mascot dat parser

src/CMakeLists.txt

@@ -66,6 +66,7 @@ SET(CPP_FILES
   core/automaticfilterparameters.cpp
   core/identificationgroup.cpp
   core/identification_sources/identificationdatasource.cpp
+  core/identification_sources/identificationmascotdatfile.cpp
   core/identification_sources/identificationpwizfile.cpp
   core/identification_sources/identificationxtandemfile.cpp
   core/labeling/label.cpp
@@ -92,6 +93,8 @@ SET(CPP_FILES
   grouping/ptm/ptmislandgroup.cpp
   grouping/ptm/ptmislandsubgroup.cpp
   grouping/ptm/ptmsamplescan.cpp
+  input/mascot/mascotdatparser.cpp
+  input/mascot/mimeparser.cpp
   input/condorqxmlsaxhandler.cpp
   input/identificationpwizreader.cpp
   input/xpipsaxhandler.cpp

src/core/identification_sources/identificationmascotdatfile.cpp

+/**
+ * \file /core/identification_sources/identificationmascotdatfile.cpp
+ * \date 17/2/2018
+ * \author Olivier Langella
+ * \brief mascot dat identification file handler
+ */
+
+/*******************************************************************************
+* Copyright (c) 2018 Olivier Langella <olivier.langella@u-psud.fr>.
+*
+* This file is part of XTPcpp.
+*
+*     XTPcpp is free software: you can redistribute it and/or modify
+*     it under the terms of the GNU General Public License as published by
+*     the Free Software Foundation, either version 3 of the License, or
+*     (at your option) any later version.
+*
+*     XTPcpp is distributed in the hope that it will be useful,
+*     but WITHOUT ANY WARRANTY; without even the implied warranty of
+*     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*     GNU General Public License for more details.
+*
+*     You should have received a copy of the GNU General Public License
+*     along with XTPcpp.  If not, see <http://www.gnu.org/licenses/>.
+*
+* Contributors:
+*     Olivier Langella <olivier.langella@u-psud.fr> - initial API and implementation
+******************************************************************************/
+
+#include "identificationmascotdatfile.h"
+
+#include <pappsomspp/pappsoexception.h>
+#include "../project.h"
+#include "../../input/mascot/mascotdatparser.h"
+
+IdentificationMascotDatFile::IdentificationMascotDatFile(const QFileInfo & mascot_dat_file) : IdentificationDataSource(mascot_dat_file.absoluteFilePath()), _mascot_dat_file(mascot_dat_file)
+{
+    _engine = IdentificationEngine::mascot;
+}
+
+IdentificationMascotDatFile::IdentificationMascotDatFile(const IdentificationMascotDatFile& other) : IdentificationDataSource(other),_mascot_dat_file (other._mascot_dat_file)
+{
+    _engine = IdentificationEngine::mascot;
+}
+
+IdentificationMascotDatFile::~IdentificationMascotDatFile()
+{
+
+}
+
+bool IdentificationMascotDatFile::operator==(const IdentificationMascotDatFile& other) const
+{
+
+}
+
+pappso::SpectrumSp IdentificationMascotDatFile::getSpectrumSp(unsigned int scan_number) const {
+    pappso::SpectrumSp spectrum_sp = IdentificationDataSource::getSpectrumSp(scan_number);
+    return spectrum_sp;
+}
+
+
+void IdentificationMascotDatFile::parseTo(Project* p_project) {
+    qDebug() << "IdentificationMascotDatFile::parseTo begin";
+
+    qDebug() << "Read Mascot dat result file '" << _mascot_dat_file.absoluteFilePath() << "'";
+
+    MsRunSp msrun_sp = p_project->getMsRunStore().getInstance(QFileInfo(_mascot_dat_file).baseName());
+    setMsRunSp(msrun_sp);
+    std::vector<IdentificationGroup *> identification_list = p_project->getIdentificationGroupList();
+    IdentificationGroup * identification_group_p = nullptr;
+    if (p_project->getProjectMode() == ProjectMode::combined) {
+        if (identification_list.size() == 0) {
+            identification_group_p = p_project->newIdentificationGroup();
+        }
+        else {
+            identification_group_p = identification_list[0];
+        }
+    }
+    else {
+        for (IdentificationGroup * identification_p_flist : identification_list) {
+            if (identification_p_flist->containSample(msrun_sp.get()->getSampleName())) {
+                identification_group_p = identification_p_flist;
+                break;
+            }
+        }
+        if (identification_group_p == nullptr) {
+            identification_group_p = p_project->newIdentificationGroup();
+        }
+    }
+
+    identification_group_p->addIdentificationDataSourceP(this);
+    MascotDatParser mascot_parser(p_project, identification_group_p, this);
+
+    QFile qfile(_mascot_dat_file.absoluteFilePath());
+    mascot_parser.parse(&qfile);
+    
+    qfile.close();
+
+    qDebug() << "IdentificationMascotDatFile::parseTo end";
+}

