libpappsomspp
Library for mass spectrometry
msrunreader.h
Go to the documentation of this file.
1/**
2 * \file pappsomspp/msrun/msrunreader.h
3 * \date 29/05/2018
4 * \author Olivier Langella
5 * \brief base interface to read MSrun files
6 */
7
8/*******************************************************************************
9 * Copyright (c) 2018 Olivier Langella <Olivier.Langella@u-psud.fr>.
10 *
11 * This file is part of the PAPPSOms++ library.
12 *
13 * PAPPSOms++ is free software: you can redistribute it and/or modify
14 * it under the terms of the GNU General Public License as published by
15 * the Free Software Foundation, either version 3 of the License, or
16 * (at your option) any later version.
17 *
18 * PAPPSOms++ is distributed in the hope that it will be useful,
19 * but WITHOUT ANY WARRANTY; without even the implied warranty of
20 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
21 * GNU General Public License for more details.
22 *
23 * You should have received a copy of the GNU General Public License
24 * along with PAPPSOms++. If not, see <http://www.gnu.org/licenses/>.
25 *
26 ******************************************************************************/
27
28#pragma once
29
30
31/////////////////////// StdLib includes
32#include <memory>
33#include <map>
34
35
36/////////////////////// Qt includes
37#include <QMutex>
38
39
40/////////////////////// pappsomspp includes
41#include "../trace/maptrace.h"
42
43/////////////////////// Local includes
44#include "msrunid.h"
45#include "../massspectrum/qualifiedmassspectrum.h"
46#include "../msfile/msfilereader.h"
47#include "../exportinmportconfig.h"
48#include "xiccoord/xiccoord.h"
49
50namespace pappso
51{
52
53/** @brief interface to collect spectrums from the MsRunReader class
54 */
56{
57 public:
58 virtual void
60
61 /** @brief tells if we need the peak list (if we want the binary data) for
62 * each spectrum
63 */
64 virtual bool needPeakList() const = 0;
65
66 /** @brief tells if we need the peak list (if we want the binary data) for
67 * each spectrum, given an MS level
68 */
69 virtual bool needMsLevelPeakList(unsigned int ms_level) const final;
70
71 /** @brief tells if we need the peak list given
72 */
73 virtual void setNeedMsLevelPeakList(unsigned int ms_level,
74 bool want_peak_list) final;
75 virtual bool shouldStop();
76 virtual void loadingEnded();
77 virtual void spectrumListHasSize(std::size_t size);
78
79
80 /** @brief use threads to read a spectrum by batch of batch_size
81 * @param is_read_ahead boolean to use threads or not
82 */
83 virtual void setReadAhead(bool is_read_ahead) final;
84
85 /** @brief tells if we want to read ahead spectrum
86 */
87 virtual bool isReadAhead() const;
88
89 private:
90 bool m_isReadAhead = false;
91
92 std::vector<bool> m_needPeakListByMsLevel = {true,
93 true,
94 true,
95 true,
96 true,
97 true,
98 true,
99 true,
100 true,
101 true,
102 true,
103 true,
104 true,
105 true,
106 true};
107};
108
109
110/** @brief example of interface to count MS levels of all spectrum in an MSrun
111 */
114{
115 private:
116 std::vector<unsigned long> m_countMsLevelSpectrum;
117
118 public:
119 virtual void
120 setQualifiedMassSpectrum(const QualifiedMassSpectrum &spectrum) override;
121 virtual bool needPeakList() const override;
122 virtual void loadingEnded() override;
123
124 unsigned long getMsLevelCount(unsigned int ms_level) const;
125
126 unsigned long getTotalCount() const;
127};
128
129/** @brief provides a multimap to find quickly spectrum index from scan number
130 */
133{
134 private:
135 std::multimap<std::size_t, std::size_t> m_mmap_scan2index;
136
137 public:
140 virtual void
141 setQualifiedMassSpectrum(const QualifiedMassSpectrum &spectrum) override;
142 virtual bool needPeakList() const override;
143
144 std::size_t getSpectrumIndexFromScanNumber(std::size_t scan_number) const;
145};
