect(){
// prepare expect sets
FactExpect=TFSet()|wesyNot|wesyIncl|wesyExcl|wesyLParen|wesyDQuote|wesyWord;
TermExpect=FactExpect;
ExpExpect=TermExpect;
// parse the expression string
TWixExpLx Lx(ExpStr);
try {
Lx.GetSym(ExpExpect);
ExpItem=ParseExp(Lx, TFSet()|wesyEof);
}
catch (EWixExp E){
Ok=false; ErrMsg=E.GetErrMsg();
}
}
void TNmObjBs::LoadCustSwSet(const PSIn& SIn){
if (SIn.Empty()){return;}
TILx Lx(SIn, TFSet(iloCmtAlw, iloRetEoln, iloExcept));
// traverse lines
Lx.GetSym(syLn, syEof);
while (Lx.Sym!=syEof){
// get stop-phrase string
TStr WordStrVStr=Lx.Str;
// split phrase to words
TStrV WordStrV; WordStrVStr.SplitOnWs(WordStrV);
if (!WordStrV.Empty()){
// define phrase as stop-word
WordStrVToNmObjAttrSetH.AddDat(WordStrV).Incl(noaIgnore);
}
// get next symbol
Lx.GetSym(syLn, syEof);
}
}
void TWebTravelCmuPww::StartTravel(){
PutConstrs();
TStrV UrlStrV(1000, 0);
PSIn SIn=PSIn(new TFIn(InFNm));
TILx Lx(SIn, TFSet()|iloRetEoln);
Lx.GetSym(syQStr, syEof);
while (Lx.Sym!=syEof){
TStr UrlStr=Lx.Str;
if (IsUrlOk(UrlStr)){
UrlStrV.Add(UrlStr);}
Lx.GetStrToEoln();
Lx.GetEoln();
Lx.GetSym(syQStr, syEof);
}
for (int UrlStrN=0; UrlStrN<UrlStrV.Len(); UrlStrN++){
Go(UrlStrV[UrlStrN]);
}
}
TStr TCordisEuPart::GetNrOrgNm(const TStr& Str){
TStrV NrStrV; TChA ChA;
PSwSet SwSet=TSwSet::GetSwSet(swstEnglish523);
PSIn HtmlSIn=TStrIn::New(Str);
THtmlLx Lx(HtmlSIn);
while (Lx.Sym!=hsyEof){
if ((Lx.Sym==hsyStr)||(Lx.Sym==hsyNum)){
Lx.UcChA.ChangeCh('Â', 'A');
Lx.UcChA.ChangeCh('Á', 'A');
Lx.UcChA.ChangeCh('Ä', 'A');
Lx.UcChA.ChangeCh('Ã', 'A');
Lx.UcChA.ChangeCh('À', 'A');
Lx.UcChA.ChangeCh('Ç', 'C');
Lx.UcChA.ChangeCh('É', 'E');
Lx.UcChA.ChangeCh('È', 'E');
Lx.UcChA.ChangeCh('Ê', 'E');
Lx.UcChA.ChangeCh('Í', 'I');
Lx.UcChA.ChangeCh('Ó', 'O');
Lx.UcChA.ChangeCh('Ö', 'O');
Lx.UcChA.ChangeCh('Ñ', 'N');
Lx.UcChA.ChangeCh('Ü', 'U');
Lx.UcChA.ChangeCh('Ú', 'U');
TStr NrStr=THtmlLx::GetAsciiStr(Lx.UcChA);
if ((NrStr.Len()==1)||(SwSet->IsIn(NrStr))||
(NrStr=="AG")||(NrStr=="GMBH")||(NrStr=="LTD")||(NrStr=="SRL")||
(NrStr=="SA")||(NrStr=="EV")||(NrStr=="AS")||(NrStr=="AB")||
(NrStr=="NV")||(NrStr=="BV")||
(NrStr=="AKTIENGESELLSCHAFT")||(NrStr=="LIMITED")||
(NrStr=="FUER")||(NrStr=="FUR")){
NrStr="";}
if (!NrStr.Empty()){NrStrV.Add(NrStr);}
}
Lx.GetSym();
}
TChA NrChA;
for (int NrStrN=0; NrStrN<NrStrV.Len(); NrStrN++){
if (NrStrN>0){NrChA+=' ';}
NrChA+=NrStrV[NrStrN];
}
if (NrChA.Empty()){
NrChA=Str.GetUc();
}
return NrChA;
}
PTb TTb::LoadTxtCsv(const TStr& FNm, const PTb& Tb){
// open file
PSIn SIn=TFIn::New(FNm);
TILx Lx(SIn, TFSet()|iloRetEoln|iloSigNum|iloUniStr);
// read variable names
Lx.GetStr();
