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;
}
// Test node subgraph conversion
void TestConvertSubGraphs() {
PNGraph NGraph;
PUNGraph UNGraph;
int N1, N2, N3;
int E1, E2, E3;
TIntV NIdV;
int i;
NGraph = GetTestTNGraph();
N1 = NGraph->GetNodes();
E1 = NGraph->GetEdges();
for (i = 0; i < 20; i += 2) {
NIdV.Add(i);
}
// TODO: fix TSnap::ConvertSubGraph<PUNGraph>(NGraph, NIdV, true), it fails
// UNGraph = TSnap::ConvertSubGraph<PUNGraph>(NGraph, NIdV, true);
UNGraph = TSnap::ConvertSubGraph<PUNGraph>(NGraph, NIdV);
N2 = UNGraph->GetNodes();
E2 = UNGraph->GetEdges();
NGraph = TSnap::ConvertSubGraph<PNGraph>(UNGraph, NIdV);
N3 = NGraph->GetNodes();
E3 = NGraph->GetEdges();
printf("---- TestConvertSubGraphs -----\n");
printf("nodes: %d,%d,%d, edges: %d,%d,%d\n", N1, N2, N3, E1, E2, E3);
printf("\n");
}
///Generate graph using the AGM model. CProbV = vector of Pc
PUNGraph TAGM::GenAGM(TVec<TIntV>& CmtyVV, const TFltV& CProbV, TRnd& Rnd, const double PNoCom) {
PUNGraph G = TUNGraph::New(100 * CmtyVV.Len(), -1);
printf("AGM begins\n");
for (int i = 0; i < CmtyVV.Len(); i++) {
TIntV& CmtyV = CmtyVV[i];
for (int u = 0; u < CmtyV.Len(); u++) {
if ( G->IsNode(CmtyV[u])) {
continue;
}
G->AddNode(CmtyV[u]);
}
double Prob = CProbV[i];
RndConnectInsideCommunity(G, CmtyV, Prob, Rnd);
}
if (PNoCom > 0.0) { //if we want to connect nodes that do not share any community
TIntSet NIDS;
for (int c = 0; c < CmtyVV.Len(); c++) {
for (int u = 0; u < CmtyVV[c].Len(); u++) {
NIDS.AddKey(CmtyVV[c][u]);
}
}
TIntV NIDV;
NIDS.GetKeyV(NIDV);
RndConnectInsideCommunity(G,NIDV,PNoCom,Rnd);
}
printf("AGM completed (%d nodes %d edges)\n",G->GetNodes(),G->GetEdges());
G->Defrag();
return G;
}
/// Clique Percolation method communities
void TCliqueOverlap::GetCPMCommunities(const PUNGraph& G, int MinMaxCliqueSize, TVec<TIntV>& NIdCmtyVV) {
printf("Clique Percolation Method\n");
TExeTm ExeTm;
TVec<TIntV> MaxCliques;
TCliqueOverlap::GetMaxCliques(G, MinMaxCliqueSize, MaxCliques);
// op RS 2012/05/15, commented out next line, a parameter is missing,
// creating a warning on OS X
// printf("...%d cliques found\n");
// get clique overlap matrix (graph)
PUNGraph OverlapGraph = TCliqueOverlap::CalculateOverlapMtx(MaxCliques, MinMaxCliqueSize-1);
printf("...overlap matrix (%d, %d)\n", G->GetNodes(), G->GetEdges());
// connected components are communities
TCnComV CnComV;
TSnap::GetWccs(OverlapGraph, CnComV);
NIdCmtyVV.Clr(false);
TIntSet CmtySet;
for (int c = 0; c < CnComV.Len(); c++) {
CmtySet.Clr(false);
for (int i = 0; i <CnComV[c].Len(); i++) {
const TIntV& CliqueNIdV = MaxCliques[CnComV[c][i]];
CmtySet.AddKeyV(CliqueNIdV);
}
NIdCmtyVV.Add();
CmtySet.GetKeyV(NIdCmtyVV.Last());
NIdCmtyVV.Last().Sort();
}
printf("done [%s].\n", ExeTm.GetStr());
}
double TAGMUtil::GetConductance(const PUNGraph& Graph, const TIntSet& CmtyS, const int Edges) {
const int Edges2 = Edges >= 0 ? 2*Edges : Graph->GetEdges();
int Vol = 0, Cut = 0;
double Phi = 0.0;
for (int i = 0; i < CmtyS.Len(); i++) {
if (! Graph->IsNode(CmtyS[i])) {
continue;
}
TUNGraph::TNodeI NI = Graph->GetNI(CmtyS[i]);
for (int e = 0; e < NI.GetOutDeg(); e++) {
if (! CmtyS.IsKey(NI.GetOutNId(e))) {
Cut += 1;
}
}
Vol += NI.GetOutDeg();
}
// get conductance
if (Vol != Edges2) {
if (2 * Vol > Edges2) {
Phi = Cut / double (Edges2 - Vol);
}
else if (Vol == 0) {
Phi = 0.0;
}
else {
Phi = Cut / double(Vol);
}
} else {
if (Vol == Edges2) {
Phi = 1.0;
}
}
return Phi;
}
/////////////////////////////////////////////////
// Spectral graph properties
void PlotEigValRank(const PUNGraph& Graph, const int& EigVals, const TStr& FNmPref, TStr DescStr) {
TFltV EigValV;
TSnap::GetEigVals(Graph, EigVals, EigValV);
EigValV.Sort(false);