146
147
148/** @brief collect retention times along MS run */
151{
152 private:
153 std::vector<double> m_retention_time_list;
154
155 public:
158 virtual void
159 setQualifiedMassSpectrum(const QualifiedMassSpectrum &spectrum) override;
160 virtual bool needPeakList() const override;
161
162 const std::vector<double> &getRetentionTimeLine() const;
163};
164
165
166/** @brief calculate a TIC chromatogram */
169{
170 public:
173 virtual void setQualifiedMassSpectrum(
174 const QualifiedMassSpectrum &qualified_mass_spectrum) override;
175 virtual bool needPeakList() const override;
176
177 Trace getTicChromatogram() const;
178
179 private:
181};
182
183
185typedef std::shared_ptr<MsRunReader> MsRunReaderSPtr;
186typedef std::shared_ptr<const MsRunReader> MsRunReaderCstSPtr;
187
188/** @brief base class to read MSrun
189 * the only way to build a MsRunReader object is to use the MsRunReaderFactory
190 */
192{
193
194 friend class MsFileAccessor;
195
196 public:
197 MsRunReader(MsRunIdCstSPtr &ms_run_id);
198 MsRunReader(const MsRunReader &other);
199 virtual ~MsRunReader();
200
201 const MsRunIdCstSPtr &getMsRunId() const;
202
203 /** @brief get a MassSpectrumSPtr class given its spectrum index
204 */
205 virtual MassSpectrumSPtr massSpectrumSPtr(std::size_t spectrum_index) = 0;
206 virtual MassSpectrumCstSPtr
207 massSpectrumCstSPtr(std::size_t spectrum_index) = 0;
208
209 /** @brief get a QualifiedMassSpectrum class given its scan number
210 */
212 qualifiedMassSpectrum(std::size_t spectrum_index,
213 bool want_binary_data = true) const = 0;
214
215
216 /** @brief get a xic coordinate object from a given spectrum index
217 */
218 virtual XicCoordSPtr
219 newXicCoordSPtrFromSpectrumIndex(std::size_t spectrum_index,
220 PrecisionPtr precision) const = 0;
221
222 /** @brief get a xic coordinate object from a given spectrum
223 */
225 const QualifiedMassSpectrum &mass_spectrum,
226 PrecisionPtr precision) const = 0;
227
228 /** @brief get the totat number of spectrum conained in the MSrun data file
229 */
230 virtual std::size_t spectrumListSize() const = 0;
231
232 /** @brief function to visit an MsRunReader and get each Spectrum in a
233 * spectrum collection handler
234 */
235 virtual void
237
238
239 /** @brief function to visit an MsRunReader and get each Spectrum in a
240 * spectrum collection handler by Ms Levels
241 */
242 virtual void
244 unsigned int ms_level) = 0;
245
246
247 /** @brief if possible, converts a scan number into a spectrum index
248 * This is a convenient function to help transition from the old scan number
249 * (not implemented by all vendors) to more secure spectrum index (not vendor
250 * dependant).
251 * It is better to not rely on this function.
252 */
253 virtual std::size_t scanNumber2SpectrumIndex(std::size_t scan_number);
254
255 /** @brief tells if spectra can be accessed using scan numbers
256 * by default, it returns false. Only overrided functions can check if scan
257 * numbers are available in the current file
258 */
259 virtual bool hasScanNumbers() const;
260
261
262 /** @brief release data back end device
263 * if a the data back end is released, the developper has to use acquireDevice
264 * before using the msrunreader object
265 * @return bool true if done
266 */
267 virtual bool releaseDevice() = 0;
268
269 /** @brief acquire data back end device
270 * @return bool true if done
271 */
272 virtual bool acquireDevice() = 0;
273
274 /** @brief retention timeline
275 * get retention times along the MSrun in seconds
276 * @return vector of retention times (seconds)
277 */
278 virtual std::vector<double> getRetentionTimeLine();
279
280 /** @brief get a TIC chromatogram
281 *
282 * for each retention time, computes the sum of all intensities.
283 * For IM-MS, combines the mobility spectra
284 *
285 * Note that, formally, a TIC chromatogram is computed only for MS1 spectra.
