tandemparameters.cpp

/**
 * \file /core/tandem_run/tandemparameters.cpp
 * \date 19/9/2017
 * \author Olivier Langella
 * \brief handles X!Tandem parameters
 */

/*******************************************************************************
 * Copyright (c) 2017 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 "tandemparameters.h"
#include <QDebug>
#include <pappsomspp/exception/exceptionnotfound.h>
#include <pappsomspp/exception/exceptionnotpossible.h>

/*
 *
specific parameters for the modified TPP tandem version :
http://tools.proteomecenter.org/wiki/index.php?title=TPP:X%21Tandem_and_the_TPP

<note label="scoring, algorithm" type="input">k-score</note>
<note label="spectrum, use conditioning" type="input">no</note>
<note label="scoring, minimum ion count" type="input">1</note>
new labels :
"protein, taxon" https://www.thegpm.org/TANDEM/api/pt.html
"protein, homolog management"
"refine, sequence path" https://www.thegpm.org/TANDEM/api/rsp.html
 "refine, tic percent" https://www.thegpm.org/TANDEM/api/rtp.html
 "output, log path" https://www.thegpm.org/TANDEM/api/olp.html
 "output, message" https://www.thegpm.org/TANDEM/api/om.html
  "output, sequence path" https://www.thegpm.org/TANDEM/api/osp.html
   "output, path" https://www.thegpm.org/TANDEM/api/opath.html
  "protein, use annotations" https://www.thegpm.org/TANDEM/api/pua.html


   obsolete, not documented :
  "refine, maximum missed cleavage sites" 3
   "protein, use minimal annotations" yes no
   "scoring, pluggable scoring" yes no
*/

TandemParameters::TandemParameters()
{
}
TandemParameters::~TandemParameters()
{
}

TandemParameters::TandemParameters(const TandemParameters &other)
{
    qDebug() << "TandemParameters::TandemParameters copy begin";
    //_map_label_value = other._map_label_value;
    _method_name = other._method_name;

    QMap<QString, QString>::const_iterator i =
        other._map_label_value.constBegin();
    while(i != other._map_label_value.constEnd())
    {
        qDebug() << "TandemParameters::TandemParameters copy " << i.key() << " "
                 << i.value();
        _map_label_value.insert(i.key(), i.value());
        i++; // added
    }
    qDebug() << "TandemParameters::TandemParameters copy end";
}


bool
TandemParameters::equals(const TandemParameters &other) const
{
    if(_method_name != other._method_name)
    {
        return false;
    }
    if(_map_label_value.size() != other._map_label_value.size())
    {
        return false;
    }
    QMap<QString, QString>::const_iterator i = _map_label_value.constBegin();
    while(i != _map_label_value.constEnd())
    {
        if(i.value() != other.getValue(i.key()))
        {
            return false;
        }
        i++; // added
    }
    return true;
}

const QString
TandemParameters::getLabelCategory(const QString &value) const
{

    qDebug() << __FILE__ << " " << __FUNCTION__ << " " << __LINE__ << " "
             << value;
    return value.split(", ").at(0);
}
const QString &
TandemParameters::getValue(const QString &label) const
{
    qDebug() << __FILE__ << " " << __FUNCTION__ << " " << __LINE__ << " "
             << label;
    QMap<QString, QString>::const_iterator it = _map_label_value.constFind(label);
    if(it == _map_label_value.end())
    {
        throw pappso::ExceptionNotFound(
            QObject::tr("X!Tandem preset label \"%1\" not found in method \"%2\"")
            .arg(label)
            .arg(_method_name));
    }
    qDebug() << __FILE__ << " " << __FUNCTION__ << " " << __LINE__;
    return it.value();
}
void
TandemParameters::setParamLabelValue(const QString &label, const QString &value)
{
    qDebug() << __FILE__ << " " << __FUNCTION__ << " " << __LINE__ << " "
             << label;
    // list path
    //<note type="input" label="list path, default
    // parameters">/gorgone/pappso/tmp/temp_condor_job24872841484824316495370334631825647/QExactive_analysis_FDR_nosemi.xml</note>
    //<note type="input" label="list path, taxonomy
    // information">/gorgone/pappso/tmp/temp_condor_job24872841484824316495370334631825647/database.xml</note>
    if(getLabelCategory(label) == "list path")
    {
        return;
    }

    qDebug() << __FILE__ << " " << __FUNCTION__ << " " << __LINE__ << " "
             << label;
    if((label == "spectrum, parent monoisotopic mass error units") ||
            (label == "spectrum, fragment monoisotopic mass error units"))
    {
        if((value != "Da") && (value != "Daltons") && (value != "ppm"))
        {
            throw pappso::ExceptionNotPossible(
                QObject::tr("\"%1\" value must be Daltons or ppm (not \"%2\")")
                .arg(label)
                .arg(value));
        }
    }
    else if((label == "spectrum, parent monoisotopic mass isotope error") ||
            (label == "spectrum, use neutral loss window") ||
            (label == "spectrum, use contrast angle") ||
            (label == "protein, cleavage semi") ||
            (label == "protein, quick acetyl") ||
            (label == "protein, quick pyrolidone") ||
            (label == "protein, stP bias") ||
            (label == "scoring, cyclic permutation") ||
            (label == "scoring, include reverse") ||
            (label == "scoring, y ions") || (label == "scoring, b ions") ||
            (label == "scoring, c ions") || (label == "scoring, z ions") ||
            (label == "scoring, a ions") || (label == "scoring, x ions") ||
            (label == "refine") ||
            (label ==
             "refine, use potential modifications for full refinement") ||
            (label == "refine, cleavage semi") ||
            (label == "refine, unanticipated cleavage") ||
            (label == "refine, spectrum synthesis") ||
            (label == "refine, point mutations") ||
            (label == "output, spectra") || (label == "output, proteins") ||
            (label == "output, sequences") ||
            (label == "output, one sequence copy") ||
            (label == "output, parameters") || (label == "output, performance") ||
            (label == "output, histograms") ||
            (label == "output, path hashing") ||
            (label == "spectrum, use noise suppression") ||
            (label == "spectrum, use conditioning") ||
            (label == "protein, homolog management") ||
            (label == "protein, use minimal annotations") ||
            (label == "protein, use annotations") ||
            (label == "scoring, pluggable scoring"))
    {
        if((value != "yes") && (value != "no"))
        {
            throw pappso::ExceptionNotPossible(
                QObject::tr("%1 value must be yes or no (not %2)")
                .arg(label)
                .arg(value));
        }
    }

