// DataIO_Std::WriteData()
int DataIO_Std::WriteData(FileName const& fname, DataSetList const& SetList)
{
int err = 0;
if (!SetList.empty()) {
// Open output file.
CpptrajFile file;
if (file.OpenWrite( fname )) return 1;
// Base write type off first data set dimension FIXME
if (SetList[0]->Group() == DataSet::CLUSTERMATRIX) {
// Special case of 2D - may have sieved frames.
err = WriteCmatrix(file, SetList);
} else if (SetList[0]->Ndim() == 1) {
if (group_ == NO_TYPE) {
if (isInverted_)
err = WriteDataInverted(file, SetList);
else
err = WriteDataNormal(file, SetList);
} else
err = WriteByGroup(file, SetList, group_);
} else if (SetList[0]->Ndim() == 2)
err = WriteData2D(file, SetList);
else if (SetList[0]->Ndim() == 3)
err = WriteData3D(file, SetList);
file.CloseFile();
}
return err;
}
// DataSetList::GetMultipleSets()
DataSetList DataSetList::GetMultipleSets( std::string const& dsargIn ) const {
DataSetList dsetOut = SelectSets(dsargIn);
if ( dsetOut.empty() ) {
mprintf("Warning: '%s' selects no data sets.\n", dsargIn.c_str());
PendingWarning();
}
return dsetOut;
}
// DataSetList::FindSetOfType()
DataSet* DataSetList::FindSetOfType(std::string const& nameIn, DataSet::DataType typeIn) const
{
DataSetList dsetOut = SelectSets( nameIn, typeIn );
if (dsetOut.empty())
return 0;
else if (dsetOut.size() > 1)
mprintf("Warning: '%s' selects multiple sets. Only using first.\n");
return dsetOut[0];
}
// DataSetList::GetDataSet()
DataSet* DataSetList::GetDataSet( std::string const& nameIn ) const {
DataSetList dsetOut = SelectSets( nameIn );
if (dsetOut.empty()) {
mprintf("Warning: Data set '%s' not found.\n", nameIn.c_str());
PendingWarning();
return 0;
} else if (dsetOut.size() > 1)
mprintf("Warning: '%s' selects multiple sets, only using first set.\n");
return dsetOut[0];
}
/** Search for a COORDS DataSet. If no name specified, create a default
* COORDS DataSet named _DEFAULTCRD_.
*/
DataSet* DataSetList::FindCoordsSet(std::string const& setname) {
DataSet* outset = 0;
if (setname.empty()) {
// crdset not given, search for the default set
outset = FindSetOfType("_DEFAULTCRD_", DataSet::COORDS);
if (outset == 0) {
// No default set; create one.
outset = AddSet(DataSet::COORDS, "_DEFAULTCRD_", "CRD");
}
} else {
DataSetList dslist = SelectGroupSets(setname, DataSet::COORDINATES);
if (!dslist.empty()) outset = dslist[0];
}
return outset;
}
// DataIO_Std::WriteDataInverted()
int DataIO_Std::WriteDataInverted(CpptrajFile& file, DataSetList const& Sets)
{
if (Sets.empty() || CheckAllDims(Sets, 1)) return 1;
// Determine size of largest DataSet.
size_t maxFrames = DetermineMax( Sets );
// Write each set to a line.
DataSet::SizeArray positions(1);
// Set up x column format
DataSetList::const_iterator set = Sets.begin();
TextFormat x_col_format;
if (hasXcolumn_)
x_col_format = XcolFmt();
else
x_col_format = (*set)->Format();
for (; set != Sets.end(); ++set) {
// Write dataset name as first column.
WriteNameToBuffer( file, (*set)->Meta().Legend(), x_col_format.ColumnWidth(), false);
// Write each frame to subsequent columns
for (positions[0] = 0; positions[0] < maxFrames; positions[0]++)
(*set)->WriteBuffer(file, positions);
file.Printf("\n");
}
return 0;
}
// Exec_SortEnsembleData::SortData()
int Exec_SortEnsembleData::SortData(DataSetList const& setsToSort, DataSetList& OutputSets)
const
{
int err = 0;
if (setsToSort.empty()) {
rprinterr("Error: No sets selected.\n");
err = 1;
}
if (CheckError(err)) return 1;
mprintf("\tSorting the following sets:\n");
setsToSort.List();
# ifdef MPI
// Number of sets to sort should be equal to # members I am responsible for.
