3
* $Date: 2007-01-04 18:46:10 +0100 (Thu, 04 Jan 2007) $
6
* Copyright (C) 2003-2007 The Chemistry Development Kit (CDK) project
8
* Contact: cdk-devel@lists.sourceforge.net
10
* This library is free software; you can redistribute it and/or
11
* modify it under the terms of the GNU Lesser General Public
12
* License as published by the Free Software Foundation; either
13
* version 2.1 of the License, or (at your option) any later version.
15
* This library is distributed in the hope that it will be useful,
16
* but WITHOUT ANY WARRANTY; without even the implied warranty of
17
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18
* Lesser General Public License for more details.
20
* You should have received a copy of the GNU Lesser General Public
21
* License along with this library; if not, write to the Free Software
22
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
24
package org.openscience.cdk.io;
26
import java.io.BufferedReader;
27
import java.io.IOException;
28
import java.io.InputStream;
29
import java.io.InputStreamReader;
30
import java.io.Reader;
31
import java.io.StreamTokenizer;
32
import java.io.StringReader;
34
import javax.vecmath.Point3d;
36
import org.openscience.cdk.interfaces.IAtom;
37
import org.openscience.cdk.interfaces.IChemFile;
38
import org.openscience.cdk.interfaces.IChemModel;
39
import org.openscience.cdk.interfaces.IChemObject;
40
import org.openscience.cdk.interfaces.IChemSequence;
41
import org.openscience.cdk.interfaces.IMolecule;
42
import org.openscience.cdk.PhysicalConstants;
43
import org.openscience.cdk.interfaces.IMoleculeSet;
44
import org.openscience.cdk.exception.CDKException;
45
import org.openscience.cdk.io.formats.GamessFormat;
46
import org.openscience.cdk.io.formats.IResourceFormat;
50
* A reader for GAMESS log file.
52
* <p><b>Expected behaviour</b>:
53
* <br>The "GamessReader" object is able to read GAMESS output log file format.
55
* <p><b>Limitations</b>: <br>This reader was developed from a small set of
56
* example log files, and therefore, is not guaranteed to properly read all
57
* GAMESS output. If you have problems, please contact the author of this code,
58
* not the developers of GAMESS.
60
* <!-- <p><b>State information</b>: <br> [] -->
61
* <!-- <p><b>Dependencies</b>: <br> [all OS/Software/Hardware dependencies] -->
63
* <p><b>Implementation</b>
64
* <br>Available feature(s):
66
* <li><b>Molecular coordinates</b>: Each set of coordinates is added to the ChemFile in the order they are found.</li>
68
* Unavailable feature(s):
70
* <!-- <li><b>GAMESS version number</b>: The version number can be retrieved.</li> -->
71
* <!-- <li><b>Point group symetry information</b>: The point group is associated with the set of molecules.</li> -->
72
* <!-- <li><b>MOPAC charges</b>: The point group is associated with the set of molecules.</li> -->
73
* <li><b>Energies</b>: They are associated with the previously read set of coordinates.</li>
74
* <li><b>Normal coordinates of vibrations</b>: They are associated with the previously read set of coordinates.</li>
77
* <!-- <p><b>Security:</b> -->
79
* <p><b>References</b>:
80
* <br><a href="http://www.msg.ameslab.gov/GAMESS/GAMESS.html">GAMESS</a> is a
81
* quantum chemistry program by Gordon research group atIowa State University.
83
* @cdk.module experimental
85
* @cdk.keyword file format
87
* @cdk.keyword log file
89
* @author Bradley A. Smith
91
* <!-- @see #GamessWriter(Reader) -->
93
//TODO Update class comments with appropriate information.
94
//TODO Update "see" tag with reference to GamessWriter when it will be implemented.
95
//TODO Update "author" tag with appropriate information.
96
public class GamessReader extends DefaultChemObjectReader {
99
* Boolean constant used to specify that the coordinates are given in Bohr units.
101
public static final boolean BOHR_UNIT = true;
104
* Double constant that contains the convertion factor from Bohr unit to
105
* Ångstrom unit.
107
//TODO Check the accuracy of this comment.
108
public static final double BOHR_TO_ANGSTROM = 0.529177249;
111
* Boolean constant used to specify that the coordinates are given in Ångstrom units.
