#include <vector>
#include <boost/numeric/mtl/mtl.hpp>
#include "Assembler.h"
#include "SecondOrderAssembler.h"
#include "Operator.h"
#include "QPsiPhi.h"
#include "FiniteElemSpace.h"
#include "Quadrature.h"
#include "DOFVector.h"
#include "OpenMP.h"

namespace AMDiS {

  std::vector<SubAssembler*> SecondOrderAssembler::optimizedSubAssemblers; 
  std::vector<SubAssembler*> SecondOrderAssembler::standardSubAssemblers;

  SecondOrderAssembler::SecondOrderAssembler(Operator *op,
					     Assembler *assembler,
					     Quadrature *quad,
					     bool optimized)
    : SubAssembler(op, assembler, quad, 2, optimized)
  {}


  SecondOrderAssembler* SecondOrderAssembler::getSubAssembler(Operator* op,
							      Assembler *assembler,
							      Quadrature *quad,
							      bool optimized)
  {
    int myRank = omp_get_thread_num();

    // check if a assembler is needed at all
    if (op->secondOrder[myRank].size() == 0)
      return NULL;

    SecondOrderAssembler *newAssembler;

    std::vector<SubAssembler*> *subAssemblers =
      optimized ?
      &optimizedSubAssemblers :
      &standardSubAssemblers;

    std::vector<OperatorTerm*> opTerms = op->zeroOrder[myRank];

    sort(opTerms.begin(), opTerms.end());

    // check if a new assembler is needed
    for (int i = 0; i < static_cast<int>(subAssemblers->size()); i++) {
      std::vector<OperatorTerm*> assTerms = *((*subAssemblers)[i]->getTerms());
    
      sort(assTerms.begin(), assTerms.end());

      if (opTerms == assTerms && (*subAssemblers)[i]->getQuadrature() == quad)	
	return dynamic_cast<SecondOrderAssembler*>((*subAssemblers)[i]);
    }

    // check if all terms are pw_const
    bool pwConst = true;
    for (int i = 0; i < static_cast<int>(op->secondOrder[myRank].size()); i++) {
      if (!op->secondOrder[myRank][i]->isPWConst()) {
	pwConst = false;
	break;
      }
    }  

    // create new assembler
    if (!optimized) {
      newAssembler = new Stand2(op, assembler, quad);
    } else {
      if (pwConst) {
	newAssembler = new Pre2(op, assembler, quad);
      } else {
	newAssembler = new Quad2(op, assembler, quad);
      }
    }

    subAssemblers->push_back(newAssembler);
    return newAssembler;
  }


  Pre2::Pre2(Operator *op, Assembler *assembler, Quadrature *quad) 
    : SecondOrderAssembler(op, assembler, quad, true)
  {
    name = "precalculated second order assembler";

    q11 = Q11PsiPhi::provideQ11PsiPhi(rowFeSpace->getBasisFcts(), 
				      colFeSpace->getBasisFcts(), 
				      quadrature);
    tmpLALt.resize(omp_get_overall_max_threads());
    for (int i = 0; i < omp_get_overall_max_threads(); i++) {
      tmpLALt[i] = new DimMat<double>*;
      *(tmpLALt[i]) = new DimMat<double>(dim, NO_INIT);
    }
  }


  Pre2::~Pre2()
  {
    for (int i = 0; i < static_cast<int>(tmpLALt.size()); i++) {
      delete *(tmpLALt[i]);
      delete tmpLALt[i];
    }
  }


  void Pre2::calculateElementMatrix(const ElInfo *elInfo, ElementMatrix& mat)
  {
    const int **nEntries;
    const int *k, *l;
    const double *values;

    int myRank = omp_get_thread_num();
    DimMat<double> **LALt = tmpLALt[0];
    DimMat<double> &tmpMat = *LALt[0];
    tmpMat.set(0.0);

    for (int i = 0; i < static_cast<int>(terms[myRank].size()); i++) {
      (static_cast<SecondOrderTerm*>(terms[myRank][i]))->getLALt(elInfo, 1, LALt);
    }

