C++ THash类代码示例

int HashFile(char **Return, char *Type, char *Path, int Encoding)
{
THash *Hash;
STREAM *S;
char *Tempstr=NULL;
int result;

S=STREAMOpenFile(Path,O_RDONLY);
if (! S) return(FALSE);

Hash=HashInit(Type);
if (! Hash) 
{
	STREAMClose(S);
	return(FALSE);
}


Tempstr=SetStrLen(Tempstr,4096);
result=STREAMReadBytes(S,Tempstr,4096);
while (result !=EOF)
{
	Hash->Update(Hash, Tempstr, result);
	result=STREAMReadBytes(S,Tempstr,4096);
}

DestroyString(Tempstr);
STREAMClose(S);

result=Hash->Finish(Hash, Encoding, Return);
HashDestroy(Hash);

return(result);
}

void TAGM::GetNodeMembership(THash<TInt,TIntV >& NIDComVH, const TVec<TIntV>& CmtyVV) {
	THash<TInt,TIntV> CmtyVH;
	for(int i=0;i<CmtyVV.Len();i++) {
		CmtyVH.AddDat(i,CmtyVV[i]);
	}
	GetNodeMembership(NIDComVH,CmtyVH);
}

/// Shingles by words
void LSH::HashShinglesOfClusters(TQuoteBase *QuoteBase,
    TClusterBase *ClusterBase, TIntV& ClusterIds, TInt ShingleLen,
    THash<TMd5Sig, TIntV>& ShingleToClusterIds) {
  Err("Hashing shingles of clusters...\n");
  for (int i = 0; i < ClusterIds.Len(); i++) {
    if (i % 1000 == 0) {
      fprintf(stderr, "%d out of %d completed\n", i, ClusterIds.Len());
    }
    TCluster C;
    ClusterBase->GetCluster(ClusterIds[i], C);
    //fprintf(stderr, "%d vs. %d\n", ClusterIds[i].Val, C.GetId().Val);

    // Put x-word shingles into hash table; x is specified by ShingleLen parameter
    THashSet < TMd5Sig > CHashedShingles;
    GetHashedShinglesOfCluster(QuoteBase, C, ShingleLen, CHashedShingles);
    for (THashSet<TMd5Sig>::TIter Hash = CHashedShingles.BegI();
        Hash < CHashedShingles.EndI(); Hash++) {
      TIntV ShingleClusterIds;
      if (ShingleToClusterIds.IsKey(*Hash)) {
        ShingleClusterIds = ShingleToClusterIds.GetDat(*Hash);
      }
      ShingleClusterIds.Add(ClusterIds[i]);
      ShingleToClusterIds.AddDat(*Hash, ShingleClusterIds);
    }
  }
  Err("Done hashing!\n");
}

// IN-OUT edges are swapped (so that the prog runs faster)
// Send message via IN edge proportional to the OUT edge weight
void TWgtNet::ReinforceEdges(const int& NIters) {
  THash<TInt, TFlt> OutWgtSumH;
  for (TNodeI NI = BegNI(); NI < EndNI(); NI++) {
    double wgt = 0;
    for (int e = 0; e < NI.GetOutDeg(); e++) { 
      wgt += NI.GetOutEDat(e); }
    OutWgtSumH.AddDat(NI.GetId(), wgt);
  }
  printf("Reinforcing edges for %d iterations\n", NIters);
  // iterate
  TExeTm ExeTm;
  for (int iter = 0; iter < NIters; iter++) {
    for (TNodeI NI = BegNI(); NI < EndNI(); NI++) {
      const double X = TInt::Rnd.GetUniDev() * OutWgtSumH.GetDat(NI.GetId());
      double x = 0;  int e = 0;
      for ( ; x + NI.GetOutEDat(e) < X; e++) {
        x += NI.GetOutEDat(e); }
      IAssert(IsEdge(NI.GetOutNId(e), NI.GetId()));
      GetEDat(NI.GetOutNId(e), NI.GetId()) += 1; // reinforce the edge
      OutWgtSumH.GetDat(NI.GetOutNId(e)) += 1; 
    }
    if (iter % (NIters/100) == 0) {
      printf("\r%d [%s]", iter, ExeTm.GetStr()); 
    }
  }
  printf(" done.\n");
}