if (Parallel::N_Ens_Members() != (int)setsToSort.size()) {
rprinterr("Internal Error: Number of ensemble members (%i) != # sets to sort (%zu)\n",
Parallel::N_Ens_Members(), setsToSort.size());
return 1;
}
# endif
DataSet::DataType dtype = setsToSort[0]->Type();
unsigned int maxSize = 0;
for (DataSetList::const_iterator ds = setsToSort.begin(); ds != setsToSort.end(); ++ds) {
if ((*ds)->Size() < 1) { //TODO check sizes match
rprinterr("Error: Set '%s' is empty.\n", (*ds)->legend());
err = 1;
break;
}
if (ds == setsToSort.begin())
maxSize = (*ds)->Size();
else if ((*ds)->Size() < maxSize) {
rprintf("Warning: Set '%s' has fewer frames (%zu) than previous set(s) (%u)\n"
"Warning: Only using the first %zu frames of all sets.\n",
(*ds)->legend(), (*ds)->Size(), maxSize, (*ds)->Size());
maxSize = (unsigned int)(*ds)->Size();
} else if ((*ds)->Size() > maxSize) {
rprintf("Warning: Set '%s' has more frames (%zu) than previous set(s) (%u)\n"
"Warning: Only using the first %u frames of all sets.\n",
(*ds)->legend(), (*ds)->Size(), maxSize, maxSize);
}
if (dtype != (*ds)->Type()) {
rprinterr("Error: Set '%s' has different type than first set.\n", (*ds)->legend());
err = 1;
break;
}
}
if (CheckError(err)) return 1;
# ifdef MPI
unsigned int threadSize = maxSize;
comm_.AllReduce( &maxSize, &threadSize, 1, MPI_UNSIGNED, MPI_MIN );
typedef std::vector<int> Iarray;
Iarray Dtypes( comm_.Size(), -1 );
if ( comm_.AllGather( &dtype, 1, MPI_INT, &Dtypes[0] ) ) return 1;
for (int rank = 1; rank < comm_.Size(); rank++)
if (Dtypes[0] != Dtypes[rank]) {
rprinterr("Error: Set types on rank %i do not match types on rank 0.\n", rank);
err = 1;
break;
}
if (comm_.CheckError( err )) return 1;
# endif
// Only work for pH data for now.
if (dtype != DataSet::PH_EXPL && dtype != DataSet::PH_IMPL) {
rprinterr("Error: Only works for pH REMD data for now.\n");
return 1;
}
err = Sort_pH_Data( setsToSort, OutputSets, maxSize );
return err;
}
// Exec_DataSetCmd::Remove()
Exec::RetType Exec_DataSetCmd::Remove(CpptrajState& State, ArgList& argIn) {
std::string status;
// Get criterion type
CriterionType criterion = UNKNOWN_C;
for (int i = 1; i < (int)N_C; i++)
if (argIn.hasKey( CriterionKeys[i] )) {
criterion = (CriterionType)i;
status.assign( CriterionKeys[i] );
break;
}
if (criterion == UNKNOWN_C) {
mprinterr("Error: No criterion specified for 'remove'.\n");
return CpptrajState::ERR;
}
// Get select type
SelectType select = UNKNOWN_S;
std::string val1, val2;
for (const SelectPairType* ptr = SelectKeys; ptr->key_ != 0; ptr++)
if (argIn.Contains( ptr->key_ )) {
select = ptr->type_;
val1 = argIn.GetStringKey( ptr->key_ );
status.append( " " + std::string(ptr->key_) + " " + val1 );
// Get 'and' value for between/outside. TODO put nargs in SelectPairType?
if (select == BETWEEN || select == OUTSIDE) {
val2 = argIn.GetStringKey("and");
if (val2.empty()) {
mprinterr("Error: Missing 'and' value for selection '%s'\n", ptr->key_);
return CpptrajState::ERR;
}
status.append(" and " + val2);
}
break;
}
if (select == UNKNOWN_S || val1.empty()) {
mprinterr("Error: No selection specified for 'remove'.\n");
return CpptrajState::ERR;
}
if ( (criterion == SMODE || criterion == STYPE) &&
(select != EQUAL && select != NOT_EQUAL) )
{
mprinterr("Error: Specified select not valid for criterion '%s'\n", CriterionKeys[criterion]);
return CpptrajState::ERR;
}
mprintf("\tRemoving data sets");
std::string setSelectArg = argIn.GetStringNext();
if (setSelectArg.empty())
setSelectArg.assign("*");
else
mprintf(" within selection '%s'", setSelectArg.c_str());
mprintf(" %s\n", status.c_str());
DataSetList tempDSL = State.DSL().GetMultipleSets( setSelectArg );
if (tempDSL.empty()) {
mprinterr("Error: No data sets selected.\n");
return CpptrajState::ERR;
}
// Remove sets
unsigned int Nremoved = 0;
if ( criterion == AVERAGE ) {
if (!validDouble( val1 )) {
mprinterr("Error: '%s' is not a valid number\n", val1.c_str());
return CpptrajState::ERR;
}
double d_val1 = convertToDouble( val1 );
double d_val2 = d_val1;
if (!val2.empty()) {