    tmpMat *= elInfo->getDet();
    nEntries = q11->getNumberEntries();

    if (symmetric) {
      TEST_EXIT_DBG(nCol == nRow)("nCol != nRow, but symmetric assembling!\n");

      for (int i = 0; i < nRow; i++) {
	k = q11->getKVec(i, i);
	l = q11->getLVec(i, i);
	values = q11->getValVec(i, i);
	double val = 0.0;
	for (int m = 0; m < nEntries[i][i]; m++) {
	  val += values[m] * tmpMat[k[m]][l[m]];
	}
	mat[i][i] += val;

	for (int j = i + 1; j < nCol; j++) {
	  k = q11->getKVec(i, j);
	  l = q11->getLVec(i, j);
	  values = q11->getValVec(i, j);
	  val = 0.0;
	  for (int m = 0; m < nEntries[i][j]; m++) {
	    val += values[m] * tmpMat[k[m]][l[m]];
	  }
	  mat[i][j] += val;
	  mat[j][i] += val;
	}
      }
    } else {  /*  A not symmetric or psi != phi        */
      for (int i = 0; i < nRow; i++) {
	for (int j = 0; j < nCol; j++) {
	  k = q11->getKVec(i, j);
	  l = q11->getLVec(i, j);
	  values = q11->getValVec(i, j);
	  double val = 0.0;
	  for (int m = 0; m < nEntries[i][j]; m++) {
	    val += values[m] * tmpMat[k[m]][l[m]];
	  }
	  mat[i][j] += val;
	}
      }
    }
  }


  Quad2::Quad2(Operator *op, Assembler *assembler, Quadrature *quad) 
    : SecondOrderAssembler(op, assembler, quad, true)
  {
    name = "fast quadrature second order assembler";

    tmpVec.resize(omp_get_overall_max_threads());
    tmpLALt.resize(omp_get_overall_max_threads());
  }


  Quad2::~Quad2()
  {
    if (!firstCall) {
      int nPoints = quadrature->getNumPoints();
      for (int i = 0; i < static_cast<int>(tmpLALt.size()); i++) {
	for (int j = 0; j < nPoints; j++)
	  delete tmpLALt[i][j];
	
	delete [] tmpLALt[i];
      }
    }
  }


  void Quad2::calculateElementMatrix(const ElInfo *elInfo, ElementMatrix& mat)
  {
    int nPoints = quadrature->getNumPoints();
    int myRank = omp_get_thread_num();

    if (firstCall) {
      tmpVec[myRank] = new DimVec<double>(dim, NO_INIT);
      tmpLALt[myRank] = new DimMat<double>*[nPoints];
      for (int j = 0; j < nPoints; j++)
	tmpLALt[myRank][j] = new DimMat<double>(dim, NO_INIT);      

#ifdef _OPENMP
#pragma omp critical
#endif
      {   
        psiFast = 
	  updateFastQuadrature(psiFast, rowFeSpace->getBasisFcts(), INIT_GRD_PHI);
        phiFast = 
	  updateFastQuadrature(phiFast, rowFeSpace->getBasisFcts(), INIT_GRD_PHI);
      }
      firstCall = false;
    }

    DimMat<double> **LALt = tmpLALt[myRank];
    DimVec<double> &dimVec = *(tmpVec[myRank]);
    
    for (int i = 0; i < nPoints; i++)
      LALt[i]->set(0.0);

    for (int i = 0; i < static_cast<int>(terms[myRank].size()); i++)
      (static_cast<SecondOrderTerm*>(terms[myRank][i]))->getLALt(elInfo, nPoints, LALt);

