2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2010,2011,2012,2013,2014, by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
37 * Declares gmx::AnalysisData and gmx::AnalysisDataHandle.
39 * \author Teemu Murtola <teemu.murtola@gmail.com>
41 * \ingroup module_analysisdata
43 #ifndef GMX_ANALYSISDATA_ANALYSISDATA_H
44 #define GMX_ANALYSISDATA_ANALYSISDATA_H
46 #include "../utility/real.h"
48 #include "abstractdata.h"
53 class AnalysisDataHandle;
54 class AnalysisDataParallelOptions;
57 * Parallelizable data container for raw data.
59 * This is the main class used to implement parallelizable data processing in
60 * analysis tools. It is used by first creating an object and setting its
61 * properties using setDataSetCount(), setColumnCount() and setMultipoint(),
62 * and attaching necessary modules using addModule() etc. Then one or more
63 * AnalysisDataHandle objects can be created using startData(). Each data
64 * handle can then be independently used to provide data frames (each frame
65 * must be provided by a single handle, but different frames can be freely
66 * mixed between the handles). When all data has been provided, the handles
67 * are destroyed using finishData() (or AnalysisDataHandle::finishData()).
68 * The AnalysisData object takes care of internally sorting the frames and
69 * passing them to the attached modules in the order in which the modules
73 * Parallel implementation is not complete.
76 * Special note for MPI implementation: assuming that the initialization of
77 * data objects is identical in all processes, associating the data objects
78 * in different MPI processes should be possible without changes in the
80 * Alternative, more robust implementation could get a unique ID as parameter
81 * to the constructor or a separate function, but would require all tools to
82 * provide it. With the current registration mechanism in
83 * TrajectoryAnalysisModule, this should be straightforward.
87 * \ingroup module_analysisdata
89 class AnalysisData : public AbstractAnalysisData
93 * Creates an empty analysis data object.
95 * \throws std::bad_alloc if out of memory.
98 virtual ~AnalysisData();
101 * Sets the number of data sets.
103 * \param[in] dataSetCount Number of data sets (must be > 0).
104 * \throws std::bad_alloc if out of memory.
105 * \throws APIError if modules have been added that are not
106 * compatible with the new data set count.
108 * Must not be called after startData() has been called.
109 * If not called, a single data set is assumed.
110 * If called multiple times, the last call takes effect.
112 void setDataSetCount(int dataSetCount);
114 * Sets the number of columns in a data set.
116 * \param[in] dataSet Zero-based data set index.
117 * \param[in] columnCount Number of columns in the data (must be > 0).
118 * \throws APIError if modules have been added that are not
119 * compatible with the new column count.
121 * Must be called before startData() for each data set.
122 * Must not be called after startData() has been called.
123 * If called multiple times for a data set, the last call takes effect.
125 void setColumnCount(int dataSet, int columnCount);
127 * Sets whether the data contains multiple points per column per frame.
129 * \param[in] bMultipoint Whether the data will allow multiple points
130 * per column within a single frame.
131 * \throws APIError if modules have been added that are not
132 * compatible with the new setting.
134 * If this method is not called, the data is not multipoint.
136 * Must not be called after startData() has been called.
138 * \see isMultipoint()
140 void setMultipoint(bool bMultipoint);
142 virtual int frameCount() const;
145 * Create a handle for adding data.
147 * \param[in] opt Options for setting how this handle will be
149 * \returns The created handle.
150 * \throws std::bad_alloc if out of memory.
151 * \throws APIError if any attached data module is not compatible.
152 * \throws unspecified Any exception thrown by attached data modules
153 * in AnalysisDataModuleInterface::dataStarted().
155 * The caller should retain the returned handle (or a copy of it), and
156 * pass it to finishData() after successfully adding all data.
157 * The caller should discard the returned handle if an error occurs;
158 * memory allocated for the handle will be freed when the AnalysisData
159 * object is destroyed.
161 * The \p opt options should be the same for all calls to this method,
162 * and the number of calls should match the parallelization factor
165 AnalysisDataHandle startData(const AnalysisDataParallelOptions &opt);
167 * Destroy a handle after all data has been added.
169 * \param[in] handle Handle to destroy.
170 * \throws unspecified Any exception thrown by attached data modules
171 * in AnalysisDataModuleInterface::dataFinished().
173 * \p handle must have been obtained from startData() of this object.
174 * The order of the calls with respect to the corresponding startData()
175 * calls is not important.
177 * The \p handle (and any copies) are invalid after the call.
179 void finishData(AnalysisDataHandle handle);
182 virtual AnalysisDataFrameRef tryGetDataFrameInternal(int index) const;
183 virtual bool requestStorageInternal(int nframes);
187 PrivateImplPointer<Impl> impl_;
189 friend class AnalysisDataHandle;
194 class AnalysisDataHandleImpl;
195 } // namespace internal
198 * Handle for inserting data into AnalysisData.
