/**
| returns the number of indices added.
*/
unsigned NxsSetReader::InterpretTokenAsIndices(NxsToken &token,
const NxsLabelToIndicesMapper & mapper,
const char * setType,
const char * cmdName,
NxsUnsignedSet * destination)
{
try {
const std::string t = token.GetToken();
if (NxsString::case_insensitive_equals(t.c_str(), "ALL"))
{
unsigned m = mapper.GetMaxIndex();
NxsUnsignedSet s;
for (unsigned i = 0; i <= m; ++i)
s.insert(i);
destination->insert(s.begin(), s.end());
return (unsigned)s.size();
}
return mapper.GetIndicesForLabel(t, destination);
}
catch (const NxsException & x)
{
NxsString errormsg = "Error in the ";
errormsg << setType << " descriptor of a " << cmdName << " command.\n";
errormsg += x.msg;
throw NxsException(errormsg, token);
}
catch (...)
{
NxsString errormsg = "Expecting a ";
errormsg << setType << " descriptor (number or label) in the " << cmdName << ". Encountered ";
errormsg << token.GetToken();
throw NxsException(errormsg, token);
}
}
std::vector<unsigned> NxsSetReader::GetSetAsVector(const NxsUnsignedSet &s)
{
std::vector<unsigned> u;
u.reserve(s.size());
for (NxsUnsignedSet::const_iterator sIt = s.begin(); sIt != s.end(); ++sIt)
u.push_back(*sIt);
return u;
}
void writeCharactersAsGapped(ostream &out, NxsCharactersBlock * cb)
{
if (!cb )
return;
NxsTaxaBlockAPI *tb = cb->GetTaxaBlockPtr(NULL);
if (tb == NULL)
return;
NxsCharactersBlock::DataTypesEnum dt = cb->GetDataType();
const char g = cb->GetGapSymbol();
const std::string baseTitle = cb->GetTitle();
if (tb == NULL
|| dt == NxsCharactersBlock::standard
|| dt == NxsCharactersBlock::continuous
|| dt == NxsCharactersBlock::mixed
|| (g == '\0' && !gGapCodeAllPresentSites))
{
cb->WriteAsNexus(out);
writeCoreAssumptions(out, cb, baseTitle.c_str());
return;
}
bool isAutogen = cb->IsAutoGeneratedTitle();
std::string newTitle = baseTitle;
newTitle.append("GapsAsMissing");
cb->SetTitle(newTitle, isAutogen);
std::set<unsigned> gappedColumns;
std::set<unsigned> residueColumns;
out << "BEGIN CHARACTERS;\n";
cb->WriteBasicBlockCommands(out);
const unsigned ntaxTotal = tb->GetNTax();
out << " DIMENSIONS";
if (tb)
{
const unsigned wod = cb->GetNTaxWithData();
if (wod > 0 && wod != ntaxTotal)
out << " NTax=" << wod;
}
const unsigned nc = cb->GetNCharTotal();
out << " NChar=" << nc << ";\n";
cb->WriteEliminateCommand(out);
cb->SetGapSymbol('\0');
cb->WriteFormatCommand(out);
cb->SetGapSymbol(g);
cb->WriteCharStateLabelsCommand(out);
std::vector<int> intWts;
std::vector<double> dblWts;
NxsUnsignedSet activeExSet = cb->GetExcludedIndexSet();
const NxsTransformationManager &tm = cb->GetNxsTransformationManagerRef();
intWts = tm.GetDefaultIntWeights();
if (intWts.empty())
dblWts = tm.GetDefaultDoubleWeights();
unsigned width = tb->GetMaxTaxonLabelLength();
out << "Matrix\n";
unsigned begCharInd = 0;
unsigned endCharInd = nc;
for (unsigned i = 0; i < ntaxTotal; i++)
{
if (cb->TaxonIndHasData(i))
{
const std::string currTaxonLabel = NxsString::GetEscaped(tb->GetTaxonLabel(i));
out << currTaxonLabel;
unsigned currTaxonLabelLen = (unsigned)currTaxonLabel.size();
