TEST(TQQueueTest, Shuffle) {
try {
TQQueue<TInt> Q(64, -1);
ASSERT_TRUE(Q.Empty());
for (int i = 0; i < 200; i++) {
Q.Push(i);
ASSERT_EQ(Q.Len(), i+1);
ASSERT_EQ(Q.Front(), 0);
ASSERT_EQ(Q.Back(), i);
ASSERT_EQ(Q[i], i);
}
TRnd Rnd;
Q.Shuffle(Rnd);
TIntV QVec; Q.GetSubValVec(0, 199, QVec);
QVec.Sort(true);
ASSERT_EQ(QVec.Len(), 200);
for (int i = 0; i < 200; i++) {
ASSERT_EQ(QVec[i], i);
}
} catch (PExcept& Except) {
printf("Error: %s", Except->GetStr());
throw Except;
}
}
void TBlobBs::GenBlockLenV(TIntV& BlockLenV){
BlockLenV.Clr();
for (int P2Exp=0; P2Exp<TB4Def::MxP2Exp; P2Exp++){
BlockLenV.Add(TInt(TB4Def::GetP2(P2Exp)));}
EAssert(int(BlockLenV.Last())<2000000000);
{for (int Len=10; Len<100; Len+=10){BlockLenV.Add(Len);}}
{for (int Len=100; Len<10000; Len+=100){BlockLenV.Add(Len);}}
{for (int Len=10000; Len<100000; Len+=1000){BlockLenV.Add(Len);}}
{for (int Len=100000; Len<1000000; Len+=25000){BlockLenV.Add(Len);}}
{for (int Len=1000000; Len<10000000; Len+=1000000){BlockLenV.Add(Len);}}
{for (int Len=10000000; Len<100000000; Len+=10000000){BlockLenV.Add(Len);}}
BlockLenV.Sort();
}
int main(int argc, char *argv[]) {
TStr BaseString = "/lfs/1/tmp/curis/week/QBDB.bin";
TFIn BaseFile(BaseString);
TQuoteBase *QB = new TQuoteBase;
TDocBase *DB = new TDocBase;
QB->Load(BaseFile);
DB->Load(BaseFile);
TIntV QuoteIds;
QB->GetAllQuoteIds(QuoteIds);
int NumQuotes = QuoteIds.Len();
THash<TInt, TStrSet> PeakCounts;
for (int i = 0; i < NumQuotes; i++) {
TQuote CurQuote;
if (QB->GetQuote(QuoteIds[i], CurQuote)) {
TVec<TSecTm> Peaks;
CurQuote.GetPeaks(DB, Peaks);
TStr QuoteString;
CurQuote.GetParsedContentString(QuoteString);
TStrSet StringSet;
if (PeakCounts.IsKey(Peaks.Len())) {
StringSet = PeakCounts.GetDat(Peaks.Len());
}
StringSet.AddKey(QuoteString);
PeakCounts.AddDat(Peaks.Len(), StringSet);
}
}
TIntV PeakCountKeys;
PeakCounts.GetKeyV(PeakCountKeys);
PeakCountKeys.Sort(true);
for (int i = 0; i < PeakCountKeys.Len(); i++) {
TStrSet CurSet = PeakCounts.GetDat(PeakCountKeys[i]);
if (CurSet.Len() > 0) {
printf("QUOTES WITH %d PEAKS\n", PeakCountKeys[i].Val);
printf("#########################################\n");
THashSet<TStr> StringSet = PeakCounts.GetDat(PeakCountKeys[i]);
for (THashSet<TStr>::TIter l = StringSet.BegI(); l < StringSet.EndI(); l++) {
printf("%s\n", l.GetKey().CStr());
}
printf("\n");
}
}
delete QB;
delete DB;
return 0;
}
// vector benchmark with integer values
void VecBench(const int& n) {
TIntV Vec;
float ft0, ft1;
int x;
int i;
int Found;
int NotFound;
int Id;
// build the vector
ft0 = GetCPUTime();
for (i = 0; i < n; i++) {
x = (int) (drand48() * 100000000);
Vec.Add(x);
}
printf("vec : size %d\n", Vec.Len());
ft1 = GetCPUTime();
printf("vec : %7.3fs inserting %d numbers\n",ft1-ft0,i);
// sort the vector
ft0 = GetCPUTime();
Vec.Sort();
printf("vec : size %d\n", Vec.Len());
ft1 = GetCPUTime();
printf("vec : %7.3fs sorting %d numbers\n",ft1-ft0,i);
// search the vector
ft0 = GetCPUTime();
Found = 0;
NotFound = 0;
for (i = 0; i < n; i++) {
x = (int) (drand48() * 100000000);
Id = Vec.IsInBin(x);
if (Id == 0) {
NotFound++;
} else {
Found++;
}
}
printf("vec : found %d, notfound %d\n", Found, NotFound);
ft1 = GetCPUTime();
printf("vec : %7.3fs searching %d numbers\n",ft1-ft0,i);
}
void TGStatVec::SaveTxt(const TStr& FNmPref, const TStr& Desc) const {
FILE *F = fopen(TStr::Fmt("growth.%s.tab", FNmPref.CStr()).CStr(), "wt");
fprintf(F, "# %s\n", Desc.CStr());
fprintf(F, "# %s", TTmInfo::GetTmUnitStr(TmUnit).CStr());
TIntSet StatValSet;
for (int i = 0; i < Len(); i++) {
for (int v = gsvNone; v < gsvMx; v++) {
if (At(i)->HasVal(TGStatVal(v))) { StatValSet.AddKey(v); }
}
}
TIntV StatValV; StatValSet.GetKeyV(StatValV); StatValV.Sort();
for (int sv = 0; sv < StatValV.Len(); sv++) {
fprintf(F, "\t%s", TGStat::GetValStr(TGStatVal(StatValV[sv].Val)).CStr()); }
fprintf(F, "Time\n");
for (int i = 0; i < Len(); i++) {
const TGStat& G = *At(i);
for (int sv = 0; sv < StatValV.Len(); sv++) {
fprintf(F, "%g\t", G.GetVal(TGStatVal(StatValV[sv].Val))); }