bool IsTupNm=(Lx.UcStr=="NAME")||(Lx.UcStr=="ID");
if (IsTupNm){Lx.GetSym(syComma); Lx.GetStr();}
do {
PTbVarType VarType=PTbVarType(new TTbVarType());
PTbVar Var=PTbVar(new TGTbVar(Lx.Str, VarType));
Tb->AddVar(Var);
if (Lx.GetSym(syComma, syEoln, syEof)==syComma){
Lx.GetSym(syStr);}
} while (Lx.Sym==syStr);
// read tuples
TFSet ValSymSet=TFSet()|syQuestion|syAsterisk|sySlash|syFlt|syStr;
if (Lx.Sym!=syEof){
Lx.GetSym(TFSet(ValSymSet)|syEof);}
while (Lx.Sym!=syEof){
int TupN;
if (IsTupNm){
TupN=Tb->AddTup(Lx.GetSymStr());
Lx.GetSym(syComma); Lx.GetSym(ValSymSet);}
else {TupN=Tb->AddTup();}
for (int VarN=0; VarN<Tb->GetVars(); VarN++){
if (VarN>0){Lx.GetSym(syComma); Lx.GetSym(ValSymSet);}
switch (Lx.Sym){
case syQuestion: Tb->PutVal(TupN, VarN, TTbVal(tvtUnknw)); break;
case syAsterisk: Tb->PutVal(TupN, VarN, TTbVal(tvtUnimp)); break;
case sySlash: Tb->PutVal(TupN, VarN, TTbVal(tvtUnapp)); break;
case syStr: Tb->PutDsc(TupN, VarN, Lx.Str); break;
case syFlt: Tb->PutFlt(TupN, VarN, Lx.Flt); break;
default: Fail;
}
}
Lx.GetSym(syEoln, syEof);
while (Lx.Sym==syEoln){Lx.GetSym(TFSet(ValSymSet)|syEoln|syEof);}
}
Tb->DefVarTypes();
return Tb;
}
TTqlExp::TTqlExp(const TStr& _ExpStr):
ExpStr(_ExpStr), Ok(true), ErrMsg(), Item(),
RelOpSet(), FactExpect(), TermExpect(), ExpExpect(){
// symbol sets
RelOpSet=TFSet(tsyLss, tsyLEq, tsyEq, tsyNEq, tsyGEq, tsyGtr);
FactExpect=TFSet(tsyNot, tsyIncl, tsyExcl, tsyTitle, tsyLParen, tsyLBracket,
tsyDQuote, tsyWord);
TermExpect=FactExpect;
ExpExpect=TermExpect;
TTqlLx Lx(ExpStr);
try {
Lx.GetSym(ExpExpect);
Item=ParseExp(Lx, TFSet()|tsyEof);
}
catch (ETql E){
Ok=false; ErrMsg=E.GetErrMsg();
}
}
TBible::TBible(const TStr& FNm):
TBook("Bible", "King James Version", "Religious"),
ChpNmSecIdKdV(), PsalmV(), Bix(TBix::New()){
// open file
PSIn SIn=TFIn::New(FNm);
TILx Lx(SIn, TFSet());
// define state variables
bool InPsalm=false; TStr ChpNm; TChA SecNChA; int SecN;
TChA SSecNChA; int SSecN; TStr SSecStr;
// parsing
while (Lx.GetSym(syLn, syEof)!=syEof){
TChA& Ln=Lx.Str;
int VBarChN=Ln.SearchCh('|');
if ((VBarChN!=-1)&&(0<VBarChN)&&(VBarChN<Lx.Str.Len())&&
TCh::IsAlNum(Ln[VBarChN-1])&&TCh::IsNum(Ln[VBarChN+1])){
if (InPsalm){AddPsalm(ChpNm, SecN, SSecN, SSecStr);}
// if (PsalmV.Len()>100){return;}
InPsalm=true;
// psalm name
ChpNm=Ln.GetSubStr(0, VBarChN-1).GetTrunc();
if (ChpNmSecIdKdV.Empty()||ChpNmSecIdKdV.Last().Key!=ChpNm){
ChpNmSecIdKdV.Add(TStrIntKd(ChpNm, PsalmV.Len()));}
// section number
SecNChA.Clr(); int ChN=VBarChN+1;
while (TCh::IsNum(Ln[ChN])){SecNChA+=Ln[ChN]; ChN++;}
SecN=TStr(SecNChA).GetInt();
IAssert(Ln[ChN]==':');
// subsection number
SSecNChA.Clr(); ChN++;