if (DescStr.Empty()) { DescStr = FNmPref; }
TGnuPlot::PlotValV(EigValV, "eigVal."+FNmPref, TStr::Fmt("%s. G(%d, %d). Largest eig val = %f",
DescStr.CStr(), Graph->GetNodes(), Graph->GetEdges(), EigValV[0].Val), "Rank", "Eigen value", gpsLog10XY, false, gpwLinesPoints);
}
// Inverse participation ratio: normalize EigVec to have L2=1 and then I=sum_k EigVec[i]^4
// see Spectra of "real-world" graphs: Beyond the semicircle law by Farkas, Derenyi, Barabasi and Vicsek
void PlotInvParticipRat(const PUNGraph& Graph, const int& MaxEigVecs, const int& TimeLimit, const TStr& FNmPref, TStr DescStr) {
TFltPrV EigIprV;
GetInvParticipRat(Graph, MaxEigVecs, TimeLimit, EigIprV);
if (DescStr.Empty()) { DescStr = FNmPref; }
if (EigIprV.Empty()) { DescStr+=". FAIL"; EigIprV.Add(TFltPr(-1,-1)); return; }
TGnuPlot::PlotValV(EigIprV, "eigIPR."+FNmPref, TStr::Fmt("%s. G(%d, %d). Largest eig val = %f (%d values)",
DescStr.CStr(), Graph->GetNodes(), Graph->GetEdges(), EigIprV.Last().Val1(), EigIprV.Len()),
"Eigenvalue", "Inverse Participation Ratio of corresponding Eigenvector", gpsLog10Y, false, gpwPoints);
}
// Test the default constructor
TEST(TUNGraph, DefaultConstructor) {
PUNGraph Graph;
Graph = TUNGraph::New();
EXPECT_EQ(0,Graph->GetNodes());
EXPECT_EQ(0,Graph->GetEdges());
EXPECT_EQ(1,Graph->IsOk());
EXPECT_EQ(1,Graph->Empty());
EXPECT_EQ(0,Graph->HasFlag(gfDirected));
}
// Test small graph
TEST(TUNGraph, GetSmallGraph) {
PUNGraph Graph;
Graph = TUNGraph::GetSmallGraph();
EXPECT_EQ(5,Graph->GetNodes());
EXPECT_EQ(5,Graph->GetEdges());
EXPECT_EQ(1,Graph->IsOk());
EXPECT_EQ(0,Graph->Empty());
EXPECT_EQ(0,Graph->HasFlag(gfDirected));
}
int main(int argc, char* argv[]) {
Env = TEnv(argc, argv, TNotify::StdNotify);
Env.PrepArgs(TStr::Fmt("Cascades. build: %s, %s. Time: %s", __TIME__, __DATE__, TExeTm::GetCurTm()));
TExeTm ExeTm;
Try
const TStr InFNm = Env.GetIfArgPrefixStr("-i:", "demo", "Input undirected graph");
const TStr OutFNm = Env.GetIfArgPrefixStr("-o:", "demo", "Output file name prefix");
const double Beta = Env.GetIfArgPrefixFlt("-b:", 0.1, "Beta (infection (i.e., cascade propagation) probability)");
// load
printf("Loading %s...", InFNm.CStr());
PUNGraph Graph;
if (InFNm == "demo") { Graph = TSnap::GenRndGnm<PUNGraph>(100, 200); }
else { Graph = TSnap::LoadEdgeList<PUNGraph>(InFNm); }
printf("nodes:%d edges:%d\n", Graph->GetNodes(), Graph->GetEdges());
// Simulate SI model
Graph = TSnap::GetMxWcc(Graph);
bool DivByM = true;
TCascadeStat CascStat;
printf("\nGraph:%s -- Beta: %g\n", OutFNm.CStr(), Beta);
for (int Run = 0; Run < 10; Run++) { // number of runs
TIntH NIdInfTmH;
// incluence cascade
PNGraph InfCasc = RunSICascade(Graph, Beta, 100, NIdInfTmH);
if (InfCasc->GetNodes() < 10) { printf("."); continue; } // min cascade size
// network cascade
PNGraph NetCasc = AddSpuriousEdges(Graph, InfCasc, NIdInfTmH);
// sample the cascade
CascStat.SampleCascade(InfCasc, NetCasc, NIdInfTmH, 0.1, 10, DivByM); // div-by-M
printf(".");
}
CascStat.PlotAll(TStr::Fmt("%s-B%03d", OutFNm.CStr(), int(100*Beta)), TStr::Fmt("%s N=%d E=%d Beta=%g",
OutFNm.CStr(), Graph->GetNodes(), Graph->GetEdges(), Beta), DivByM);
Catch
printf("\nrun time: %s (%s)\n", ExeTm.GetTmStr(), TSecTm::GetCurTm().GetTmStr().CStr());
return 0;
}
int main(int argc, char* argv[]) {
Env = TEnv(argc, argv, TNotify::StdNotify);
Env.PrepArgs(TStr::Fmt("Network community detection. build: %s, %s. Time: %s", __TIME__, __DATE__, TExeTm::GetCurTm()));
TExeTm ExeTm;
Try
const TStr InFNm = Env.GetIfArgPrefixStr("-i:", "graph.txt", "Input graph (undirected graph)");
const TStr OutFNm = Env.GetIfArgPrefixStr("-o:", "communities.txt", "Output file");
const int CmtyAlg = Env.GetIfArgPrefixInt("-a:", 2, "Algorithm: 1:Girvan-Newman, 2:Clauset-Newman-Moore, 3:Infomap");