src/core/identification_sources/identificationmascotdatfile.h

+/**
+ * \file /core/identification_sources/identificationmascotdatfile.h
+ * \date 17/2/2018
+ * \author Olivier Langella
+ * \brief mascot dat identification file handler
+ */
+
+/*******************************************************************************
+* Copyright (c) 2018 Olivier Langella <olivier.langella@u-psud.fr>.
+*
+* This file is part of XTPcpp.
+*
+*     XTPcpp is free software: you can redistribute it and/or modify
+*     it under the terms of the GNU General Public License as published by
+*     the Free Software Foundation, either version 3 of the License, or
+*     (at your option) any later version.
+*
+*     XTPcpp is distributed in the hope that it will be useful,
+*     but WITHOUT ANY WARRANTY; without even the implied warranty of
+*     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*     GNU General Public License for more details.
+*
+*     You should have received a copy of the GNU General Public License
+*     along with XTPcpp.  If not, see <http://www.gnu.org/licenses/>.
+*
+* Contributors:
+*     Olivier Langella <olivier.langella@u-psud.fr> - initial API and implementation
+******************************************************************************/
+
+#ifndef IDENTIFICATIONMASCOTDATFILE_H
+#define IDENTIFICATIONMASCOTDATFILE_H
+
+#include "identificationdatasource.h"
+#include <QFileInfo>
+
+class IdentificationMascotDatFile: public IdentificationDataSource
+{
+public:
+    IdentificationMascotDatFile(const QFileInfo & mascot_dat_file);
+    IdentificationMascotDatFile(const IdentificationMascotDatFile& other);
+    ~IdentificationMascotDatFile();
+    bool operator==(const IdentificationMascotDatFile& other) const;
+
+    virtual pappso::SpectrumSp getSpectrumSp(unsigned int scan_number) const override;
+    virtual void parseTo(Project* p_project) override;
+    
+private:
+    const QFileInfo _mascot_dat_file;
+};
+
+#endif // IDENTIFICATIONMASCOTDATFILE_H

src/input/mascot/mascotdatparser.cpp

+/**
+ * \file /input/mascot/mascotdatparser.h
+ * \date 17/2/2018
+ * \author Olivier Langella
+ * \brief MASCOT dat file parser
+ */
+
+/*******************************************************************************
+* Copyright (c) 2018 Olivier Langella <olivier.langella@u-psud.fr>.
+*
+* This file is part of XTPcpp.
+*
+*     XTPcpp is free software: you can redistribute it and/or modify
+*     it under the terms of the GNU General Public License as published by
+*     the Free Software Foundation, either version 3 of the License, or
+*     (at your option) any later version.
+*
+*     XTPcpp is distributed in the hope that it will be useful,
+*     but WITHOUT ANY WARRANTY; without even the implied warranty of
+*     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*     GNU General Public License for more details.
+*
+*     You should have received a copy of the GNU General Public License
+*     along with XTPcpp.  If not, see <http://www.gnu.org/licenses/>.
+*
+* Contributors:
+*     Olivier Langella <olivier.langella@u-psud.fr> - initial API and implementation
+******************************************************************************/
+
+#include "mascotdatparser.h"
+#include "mimeparser.h"
+#include <QDebug>
+
+MascotDatParser::MascotDatParser(Project * p_project, IdentificationGroup * p_identification_group,
+                                 IdentificationDataSource * p_identification_data_source) {
+    _p_project = p_project;
+    _p_identification_group = p_identification_group;
+    _p_identification_data_source = p_identification_data_source;
+
+}
+MascotDatParser::~MascotDatParser() {
+}
+void MascotDatParser::parse(QIODevice * in_stream) {
+    qDebug() << "MascotDatParser::parse begin";
+    MimeParser mime_parser(in_stream);
+    mime_parser.open();
+
+    for(bool more=mime_parser.goToFirstFile(); more; more=mime_parser.goToNextFile()) {
+        qDebug() << "MascotDatParser::parse mimetype=" << mime_parser.getCurrentMimeType() << " filename=" << mime_parser.getCurrentFileName();
+
+        if (mime_parser.getCurrentFileName() == "proteins") {
+            while(!mime_parser.getCurrentTextStream().atEnd()) {
+                parseProteinLine( mime_parser.getCurrentTextStream().readLine());
+            }
+        }
+    }
+
+    mime_parser.close();
+    qDebug() << "MascotDatParser::parse end";
+}
+void MascotDatParser::parseProteinLine(const QString & protein_line) {
+    //02::"tr|A0A0D9SF80|A0A0D9SF80_HUMAN"=55120.88,"General transcription factor II-I repeat domain-containing protein 2A OS=Homo sapiens GN=GTF2IRD2B PE=4 SV=1"
+    QRegExp   regexp_protein("^(.*)::\"(.*)\"=([0-9]+\\.[0-9]+),\"(.*)\"$");
+    if (regexp_protein.exactMatch(protein_line)) {
+        QStringList protein_list = regexp_protein.capturedTexts();
+    }
+    else {
+        QRegExp   regexp_proteinb("^\"(.*)\"=([0-9]+\\.[0-9]+),\"(.*)\"$");
+        if (regexp_proteinb.exactMatch(protein_line)) {
+            QStringList protein_list = regexp_proteinb.capturedTexts();
+        }
+        else {
+            qDebug() << "MascotDatParser::parseProteinLine error " << protein_line;
+        }
+    }
+}