while (TCh::IsNum(Ln[ChN])){SSecNChA+=Ln[ChN]; ChN++;}
SSecN=TStr(SSecNChA).GetInt();
// text
SSecStr=Ln.GetSubStr(ChN, Ln.Len()-1).GetTrunc();
} else
if (InPsalm){
SSecStr=(SSecStr+' '+Ln).GetTrunc();
}
}
// save last psalm
if (InPsalm){AddPsalm(ChpNm, SecN, SSecN, SSecStr);}
}
void TWebTravelHomeNet::StartTravel(){
PutConstrs();
TStrV UrlStrV(300000, 0);
TIntIntH UserIdToDocsH(1000);
PSIn SIn=PSIn(new TFIn(InFNm));
TILx Lx(SIn, TFSet()|iloRetEoln);
TChA UrlStr;
Lx.GetSym(syInt, syEof);
while ((Lx.Sym!=syEof)&&(Lx.SymLnN<200000)){
// while (Lx.Sym!=syEof){
int UserId=Lx.Int; Lx.GetSym(syComma);
Lx.GetInt(); Lx.GetSym(syComma);
Lx.GetInt(); Lx.GetSym(syComma);
Lx.GetInt(); Lx.GetSym(syComma);
Lx.GetInt(); Lx.GetSym(syComma);
TStr Method=Lx.GetIdStr(); Lx.GetSym(syComma); // GET, POST
UrlStr.Clr(); UrlStr+=Lx.GetIdStr(); Lx.GetSym(syComma); // http, ftp
UrlStr+="://";
UrlStr+=Lx.GetStrToCh(','); Lx.GetSym(syComma); // domain name
UrlStr+=Lx.GetStrToEoln(); Lx.GetEoln(); // path
if ((UserId==TgUserId)&&IsUrlOk(UrlStr)&&(Method=="GET")){
UserIdToDocsH.AddDat(UserId)++;
UrlStrV.Add(UrlStr);
}
Lx.GetSym(syInt, syEof);
if (Lx.SymLnN%100000==0){OnNotify(TInt::GetStr(Lx.SymLnN)+ " docs");}
}
int UserIdToDocsP=UserIdToDocsH.FFirstKeyId();
while (UserIdToDocsH.FNextKeyId(UserIdToDocsP)){
int UserId=UserIdToDocsH.GetKey(UserIdToDocsP);
int Docs=UserIdToDocsH[UserIdToDocsP];
TStr MsgStr=TStr("User ")+TInt::GetStr(UserId)+": "+
TInt::GetStr(Docs)+" Docs.";
OnNotify(MsgStr);
}
UrlStrV.Shuffle(TRnd());
for (int UrlStrN=0; UrlStrN<UrlStrV.Len(); UrlStrN++){
Go(UrlStrV[UrlStrN]);
}
}
void BinaryRelation::validate () const
{
POMAGMA_INFO("Validating BinaryRelation");
m_lines.validate();
size_t num_pairs = 0;
DenseSet Lx(round_item_dim(), nullptr);
DenseSet Rx(round_item_dim(), nullptr);
for (Ob i = 1; i <= item_dim(); ++i) {
bool sup_i = supports(i);
Lx.init(m_lines.Lx(i));
for (Ob j = 1; j <= item_dim(); ++j) {
bool sup_ij = sup_i and supports(j);
Rx.init(m_lines.Rx(j));
bool Lx_ij = Lx.contains(j);
bool Rx_ij = Rx.contains(i);
num_pairs += Rx_ij;
POMAGMA_ASSERT(Lx_ij == Rx_ij,
"Lx,Rx disagree at " << i << "," << j
<< ", Lx is " << Lx_ij << ", Rx is " << Rx_ij );
POMAGMA_ASSERT(sup_ij or not Lx_ij,
"Lx unsupported at " << i << "," << j );
POMAGMA_ASSERT(sup_ij or not Rx_ij,
"Rx unsupported at " << i << "," << j );
}
}
size_t true_size = count_pairs();
POMAGMA_ASSERT(num_pairs == true_size,
"incorrect number of pairs: "
<< num_pairs << " should be " << true_size);
}
void TTb::SaveTxt(const TStr& FNm, const bool& SaveTupNm){
TOLx Lx(PSOut(new TFOut(FNm)), TFSet()|oloFrcEoln|oloSigNum|oloUniStr|oloTabSep);
if (SaveTupNm){Lx.PutStr("Name");}