if (!validDouble( val2 )) {
mprinterr("Error: '%s' is not a valid number\n", val2.c_str());
return CpptrajState::ERR;
}
d_val2 = convertToDouble( val2 );
}
for (DataSetList::const_iterator ds = tempDSL.begin(); ds != tempDSL.end(); ++ds)
{
if ( (*ds)->Group() != DataSet::SCALAR_1D )
mprintf("Warning: '%s' is not a valid data set for 'average' criterion.\n", (*ds)->legend());
else {
DataSet_1D const& ds1 = static_cast<DataSet_1D const&>( *(*ds) );
double avg = ds1.Avg();
bool remove = false;
switch (select) {
case EQUAL : remove = (avg == d_val1); break;
case NOT_EQUAL : remove = (avg != d_val1); break;
case LESS_THAN : remove = (avg < d_val1); break;
case GREATER_THAN : remove = (avg > d_val1); break;
case BETWEEN : remove = (avg > d_val1 && avg < d_val2); break;
case OUTSIDE : remove = (avg < d_val1 || avg > d_val2); break;
case UNKNOWN_S:
case N_S : return CpptrajState::ERR; // Sanity check
}
if (remove) {
mprintf("\t Removing set '%s' (avg is %g)\n", (*ds)->legend(), avg);
State.RemoveDataSet( *ds );
++Nremoved;
}
}
}
} else if ( criterion == SIZE ) {
if (!validInteger( val1 )) {
mprinterr("Error: '%s' is not a valid number\n", val1.c_str());
return CpptrajState::ERR;
}
unsigned int i_val1 = (unsigned int)convertToInteger( val1 );
unsigned int i_val2 = i_val1;
if (!val2.empty()) {
if (!validInteger( val2 )) {
mprinterr("Error: '%s' is not a valid number\n", val2.c_str());
return CpptrajState::ERR;
}
i_val2 = convertToInteger( val2 );
}
for (DataSetList::const_iterator ds = tempDSL.begin(); ds != tempDSL.end(); ++ds)
{
unsigned int size = (*ds)->Size();
bool remove = false;
switch ( select ) {
case EQUAL : remove = (size == i_val1); break;
case NOT_EQUAL : remove = (size != i_val1); break;
case LESS_THAN : remove = (size < i_val1); break;
case GREATER_THAN : remove = (size > i_val1); break;
case BETWEEN : remove = (size > i_val1 && size < i_val2); break;
case OUTSIDE : remove = (size < i_val1 || size > i_val2); break;
case UNKNOWN_S:
case N_S : return CpptrajState::ERR; // Sanity check
}
if (remove) {
mprintf("\t Removing set '%s' (size is %u)\n", (*ds)->legend(), size);
State.RemoveDataSet( *ds );
++Nremoved;
}
}
} else if ( criterion == SMODE ) {
MetaData::scalarMode mode_val = MetaData::ModeFromKeyword( val1 );
if (mode_val == MetaData::UNKNOWN_MODE) {
mprinterr("Error: '%s' is not a valid mode.\n", val1.c_str());
return CpptrajState::ERR;
}
for (DataSetList::const_iterator ds = tempDSL.begin(); ds != tempDSL.end(); ++ds)
{
bool remove = false;
MetaData::scalarMode mode = (*ds)->Meta().ScalarMode();
if (select == EQUAL ) remove = ( mode == mode_val );
else if (select == NOT_EQUAL) remove = ( mode != mode_val );
else return CpptrajState::ERR; // Sanity check
if (remove) {
mprintf("\t Removing set '%s' (mode is '%s')\n", (*ds)->legend(), MetaData::ModeString(mode));
State.RemoveDataSet( *ds );
++Nremoved;
}
}
} else if ( criterion == STYPE ) {
MetaData::scalarType type_val = MetaData::TypeFromKeyword( val1, MetaData::UNKNOWN_MODE );
if (type_val == MetaData::UNDEFINED) {
mprinterr("Error: '%s' is not a valid type.\n", val1.c_str());
return CpptrajState::ERR;
}
for (DataSetList::const_iterator ds = tempDSL.begin(); ds != tempDSL.end(); ++ds)
{
bool remove = false;
MetaData::scalarType type = (*ds)->Meta().ScalarType();
if (select == EQUAL ) remove = ( type == type_val );
else if (select == NOT_EQUAL) remove = ( type != type_val );
else return CpptrajState::ERR; // Sanity check
if (remove) {
mprintf("\t Removing set '%s' (typeis '%s')\n", (*ds)->legend(), MetaData::TypeString(type));
State.RemoveDataSet( *ds );
++Nremoved;
}
}
} else {
mprinterr("Internal Error: Criterion not yet implemented.\n");
return CpptrajState::ERR;
}
mprintf("\tRemoved %u of %zu sets.\n", Nremoved, tempDSL.size());
return CpptrajState::OK;
}
// DataIO_Std::WriteDataNormal()
int DataIO_Std::WriteDataNormal(CpptrajFile& file, DataSetList const& Sets) {
// Assume all 1D data sets.
if (Sets.empty() || CheckAllDims(Sets, 1)) return 1;
// For this output to work the X-dimension of all sets needs to match.
