TEST(TStr, Search) {
TStr Str = "abcdaaba";
int Len = Str.Len();
EXPECT_EQ(Str.CountCh('a'), 4);
EXPECT_EQ(Str.CountCh('b'), 2);
EXPECT_EQ(Str.CountCh('e'), 0);
EXPECT_TRUE(Str.IsChIn('a'));
EXPECT_TRUE(Str.IsChIn('b'));
EXPECT_FALSE(Str.IsChIn('e'));
EXPECT_TRUE(Str.IsStrIn(Str));
EXPECT_TRUE(Str.IsStrIn(""));
EXPECT_TRUE(Str.IsStrIn("bcd"));
EXPECT_TRUE(Str.IsStrIn("ab"));
EXPECT_FALSE(Str.IsStrIn("eba"));
EXPECT_EQ(Str.CountCh('a', 1), 3);
EXPECT_ANY_THROW(Str.CountCh('a', 10));
EXPECT_EQ(Str.CountCh('b', 2), 1);
EXPECT_EQ(Str.CountCh('e', 1), 0);
EXPECT_EQ(Str.SearchCh('a'), 0);
EXPECT_EQ(Str.SearchCh('b'), 1);
EXPECT_EQ(Str.SearchCh('e'), -1);
EXPECT_EQ(Str.SearchCh('a', 1), 4);
EXPECT_EQ(Str.SearchCh('b', 2), 6);
EXPECT_EQ(Str.SearchCh('e', 1), -1);
EXPECT_EQ(Str.SearchChBack('a'), Len - 1);
EXPECT_EQ(Str.SearchChBack('b'), Len - 2);
EXPECT_EQ(Str.SearchChBack('e'), -1);
EXPECT_EQ(Str.SearchChBack('a', Len - 2), Len - 3);
EXPECT_EQ(Str.SearchChBack('b', Len - 3), 1);;
EXPECT_EQ(Str.SearchChBack('e', 3), -1);
EXPECT_EQ(Str.SearchStr("a"), 0);
EXPECT_EQ(Str.SearchStr("b"), 1);
EXPECT_EQ(Str.SearchStr("e"), -1);
EXPECT_EQ(Str.SearchStr(""), 0);
EXPECT_EQ(Str.SearchStr("a", 1), 4);
EXPECT_EQ(Str.SearchStr("b", 2), 6);
EXPECT_EQ(Str.SearchStr("e", 1), -1);
}
int main(int argc, char* argv[]) {
Env = TEnv(argc, argv, TNotify::StdNotify);
Env.PrepArgs(TStr::Fmt("ragm. build: %s, %s. Time: %s", __TIME__, __DATE__, TExeTm::GetCurTm()));
TExeTm ExeTm;
Try
TStr OutFPrx = Env.GetIfArgPrefixStr("-o:", "", "Output Graph data prefix");
const TStr InFNm = Env.GetIfArgPrefixStr("-i:", "../as20graph.txt", "Input edgelist file name");
const TStr LabelFNm = Env.GetIfArgPrefixStr("-l:", "", "Input file name for node names (Node ID, Node label) ");
int OptComs = Env.GetIfArgPrefixInt("-c:", -1, "The number of communities to detect (-1: detect automatically)");
const int MinComs = Env.GetIfArgPrefixInt("-mc:", 5, "Minimum number of communities to try");
const int MaxComs = Env.GetIfArgPrefixInt("-xc:", 100, "Maximum number of communities to try");
const int DivComs = Env.GetIfArgPrefixInt("-nc:", 10, "How many trials for the number of communities");
const int NumThreads = Env.GetIfArgPrefixInt("-nt:", 1, "Number of threads for parallelization");
const double StepAlpha = Env.GetIfArgPrefixFlt("-sa:", 0.3, "Alpha for backtracking line search");
const double StepBeta = Env.GetIfArgPrefixFlt("-sb:", 0.3, "Beta for backtracking line search");
PUNGraph G;
TIntStrH NIDNameH;
if (InFNm.IsStrIn(".ungraph")) {
TFIn GFIn(InFNm);
G = TUNGraph::Load(GFIn);
} else {
G = TAGMUtil::LoadEdgeListStr<PUNGraph>(InFNm, NIDNameH);
}
if (LabelFNm.Len() > 0) {
TSsParser Ss(LabelFNm, ssfTabSep);
while (Ss.Next()) {
if (Ss.Len() > 0) { NIDNameH.AddDat(Ss.GetInt(0), Ss.GetFld(1)); }
}
}
else {
}
printf("Graph: %d Nodes %d Edges\n", G->GetNodes(), G->GetEdges());