for (int VarN=0; VarN<GetVars(); VarN++){
Lx.PutStr(GetVar(VarN)->GetNm());}
Lx.PutLn();
for (int TupN=0; TupN<GetTups(); TupN++){
if (SaveTupNm){Lx.PutStr(GetTupNm(TupN));}
for (int VarN=0; VarN<GetVars(); VarN++){
TTbVal Val=GetVal(TupN, VarN);
switch (Val.GetValTag()){
case tvtUnknw: Lx.PutSym(syQuestion); break;
case tvtUnimp: Lx.PutSym(syAsterisk); break;
case tvtUnapp: Lx.PutSym(sySlash); break;
case tvtDsc: Lx.PutStr(GetVar(VarN)->GetVarType()->GetValStr(Val)); break;
case tvtFlt: Lx.PutFlt(Val.GetFlt()); break;
default: Fail;
}
}
Lx.PutLn();
}
}
PPlBs TPlBs::LoadTxtPl(const TStr& FNmWc){
PPlBs PlBs=TPlBs::New();
printf("Loading Prolog from '%s'...\n", FNmWc.CStr());
TFFile FFile(FNmWc); TStr FNm;
while (FFile.Next(FNm)){
printf(" ...loading '%s'\n", FNm.CStr());
// process prolog-file
PSIn SIn=TFIn::New(FNm);
TILx Lx(SIn, TFSet()|/*iloList|*/iloCmtAlw|iloCsSens|iloUniStr|iloExcept);
Lx.GetSym(TFSet(TupExpect)|syEof);
while (Lx.Sym!=syEof){
TPlVal TupVal=TPlBs::ParseTup(Lx, TFSet()|syPeriod, PlBs);
int FuncId=PlBs->TupV[TupVal.GetTupId()]->GetFuncId();
int Arity=PlBs->TupV[TupVal.GetTupId()]->GetVals();
PlBs->FuncIdArityPrToTupIdVH.
AddDat(TIntPr(FuncId, Arity)).Add(TupVal.GetTupId());
Lx.GetSym(TFSet(TupExpect)|syEof);
}
//break;
}
printf("Done.\n");
return PlBs;
}
void base_bin_rel::validate() const
{
m_support.validate();
// TODO validate Lx, Rx agree with support
// (move code over from dense_bin_rel::validate()
if (_symmetric()) {
// check emptiness outside of support
dense_set set(item_dim(), NULL);
dense_set round_set(m_round_item_dim, NULL);
for (oid_t i = 0; i < m_round_item_dim; ++i) {
if (1 <= i and i <= item_dim()) {
set.init(Lx(i));
set.validate();
} else {
round_set.init(m_Lx_lines + m_round_word_dim * i);
round_set.validate();
POMAGMA_ASSERT(round_set.empty(),
"unsupported Lx(" << i << ") has " <<
round_set.count_items() << " items");
}
}
// check for Lx/Rx agreement
for (oid_t i = 1; i <= item_dim(); ++i) {
for (oid_t j = i; j <= item_dim(); ++j) {
POMAGMA_ASSERT(Lx(i, j) == Rx(i, j),
"Lx, Rx disagree at " << i << ',' << j);
}}
} else {
// check emptiness outside of support
dense_set set(item_dim(), NULL);
dense_set round_set(m_round_item_dim, NULL);
for (oid_t i = 0; i < m_round_item_dim; ++i) {
if (1 <= i and i <= item_dim()) {
set.init(Lx(i));
set.validate();
set.init(Rx(i));
set.validate();
} else {
round_set.init(m_Lx_lines + m_round_word_dim * i);
round_set.validate();
POMAGMA_ASSERT(round_set.empty(),
"unsupported Lx(" << i << ") has " <<
round_set.count_items() << " items");
round_set.init(m_Rx_lines + m_round_word_dim * i);
round_set.validate();
POMAGMA_ASSERT(round_set.empty(),
