DoubleVectorVector QuanMSMSXMLData::getFormulaPurityCoefficients ( const string& name ) const
{
DoubleVectorVector dvv;
StringVector f = getQuanMSMSInfo ( name ).formulae;
DoubleVector m = getQuanMSMSInfo ( name ).masses;
for ( int i = 0 ; i < f.size () ; i++ ) {
DoubleVector dv;
if ( f [i] != "" ) {
IsotopePeakStats ips ( f [i], 1 );
int index = ips.getProbabilityIndexForMass ( m [i] );
if ( index >= 2 ) dv.push_back ( ips.getProbability ( index - 2 ) );
else dv.push_back ( 0.0 );
if ( index >= 1 ) dv.push_back ( ips.getProbability ( index - 1 ) );
else dv.push_back ( 0.0 );
dv.push_back ( ips.getProbability ( index + 1 ) );
dv.push_back ( ips.getProbability ( index + 2 ) );
}
else { // No formulae so assume no correction
for ( int j = 0 ; j < 4 ; j++ ) {
dv.push_back ( 0.0 );
}
}
dvv.push_back ( dv );
}
return dvv;
}
pair<DoubleVectorVector,IntVector> readJF(string filename) {
igzstream is; is.open(filename.c_str());
if(!is.good()) {
ERR << "Reading from " << filename << endl;
exit(20);
}
DoubleVectorVector resultData;
IntVector resultClasses;
int maxCls;
uint dim;
int cls=0;
DoubleVector *tmpDblVec;
if(is.good() && is) {
is >> maxCls >> dim;
while(cls != -1) {
double tmpDbl;
tmpDblVec=new DoubleVector(dim);
is >> cls;
if(cls !=-1) {
for(uint i=0;i<dim;++i) {
is >> tmpDbl;
(*tmpDblVec)[i]=tmpDbl;
}
resultClasses.push_back(cls);
resultData.push_back(tmpDblVec);
}
}
is.close();
}
ResultVector cluster(vector<LocalFeatures*> &locfeat, bool clusterWithPositions, string splitMode, uint maxSplit, uint stopWithNClusters, string disturbMode, string poolMode, uint dontSplitBelow, uint iterationsBetweenSplits, uint minObservationsPerCluster, double epsilon, string distanceMaker, string saveModelTo, bool saveBeforeSplits, string loadModelFrom)
{
BaseClusterer *clusterer = new EM();
if(loadModelFrom!="")
{
clusterer->loadModel(loadModelFrom);
}
setupClusterer(dynamic_cast<EM*>(clusterer), splitMode, maxSplit, stopWithNClusters, disturbMode, poolMode, dontSplitBelow,
iterationsBetweenSplits, minObservationsPerCluster, epsilon, distanceMaker, saveBeforeSplits);
ResultVector clusterinformation;
if (clusterWithPositions)
{
DBG(10) << "clustering with position information included" << endl;
}
DoubleVectorVector localFeaturesData;
for (uint j = 0; j < locfeat.size(); j++)
{
LocalFeatures* localFeatures = locfeat[j];
int xSize = localFeatures->imageSizeX();
int ySize = localFeatures->imageSizeY();
for (uint i = 0; i < localFeatures->numberOfFeatures(); i++)
{
DoubleVector* lfvector = &((*localFeatures)[i]);
if (clusterWithPositions)
{
pair<double, double> pos = localFeatures->relativePosition(i);
lfvector->push_back((double) localFeatures->position(i).x / (double) xSize);
lfvector->push_back((double) localFeatures->position(i).y / (double) ySize);
}
localFeaturesData.push_back( lfvector );
}
}
if(saveModelTo!="")
{
dynamic_cast<EM*>(clusterer)->run(localFeaturesData, clusterinformation, saveModelTo);
clusterer->saveModel(saveModelTo);
}
else
{
clusterer->run(localFeaturesData, clusterinformation);
}
delete clusterer;
return clusterinformation;
}
static void initialiseIndexScore ()
{
int numEntries;
char* info = getFileInfo ( MsparamsDir::instance ().getParamPath ( "indicies.txt" ), '>', 1, true, &numEntries );
for ( int i = 0 ; i < numEntries ; i++ ) {
names.push_back ( ( i == 0 ) ? strtok ( info, "\n" ) : strtok ( NULL, "\n" ) );
DoubleVector dv (52);
indexV.push_back ( dv );
fill ( dv.begin (), dv.end (), 0.0 );
for ( ; ; ) {
char aa;
double value;
char* line = strtok ( NULL, "\n" );
if ( !strcmp ( line, ">" ) ) break;
sscanf ( line, "%c %lf", &aa, &value );
indexV [i][aa-'A'] = value;
}
}
initialised = true;
}