PUNGraph Graph = TSnap::LoadEdgeList<PUNGraph>(InFNm, false);
//PUNGraph Graph = TSnap::LoadEdgeList<PUNGraph>("../as20graph.txt", false);
//PUNGraph Graph = TSnap::GenRndGnm<PUNGraph>(5000, 10000); // generate a random graph
TSnap::DelSelfEdges(Graph);
TCnComV CmtyV;
double Q = 0.0;
TStr CmtyAlgStr;
if (CmtyAlg == 1) {
CmtyAlgStr = "Girvan-Newman";
Q = TSnap::CommunityGirvanNewman(Graph, CmtyV); }
else if (CmtyAlg == 2) {
CmtyAlgStr = "Clauset-Newman-Moore";
Q = TSnap::CommunityCNM(Graph, CmtyV); }
else if (CmtyAlg == 3) {
CmtyAlgStr = "Infomap";
Q = TSnap::Infomap(Graph, CmtyV); }
else { Fail; }
FILE *F = fopen(OutFNm.CStr(), "wt");
fprintf(F, "# Input: %s\n", InFNm.CStr());
fprintf(F, "# Nodes: %d Edges: %d\n", Graph->GetNodes(), Graph->GetEdges());
fprintf(F, "# Algoritm: %s\n", CmtyAlgStr.CStr());
if (CmtyAlg!=3) {
fprintf(F, "# Modularity: %f\n", Q);
} else {
fprintf(F, "# Average code length: %f\n", Q);
}
fprintf(F, "# Communities: %d\n", CmtyV.Len());
fprintf(F, "# NId\tCommunityId\n");
for (int c = 0; c < CmtyV.Len(); c++) {
for (int i = 0; i < CmtyV[c].Len(); i++) {
fprintf(F, "%d\t%d\n", CmtyV[c][i].Val, c);
}
}
fclose(F);
Catch
printf("\nrun time: %s (%s)\n", ExeTm.GetTmStr(), TSecTm::GetCurTm().GetTmStr().CStr());
return 0;
}
// Test graph conversion
TEST(subgraph, TestConvertGraphs) {
PNGraph NGraph;
PUNGraph UNGraph;
NGraph = GetTestTNGraph();
EXPECT_EQ(20,NGraph->GetNodes());
EXPECT_EQ(60,NGraph->GetEdges());
UNGraph = TSnap::ConvertGraph<PUNGraph>(NGraph);
EXPECT_EQ(20,UNGraph->GetNodes());
EXPECT_EQ(60,UNGraph->GetEdges());
NGraph = TSnap::ConvertGraph<PNGraph>(UNGraph);
EXPECT_EQ(20,NGraph->GetNodes());
EXPECT_EQ(120,NGraph->GetEdges());
}
double GetAsstyCor(const PUNGraph& Graph) {
TIntFltH deg(Graph->GetNodes()), deg_sq(Graph->GetNodes());
for (TUNGraph::TNodeI NI = Graph->BegNI(); NI < Graph->EndNI(); NI++) {
deg.AddDat(NI.GetId()) = NI.GetOutDeg();
deg_sq.AddDat(NI.GetId()) = NI.GetOutDeg() * NI.GetOutDeg();
}
double m = Graph->GetEdges(), num1 = 0.0, num2 = 0.0, den1 = 0.0;
for (TUNGraph::TEdgeI EI = Graph->BegEI(); EI < Graph->EndEI(); EI++) {
double t1 = deg.GetDat(EI.GetSrcNId()).Val, t2 = deg.GetDat(EI.GetDstNId()).Val;
num1 += t1 * t2;
num2 += t1 + t2;
den1 += deg_sq.GetDat(EI.GetSrcNId()).Val + deg_sq.GetDat(EI.GetDstNId()).Val;
}
num1 /= m;
den1 /= (2.0 * m);
num2 = (num2 / (2.0 * m)) * (num2 / (2.0 * m));
return (num1 - num2) / (den1 - num2);
}
/// Rewire the network. Keeps node degrees as is but randomly rewires the edges.
/// Use this function to generate a random graph with the same degree sequence
/// as the OrigGraph.
/// See: On the uniform generation of random graphs with prescribed degree
/// sequences by R. Milo, N. Kashtan, S. Itzkovitz, M. E. J. Newman, U. Alon
/// URL: http://arxiv.org/abs/cond-mat/0312028
PUNGraph GenRewire(const PUNGraph& OrigGraph, const int& NSwitch, TRnd& Rnd) {
const int Nodes = OrigGraph->GetNodes();
const int Edges = OrigGraph->GetEdges();
PUNGraph GraphPt = TUNGraph::New();
TUNGraph& Graph = *GraphPt;
Graph.Reserve(Nodes, -1);
TExeTm ExeTm;
// generate a graph that satisfies the constraints
printf("Randomizing edges (%d, %d)...\n", Nodes, Edges);
TIntPrSet EdgeSet(Edges);
for (TUNGraph::TNodeI NI = OrigGraph->BegNI(); NI < OrigGraph->EndNI(); NI++) {
const int NId = NI.GetId();
for (int e = 0; e < NI.GetOutDeg(); e++) {
if (NId <= NI.GetOutNId(e)) { continue; }
EdgeSet.AddKey(TIntPr(NId, NI.GetOutNId(e)));
}
Graph.AddNode(NI.GetId());
}
// edge switching
uint skip=0;
for (uint swps = 0; swps < 2*uint(Edges)*uint(NSwitch); swps++) {
const int keyId1 = EdgeSet.GetRndKeyId(Rnd);
const int keyId2 = EdgeSet.GetRndKeyId(Rnd);
