void MzidentmlReader::getMaxMinCharge(const std::string &fn, bool isDecoy) {
ifstream ifs;
ifs.exceptions(ifstream::badbit | ifstream::failbit);
try {
ifs.open(fn.c_str());
parser p;
bool schemaVal = true;
xml_schema::dom::auto_ptr<xercesc_3_1::DOMDocument> doc
(p.start(ifs, fn.c_str(), schemaVal, schemaDefinition, schema_major,
schema_minor, scheme_namespace));
// MT: This seems to be a bit slow for doing nothing
for (doc = p.next(); doc.get() != 0
&& !XMLString::equals(spectrumIdentificationResultStr, doc->getDocumentElement()->getTagName()); doc = p.next()) {
// Let's skip some sub trees that we are not interested, e.g. AnalysisCollection
}
// For each SpectrumIdentificationResult
int itemCount = 1;
for (; doc.get() != 0 && XMLString::equals(spectrumIdentificationResultStr,
doc->getDocumentElement()->getTagName()); doc = p.next()) {
::mzIdentML_ns::SpectrumIdentificationResultType specIdResult(*doc->getDocumentElement());
// For each SpectrumIdentificationItem
BOOST_FOREACH(const ::mzIdentML_ns::SpectrumIdentificationItemType & item, specIdResult.SpectrumIdentificationItem()) {
minCharge = (std::min)(item.chargeState(), minCharge);
maxCharge = (std::max)(item.chargeState(), maxCharge);
searchEngineSpecificParsing(item, itemCount); // Virtual function that potentially checks the features
++itemCount;
}
}
} catch (ifstream::failure e) {
cerr << "Exception opening/reading file :" << fn << endl;
} catch (const xercesc::DOMException& e) {
char * tmpStr = XMLString::transcode(e.getMessage());
std::cerr << "catch xercesc_3_1::DOMException=" << tmpStr << std::endl;
XMLString::release(&tmpStr);
} catch (const xml_schema::exception& e) {
cerr << "XML schema exception in getMaxMinCharge: " << e << endl;
} catch (MyException e) {
cerr << e.what() << endl;
exit(1);
} catch (std::exception e) {
cerr << "Some unknown exception in getMaxMinCharge: " << e.what() << endl;
}
ifs.close();
return;
}
void MzidentmlReader::read(const std::string &fn, bool isDecoy, boost::shared_ptr<FragSpectrumScanDatabase> database)
{
namespace xml = xsd::cxx::xml;
scanNumberMapType scanNumberMap;
ifstream ifs;
try
{
ifs.exceptions(ifstream::badbit | ifstream::failbit);
ifs.open(fn.c_str());
parser p;
xml_schema::dom::auto_ptr<xercesc_3_1::DOMDocument> doc
(p.start(ifs, fn.c_str(), true, schemaDefinition, schema_major, schema_minor, scheme_namespace));
//NOTE wouldnt be better to use the get tag by Name to jump SequenceCollenction directly?
while (doc.get() != 0 && !XMLString::equals(sequenceCollectionStr,
doc->getDocumentElement()->getTagName())) {
doc = p.next(); // Let's skip some sub trees that we are not interested, e.g. AuditCollection
}
assert(doc.get());
mzIdentML_ns::SequenceCollectionType sequenceCollection(*doc->getDocumentElement());
peptideMap.clear();
proteinMap.clear();
peptideEvidenceMap.clear();
//NOTE probably I can get rid of these hash tables with a proper access to elements by tag and id
BOOST_FOREACH(const mzIdentML_ns::SequenceCollectionType::Peptide_type &peptide, sequenceCollection.Peptide()) {
//PEPTIDE
mzIdentML_ns::SequenceCollectionType::Peptide_type *pept =
new mzIdentML_ns::SequenceCollectionType::Peptide_type(peptide);
peptideMap.insert(std::make_pair(peptide.id(), pept));
}
BOOST_FOREACH(const mzIdentML_ns::SequenceCollectionType::DBSequence_type &protein, sequenceCollection.DBSequence()) {
//PROTEIN
mzIdentML_ns::SequenceCollectionType::DBSequence_type *prot =
new mzIdentML_ns::SequenceCollectionType::DBSequence_type(protein);
proteinMap.insert(std::make_pair(protein.id(), prot));
}
BOOST_FOREACH(const ::mzIdentML_ns::PeptideEvidenceType &peptideE, sequenceCollection.PeptideEvidence()) {
//PEPTIDE EVIDENCE
::mzIdentML_ns::PeptideEvidenceType *peptE = new mzIdentML_ns::PeptideEvidenceType(peptideE);
peptideEvidenceMap.insert(std::make_pair(peptideE.id(), peptE));
}
for (doc = p.next(); doc.get() != 0 && !XMLString::equals(spectrumIdentificationResultStr,
doc->getDocumentElement()->getTagName()); doc = p.next()) {
// Let's skip some sub trees that we are not interested, e.g. AnalysisCollection
}
unsigned scanNumber = 0;
bool useRankedScanNumbers = false; /* True scan numbers are used,
if they can't be found, use ranked scan numbers from 1 and up. */
for (; doc.get() != 0 && XMLString::equals(spectrumIdentificationResultStr,
doc->getDocumentElement()->getTagName()); doc = p.next()) {
::mzIdentML_ns::SpectrumIdentificationResultType specIdResult(*doc->getDocumentElement());
assert(specIdResult.SpectrumIdentificationItem().size() > 0);
unsigned numberHitsSpectra = 0;
//Find scan number from the cvParam element in spetrumIdentificationResults
if(!useRankedScanNumbers) {
bool foundScanNumber = false; // Indicates whether a proper scan number was found
BOOST_FOREACH(const ::mzIdentML_ns::CVParamType & cv, specIdResult.cvParam()) {
std::string param_name(cv.name().c_str());
std::string expected_name("scan number(s)");
if (param_name == expected_name) {
scanNumber = boost::lexical_cast<unsigned>(cv.value().get().c_str());
foundScanNumber = true;
}
}
if(!foundScanNumber || scanNumber == 0) {
std::cerr << "No scan number was found for a PSM (or it equaled 0), scans are ranked from 1 and up" << std::endl;
useRankedScanNumbers = true;
}
}
// If no scan numbers were found, or a scan of 0 was found, just rank them
if(useRankedScanNumbers) {
++scanNumber;
}
BOOST_FOREACH(const ::mzIdentML_ns::SpectrumIdentificationItemType & item, specIdResult.SpectrumIdentificationItem()) {
if(++numberHitsSpectra <= po->hitsPerSpectrum) {
assert(item.experimentalMassToCharge());
int charge = item.chargeState();
::percolatorInNs::fragSpectrumScan::experimentalMass_type experimentalMass = item.experimentalMassToCharge()*charge - proton_mass*charge;
createPSM(item, experimentalMass, isDecoy, scanNumber, database, fn);
}
}
}
cleanHashMaps();
ifs.close();
}