fprintf(F, "%s\n", G.GetTmStr().CStr());
}
fclose(F);
}
void TTable::Unique(TStr Col){
if(!ColTypeMap.IsKey(Col)){TExcept::Throw("no such column " + Col);}
TIntV RemainingRows = TIntV(NumValidRows,0);
// group by given column (keys) and keep only first row for each key
switch(GetColType(Col)){
case INT:{
THash<TInt,TIntV> T; // can't really estimate the size of T for constructor hinting
GroupByIntCol(Col, T, TIntV(0), true);
for(THash<TInt,TIntV>::TIter it = T.BegI(); it < T.EndI(); it++){
RemainingRows.Add(it->Dat[0]);
}
break;
}
case FLT:{
THash<TFlt,TIntV> T;
GroupByFltCol(Col, T, TIntV(0), true);
for(THash<TFlt,TIntV>::TIter it = T.BegI(); it < T.EndI(); it++){
RemainingRows.Add(it->Dat[0]);
}
break;
}
case STR:{
THash<TStr,TIntV> T;
GroupByStrCol(Col, T, TIntV(0), true);
for(THash<TStr,TIntV>::TIter it = T.BegI(); it < T.EndI(); it++){
RemainingRows.Add(it->Dat[0]);
}
break;
}
}
// with the current implementation of GroupByX, RemainingRows is sorted:
// GroupByX returns a hash Table T:X-->TIntV. In the current implementation,
// if key X1 appears before key X2 in T Then T(X1)[0] <= T(X2)[0]
// Not sure if we could always make this assumption. Might want to remove this sorting..
RemainingRows.Sort();
KeepSortedRows(RemainingRows);
}
void TGreedyAlg::runGreedyAlgorithm() {
outputGraph = TKColourNet::New();
for (THash<TInt, TNodeInfo>::TIter NI = nodeNmH.BegI(); NI < nodeNmH.EndI(); NI++) {
outputGraph->AddNode(NI.GetKey(), TKColourNode());
printf("Added node %d to output graph\n", (int) NI.GetKey());
}
// for each node i
for (THash<TInt, TNodeInfo>::TIter NI = nodeNmH.BegI(); NI < nodeNmH.EndI(); NI++) {
int nodeI = NI.GetKey();
printf("***** Considering node i: %d *****\n", nodeI);
// initialise unaccounted cascades U
TVec<TCascade> cascades = cascadeV;
// initialise parental neighbourhood
TIntV parentalNeighbourhood;
bool uselessUnaccountedCasacdesLeft = false;
while ((cascades.Len() != 0) && (!uselessUnaccountedCasacdesLeft)) {
// printf("%d Casacdes left...\n", cascades.Len());
// find node j that could possible have infected node i that has been observed for largest number of cascades
int argmax = -1;
int maxNoCascades = 0;
bool areUnaccountedCascadesUseless = true;
for (THash<TInt, TNodeInfo>::TIter LNI = nodeNmH.BegI(); LNI < nodeNmH.EndI(); LNI++) {
int nodeJ = LNI.GetKey();
// printf("nodeJ: %d\n", nodeJ);
int countPotentialNoCascades = 0;
if (nodeI != nodeJ) {
for (int c = 0; c < cascades.Len(); c++) {
TCascade cascade = cascades[c];
if (cascade.IsNode(nodeJ) && cascade.IsNode(nodeI)) {
// if (cascade.GetTm(nodeI) > cascade.GetTm(nodeJ)) {
if (cascade.GetTm(nodeJ) == (cascade.GetTm(nodeI) - 1)) {
countPotentialNoCascades++;
}
}
}
if (countPotentialNoCascades > maxNoCascades) {
maxNoCascades = countPotentialNoCascades;
argmax = nodeJ;
areUnaccountedCascadesUseless = false;
}
else {
areUnaccountedCascadesUseless = areUnaccountedCascadesUseless && true;
}
}
}
if (areUnaccountedCascadesUseless) {
uselessUnaccountedCasacdesLeft = true;
}
if (argmax != -1) {
// printf("argmax (k) = %d, noCasacdesAppearedIn = %d\n", argmax, maxNoCascades);
// add arg max (k) to the set of parental neighbours
parentalNeighbourhood.Add(argmax);
}
// remove the cascades which k belongs to
TIntV cascadesToRemove;
for (int c = 0; c < cascades.Len(); c++) {
TCascade cascade = cascades[c];
if (cascade.IsNode(argmax) && cascade.IsNode(nodeI)) {
if (cascade.GetTm(nodeI) > cascade.GetTm(argmax)) {
cascadesToRemove.Add(c);
}
}
}
cascadesToRemove.Sort();
// printf("cascadesToRemove: ");
for (int i = 0; i < cascadesToRemove.Len(); i++) {
// printf("%d, ", (int) cascadesToRemove[i]);
cascades.Del(cascadesToRemove[i]-i);
}
// printf("\n");
}
// add edges from node i to each of parent
for (int i = 0; i < parentalNeighbourhood.Len(); i++) {
int srcNodeId = parentalNeighbourhood[i];
int dstNodeId = nodeI;
outputGraph->AddEdge(srcNodeId, dstNodeId);
// printf("##### Added Edge: %d -> %d #####\n", srcNodeId, dstNodeId);
}
}
}