unsigned diff = width - currTaxonLabelLen;
for (unsigned k = 0; k < diff + 5; k++)
out << ' ';
const NxsDiscreteStateRow & row = cb->GetDiscreteMatrixRow(i);
if (!row.empty())
{
const NxsDiscreteDatatypeMapper * dm = cb->GetDatatypeMapperForChar(0);
if (dm == NULL)
throw NxsNCLAPIException("No DatatypeMapper in WriteStatesForTaxonAsNexus");
if (cb->IsMixedType())
throw NxsNCLAPIException("Mixed datatypes are not supported by the NEXUSgapcode program.");
if (cb->IsTokens())
{
for (unsigned charInd = begCharInd; charInd < endCharInd; ++charInd)
{
int sc = row[charInd];
out << ' ';
if (sc == NXS_GAP_STATE_CODE)
out << dm->GetMissingSymbol();
else
{
NxsString sl = cb->GetStateLabel(charInd, sc); /*v2.1to2.2 2 */
if (sl == " ")
{
NxsString errormsg = "Writing character state ";
errormsg << 1 + sc << " for character " << 1+charInd << ", but no appropriate chararcter label or symbol was found.";
throw NxsNCLAPIException(errormsg);
}
else
out << NxsString::GetEscaped(sl);
}
}
}
else
{
std::vector<NxsDiscreteStateCell>::const_iterator endIt = row.begin() + begCharInd;
std::vector<NxsDiscreteStateCell>::const_iterator begIt = endIt;
if (endCharInd == row.size())
endIt = row.end();
else
endIt += endCharInd - begCharInd;
unsigned j = begCharInd;
for (; begIt != endIt; ++begIt, ++j)
{
NxsDiscreteStateCell c;
if (*begIt == NXS_GAP_STATE_CODE)
{
c = NXS_MISSING_CODE;
gappedColumns.insert(j);
}
else
{
c = *begIt;
if (gGapCodeAllPresentSites)
gappedColumns.insert(j);
residueColumns.insert(j);
}
dm->WriteStateCodeAsNexusString(out, c, true);
}
}
}
out << '\n';
}
}
out << ";\n";
cb->WriteSkippedCommands(out);
out << "END;\n";
writeCoreAssumptions(out, cb, newTitle.c_str());
const unsigned nGappedCols = gappedColumns.size();
if (nGappedCols > 0)
{
newTitle = baseTitle;
newTitle.append("GapsAsBinary");
cb->SetTitle(newTitle, isAutogen);
out << "BEGIN CHARACTERS;\n";
cb->WriteBasicBlockCommands(out);
out << " DIMENSIONS";
if (tb)
{
const unsigned wod = cb->GetNTaxWithData();
if (wod > 0 && wod != ntaxTotal)
out << " NTax=" << wod;
}
out << " NChar=" << nGappedCols << ";\n";
out << " CharStateLabels " ;
unsigned currChNumber = 1;
std::set<unsigned>::iterator gcIt = gappedColumns.begin();
out << currChNumber++ << " col_" << (1 + *gcIt);
for (++gcIt ; gcIt != gappedColumns.end(); ++gcIt)
out << ",\n " << currChNumber++ << " col_" << (1 + *gcIt);
out << " ;\n" ;
out << "Format Datatype = Standard Symbols=\"01\" missing = '?' ;\n";
out << "Matrix\n";
for (unsigned i = 0; i < ntaxTotal; i++)
{
if (cb->TaxonIndHasData(i))
{
const std::string currTaxonLabel = NxsString::GetEscaped(tb->GetTaxonLabel(i));
out << currTaxonLabel;
unsigned currTaxonLabelLen = (unsigned)currTaxonLabel.size();
unsigned diff = width - currTaxonLabelLen;
for (unsigned k = 0; k < diff + 5; k++)
out << ' ';
const NxsDiscreteStateRow & row = cb->GetDiscreteMatrixRow(i);
for (std::set<unsigned>::iterator cIt = gappedColumns.begin(); cIt != gappedColumns.end(); ++cIt)
{
NxsDiscreteStateCell sc = row[*cIt];
if (sc == NXS_GAP_STATE_CODE)