src/input/mascot/mascotdatparser.h

+/**
+ * \file /input/mascot/mascotdatparser.h
+ * \date 17/2/2018
+ * \author Olivier Langella
+ * \brief MASCOT dat file parser
+ */
+
+/*******************************************************************************
+* Copyright (c) 2018 Olivier Langella <olivier.langella@u-psud.fr>.
+*
+* This file is part of XTPcpp.
+*
+*     XTPcpp is free software: you can redistribute it and/or modify
+*     it under the terms of the GNU General Public License as published by
+*     the Free Software Foundation, either version 3 of the License, or
+*     (at your option) any later version.
+*
+*     XTPcpp is distributed in the hope that it will be useful,
+*     but WITHOUT ANY WARRANTY; without even the implied warranty of
+*     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*     GNU General Public License for more details.
+*
+*     You should have received a copy of the GNU General Public License
+*     along with XTPcpp.  If not, see <http://www.gnu.org/licenses/>.
+*
+* Contributors:
+*     Olivier Langella <olivier.langella@u-psud.fr> - initial API and implementation
+******************************************************************************/
+
+#ifndef MASCOTDATPARSER_H
+#define MASCOTDATPARSER_H
+
+#include "../../core/project.h"
+
+class MascotDatParser
+{
+public:
+    MascotDatParser(Project * p_project, IdentificationGroup * p_identification_group,
+                    IdentificationDataSource * p_identification_data_source);
+    virtual ~MascotDatParser();
+    void parse(QIODevice * in_stream);
+private:
+    void parseProteinLine(const QString & protein_line);
+private:
+    Project * _p_project;
+    IdentificationGroup * _p_identification_group;
+    IdentificationDataSource * _p_identification_data_source;
+};
+
+#endif // MASCOTDATPARSER_H

src/input/mascot/mimeparser.cpp

@@ -29,16 +29,34 @@
 
 #include "mimeparser.h"
 #include <QRegExp>
+#include <QDebug>
+#include <pappsomspp/pappsoexception.h>
 
 MimeParser::MimeParser(QIODevice * p_inputstream) {
     _p_inputstream = p_inputstream;
 }
+
+MimeParser::~MimeParser() {
+    close();
+}
 bool MimeParser::open() {
-    _p_inputstream->open(QIODevice::ReadOnly);
-    _real_in = new QTextStream(_p_inputstream);
+    qDebug() << "MimeParser::open begin";
+    if (_p_inputstream->open(QIODevice::ReadOnly)) {
+        _real_in = new QTextStream(_p_inputstream);
+    }
+    else {
+        throw pappso::PappsoException(QObject::tr("Error opening mime QIODevice"));
+    }
+    qDebug() << "MimeParser::open end";
 
 }
 bool MimeParser::close() {
+    qDebug() << "MimeParser::close begin";
+
+    _priv_file_string = "";
+    if (_p_current_file_content != nullptr) {
+        delete _p_current_file_content;
+    }
     if (_real_in !=nullptr) {
         delete _real_in;
         _real_in =nullptr;
@@ -49,8 +67,10 @@ bool MimeParser::close() {
         delete _p_inputstream;
         _p_inputstream =nullptr;
     }
+    qDebug() << "MimeParser::close end";
 }
 bool MimeParser::goToFirstFile() {
+    qDebug() << "MimeParser::goToFirstFile begin";
 