286 *
287 * @return a trace (x=rt, y=intensities)
288 */
289 virtual Trace getTicChromatogram();
290
291
292 /** @brief set only one is_mono_thread to true
293 *
294 * this avoid to use qtconcurrent
295 */
296 void setMonoThread(bool is_mono_thread);
297
298 bool isMonoThread() const;
299
300 protected:
302 MsRunReaderScanNumberMultiMap *mpa_multiMapScanNumber = nullptr;
303
304 virtual void initialize() = 0;
305
306 /** @brief tells if the reader is able to handle this file
307 * must be implemented by private MS run reader, specific of one or more file
308 * format
309 */
310 virtual bool accept(const QString &file_name) const = 0;
311
312 private:
313 bool m_isMonoThread = false;
314};
315
316
317} // namespace pappso
318
collect retention times along MS run
Definition: msrunreader.h:151
std::vector< double > m_retention_time_list
Definition: msrunreader.h:153
provides a multimap to find quickly spectrum index from scan number
Definition: msrunreader.h:133
std::multimap< std::size_t, std::size_t > m_mmap_scan2index
Definition: msrunreader.h:135
calculate a TIC chromatogram
Definition: msrunreader.h:169
base class to read MSrun the only way to build a MsRunReader object is to use the MsRunReaderFactory
Definition: msrunreader.h:192
virtual MassSpectrumCstSPtr massSpectrumCstSPtr(std::size_t spectrum_index)=0
virtual std::size_t spectrumListSize() const =0
get the totat number of spectrum conained in the MSrun data file
MsRunIdCstSPtr mcsp_msRunId
Definition: msrunreader.h:301
virtual bool acquireDevice()=0
acquire data back end device
virtual bool accept(const QString &file_name) const =0
tells if the reader is able to handle this file must be implemented by private MS run reader,...
virtual XicCoordSPtr newXicCoordSPtrFromSpectrumIndex(std::size_t spectrum_index, PrecisionPtr precision) const =0
get a xic coordinate object from a given spectrum index
virtual bool releaseDevice()=0
release data back end device if a the data back end is released, the developper has to use acquireDev...
virtual void initialize()=0
virtual XicCoordSPtr newXicCoordSPtrFromQualifiedMassSpectrum(const QualifiedMassSpectrum &mass_spectrum, PrecisionPtr precision) const =0
get a xic coordinate object from a given spectrum
virtual MassSpectrumSPtr massSpectrumSPtr(std::size_t spectrum_index)=0
get a MassSpectrumSPtr class given its spectrum index
virtual void readSpectrumCollection(SpectrumCollectionHandlerInterface &handler)=0
function to visit an MsRunReader and get each Spectrum in a spectrum collection handler
virtual QualifiedMassSpectrum qualifiedMassSpectrum(std::size_t spectrum_index, bool want_binary_data=true) const =0
get a QualifiedMassSpectrum class given its scan number
virtual void readSpectrumCollectionByMsLevel(SpectrumCollectionHandlerInterface &handler, unsigned int ms_level)=0
function to visit an MsRunReader and get each Spectrum in a spectrum collection handler by Ms Levels
example of interface to count MS levels of all spectrum in an MSrun
Definition: msrunreader.h:114
std::vector< unsigned long > m_countMsLevelSpectrum
Definition: msrunreader.h:116
Class representing a fully specified mass spectrum.
interface to collect spectrums from the MsRunReader class
Definition: msrunreader.h:56
virtual bool needPeakList() const =0
tells if we need the peak list (if we want the binary data) for each spectrum
virtual void setQualifiedMassSpectrum(const QualifiedMassSpectrum &spectrum)=0
A simple container of DataPoint instances.
Definition: trace.h:148
#define PMSPP_LIB_DECL
Q_DECLARE_METATYPE(pappso::MsRunReaderSPtr)
int msRunReaderSPtrMetaTypeId
Definition: msrunreader.cpp:35
tries to keep as much as possible monoisotopes, removing any possible C13 peaks and changes multichar...
Definition: aa.cpp:39
std::shared_ptr< MsRunReader > MsRunReaderSPtr
Definition: msrunreader.h:185
std::shared_ptr< const MsRunReader > MsRunReaderCstSPtr
Definition: msrunreader.h:186
class PMSPP_LIB_DECL MsRunReader
Definition: msrunreader.h:184
std::shared_ptr< const MsRunId > MsRunIdCstSPtr
Definition: msrunid.h:46
std::shared_ptr< const MassSpectrum > MassSpectrumCstSPtr
Definition: massspectrum.h:55
std::shared_ptr< MassSpectrum > MassSpectrumSPtr
Definition: massspectrum.h:54
std::shared_ptr< XicCoord > XicCoordSPtr
Definition: xiccoord.h:43
XIC coordinate in MSrun.