{
if (gGapCodeAllAbsentSites || (residueColumns.find(*cIt) != residueColumns.end()))
out << '0';
else
out << '?';
}
else if (sc == NXS_MISSING_CODE)
{
out << '?';
}
else
{
out << '1';
}
}
out << '\n';
}
}
out << ";\n";
cb->WriteSkippedCommands(out);
out << "END;\n";
}
std::vector<int> gapintWts;
std::vector<double> gapdblWts;
std::set<unsigned> gapactiveExSet;
if (!activeExSet.empty())
{
unsigned gapind = 0;
for (std::set<unsigned>::iterator cIt = gappedColumns.begin(); cIt != gappedColumns.end(); ++cIt, ++gapind)
{
if (activeExSet.find(*cIt) != activeExSet.end())
gapactiveExSet.insert(gapind);
}
}
if (!intWts.empty())
{
for (std::set<unsigned>::iterator cIt = gappedColumns.begin(); cIt != gappedColumns.end(); ++cIt)
gapintWts.push_back(intWts[*cIt]);
}
if (!dblWts.empty())
{
for (std::set<unsigned>::iterator cIt = gappedColumns.begin(); cIt != gappedColumns.end(); ++cIt)
gapdblWts.push_back(dblWts[*cIt]);
}
if (!(gapactiveExSet.empty() && gapintWts.empty() && gapdblWts.empty()))
{
out << "BEGIN ASSUMPTIONS; \n LINK CHARACTERS = ";
out << NxsString::GetEscaped(newTitle) << " ;\n";
if (!gapactiveExSet.empty())
{
NxsString exsetName;
exsetName << newTitle;
exsetName.append("GapExSet");
NxsUnsignedSetMap m;
m[exsetName] = gapactiveExSet;
NxsWriteSetCommand("EXSET", m, out, exsetName.c_str());;
}
if (!(gapintWts.empty() && gapdblWts.empty()))
{
NxsTransformationManager &cbntm = cb->GetNxsTransformationManagerRef();
std::string wtSetName = cbntm.GetDefaultWeightSetName();
wtSetName.append("GapWtSet");
NxsTransformationManager ntm;
if (!gapintWts.empty())
{
NxsTransformationManager::ListOfIntWeights iw;
vecToListOfIntWeights(gapintWts, iw);
ntm.AddIntWeightSet(wtSetName, iw, true);
}
else
{
NxsTransformationManager::ListOfDblWeights dw;
vecToListOfDblWeights(gapdblWts, dw);
ntm.AddRealWeightSet(wtSetName, dw, true);
}
ntm.WriteWtSet(out);
}
out << "END;\n";
}
cb->SetTitle(baseTitle, isAutogen);
}
void NxsSetReader::ReadSetDefinition(
NxsToken &token,
const NxsLabelToIndicesMapper & mapper,
const char * setType, /* "TAXON" or "CHARACTER" -- for error messages only */
const char * cmdName, /* command name -- "TAXSET" or "EXSET"-- for error messages only */
NxsUnsignedSet * destination, /** to be filled */
const NxsUnsignedSet * taboo)
{
NxsString errormsg;
NxsUnsignedSet tmpset;
NxsUnsignedSet dummy;
if (destination == NULL)
destination = & dummy;
unsigned previousInd = UINT_MAX;
std::vector<unsigned> intersectVec;
while (!token.Equals(";"))
{
if (taboo && token.Equals(","))
return;
if (token.Equals("-"))
{
if (previousInd == UINT_MAX)
{
errormsg = "The '-' must be preceded by number or a ";
errormsg << setType << " label in the " << cmdName << " command.";
throw NxsException(errormsg, token);
}
token.GetNextToken();
if (token.Equals(";") || token.Equals("\\"))
{
errormsg = "Range in the ";
errormsg << setType << " set definition in the " << cmdName << " command must be closed with a number or label.";
throw NxsException(errormsg, token);
}
unsigned endpoint;
if (token.Equals("."))