THash<TInt, TInt> * choose_seeds (const PUNGraph g, const int num, const int * infection_state, const int infect) {

  THash<TInt, TInt> choices; 
  THash<TInt, TUNGraph::TNode> nodes;
  THash<TInt, TInt> * output = new THash<TInt, TInt> ();
  TInt weight = 0;
  TInt num_total = 0;
  for (TUNGraph::TNodeI n = g->BegNI(); n != g->EndNI(); n++) {
    //cout << "nodeID: " << n.GetId() << ",\tStatus: " << infection_state[n.GetId () - 1] << endl;
    if (infection_state[n.GetId () - 1] != infect) {
      weight += n.GetDeg ();
      choices.AddDat (num_total, weight);
      nodes.AddDat (num_total, n.GetId());
      num_total++;
    }
  }
  //  TRnd random ((int) time(NULL));
  // TRnd random (0);
  TInt num_chosen = 0;
  while (num_chosen < num) {
    TInt choice = my_random.GetUniDevInt (weight);
    TUNGraph::TNode node_choice = nodes[find (choice, choices, 0,  num_total-1)];
    if (!output->IsKey(node_choice.GetId())) {
      num_chosen++;
      // cout << node_choice.GetId () << "\n";
      output->AddDat(node_choice.GetId (), 1);
    }
  }
  return output;
}

void InitListViewContorol(HWND hWnd)
{
	HWND hWndLV = GetDlgItem(hWnd, IDC_LIST_MAILBOX);
	LV_COLUMN	lc;
	LV_ITEM		li;
	HIMAGELIST	hImgList;
	RECT		rect;

	GetClientRect(hWndLV, &rect);
	
	// イメージリストの作成
	hImgList = ImageList_Create(16, 16, ILC_COLOR | ILC_MASK, 1 , 1);
	ImageList_AddIcon(hImgList, LoadIcon(g_hInstance, "IDI_MAILBOX"));

	ListView_SetImageList(hWndLV, hImgList, LVSIL_SMALL);
	ListView_SetExtendedListViewStyle(hWndLV, LVS_EX_FULLROWSELECT | LVS_EX_CHECKBOXES);

	// ヘッダを追加 (表示しない)
	lc.mask = LVCF_FMT | LVCF_SUBITEM | LVCF_TEXT | LVCF_WIDTH;
	lc.fmt = LVCFMT_LEFT;

	lc.pszText = "メールボックス";
	lc.cx = rect.right;
	lc.iSubItem = 0;
	lc.cchTextMax = strlen(lc.pszText);
	ListView_InsertColumn(hWndLV, 0, &lc);

	lc.pszText = "フォルダ名";
	lc.cx = 0;
	lc.iSubItem = 1;
	lc.cchTextMax = strlen(lc.pszText);
	ListView_InsertColumn(hWndLV, 1, &lc);
	
	// メールボックスを列挙
	THash MailBoxes;
	EnumMailBox(MailBoxes);

	// メールボックスを追加
	int i = 0;
	for (THash::iterator itr = MailBoxes.begin(); itr != MailBoxes.end(); ++itr) {
		li.mask = LVIF_TEXT | LVIF_IMAGE;
		li.iItem = i;
		li.iImage = 0;
		// 表示名
		li.pszText = (char *)((*itr).second).c_str();
		li.iSubItem = 0;
		ListView_InsertItem(hWndLV, &li);
		// メールボックスのディレクトリ名
		li.pszText = (char *)((*itr).first).c_str();
		li.iSubItem = 1;
		ListView_SetItem(hWndLV, &li);
		