TVec<TIntV> EstCmtyVV;
TExeTm RunTm;
TAGMFast RAGM(G, 10, 10);
if (OptComs == -1) {
printf("finding number of communities\n");
OptComs = RAGM.FindComsByCV(NumThreads, MaxComs, MinComs, DivComs, OutFPrx, StepAlpha, StepBeta);
}
RAGM.NeighborComInit(OptComs);
if (NumThreads == 1 || G->GetEdges() < 1000) {
RAGM.MLEGradAscent(0.0001, 1000 * G->GetNodes(), "", StepAlpha, StepBeta);
} else {
RAGM.MLEGradAscentParallel(0.0001, 1000, NumThreads, "", StepAlpha, StepBeta);
}
RAGM.GetCmtyVV(EstCmtyVV);
TAGMUtil::DumpCmtyVV(OutFPrx + "cmtyvv.txt", EstCmtyVV, NIDNameH);
TAGMUtil::SaveGephi(OutFPrx + "graph.gexf", G, EstCmtyVV, 1.5, 1.5, NIDNameH);
Catch
printf("\nrun time: %s (%s)\n", ExeTm.GetTmStr(), TSecTm::GetCurTm().GetTmStr().CStr());
return 0;
}
int Intersect(TUNGraph::TNodeI Node, TStr NNodes){
int br = 0;
TInt digi = -1;
TStr buf = "";
for (int i = 0; i<Node.GetDeg(); i++)
{
digi = Node.GetNbrNId(i);
TStr buf = digi.GetStr();
if (NNodes.IsStrIn(buf.CStr()))
br++;
}
digi = Node.GetId();
buf = digi.GetStr();
if (NNodes.IsStrIn(buf.CStr()))
br++;
return br;
}
void TVizMapContext::PaintMgGlass(PGks Gks, const int& KeyWdFontSize) {
// drawing the dark circle
TFltRect ZoomRect = GetZoomRect();
int SizeX = TFlt::Round((MgGlassSize/ZoomRect.GetXLen()) * Gks->GetWidth());
int SizeY = TFlt::Round((MgGlassSize/ZoomRect.GetYLen()) * Gks->GetHeight());
Gks->SetBrush(TGksBrush::New(ColorMgGlass));
Gks->FillEllipse(ScreenX-SizeX, ScreenY-SizeY, ScreenX+SizeX, ScreenY+SizeY);
// drawing the keywords
if (MgGlassKeyWdV.Len() > 0) {
// prepare the string
Gks->SetFont(TGksFont::New("ARIAL", KeyWdFontSize, TGksColor::GetBlack(), TFSet()|gfsBold));
TStr KeyWdStr = Gks->BreakTxt(MgGlassKeyWdV, ", ", ",", MgGlassWindowWidth);
TStr NearPointStr;
if (NearPointN != -1) {
PVizMapPoint NearPoint = VizMapFrame->GetPoint(NearPointN);
if (NearPoint->IsPointNm()) {
TStr NearPointNm = NearPoint->GetPointNm();
if (NearPointNm.IsStrIn("[[")) {
const int StartPos = NearPointNm.SearchStr("[[");
NearPointNm = NearPointNm.Left(StartPos - 1);
}
NearPointStr = Gks->BreakTxt(NearPointNm,
" ", "", MgGlassWindowWidth, 1);
NearPointStr.DelChAll('\n');
NearPointStr += "\n";
}
}
TStr DocCountStr = "#documents = " + MgGlassPoints.GetStr() + "\n";
// compose the final message
KeyWdStr = NearPointStr + DocCountStr + KeyWdStr;
// find position of the window
int WndWidth = Gks->GetTxtWidth(KeyWdStr) + 6;
int WndHeight = Gks->GetTxtHeight(KeyWdStr) + 6;
int PosX = ScreenX + 20, PosY = ScreenY + 20;
if (PosX + WndWidth > Gks->GetWidth()) {
PosX = ScreenX - 20 - WndWidth; }
if (PosY + WndHeight > Gks->GetHeight()) {
PosY = ScreenY - 20 - WndHeight; }
// draw the keyword string
Gks->SetBrush(TGksBrush::New(ColorMgGlassWndShadow));
Gks->FillRect(PosX + 5, PosY + 5,
PosX + WndWidth + 5, PosY + WndHeight + 5);
Gks->SetBrush(TGksBrush::New(ColorMgGlassWnd));
Gks->SetPen(TGksPen::New(ColorMgGlassWndFrm));