if (keyId1 == keyId2) { skip++; continue; }
const TIntPr& E1 = EdgeSet[keyId1];
const TIntPr& E2 = EdgeSet[keyId2];
TIntPr NewE1(E1.Val1, E2.Val1), NewE2(E1.Val2, E2.Val2);
if (NewE1.Val1 > NewE1.Val2) { Swap(NewE1.Val1, NewE1.Val2); }
if (NewE2.Val1 > NewE2.Val2) { Swap(NewE2.Val1, NewE2.Val2); }
if (NewE1!=NewE2 && NewE1.Val1!=NewE1.Val2 && NewE2.Val1!=NewE2.Val2 && ! EdgeSet.IsKey(NewE1) && ! EdgeSet.IsKey(NewE2)) {
EdgeSet.DelKeyId(keyId1); EdgeSet.DelKeyId(keyId2);
EdgeSet.AddKey(TIntPr(NewE1));
EdgeSet.AddKey(TIntPr(NewE2));
} else { skip++; }
if (swps % Edges == 0) {
printf("\r %uk/%uk: %uk skip [%s]", swps/1000u, 2*uint(Edges)*uint(NSwitch)/1000u, skip/1000u, ExeTm.GetStr());
if (ExeTm.GetSecs() > 2*3600) { printf(" *** Time limit!\n"); break; } // time limit 2 hours
}
}
printf("\r total %uk switchings attempted, %uk skiped [%s]\n", 2*uint(Edges)*uint(NSwitch)/1000u, skip/1000u, ExeTm.GetStr());
for (int e = 0; e < EdgeSet.Len(); e++) {
Graph.AddEdge(EdgeSet[e].Val1, EdgeSet[e].Val2); }
return GraphPt;
}
void Init(const PUNGraph& Graph) {
const double M = 0.5/Graph->GetEdges(); // 1/2m
Q = 0.0;
for (TUNGraph::TNodeI NI = Graph->BegNI(); NI < Graph->EndNI(); NI++) {
CmtyIdUF.Add(NI.GetId());
const int OutDeg = NI.GetOutDeg();
if (OutDeg == 0) { continue; }
TCmtyDat& Dat = CmtyQH.AddDat(NI.GetId(), TCmtyDat(M * OutDeg, OutDeg));
for (int e = 0; e < NI.GetOutDeg(); e++) {
const int DstNId = NI.GetOutNId(e);
const double DstMod = 2 * M * (1.0 - OutDeg * Graph->GetNI(DstNId).GetOutDeg() * M);
Dat.AddQ(DstNId, DstMod);
}
Q += -1.0*TMath::Sqr(OutDeg*M);
if (NI.GetId() < Dat.GetMxQNId()) {
MxQHeap.Add(TFltIntIntTr(Dat.GetMxQ(), NI.GetId(), Dat.GetMxQNId())); }
}
MxQHeap.MakeHeap();
}
void PlotEigValDistr(const PUNGraph& Graph, const int& EigVals, const TStr& FNmPref, TStr DescStr) {
const int NBuckets = 50;
TFltV EigValV;
for (int f = 1; EigValV.Empty() && f < 4; f++) {
TSnap::GetEigVals(Graph, f*EigVals, EigValV);
}
EigValV.Sort(true);
THash<TFlt, TFlt> BucketCntH;
double Step = (EigValV.Last()-EigValV[0]) / double(NBuckets-1);
for (int i = 0; i < NBuckets; i++) {
BucketCntH.AddDat(EigValV[0]+Step*(i+0.5), 0);
}
for (int i = 0; i < EigValV.Len(); i++) {
const int Bucket = (int) floor((EigValV[i]-EigValV[0]) / Step);
BucketCntH[Bucket] += 1;
}
TFltPrV EigCntV;
BucketCntH.GetKeyDatPrV(EigCntV);
if (DescStr.Empty()) { DescStr = FNmPref; }
TGnuPlot::PlotValV(EigCntV, "eigDistr."+FNmPref, TStr::Fmt("%s. G(%d, %d). Largest eig val = %f", DescStr.CStr(),
Graph->GetNodes(), Graph->GetEdges(), EigValV.Last().Val), "Eigen value", "Count", gpsAuto, false, gpwLinesPoints);
}
// Maximum modularity clustering by Girvan-Newman algorithm (slow)
// Girvan M. and Newman M. E. J., Community structure in social and biological networks, Proc. Natl. Acad. Sci. USA 99, 7821–7826 (2002)
double CommunityGirvanNewman(PUNGraph& Graph, TCnComV& CmtyV) {
TIntH OutDegH;
const int NEdges = Graph->GetEdges();
for (TUNGraph::TNodeI NI = Graph->BegNI(); NI < Graph->EndNI(); NI++) {
OutDegH.AddDat(NI.GetId(), NI.GetOutDeg());
}
double BestQ = -1; // modularity
TCnComV CurCmtyV;
CmtyV.Clr();
TIntV Cmty1, Cmty2;
while (true) {
CmtyGirvanNewmanStep(Graph, Cmty1, Cmty2);
const double Q = _GirvanNewmanGetModularity(Graph, OutDegH, NEdges, CurCmtyV);
//printf("current modularity: %f\n", Q);
if (Q > BestQ) {
BestQ = Q;
CmtyV.Swap(CurCmtyV);
}
if (Cmty1.Len()==0 || Cmty2.Len() == 0) { break; }
}
return BestQ;
}
PUNGraph TAGM::GenAGM(TVec<TIntV>& CmtyVV, const double& DensityCoef, const double& ScaleCoef, TRnd& Rnd){
TFltV CProbV;
double Prob;
for(int i=0;i<CmtyVV.Len();i++) {
Prob = ScaleCoef*pow(double(CmtyVV[i].Len()),-DensityCoef);
if(Prob>1){Prob = 1;}
CProbV.Add(Prob);
}