    VectorOfFixVecs< DimVec<double> > *grdPsi, *grdPhi;

    if (symmetric) {
      // === Symmetric assembling. ===
      TEST_EXIT_DBG(nCol == nRow)("nCol != nRow, but symmetric assembling!\n");

      for (int iq = 0; iq < nPoints; iq++) {
	(*LALt[iq]) *= elInfo->getDet();
	grdPsi = psiFast->getGradient(iq);
	grdPhi = phiFast->getGradient(iq);

	for (int i = 0; i < nCol; i++) {	  
	  amv(quadrature->getWeight(iq), (*LALt[iq]), (*grdPhi)[i], dimVec);

	  mat[i][i] += (*grdPsi)[i] * dimVec;
	  for (int j = i + 1; j < nRow; j++) {
	    double tmp = (*grdPhi)[j] * dimVec;
	    mat[i][j] += tmp;
	    mat[j][i] += tmp;
	  }
	}
      }
    } else {      
      // === Non symmetric assembling. ===
      for (int iq = 0; iq < nPoints; iq++) {
	(*LALt[iq]) *= elInfo->getDet();
	grdPsi = psiFast->getGradient(iq);
	grdPhi = phiFast->getGradient(iq);

	for (int i = 0; i < nRow; i++)
	  for (int j = 0; j < nCol; j++)
	    mat[i][j] += quadrature->getWeight(iq) * 
	      xAy((*grdPsi)[i], (*LALt[iq]), (*grdPhi)[j]);
      }
    }
  }


  Stand2::Stand2(Operator *op, Assembler *assembler, Quadrature *quad) 
    : SecondOrderAssembler(op, assembler, quad, false)
  {
    name = "standard second order assembler";
  }


  void Stand2::calculateElementMatrix(const ElInfo *elInfo, ElementMatrix& mat)
  {
    DimVec<double> grdPsi(dim, NO_INIT);
    VectorOfFixVecs<DimVec<double> > grdPhi(dim, nCol, NO_INIT);

    const BasisFunction *psi = rowFeSpace->getBasisFcts();
    const BasisFunction *phi = colFeSpace->getBasisFcts();

    int nPoints = quadrature->getNumPoints();

    DimMat<double> **LALt = new DimMat<double>*[nPoints];
    for (int iq = 0; iq < nPoints; iq++) {
      LALt[iq] = new DimMat<double>(dim, NO_INIT);
      LALt[iq]->set(0.0);
    }

    int myRank = omp_get_thread_num();

    for (int i = 0; i < static_cast<int>(terms[myRank].size()); i++)
      (static_cast<SecondOrderTerm*>(terms[myRank][i]))->getLALt(elInfo, nPoints, LALt);    

    if (symmetric) {
      TEST_EXIT_DBG(nCol == nRow)("nCol != nRow, but symmetric assembling!\n");

      for (int iq = 0; iq < nPoints; iq++) {
	(*LALt[iq]) *= elInfo->getDet();

	for (int i = 0; i < nCol; i++)
	  (*(phi->getGrdPhi(i)))(quadrature->getLambda(iq), grdPhi[i]);

	for (int i = 0; i < nRow; i++) {
	  (*(psi->getGrdPhi(i)))(quadrature->getLambda(iq), grdPsi);

	  mat[i][i] += quadrature->getWeight(iq) * 
	    (grdPsi * ((*LALt[iq]) * grdPhi[i]));

	  for (int j = i + 1; j < nCol; j++) {
	    double val = quadrature->getWeight(iq) * 
	      (grdPsi * ((*LALt[iq]) * grdPhi[j]));
	    mat[i][j] += val;
	    mat[j][i] += val;
	  }
	}
      }
    } else {      /*  non symmetric assembling   */
      for (int iq = 0; iq < nPoints; iq++) {
	(*LALt[iq]) *= elInfo->getDet();

	for (int i = 0; i < nCol; i++)
	  (*(phi->getGrdPhi(i)))(quadrature->getLambda(iq), grdPhi[i]);

	for (int i = 0; i < nRow; i++) {
	  (*(psi->getGrdPhi(i)))(quadrature->getLambda(iq), grdPsi);
	  for (int j = 0; j < nCol; j++) {
	    mat[i][j] += quadrature->getWeight(iq) *
	      (grdPsi * ((*LALt[iq]) * grdPhi[j]));
	  }
	}
      }
    }

    for (int iq = 0; iq < nPoints; iq++) 
      delete LALt[iq];

    delete [] LALt;
  }

}