Gks->Rectangle(PosX, PosY,
PosX + WndWidth, PosY + WndHeight);
Gks->PutTxt(KeyWdStr, PosX+3, PosY+3);
}
}
TFltRect TVizMapContext::PaintPointNm(PGks Gks, PVizMapPoint Point, const int& X,
const int& Y, const int& PointFontSize, const int& PointNmFontScale,
const bool& SelPointP, const bool& IsCatP) {
// get and clean point name
TStr PointNm = Point->GetPointNm();
PointNm.ChangeChAll('_', ' ');
if (PointNm.IsStrIn("[[")) {
const int StartPos = PointNm.SearchStr("[[");
PointNm = PointNm.Left(StartPos - 1);
}
// set font
TGksColor FontColor = SelPointP ? ColorSelPointFont : ColorPointFont;
const int FontSize = PointFontSize + TFlt::Round(Point->GetWgt()*PointNmFontScale);
//TFSet FontStyle = IsCatP ? (TFSet() | gfsBold) : TFSet();
//Gks->SetFont(TGksFont::New("ARIAL", FontSize, FontColor, FontStyle));
Gks->SetFont(TGksFont::New("ARIAL", FontSize, FontColor));
// refit it for the screen
TStr ScreenPointNm = Gks->BreakTxt(PointNm, " ", "", PointNmWidth, PointNmMxLines);
// calculate string position on the screen
const int HalfTxtWidth = Gks->GetTxtWidth(ScreenPointNm) / 2;
const int HalfTxtHeight = Gks->GetTxtHeight(ScreenPointNm) / 2;
// draw it!
const int MnX = X - HalfTxtWidth;
int CurrY = Y - HalfTxtHeight;
TStrV LineV; ScreenPointNm.SplitOnAllCh('\n', LineV);
for (int LineN = 0; LineN < LineV.Len(); LineN++) {
const int HalfLineWidth = Gks->GetTxtWidth(LineV[LineN]) / 2;
const int LineHeight = Gks->GetTxtHeight(LineV[LineN]);
Gks->PutTxt(LineV[LineN], MnX + (HalfTxtWidth - HalfLineWidth), CurrY);
CurrY += LineHeight-3;
}
// finish
return TFltRect(X - HalfTxtWidth, Y - HalfTxtHeight,
X + HalfTxtWidth, Y + HalfTxtHeight - LineV.Len()*3);
}
int main(int argc, char* argv[]) {
Env = TEnv(argc, argv, TNotify::StdNotify);
Env.PrepArgs(TStr::Fmt("cesna. build: %s, %s. Time: %s", __TIME__, __DATE__, TExeTm::GetCurTm()));
TExeTm ExeTm;
Try
TStr OutFPrx = Env.GetIfArgPrefixStr("-o:", "", "Output Graph data prefix");
const TStr InFNm = Env.GetIfArgPrefixStr("-i:", "./1912.edges", "Input edgelist file name");
const TStr LabelFNm = Env.GetIfArgPrefixStr("-l:", "", "Input file name for node names (Node ID, Node label) ");
const TStr AttrFNm = Env.GetIfArgPrefixStr("-a:", "./1912.nodefeat", "Input node attribute file name");
const TStr ANameFNm = Env.GetIfArgPrefixStr("-n:", "./1912.nodefeatnames", "Input file name for node attribute names");
int OptComs = Env.GetIfArgPrefixInt("-c:", 10, "The number of communities to detect (-1: detect automatically)");
const int MinComs = Env.GetIfArgPrefixInt("-mc:", 3, "Minimum number of communities to try");
const int MaxComs = Env.GetIfArgPrefixInt("-xc:", 20, "Maximum number of communities to try");
const int DivComs = Env.GetIfArgPrefixInt("-nc:", 5, "How many trials for the number of communities");
const int NumThreads = Env.GetIfArgPrefixInt("-nt:", 4, "Number of threads for parallelization");