PUNGraph G = TUNGraph::New();
printf("AGM begins\n");
for(int i=0;i<CmtyVV.Len();i++) {
TIntV& CmtyV = CmtyVV[i];
for(int u=0;u<CmtyV.Len();u++) {
G->AddNode(CmtyV[u]);
}
Prob = CProbV[i];
printf("\r%d(%d)/%d",i,CmtyVV[i].Len(),CmtyVV.Len());
RndConnectInsideCommunity(G,CmtyV,Prob,Rnd);
}
printf("AGM completed (%d nodes %d edges)\n",G->GetNodes(),G->GetEdges());
return G;
}
// Test graph conversion
void TestConvertGraphs() {
PNGraph NGraph;
PUNGraph UNGraph;
int N1, N2, N3;
int E1, E2, E3;
NGraph = GetTestTNGraph();
N1 = NGraph->GetNodes();
E1 = NGraph->GetEdges();
UNGraph = TSnap::ConvertGraph<PUNGraph>(NGraph);
N2 = UNGraph->GetNodes();
E2 = UNGraph->GetEdges();
NGraph = TSnap::ConvertGraph<PNGraph>(UNGraph);
N3 = NGraph->GetNodes();
E3 = NGraph->GetEdges();
printf("---- TestConvertGraphs -----\n");
printf("nodes: %d,%d,%d, edges: %d,%d,%d\n", N1, N2, N3, E1, E2, E3);
printf("\n");
}
// Test subgraphs
TEST(subgraph, TestSubTUNGraphs) {
PUNGraph Graph;
PUNGraph Graph1;
PUNGraph Graph2;
PUNGraph Graph3;
PUNGraph Graph4;
int i;
TIntV NIdV;
TIntV NIdV1;
Graph = GetTestTUNGraph();
EXPECT_EQ(20,Graph->GetNodes());
EXPECT_EQ(60,Graph->GetEdges());
for (i = 10; i < 15; i++) {
NIdV.Add(i);
}
Graph1 = TSnap::GetSubGraph(Graph, NIdV);
EXPECT_EQ(5,Graph1->GetNodes());
EXPECT_EQ(9,Graph1->GetEdges());
Graph2 = TSnap::GetSubGraph(Graph, NIdV, true);
EXPECT_EQ(5,Graph2->GetNodes());
EXPECT_EQ(9,Graph2->GetEdges());
for (i = 0; i < 20; i += 2) {
NIdV1.Add(i);
}
Graph3 = TSnap::GetSubGraph(Graph, NIdV1, true);
EXPECT_EQ(10,Graph3->GetNodes());
EXPECT_EQ(10,Graph3->GetEdges());
Graph4 = TSnap::GetSubGraphRenumber(Graph, NIdV1);
EXPECT_EQ(10,Graph4->GetNodes());
EXPECT_EQ(10,Graph4->GetEdges());
}
// Test node subgraph conversion
TEST(subgraph, TestConvertSubGraphs) {
PNGraph NGraph;
PUNGraph UNGraph;
TIntV NIdV;
int i;
NGraph = GetTestTNGraph();
EXPECT_EQ(20,NGraph->GetNodes());
EXPECT_EQ(60,NGraph->GetEdges());
for (i = 0; i < 20; i += 2) {
NIdV.Add(i);
}
// TODO: fix TSnap::ConvertSubGraph<PUNGraph>(NGraph, NIdV, true), it fails
// UNGraph = TSnap::ConvertSubGraph<PUNGraph>(NGraph, NIdV, true);
UNGraph = TSnap::ConvertSubGraph<PUNGraph>(NGraph, NIdV);
EXPECT_EQ(10,UNGraph->GetNodes());
EXPECT_EQ(10,UNGraph->GetEdges());
NGraph = TSnap::ConvertSubGraph<PNGraph>(UNGraph, NIdV);
EXPECT_EQ(10,NGraph->GetNodes());
EXPECT_EQ(20,NGraph->GetEdges());
}
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;
}
// Print graph statistics
void PrintGStats(const char s[], PUNGraph Graph) {
printf("graph %s, nodes %d, edges %d, empty %s\n",
s, Graph->GetNodes(), Graph->GetEdges(),
Graph->Empty() ? "yes" : "no");
}
PUNGraph TAGM::GenAGM(TVec<TIntV>& CmtyVV, const double& DensityCoef, const int TargetEdges, TRnd& Rnd) {
PUNGraph TryG = TAGM::GenAGM(CmtyVV, DensityCoef, 1.0, Rnd);
const double ScaleCoef = (double) TargetEdges / (double) TryG->GetEdges();
return TAGM::GenAGM(CmtyVV, DensityCoef, ScaleCoef, Rnd);
}
/// estimate number of communities using AGM
int TAGMUtil::FindComsByAGM(const PUNGraph& Graph, const int InitComs, const int MaxIter, const int RndSeed, const double RegGap, const double PNoCom, const TStr PltFPrx) {
TRnd Rnd(RndSeed);
int LambdaIter = 100;
if (Graph->GetNodes() < 200) {
LambdaIter = 1;
}
if (Graph->GetNodes() < 200 && Graph->GetEdges() > 2000) {
LambdaIter = 100;
}
//Find coms with large C
TAGMFit AGMFitM(Graph, InitComs, RndSeed);
if (PNoCom > 0.0) {
AGMFitM.SetPNoCom(PNoCom);
}
AGMFitM.RunMCMC(MaxIter, LambdaIter, "");
int TE = Graph->GetEdges();
TFltV RegV;
RegV.Add(0.3 * TE);
for (int r = 0; r < 25; r++) {
RegV.Add(RegV.Last() * RegGap);
}
TFltPrV RegComsV, RegLV, RegBICV;
TFltV LV, BICV;
//record likelihood and number of communities with nonzero P_c