     /*
     MIME-Version: 1.0 (Generated by Mascot version 1.0)
@@ -63,18 +83,71 @@ bool MimeParser::goToFirstFile() {
 
     while(!_real_in->atEnd()) {
         QString line = _real_in->readLine();
+        qDebug() << "MimeParser::goToFirstFile line=" << line;
         if (regexp_boundary.exactMatch(line)) {
             QStringList boundary_list = regexp_boundary.capturedTexts();
-            _boundary = boundary_list[1];
+            _boundary = QString("--%1").arg(boundary_list[1]);
             break;
         }
     }
 
+    if (_boundary.isEmpty()) {
+        qDebug() << "MimeParser::goToFirstFile _boundary.isEmpty";
+        return false;
+    }
+
+    while(!_real_in->atEnd()) {
+        QString line = _real_in->readLine();
+        if (line == _boundary) break;
+    }
+    qDebug() << "MimeParser::goToFirstFile end";
+    return readFile();
+}
+bool MimeParser::readFile() {
+    qDebug() << "MimeParser::readFile begin";
+    _current_mime_type = "";
+    _current_file_name = "";
+    _priv_file_string = "";
+
+    if (_p_current_file_content != nullptr) {
+        delete _p_current_file_content;
+    }
+    _p_current_file_content = new QTextStream(&_priv_file_string);
+
+    QString line = _real_in->readLine();
+    if (line.isEmpty()) return false;
+    //Content-Type: application/x-Mascot; name="masses"
+    QRegExp   regexp_mimetype("^Content-Type:\\s(.*);\\sname=\"(.*)\"$");
+    if (regexp_mimetype.exactMatch(line)) {
+        QStringList mimetype_list = regexp_mimetype.capturedTexts();
+        _current_mime_type = mimetype_list[1];
+        _current_file_name = mimetype_list[2];
+    }
+    else {
+        return false;
+    }
+    _real_in->readLine();
+    while(!_real_in->atEnd()) {
+        QString line = _real_in->readLine();
+        if (line == _boundary) break;
+        *_p_current_file_content << line << endl;
+    }
+    qDebug() << "MimeParser::readFile end";
+    return true;
 }
 bool MimeParser::goToNextFile() {
+    return readFile();
 }
-const QString & MimeParser::getCurrentFileName() {
+const QString & MimeParser::getCurrentMimeType() const {
+    return _current_mime_type;
+}
+const QString & MimeParser::getCurrentFileName() const {
     return _current_file_name;
 }
 QTextStream & MimeParser::getCurrentTextStream() {
+    if (_p_current_file_content == nullptr) {
+        throw pappso::PappsoException(QObject::tr("Error current file content empty"));
+    }
+    _p_current_file_content->reset();
+    return *_p_current_file_content;
 }

src/input/mascot/mimeparser.h

@@ -37,18 +37,26 @@ class MimeParser
 {
 public:
     MimeParser(QIODevice * p_inputstream);
+    virtual ~MimeParser();
     bool open();
     bool close();
     bool goToFirstFile();
     bool goToNextFile();
-    const QString & getCurrentFileName();
+    const QString & getCurrentMimeType() const;
+    const QString & getCurrentFileName() const;
     QTextStream & getCurrentTextStream();
     
+private:
+    bool readFile();
+    
 private:
     QString _current_file_name;
+    QString _current_mime_type;
     QIODevice * _p_inputstream = nullptr;
     QTextStream * _real_in = nullptr;
     QString _boundary;
+    QTextStream * _p_current_file_content = nullptr;
+    QString _priv_file_string;
 };
 
 #endif // MIMEPARSER_H

src/utils/identificationdatasourcestore.cpp

@@ -30,6 +30,7 @@
 
 #include "identificationdatasourcestore.h"
 #include "../core/identification_sources/identificationxtandemfile.h"
+#include "../core/identification_sources/identificationmascotdatfile.h"
 #include "../core/identification_sources/identificationpwizfile.h"
 #include <pappsomspp/pappsoexception.h>
 #include <pappsomspp/utils.h>
@@ -98,6 +99,9 @@ IdentificationDataSourceSp IdentificationDataSourceStore::getInstance(const QStr
         } else if (ext.toLower() == "pep") {
             //pep xml file
             p_identfile = std::make_shared<IdentificationPwizFile>(location_file);
+        } else if (ext.toLower() == "dat") {
+            //MASCOT dat file
+            p_identfile = std::make_shared<IdentificationMascotDatFile>(location_file);
         } else {
             p_identfile = std::make_shared<IdentificationPwizFile>(location_file);
         }