// The most important things for output are min and step so just check that.
// Use X dimension of set 0 for all set dimensions.
CheckXDimension( Sets );
// TODO: Check for empty dim.
DataSet* Xdata = Sets[0];
Dimension const& Xdim = static_cast<Dimension const&>( Xdata->Dim(0) );
int xcol_width = Xdim.Label().size() + 1; // Only used if hasXcolumn_
if (xcol_width < 8) xcol_width = 8;
int xcol_precision = 3;
// Determine size of largest DataSet.
size_t maxFrames = DetermineMax( Sets );
// Set up X column.
TextFormat x_col_format(XcolFmt());
if (hasXcolumn_) {
if (XcolPrecSet()) {
xcol_width = XcolWidth();
x_col_format = TextFormat(XcolFmt(), XcolWidth(), XcolPrec());
} else {
// Create format string for X column based on dimension in first data set.
// Adjust X col precision as follows: if the step is set and has a
// fractional component set the X col width/precision to either the data
// width/precision or the current width/precision, whichever is larger. If
// the set is XYMESH but step has not been set (so we do not know spacing
// between X values) use default precision. Otherwise the step has no
// fractional component so make the precision zero.
double step_i;
double step_f = modf( Xdim.Step(), &step_i );
double min_f = modf( Xdim.Min(), &step_i );
if (Xdim.Step() > 0.0 && (step_f > 0.0 || min_f > 0.0)) {
xcol_precision = std::max(xcol_precision, Xdata->Format().Precision());
xcol_width = std::max(xcol_width, Xdata->Format().Width());
} else if (Xdata->Type() != DataSet::XYMESH)
xcol_precision = 0;
x_col_format.SetCoordFormat( maxFrames, Xdim.Min(), Xdim.Step(), xcol_width, xcol_precision );
}
} else {
// If not writing an X-column, no leading space for the first dataset.
Xdata->SetupFormat().SetFormatAlign( TextFormat::RIGHT );
}
// Write header to buffer
std::vector<int> Original_Column_Widths;
if (writeHeader_) {
// If x-column present, write x-label
if (hasXcolumn_)
WriteNameToBuffer( file, Xdim.Label(), xcol_width, true );
// To prevent truncation of DataSet legends, adjust the width of each
// DataSet if necessary.
for (DataSetList::const_iterator ds = Sets.begin(); ds != Sets.end(); ++ds) {
// Record original column widths in case they are changed.
Original_Column_Widths.push_back( (*ds)->Format().Width() );
int colLabelSize;
if (ds == Sets.begin() && !hasXcolumn_)
colLabelSize = (int)(*ds)->Meta().Legend().size() + 1;
else
colLabelSize = (int)(*ds)->Meta().Legend().size();
//mprintf("DEBUG: Set '%s', fmt width= %i, colWidth= %i, colLabelSize= %i\n",
// (*ds)->legend(), (*ds)->Format().Width(), (*ds)->Format().ColumnWidth(),
// colLabelSize);
if (colLabelSize >= (*ds)->Format().ColumnWidth())
(*ds)->SetupFormat().SetFormatWidth( colLabelSize );
}
// Write dataset names to header, left-aligning first set if no X-column
DataSetList::const_iterator set = Sets.begin();
if (!hasXcolumn_)
WriteNameToBuffer( file, (*set)->Meta().Legend(), (*set)->Format().ColumnWidth(), true );
else
WriteNameToBuffer( file, (*set)->Meta().Legend(), (*set)->Format().ColumnWidth(), false );
++set;
for (; set != Sets.end(); ++set)
WriteNameToBuffer( file, (*set)->Meta().Legend(), (*set)->Format().ColumnWidth(), false );
file.Printf("\n");
}
// Write Data
DataSet::SizeArray positions(1);
for (positions[0] = 0; positions[0] < maxFrames; positions[0]++) {
// Output Frame for each set
if (hasXcolumn_)
file.Printf( x_col_format.fmt(), Xdata->Coord(0, positions[0]) );
for (DataSetList::const_iterator set = Sets.begin(); set != Sets.end(); ++set)
(*set)->WriteBuffer(file, positions);
file.Printf("\n");
}
// Restore original column widths if necessary
if (!Original_Column_Widths.empty())
for (unsigned int i = 0; i != Original_Column_Widths.size(); i++)
Sets[i]->SetupFormat().SetFormatWidth( Original_Column_Widths[i] );
return 0;
}