		// 設定されているメールボックスにチェックをつける
		if (g_temporary_mailbox.count((*itr).first) == 1) {
			ListView_SetCheckState(hWndLV, i, 1);
		}

		i++;
	}
}

int ComputeKCore(const PUNGraph& G) {
  int cnt = 0;
  for(TUNGraph::TNodeI NI = G->BegNI(); NI < G->EndNI(); NI++)
    cnt = max(cnt, NI.GetOutDeg());
  THashSet <TInt> D[cnt+1];
  THash <TInt, TInt> deg;
  for(TUNGraph::TNodeI NI = G->BegNI(); NI < G->EndNI(); NI++) { 
    TInt tmp = NI.GetOutDeg() - G->IsEdge(NI.GetId(), NI.GetId() );
    D[tmp.Val].AddKey(NI.GetId());
    deg.AddDat(NI.GetId()) = tmp;
  }
  int max_k = 0;
  for(int num_iters = 0;num_iters < G->GetNodes(); num_iters++)
    for(int i = 0; i < cnt; i++)
      if(D[i].Empty() == 0) {
        max_k = max(max_k, i);
        TInt a = *(D[i].BegI());
        D[i].DelKey(a);
        deg.AddDat(a.Val) = -1; // Hope overwriting works
        TUNGraph::TNodeI NI = G->GetNI(a.Val);
        for(int e = 0; e < NI.GetOutDeg(); e++) {
          TInt b = NI.GetOutNId(e);
          if(deg.GetDat(b) >= 0) {
            int Id = deg.GetKeyId(b); 
            D[deg[Id].Val].DelKey(b);
            deg[Id] = deg[Id] - 1;  //Hope the overwriting works
            D[deg[Id]].AddKey(b);
          }
        }
        break;
      }
  return max_k;
}

void LSH::ElCheapoHashing(TQuoteBase *QuoteBase, TInt ShingleLen,
    THash<TMd5Sig, TIntSet>& ShingleToQuoteIds) {
  fprintf(stderr, "Hashing shingles the el cheapo way...\n");
  TIntV QuoteIds;
  QuoteBase->GetAllQuoteIds(QuoteIds);
  for (int qt = 0; qt < QuoteIds.Len(); qt++) {
    if (qt % 1000 == 0) {
      fprintf(stderr, "%d out of %d completed\n", qt, QuoteIds.Len());
    }
    TQuote Q;
    QuoteBase->GetQuote(QuoteIds[qt], Q);

    // Put x-character (or x-word) shingles into hash table; x is specified by ShingleLen parameter
    TStr QContentStr;
    Q.GetParsedContentString(QContentStr);
    TChA QContentChA = TChA(QContentStr);

    for (int i = 0; i < QContentChA.Len() - ShingleLen + 1; i++) {
      TChA ShingleChA = TChA();
      for (int j = 0; j < ShingleLen; j++) {
        ShingleChA.AddCh(QContentChA.GetCh(i + j));
      }
      TStr Shingle = TStr(ShingleChA);
      const TMd5Sig ShingleMd5(Shingle);
      TIntSet ShingleQuoteIds;
      if (ShingleToQuoteIds.IsKey(ShingleMd5)) {
        ShingleQuoteIds = ShingleToQuoteIds.GetDat(ShingleMd5);
      }

      ShingleQuoteIds.AddKey(QuoteIds[qt]);
      ShingleToQuoteIds.AddDat(ShingleMd5, ShingleQuoteIds);
    }
  }
  Err("Done with el cheapo hashing!\n");
}

void HMACUpdate(THash *HMAC, char *Data, int Len)
{
THash *Hash;

Hash=(THash *) HMAC->Ctx;
Hash->Update(Hash,Data,Len);
}

void QuoteGraph::CompareUsingShingles(THash<TMd5Sig, TIntSet>& Shingles) {
  int Count = 0;
  int RealCount = 0;
  TVec<TMd5Sig> ShingleKeys;
  Shingles.GetKeyV(ShingleKeys);
  THashSet<TIntPr> EdgeCache;

  for (int i = 0; i < ShingleKeys.Len(); i++) {
    if (i % 100 == 0) {
      Err("Processed %d out of %d shingles, count = %d\n", i, ShingleKeys.Len(), Count);
    }
    TIntSet Bucket;
    Shingles.IsKeyGetDat(ShingleKeys[i], Bucket);