const double AttrWeight = Env.GetIfArgPrefixFlt("-aw:", 0.5, "We maximize (1 - aw) P(Network) + aw * P(Attributes)");
const double LassoWeight = Env.GetIfArgPrefixFlt("-lw:", 1.0, "Weight for l-1 regularization on learning the logistic model parameters");
const double StepAlpha = Env.GetIfArgPrefixFlt("-sa:", 0.05, "Alpha for backtracking line search");
const double StepBeta = Env.GetIfArgPrefixFlt("-sb:", 0.3, "Beta for backtracking line search");
const double MinFeatFrac = Env.GetIfArgPrefixFlt("-mf:", 0.0, "If the fraction of nodes with positive values for an attribute is smaller than this, we ignore that attribute");
#ifdef USE_OPENMP
omp_set_num_threads(NumThreads);
#endif
PUNGraph G;
TIntStrH NIDNameH;
TStrHash<TInt> NodeNameH;
TVec<TFltV> Wck;
TVec<TIntV> EstCmtyVV;
if (InFNm.IsStrIn(".ungraph")) {
TFIn GFIn(InFNm);
G = TUNGraph::Load(GFIn);
} else {
G = TAGMUtil::LoadEdgeListStr<PUNGraph>(InFNm, NodeNameH);
NIDNameH.Gen(NodeNameH.Len());
for (int s = 0; s < NodeNameH.Len(); s++) { NIDNameH.AddDat(s, NodeNameH.GetKey(s)); }
}
if (LabelFNm.Len() > 0) {
TSsParser Ss(LabelFNm, ssfTabSep);
while (Ss.Next()) {
if (Ss.Len() > 0) { NIDNameH.AddDat(Ss.GetInt(0), Ss.GetFld(1)); }
}
}
printf("Graph: %d Nodes %d Edges\n", G->GetNodes(), G->GetEdges());
//load attribute
TIntV NIDV;
G->GetNIdV(NIDV);
THash<TInt, TIntV> RawNIDAttrH, NIDAttrH;
TIntStrH RawFeatNameH, FeatNameH;
if (ANameFNm.Len() > 0) {
TSsParser Ss(ANameFNm, ssfTabSep);
while (Ss.Next()) {
if (Ss.Len() > 0) { RawFeatNameH.AddDat(Ss.GetInt(0), Ss.GetFld(1)); }
}
}
TCesnaUtil::LoadNIDAttrHFromNIDKH(NIDV, AttrFNm, RawNIDAttrH, NodeNameH);
TCesnaUtil::FilterLowEntropy(RawNIDAttrH, NIDAttrH, RawFeatNameH, FeatNameH, MinFeatFrac);
TExeTm RunTm;
TCesna CS(G, NIDAttrH, 10, 10);
if (OptComs == -1) {
printf("finding number of communities\n");
OptComs = CS.FindComs(NumThreads, MaxComs, MinComs, DivComs, "", false, 0.1, StepAlpha, StepBeta);
}
CS.NeighborComInit(OptComs);
CS.SetWeightAttr(AttrWeight);
CS.SetLassoCoef(LassoWeight);
if (NumThreads == 1 || G->GetEdges() < 1000) {
CS.MLEGradAscent(0.0001, 1000 * G->GetNodes(), "", StepAlpha, StepBeta);
} else {
CS.MLEGradAscentParallel(0.0001, 1000, NumThreads, "", StepAlpha, StepBeta);
}
CS.GetCmtyVV(EstCmtyVV, Wck);
TAGMUtil::DumpCmtyVV(OutFPrx + "cmtyvv.txt", EstCmtyVV, NIDNameH);
FILE* F = fopen((OutFPrx + "weights.txt").CStr(), "wt");
if (FeatNameH.Len() == Wck[0].Len()) {
fprintf(F, "#");
for (int k = 0; k < FeatNameH.Len(); k++) {
fprintf(F, "%s", FeatNameH[k].CStr());
if (k < FeatNameH.Len() - 1) { fprintf(F, "\t"); }
}
fprintf(F, "\n");
}
for (int c = 0; c < Wck.Len(); c++) {
for (int k = 0; k < Wck[c].Len(); k++) {
fprintf(F, "%f", Wck[c][k].Val);
if (k < Wck[c].Len() - 1) { fprintf(F, "\t"); }
}
fprintf(F, "\n");
}
fclose(F);
Catch
printf("\nrun time: %s (%s)\n", ExeTm.GetTmStr(), TSecTm::GetCurTm().GetTmStr().CStr());
return 0;
}