    else if((label == "spectrum, parent monoisotopic mass error minus") ||
            (label == "spectrum, parent monoisotopic mass error plus") ||
            (label == "spectrum, dynamic range") ||
            (label == "spectrum, minimum parent m+h") ||
            (label == "protein, cleavage C-terminal mass change") ||
            (label == "protein, cleavage N-terminal mass change") ||
            (label == "protein, C-terminal residue modification mass") ||
            (label == "protein, N-terminal residue modification mass") ||
            (label == "refine, maximum valid expectation value") ||
            (label == "spectrum, fragment monoisotopic mass error") ||
            (label == "spectrum, neutral loss mass") ||
            (label == "spectrum, neutral loss window") ||
            (label == "spectrum, minimum fragment mz") ||
            (label == "output, maximum valid expectation value") ||
            (label == "output, maximum valid protein expectation value") ||
            (label == "refine, tic percent")


           )

    {

        bool ok;
        QString value_bis = value.simplified();
        if(value_bis.isEmpty())
        {
            value_bis = "0.0";
        }
        value_bis.toDouble(&ok);
        if(!ok)
        {   // not an integer
            throw pappso::ExceptionNotPossible(
                QObject::tr("\"%1\" value must be a number (not \"%2\")")
                .arg(label)
                .arg(value));
        }
    }
    else if((label == "protein, cleavage site") ||
            (label == "protein, modified residue mass file") ||
            (label == "residue, modification mass") ||
            (label == "residue, modification mass 1") ||
            (label == "residue, modification mass 2") ||
            (label == "residue, potential modification mass") ||
            (label == "residue, potential modification motif") ||
            (label == "refine, potential N-terminus modifications") ||
            (label == "refine, potential C-terminus modifications") ||
            (label == "refine, modification mass") ||
            (label == "refine, modification mass 1") ||
            (label == "refine, modification mass 2") ||
            (label == "refine, potential modification mass") ||
            (label == "refine, potential modification mass 1") ||
            (label == "refine, potential modification mass 2") ||
            (label == "refine, potential modification motif") ||
            (label == "refine, potential modification motif 1") ||
            (label == "refine, potential modification motif 2") ||
            (label == "output, sort results by") ||
            (label == "output, xsl path") || (label == "spectrum, path") ||
            (label == "scoring, algorithm") || (label == "protein, taxon") ||
            (label == "refine, sequence path") || (label == "output, log path") ||
            (label == "output, message") || (label == "output, sequence path") ||
            (label == "output, path"))
    {
        // string
    }
    else if((label == "spectrum, maximum parent charge") ||
            (label == "spectrum, total peaks") ||
            (label == "spectrum, minimum peaks") ||
            (label == "spectrum, sequence batch size") ||
            (label == "spectrum, threads") ||
            (label == "scoring, minimum ion count") ||
            (label == "scoring, maximum missed cleavage sites") ||
            (label == "output, histogram column width") ||
            (label == "refine, maximum missed cleavage sites"))
    {

        bool ok;
        value.toInt(&ok);
        if(!ok)
        {   // not an integer
            throw pappso::ExceptionNotPossible(
                QObject::tr("\"%1\" value must be an integer (not \"%2\")")
                .arg(label)
                .arg(value));
        }
    }
    else if(label == "spectrum, fragment mass type")
    {
        if((value != "monoisotopic") && (value != "average"))
        {
            throw pappso::ExceptionNotPossible(
                QObject::tr(
                    "\"%1\" value must be monoisotopic or average (not \"%2\")")
                .arg(label)
                .arg(value));
        }
    }
    // tandem input parameter label "output, results" for value "valid" not known.

    else if(label == "output, results")
    {
        /*
        if((value != "all") && (value != "valid"))
        {
          throw pappso::ExceptionNotPossible(
            QObject::tr("\"%1\" value must be valid or all (not \"%2\")")
              .arg(label)
              .arg(value));
        }
        */
    }
    else
    {

        throw pappso::ExceptionNotPossible(
            QObject::tr(
                "tandem input parameter label \"%1\" for value \"%2\" not known.")
            .arg(label)
            .arg(value));
    }


    // text "protein, cleavage site" "protein, modified residue mass file"
    // "residue, modification mass"  || (label == "refine, potential N-terminus
    // modifications") || (label == "refine, potential C-terminus modifications")
    // spectrum, neutral loss mass
    // spectrum, neutral loss window
    qDebug() << __FILE__ << " " << __FUNCTION__ << " " << __LINE__ << " "
             << label;
    _map_label_value.insert(label, value);
    qDebug() << __FILE__ << " " << __FUNCTION__ << " " << __LINE__ << " "
             << label;
}
void
TandemParameters::setMethodName(const QString &method)
{
    _method_name = method;
}
const QString &
TandemParameters::getMethodName() const
{
    return _method_name;
}
const QMap<QString, QString> &
TandemParameters::getMapLabelValue() const
{
    return _map_label_value;
}