for (int r = 0; r < RegV.Len(); r++) {
double RegCoef = RegV[r];
AGMFitM.SetRegCoef(RegCoef);
AGMFitM.MLEGradAscentGivenCAG(0.01, 1000);
AGMFitM.SetRegCoef(0.0);
TVec<TIntV> EstCmtyVV;
AGMFitM.GetCmtyVV(EstCmtyVV, 0.99);
int NumLowQ = EstCmtyVV.Len();
RegComsV.Add(TFltPr(RegCoef, (double) NumLowQ));
if (EstCmtyVV.Len() > 0) {
TAGMFit AFTemp(Graph, EstCmtyVV, Rnd);
AFTemp.MLEGradAscentGivenCAG(0.001, 1000);
double CurL = AFTemp.Likelihood();
LV.Add(CurL);
BICV.Add(-2.0 * CurL + (double) EstCmtyVV.Len() * log((double) Graph->GetNodes() * (Graph->GetNodes() - 1) / 2.0));
}
else {
break;
}
}
// if likelihood does not exist or does not change at all, report the smallest number of communities or 2
if (LV.Len() == 0) {
return 2;
}
else if (LV[0] == LV.Last()) {
return (int) TMath::Mx<TFlt>(2.0, RegComsV[LV.Len() - 1].Val2);
}
//normalize likelihood and BIC to 0~100
int MaxL = 100;
{
TFltV& ValueV = LV;
TFltPrV& RegValueV = RegLV;
double MinValue = TFlt::Mx, MaxValue = TFlt::Mn;
for (int l = 0; l < ValueV.Len(); l++) {
if (ValueV[l] < MinValue) {
MinValue = ValueV[l];
}
if (ValueV[l] > MaxValue) {
MaxValue = ValueV[l];
}
}
while (ValueV.Len() < RegV.Len()) {
ValueV.Add(MinValue);
}
double RangeVal = MaxValue - MinValue;
for (int l = 0; l < ValueV.Len(); l++) {
RegValueV.Add(TFltPr(RegV[l], double(MaxL) * (ValueV[l] - MinValue) / RangeVal));
}
}
{
TFltV& ValueV = BICV;
TFltPrV& RegValueV = RegBICV;
double MinValue = TFlt::Mx, MaxValue = TFlt::Mn;
for (int l = 0; l < ValueV.Len(); l++) {
if (ValueV[l] < MinValue) {
MinValue = ValueV[l];
}
if (ValueV[l] > MaxValue) {
MaxValue = ValueV[l];
}
}
while (ValueV.Len() < RegV.Len()) {
ValueV.Add(MaxValue);
}
double RangeVal = MaxValue - MinValue;
for (int l = 0; l < ValueV.Len(); l++) {
RegValueV.Add(TFltPr(RegV[l], double(MaxL) * (ValueV[l] - MinValue) / RangeVal));
}
}
//fit logistic regression to normalized likelihood.
TVec<TFltV> XV(RegLV.Len());
TFltV YV (RegLV.Len());
for (int l = 0; l < RegLV.Len(); l++) {
XV[l] = TFltV::GetV(log(RegLV[l].Val1));
YV[l] = RegLV[l].Val2 / (double) MaxL;
}
TFltPrV LRVScaled, LRV;
TLogRegFit LRFit;
PLogRegPredict LRMd = LRFit.CalcLogRegNewton(XV, YV, PltFPrx);
for (int l = 0; l < RegLV.Len(); l++) {
LRV.Add(TFltPr(RegV[l], LRMd->GetCfy(XV[l])));
LRVScaled.Add(TFltPr(RegV[l], double(MaxL) * LRV.Last().Val2));
}
//estimate # communities from fitted logistic regression
int NumComs = 0, IdxRegDrop = 0;
double LRThres = 1.1, RegDrop; // 1 / (1 + exp(1.1)) = 0.25
double LeftReg = 0.0, RightReg = 0.0;
TFltV Theta;
LRMd->GetTheta(Theta);
RegDrop = (- Theta[1] - LRThres) / Theta[0];
if (RegDrop <= XV[0][0]) {
NumComs = (int) RegComsV[0].Val2;
}
else if (RegDrop >= XV.Last()[0]) {
NumComs = (int) RegComsV.Last().Val2;
}
else { //interpolate for RegDrop
for (int i = 0; i < XV.Len(); i++) {
if (XV[i][0] > RegDrop) {
IdxRegDrop = i;
break;
}
}
if (IdxRegDrop == 0) {
printf("Error!! RegDrop:%f, Theta[0]:%f, Theta[1]:%f\n", RegDrop, Theta[0].Val, Theta[1].Val);
for (int l = 0; l < RegLV.Len(); l++) {
printf("X[%d]:%f, Y[%d]:%f\n", l, XV[l][0].Val, l, YV[l].Val);
}
}
IAssert(IdxRegDrop > 0);
LeftReg = RegDrop - XV[IdxRegDrop - 1][0];
RightReg = XV[IdxRegDrop][0] - RegDrop;
NumComs = (int) TMath::Round( (RightReg * RegComsV[IdxRegDrop - 1].Val2 + LeftReg * RegComsV[IdxRegDrop].Val2) / (LeftReg + RightReg));
}
//printf("Interpolation coeff: %f, %f, index at drop:%d (%f), Left-Right Vals: %f, %f\n", LeftReg, RightReg, IdxRegDrop, RegDrop, RegComsV[IdxRegDrop - 1].Val2, RegComsV[IdxRegDrop].Val2);
printf("Num Coms:%d\n", NumComs);
if (NumComs < 2) {
NumComs = 2;
}
if (PltFPrx.Len() > 0) {
TStr PlotTitle = TStr::Fmt("N:%d, E:%d ", Graph->GetNodes(), TE);
TGnuPlot GPC(PltFPrx + ".l");