200 * This class provides an interface for adding data frames into an AnalysisData
201 * object. After a handle is obtained from AnalysisData::startData(), new
202 * frames can be added using startFrame(). Then values for that frame are set
203 * using provided methods (see below), and finishFrame() is called. After all
204 * frames have been added, finishData() (or AnalysisData::finishData()) must be
207 * For simple (non-multipoint) data, within a frame values can be set using
208 * selectDataSet(), setPoint() and setPoints(). Setting the same column in the
209 * same data set multiple times overrides previously set values.
210 * When the frame is finished, attached modules are notified.
212 * Multipoint data works otherwise similarly, but requires finishPointSet() to
213 * be called for each set of points for which the modules need to be notified.
214 * Each point set starts empty (after startFrame() or finishPointSet()), and
215 * values can be set using setPoint()/setPoints().
216 * A single point set can contain values only for a single data set, which must
217 * be selected with selectDataSet() before setting any values.
218 * finishPointSet() must also be called for the last point set just before
221 * This class works like a pointer type: copying and assignment is lightweight,
222 * and all copies work interchangeably, accessing the same internal handle.
223 * However, normally you should only keep one copy of a handle, i.e., treat
224 * this type as movable.
225 * Several handles created from the same AnalysisData object can exist
226 * concurrently, but must currently operate on separate frames.
229 * \ingroup module_analysisdata
231 class AnalysisDataHandle
235 * Constructs an invalid data handle.
237 * This constructor is provided for convenience in cases where it is
238 * easiest to declare an AnalysisDataHandle without immediately
239 * assigning a value to it. Any attempt to call methods without first
240 * assigning a value from AnalysisData::startData() to the handle
245 AnalysisDataHandle();
247 //! Returns whether this data handle is valid.
248 bool isValid() const { return impl_ != NULL; }
251 * Start data for a new frame.
253 * \param[in] index Zero-based index for the frame to start.
254 * \param[in] x x value for the frame.
255 * \param[in] dx Error in x for the frame if applicable.
257 * \throws unspecified Any exception thrown by attached data
258 * modules in AnalysisDataModuleInterface::frameStarted().
260 * Each \p index value 0, 1, ..., N (where N is the total number of
261 * frames) should be started exactly once by exactly one handle of an
262 * AnalysisData object. The frames may be started out of order, but
263 * currently the implementation places some limitations on how far
264 * the index can be in the future (as counted from the first frame that
267 void startFrame(int index, real x, real dx = 0.0);
269 * Selects a data set for subsequent setPoint()/setPoints() calls.
271 * \param[in] index Zero-based data set index.
273 * After startFrame(), the first data set is always selected.
274 * The set value is remembered until the end of the current frame, also
275 * across finishPointSet() calls.
279 void selectDataSet(int index);
281 * Set a value for a single column for the current frame.
283 * \param[in] column Zero-based column index.
284 * \param[in] value Value to set for the column.
285 * \param[in] bPresent Present flag to set for the column.
287 * If called multiple times for a column (within one point set for
288 * multipoint data), old values are overwritten.
292 void setPoint(int column, real value, bool bPresent = true);
294 * Set a value and its error estimate for a single column for the
297 * \param[in] column Zero-based column index.
298 * \param[in] value Value to set for the column.
299 * \param[in] error Error estimate to set for the column.
300 * \param[in] bPresent Present flag to set for the column.
302 * If called multiple times for a column (within one point set for
303 * multipoint data), old values are overwritten.
307 void setPoint(int column, real value, real error, bool bPresent = true);
309 * Set values for consecutive columns for the current frame.
311 * \param[in] firstColumn Zero-based column index.
312 * \param[in] count Number of columns to set.
313 * \param[in] values Value array of \p column items.
315 * Equivalent to calling setPoint(firstColumn + i, values[i]) for
320 void setPoints(int firstColumn, int count, const real *values);
322 * Finish data for the current point set.
324 * \throws APIError if any attached data module is not compatible.
325 * \throws unspecified Any exception thrown by attached data
326 * modules in AnalysisDataModuleInterface::pointsAdded().
328 * Must be called after each point set for multipoint data, including
329 * the last (i.e., no values must be set between the last call to this
330 * method and AnalysisDataStorage::finishFrame()).
331 * Must not be called for non-multipoint data.
333 void finishPointSet();
335 * Finish data for the current frame.
337 * \throws APIError if any attached data module is not compatible.
338 * \throws unspecified Any exception thrown by attached data
339 * modules in frame notification methods.
342 //! Calls AnalysisData::finishData() for this handle.
347 * Creates a new data handle associated with \p data.
349 * \param impl Data to associate the handle with.
351 * The constructor is private because data handles should only be
352 * constructed through AnalysisData::startData().
356 explicit AnalysisDataHandle(internal::AnalysisDataHandleImpl *impl);
359 * Pointer to the internal implementation class.
361 * The memory for this object is managed by the AnalysisData object,
362 * and AnalysisDataHandle simply provides a public interface for
363 * accessing the implementation.
365 internal::AnalysisDataHandleImpl *impl_;
368 * Needed to access the non-public implementation.
370 friend class AnalysisData;