    for (TIntSet::TIter Quote1 = Bucket.BegI(); Quote1 < Bucket.EndI(); Quote1++) {
      TIntSet::TIter Quote1Copy = Quote1;
      Quote1Copy++;
      for (TIntSet::TIter Quote2 = Quote1Copy; Quote2 < Bucket.EndI(); Quote2++) {
        if (!EdgeCache.IsKey(TIntPr(Quote1.GetKey(), Quote2.GetKey())) && !EdgeCache.IsKey(TIntPr(Quote2.GetKey(), Quote1.GetKey()))) {
          EdgeCache.AddKey(TIntPr(Quote1.GetKey(), Quote2.GetKey()));
          EdgeCache.AddKey(TIntPr(Quote2.GetKey(), Quote1.GetKey()));
          RealCount++;
          AddEdgeIfSimilar(Quote1.GetKey(), Quote2.GetKey());
        }
      }
    }
    int Len = Bucket.Len() * (Bucket.Len() - 1) / 2;
    Count += Len;
  }
  fprintf(stderr, "NUMBER OF COMPARES: %d\n", Count);
  fprintf(stderr, "NUMBER OF REAL COMPARES: %d\n", RealCount);
}

void LogOutput::PrintClusterInformation(TDocBase *DB, TQuoteBase *QB, TClusterBase *CB, PNGraph& QGraph, TIntV& ClusterIds, TSecTm PresentTime, TIntV &OldTopClusters) {
  if (!ShouldLog) return;
  TStr CurDateString = PresentTime.GetDtYmdStr();
  Err("Writing cluster information...\n");

  // PREVIOUS RANKING SETUP
  THash<TInt, TInt> OldRankings;
  if (OldTopClusters.Len() > 0) {
    for (int i = 0; i < OldTopClusters.Len(); i++) {
      OldRankings.AddDat(OldTopClusters[i], i + 1);
    }
  }

  TStrV RankStr;
  TStr ClusterJSONDirectory = Directory + "/web/json/clusters/";
  for (int i = 0; i < ClusterIds.Len(); i++) {
    TStr OldRankStr;
    ComputeOldRankString(OldRankings, ClusterIds[i], i+1, OldRankStr);
    RankStr.Add(OldRankStr);

    // JSON file for each cluster!
    TPrintJson::PrintClusterJSON(QB, DB, CB, QGraph, ClusterJSONDirectory, ClusterIds[i], PresentTime);
  }

  Err("JSON Files for individual written!\n");
  TStr JSONTableFileName = Directory + "/web/json/daily/" + CurDateString + ".json";
  TPrintJson::PrintClusterTableJSON(QB, DB, CB, JSONTableFileName, ClusterIds, RankStr);
  Err("JSON Files for the cluster table written!\n");
}

/// rewire bipartite community affiliation graphs
void TAGMUtil::RewireCmtyVV(const TVec<TIntV>& CmtyVVIn, TVec<TIntV>& CmtyVVOut, TRnd& Rnd) {
    THash<TInt,TIntV> CmtyVH;
    for (int i = 0; i < CmtyVVIn.Len(); i++) {
        CmtyVH.AddDat(i, CmtyVVIn[i]);
    }
    TAGMUtil::RewireCmtyNID(CmtyVH, Rnd);
    CmtyVH.GetDatV(CmtyVVOut);
}

int HashBytes(char **Return, char *Type, char *text, int len, int Encoding)
{
THash *Hash;

Hash=HashInit(Type);
Hash->Update(Hash, text, len);
return(Hash->Finish(Hash, Encoding, Return));
}

void TTop2FriendNet::GetAvgSDevV(const THash<TFlt, TMom>& MomH, TFltTrV& ValAvgSDevV) {
  ValAvgSDevV.Clr(false);
  for (int i = 0; i < MomH.Len(); i++) {
    TMom Mom=MomH[i];
    Mom.Def();
    ValAvgSDevV.Add(TFltTr(MomH.GetKey(i), Mom.GetMean(), Mom.GetSDev()));
  }
  ValAvgSDevV.Sort();
}

void TAGMUtil::GetNodeMembership(THash<TInt,TIntSet >& NIDComVH, const THash<TInt,TIntV>& CmtyVH) {
    for (THash<TInt,TIntV>::TIter HI = CmtyVH.BegI(); HI < CmtyVH.EndI(); HI++) {
        int CID = HI.GetKey();
        for (int j = 0; j < HI.GetDat().Len(); j++) {
            int NID = HI.GetDat()[j];
            NIDComVH.AddDat(NID).AddKey(CID);
        }
    }
}