GPC.AddPlot(RegComsV, gpwLinesPoints, "C");
GPC.AddPlot(RegLV, gpwLinesPoints, "likelihood");
GPC.AddPlot(RegBICV, gpwLinesPoints, "BIC");
GPC.AddPlot(LRVScaled, gpwLinesPoints, "Sigmoid (scaled)");
GPC.SetScale(gpsLog10X);
GPC.SetTitle(PlotTitle);
GPC.SavePng(PltFPrx + ".l.png");
}
return NumComs;
}
void GetMotifCount(const PUNGraph& G, const int MotifSize, TVec <int64> & MotifV, const int num) {
if (MotifSize == 3) {
MotifV = TVec <int64> (2);
MotifV.PutAll(0);
TSnap::GetTriads(G,MotifV[mtThreeClosed],MotifV[mtThreeOpen],num);
}
else {
MotifV = TVec <int64> (6);
MotifV.PutAll(0);
TIntPrV V(G->GetEdges(), 0);
for (TUNGraph::TEdgeI EI = G->BegEI(); EI < G->EndEI(); EI++) {
V.Add(TIntPr(EI.GetSrcNId(), EI.GetDstNId()));
}
TRnd blargh;
V.Shuffle(blargh);
for (int z = 0; z < num; z++) {
int SrcNId = V[z].Val1.Val, DstNId = V[z].Val2.Val;
TUNGraph::TNodeI SrcNI = G->GetNI(SrcNId), DstNI = G->GetNI(DstNId);
TIntV SrcV(SrcNI.GetOutDeg(),0), DstV(DstNI.GetOutDeg(),0), BothV(min(SrcNI.GetOutDeg(), DstNI.GetOutDeg()),0);
SrcV.Clr(0,-1);
DstV.Clr(0,-1);
BothV.Clr(0,-1);
//Grouping the vertices into sets
for (int e = 0; e < SrcNI.GetOutDeg(); e++) {
if (SrcNI.GetOutNId(e) == DstNId) continue;
if (G->IsEdge(DstNId, SrcNI.GetOutNId(e)) ) { BothV.Add(SrcNI.GetOutNId(e)); }
else { SrcV.Add(SrcNI.GetOutNId(e)); }
}
for (int e = 0; e < DstNI.GetOutDeg(); e++) {
if (DstNI.GetOutNId(e) == SrcNId) continue;
if (G->IsEdge(SrcNId, DstNI.GetOutNId(e)) == 0) { DstV.Add(DstNI.GetOutNId(e)); }
}
//Compute Motif 0 and 1
for (int i = 0; i < SrcV.Len(); i++) {
for (int j = 0; j < DstV.Len(); j++) {
if (G->IsEdge(SrcV[i], DstV[j]) ) { MotifV[mfFourSquare]++; }
else MotifV[mfFourLine]++;
}
}
//Compute Motif 2 and 3
for (int i = 0; i < SrcV.Len(); i++) {
for (int j = i + 1; j < SrcV.Len(); j++) {
if (G->IsEdge(SrcV[i], SrcV[j]) ) { MotifV[mfFourTriangleEdge]++; }
else MotifV[mfFourStar]++;
}
}
for (int i = 0; i < DstV.Len(); i++) {
for (int j = i + 1; j < DstV.Len(); j++) {
if (G->IsEdge(DstV[i], DstV[j]) ) { MotifV[mfFourTriangleEdge]++; }
else MotifV[mfFourStar]++;
}
}
//Compute Motif 4 and 5
for (int i = 0; i < BothV.Len(); i++) {
for (int j = i + 1; j < BothV.Len(); j++) {
if (G->IsEdge(BothV[i], BothV[j]) ) { MotifV[mfFourComplete]++; }
else MotifV[mfFourSquareDiag]++;
}
}
}
MotifV[mfFourSquare] /= 4ll;
MotifV[mfFourStar] /= 3ll;
MotifV[mfFourComplete] /= 6ll;
}
}
int main(int argc, char* argv[]) {
Env = TEnv(argc, argv, TNotify::StdNotify);
Env.PrepArgs(TStr::Fmt("cpm. build: %s, %s. Time: %s", __TIME__, __DATE__, TExeTm::GetCurTm()));
TExeTm ExeTm;
Try
TStr OutFPrx = Env.GetIfArgPrefixStr("-o:", "", "Output file name prefix");
const TStr InFNm = Env.GetIfArgPrefixStr("-i:", "football.edgelist", "Input edgelist file name. DEMO: AGM with 2 communities");
const TStr LabelFNm = Env.GetIfArgPrefixStr("-l:", "football.labels", "Input file name for node names (Node ID, Node label) ");
const TInt RndSeed = Env.GetIfArgPrefixInt("-s:", 0, "Random seed for AGM");
const TFlt Epsilon = Env.GetIfArgPrefixFlt("-e:", 0, "Edge probability between the nodes that do not share any community (default (0.0): set it to be 1 / N^2)");
const TInt Coms = Env.GetIfArgPrefixInt("-c:", 0, "Number of communities (0: determine it by AGM)");
PUNGraph G = TUNGraph::New();
TVec<TIntV> CmtyVV;
TIntStrH NIDNameH;
if (InFNm == "DEMO") {
TVec<TIntV> TrueCmtyVV;
TRnd AGMRnd(RndSeed);
//generate community bipartite affiliation
const int ABegin = 0, AEnd = 70, BBegin = 30, BEnd = 100;
TrueCmtyVV.Add(TIntV());
TrueCmtyVV.Add(TIntV());
for (int u = ABegin; u < AEnd; u++) {
TrueCmtyVV[0].Add(u);
}
for (int u = BBegin; u < BEnd; u++) {
TrueCmtyVV[1].Add(u);
}
G = TAGM::GenAGM(TrueCmtyVV, 0.0, 0.2, AGMRnd);