"unsupported Rx(" << i << ") has " <<
round_set.count_items() << " items");
}
}
// check for Lx/Rx agreement
for (oid_t i = 1; i <= item_dim(); ++i) {
for (oid_t j = 1; j <= item_dim(); ++j) {
POMAGMA_ASSERT(Lx(i, j) == Rx(i, j),
"Lx, Rx disagree at " << i << ',' << j);
}}
}
}
void TExpVal::SaveTxt(const PSOut& SOut) const {
TOLx Lx(SOut, TFSet()|oloCmtAlw|oloSigNum|oloCsSens);
SaveTxt(Lx);
}
void TCpDoc::SaveAsfaToCpd(const TStr& InFPath, const TStr& OutCpdFNm){
// create output file
PSOut SOut=TFOut::New(OutCpdFNm);
// traverse files
TStrH AccessionIdH;
TFFile FFile(TStr::GetNrFPath(InFPath)+"*.Asfa"); TStr AsfaFNm;
while (FFile.Next(AsfaFNm)){
printf("Processing file '%s'\n", AsfaFNm.CStr());
PSIn SIn=TFIn::New(AsfaFNm);
TILx Lx(SIn, TFSet(iloRetEoln, iloExcept));
Lx.GetSym(syLn, syEof);
while (Lx.Sym!=syEof){
// Query Line
TStr QueryLnStr=Lx.Str;
TStrV QueryStrV; QueryLnStr.SplitOnAllCh('\t', QueryStrV, false);
IAssert(QueryStrV[0]=="Query");
// RecordNo Line
Lx.GetSym(syLn); TStr RecNoLnStr=Lx.Str;
TStrV RecNoStrV; RecNoLnStr.SplitOnAllCh('\t', RecNoStrV, false);
IAssert(RecNoStrV[0]=="RecordNo");
//int RecN=RecNoStrV[1].GetInt();
// fields (format: Short-Name Tab Long-Name Tab Value-String)
TStr TitleStr, AbstractStr, PublicationYearStr, AccessionId;
TStrV AuthorNmV; TStrV TermNmV1, TermNmV2;
while (true){
Lx.GetSym(syLn); TStr FldLnStr=Lx.Str;
TStrV FldStrV; FldLnStr.SplitOnAllCh('\t', FldStrV, false);
if (FldStrV[0]=="----"){
if (!AccessionIdH.IsKey(AccessionId)){
AccessionIdH.AddKey(AccessionId);
// create & save cpd document
PCpDoc CpDoc=TCpDoc::New();
CpDoc->DocNm=AccessionId;
CpDoc->DateStr=PublicationYearStr;
CpDoc->TitleStr=TitleStr;
CpDoc->ParStrV.Add(AbstractStr);
CpDoc->TopCdNmV=TermNmV1;
CpDoc->GeoCdNmV=TermNmV2;
CpDoc->IndCdNmV=AuthorNmV;
CpDoc->Save(*SOut);
} else {/*printf("[%s]", AccessionId.CStr());*/}
break;
} else
if (FldStrV[0]=="TI"){
TitleStr=FldStrV[2];
} else if (FldStrV[0]=="TI"){
TitleStr=FldStrV[2];
} else if (FldStrV[0]=="AU"){
FldStrV[2].SplitOnAllCh(';', AuthorNmV);
for (int StrN=0; StrN<AuthorNmV.Len(); StrN++){AuthorNmV[StrN].ToTrunc();}
} else if (FldStrV[0]=="AB"){
AbstractStr=FldStrV[2];
} else if (FldStrV[0]=="PY"){
PublicationYearStr=FldStrV[2];
} else if (FldStrV[0]=="DE"){
FldStrV[2].SplitOnAllCh(';', TermNmV1);
for (int StrN=0; StrN<TermNmV1.Len(); StrN++){TermNmV1[StrN].ToTrunc();}
} else if (FldStrV[0]=="CL"){
FldStrV[2].SplitOnAllCh(';', TermNmV2);
for (int StrN=0; StrN<TermNmV2.Len(); StrN++){TermNmV2[StrN].ToTrunc();}
} else if (FldStrV[0]=="AN"){
AccessionId=FldStrV[2];
}
}
printf("%d\r", AccessionIdH.Len());
Lx.GetSym(syLn, syEof);
}
}
}