endpoint = mapper.GetMaxIndex();
else
{
tmpset.clear();
unsigned nAdded = NxsSetReader::InterpretTokenAsIndices(token, mapper, setType, cmdName, &tmpset);
if (nAdded != 1)
{
errormsg = "End of a range in a ";
errormsg << setType << " set definition in the " << cmdName << " command must be closed with a single number or label (not a set).";
throw NxsException(errormsg, token);
}
endpoint = *(tmpset.begin());
if (endpoint < previousInd)
{
errormsg = "End of a range in a ";
errormsg << setType << " set definition in the " << cmdName << " command must be a larger index than the start of the range (found ";
errormsg << previousInd + 1 << " - " << token.GetToken();
throw NxsException(errormsg, token);
}
}
token.GetNextToken();
if (token.Equals("\\"))
{
token.GetNextToken();
NxsString t = token.GetToken();
unsigned stride = 0;
try
{
stride = t.ConvertToUnsigned();
}
catch (const NxsString::NxsX_NotANumber &)
{}
if (stride == 0)
{
errormsg = "Expecting a positive number indicating the 'stride' after the \\ in the ";
errormsg << setType << " set definition in the " << cmdName << " command. Encountered ";
errormsg << t;
throw NxsException(errormsg, token);
}
AddRangeToSet(previousInd, endpoint, stride, destination, taboo, token);
token.GetNextToken();
}
else
AddRangeToSet(previousInd, endpoint, 1, destination, taboo, token);
previousInd = UINT_MAX;
}
else
{
tmpset.clear();
const unsigned nAdded = NxsSetReader::InterpretTokenAsIndices(token, mapper, setType, cmdName, &tmpset);
if (taboo != NULL)
{
set_intersection(taboo->begin(), taboo->end(), tmpset.begin(), tmpset.end(), back_inserter(intersectVec));
if (!intersectVec.empty())
{
errormsg << "Illegal repitition of an index (" << 1 + *(intersectVec.begin()) << ") in multiple subsets.";
throw NxsException(errormsg, token);
}
}
if (nAdded == 1 )
{
previousInd = *(tmpset.begin());
destination->insert(previousInd);
}
else
{
previousInd = UINT_MAX;
destination->insert(tmpset.begin(), tmpset.end());
}
token.GetNextToken();
}
}
}
void NxsSetReader::WriteSetAsNexusValue(const NxsUnsignedSet & nxsset, std::ostream & out)
{
NxsUnsignedSet::const_iterator currIt = nxsset.begin();
const NxsUnsignedSet::const_iterator endIt = nxsset.end();
if (currIt == endIt)
return;
unsigned rangeBegin = 1 + *currIt++;
if (currIt == endIt)
{
out << ' ' << rangeBegin;
return;
}
unsigned prev = 1 + *currIt++;
if (currIt == endIt)
{
out << ' ' << rangeBegin << ' ' << prev;
return;
}
unsigned stride = prev - rangeBegin;
unsigned curr = 1 + *currIt++;
bool inRange = true;
while (currIt != endIt)
{
if (inRange)
{
if (curr - prev != stride)
{
if (prev - rangeBegin == stride)
{
out << ' ' << rangeBegin;
rangeBegin = prev;
stride = curr - prev;
}
else
{
if (stride > 1)
out << ' ' << rangeBegin << '-' << prev << " \\ " << stride;
else
out << ' ' << rangeBegin << '-' << prev ;
inRange = false;
}
}
}
else
{
inRange = true;
rangeBegin = prev;
stride = curr - prev;
}
prev = curr;
curr = 1 + *currIt;
currIt++;
}
if (inRange)
{
if (curr - prev != stride)
{
if (prev - rangeBegin == stride)
out << ' ' << rangeBegin << ' ' << prev;
else
{
if (stride > 1)
out << ' ' << rangeBegin << '-' << prev << " \\ " << stride;
else
out << ' ' << rangeBegin << '-' << prev ;
}
out << ' ' << curr;
}
else
{
if (stride > 1)
out << ' ' << rangeBegin << '-' << curr << " \\ " << stride;
else
out << ' ' << rangeBegin << '-' << curr ;
}
}
else
out << ' ' << prev << ' ' << curr;
}