}
else if (LabelFNm.Len() > 0) {
G = TSnap::LoadEdgeList<PUNGraph>(InFNm);
TSsParser Ss(LabelFNm, ssfTabSep);
while (Ss.Next()) {
if (Ss.Len() > 0) { NIDNameH.AddDat(Ss.GetInt(0), Ss.GetFld(1)); }
}
}
else {
G = TAGMUtil::LoadEdgeListStr<PUNGraph>(InFNm, NIDNameH);
}
printf("Graph: %d Nodes %d Edges\n", G->GetNodes(), G->GetEdges());
int MaxIter = 50 * G->GetNodes() * G->GetNodes();
if (MaxIter < 0) { MaxIter = TInt::Mx; }
int NumComs = Coms;
if (NumComs < 2) {
int InitComs;
if (G->GetNodes() > 1000) {
InitComs = G->GetNodes() / 5;
NumComs = TAGMUtil::FindComsByAGM(G, InitComs, MaxIter, RndSeed, 1.5, Epsilon, OutFPrx);
} else {
InitComs = G->GetNodes() / 5;
NumComs = TAGMUtil::FindComsByAGM(G, InitComs, MaxIter, RndSeed, 1.2, Epsilon, OutFPrx);
}
}
TAGMFit AGMFit(G, NumComs, RndSeed);
if (Epsilon > 0) { AGMFit.SetPNoCom(Epsilon); }
AGMFit.RunMCMC(MaxIter, 10);
AGMFit.GetCmtyVV(CmtyVV, 0.9999);
TAGMUtil::DumpCmtyVV(OutFPrx + "cmtyvv.txt", CmtyVV, NIDNameH);
TAGMUtil::SaveGephi(OutFPrx + "graph.gexf", G, CmtyVV, 1.5, 1.5, NIDNameH);
AGMFit.PrintSummary();
Catch
printf("\nrun time: %s (%s)\n", ExeTm.GetTmStr(), TSecTm::GetCurTm().GetTmStr().CStr());
return 0;
}
// Test node, edge creation
TEST(TUNGraph, ManipulateNodesEdges) {
int NNodes = 10000;
int NEdges = 100000;
const char *FName = "test.graph";
PUNGraph Graph;
PUNGraph Graph1;
PUNGraph Graph2;
int i;
int n;
int NCount;
int x,y;
int Deg, InDeg, OutDeg;
Graph = TUNGraph::New();
EXPECT_EQ(1,Graph->Empty());
// create the nodes
for (i = 0; i < NNodes; i++) {
Graph->AddNode(i);
}
EXPECT_EQ(0,Graph->Empty());
EXPECT_EQ(NNodes,Graph->GetNodes());
// create random edges
NCount = NEdges;
while (NCount > 0) {
x = (long) (drand48() * NNodes);
y = (long) (drand48() * NNodes);
// Graph->GetEdges() is not correct for the loops (x == y),
// skip the loops in this test
if (x != y && !Graph->IsEdge(x,y)) {
n = Graph->AddEdge(x, y);
NCount--;
}
}
EXPECT_EQ(NEdges,Graph->GetEdges());
EXPECT_EQ(0,Graph->Empty());
EXPECT_EQ(1,Graph->IsOk());
for (i = 0; i < NNodes; i++) {
EXPECT_EQ(1,Graph->IsNode(i));
}
EXPECT_EQ(0,Graph->IsNode(NNodes));
EXPECT_EQ(0,Graph->IsNode(NNodes+1));
EXPECT_EQ(0,Graph->IsNode(2*NNodes));
// nodes iterator
NCount = 0;
for (TUNGraph::TNodeI NI = Graph->BegNI(); NI < Graph->EndNI(); NI++) {
NCount++;
}
EXPECT_EQ(NNodes,NCount);
// edges per node iterator
NCount = 0;
for (TUNGraph::TNodeI NI = Graph->BegNI(); NI < Graph->EndNI(); NI++) {
for (int e = 0; e < NI.GetOutDeg(); e++) {
NCount++;
}
}
EXPECT_EQ(NEdges*2,NCount);
// edges iterator
NCount = 0;
for (TUNGraph::TEdgeI EI = Graph->BegEI(); EI < Graph->EndEI(); EI++) {
NCount++;
}
EXPECT_EQ(NEdges,NCount);
// node degree
for (TUNGraph::TNodeI NI = Graph->BegNI(); NI < Graph->EndNI(); NI++) {
Deg = NI.GetDeg();
InDeg = NI.GetInDeg();
OutDeg = NI.GetOutDeg();
EXPECT_EQ(Deg,InDeg);
EXPECT_EQ(Deg,OutDeg);
}
// assignment
Graph1 = TUNGraph::New();
*Graph1 = *Graph;
EXPECT_EQ(NNodes,Graph1->GetNodes());
EXPECT_EQ(NEdges,Graph1->GetEdges());
EXPECT_EQ(0,Graph1->Empty());
EXPECT_EQ(1,Graph1->IsOk());
// saving and loading
{
TFOut FOut(FName);
Graph->Save(FOut);
FOut.Flush();
}
{
TFIn FIn(FName);
Graph2 = TUNGraph::Load(FIn);
}
EXPECT_EQ(NNodes,Graph2->GetNodes());
EXPECT_EQ(NEdges,Graph2->GetEdges());
EXPECT_EQ(0,Graph2->Empty());
EXPECT_EQ(1,Graph2->IsOk());
// remove all the nodes and edges
for (i = 0; i < NNodes; i++) {
n = Graph->GetRndNId();
Graph->DelNode(n);
}
EXPECT_EQ(0,Graph->GetNodes());
EXPECT_EQ(0,Graph->GetEdges());
EXPECT_EQ(1,Graph->IsOk());
EXPECT_EQ(1,Graph->Empty());
Graph1->Clr();
EXPECT_EQ(0,Graph1->GetNodes());
EXPECT_EQ(0,Graph1->GetEdges());
EXPECT_EQ(1,Graph1->IsOk());
EXPECT_EQ(1,Graph1->Empty());
}