113
public static final boolean ANGSTROM_UNIT = false;
116
* The "BufferedReader" object used to read data from the "file system" file.
118
* @see org.openscience.cdk.io.GamessReader#GamessReader(Reader)
120
//TODO Improve field comment.
121
//TODO Answer the question : When is it opened and when is it closed?
122
private BufferedReader input;
125
* Constructs a new "GamessReader" object given a "Reader" object as input.
127
* <p>The "Reader" object may be an instantiable object from the "Reader"
129
* <br>For more detail about the "Reader" objects that are really accepted
130
* by this "GamessReader" see <code>accepts(IChemObject)</code> method
133
* @param inputReader The "Reader" object given as input parameter.
135
* @see #accepts(Class)
136
* @see java.io.Reader
139
public GamessReader(Reader inputReader) {
140
this.input = new BufferedReader(inputReader);
143
public GamessReader(InputStream input) {
144
this(new InputStreamReader(input));
147
public GamessReader() {
148
this(new StringReader(""));
151
/* (non-Javadoc) (Javadoc is automaticly inherited from the link below)
152
* @see org.openscience.cdk.io.ChemObjectIO#accepts(org.openscience.cdk.ChemObject)
154
//TODO Update comment with appropriate information to comply Constructor's documentation.
156
public IResourceFormat getFormat() {
157
return GamessFormat.getInstance();
160
public void setReader(Reader reader) throws CDKException {
161
this.input = new BufferedReader(input);
164
public void setReader(InputStream input) throws CDKException {
165
setReader(new InputStreamReader(input));
168
public boolean accepts(Class classObject) {
169
Class[] interfaces = classObject.getInterfaces();
170
for (int i=0; i<interfaces.length; i++) {
171
if (IChemFile.class.equals(interfaces[i])) return true;
176
/* (non-Javadoc) (Javadoc is automaticly inherited from the link below)
177
* @see org.openscience.cdk.io.ChemObjectReader#read(org.openscience.cdk.ChemObject)
179
public IChemObject read(IChemObject object) throws CDKException {
180
if (object instanceof IChemFile) {
182
return (IChemObject) readChemFile((IChemFile)object);
183
} catch (IOException e) {
187
throw new CDKException("Only supported is reading of ChemFile objects.");
192
* Reads data from the "file system" file through the use of the "input"
193
* field, parses data and feeds the ChemFile object with the extracted data.
195
* @return A ChemFile containing the data parsed from input.
197
* @throws IOException may be thrown buy the <code>this.input.readLine()</code> instruction.
199
* @see org.openscience.cdk.io.GamessReader#input
201
//TODO Answer the question : Is this method's name appropriate (given the fact that it do not read a ChemFile object, but return it)?
202
private IChemFile readChemFile(IChemFile file) throws IOException {
203
IChemSequence sequence = file.getBuilder().newChemSequence(); // TODO Answer the question : Is this line needed ?
204
IChemModel model = file.getBuilder().newChemModel(); // TODO Answer the question : Is this line needed ?
205
IMoleculeSet moleculeSet = file.getBuilder().newMoleculeSet();
207
model.setMoleculeSet(moleculeSet); //TODO Answer the question : Should I do this?
208
sequence.addChemModel(model); //TODO Answer the question : Should I do this?
209
file.addChemSequence(sequence); //TODO Answer the question : Should I do this?
211
String currentReadLine = this.input.readLine();
212
while (this.input.ready() == true && (currentReadLine != null)) {
215
* There are 2 types of coordinate sets:
216
* - bohr coordinates sets (if statement)
217
* - angstr???m coordinates sets (else statement)
219
if (currentReadLine.indexOf("COORDINATES (BOHR)") >= 0) {
222
* The following line do no contain data, so it is ignored.
224
this.input.readLine();
225
moleculeSet.addMolecule(this.readCoordinates(
226
file.getBuilder().newMolecule(), GamessReader.BOHR_UNIT
228
//break; //<- stops when the first set of coordinates is found.
229
} else if (currentReadLine.indexOf(" COORDINATES OF ALL ATOMS ARE (ANGS)") >= 0) {
232
* The following 2 lines do no contain data, so it are ignored.
234
this.input.readLine();
235
this.input.readLine();
237
moleculeSet.addMolecule(this.readCoordinates(
238
file.getBuilder().newMolecule(), GamessReader.ANGSTROM_UNIT
240
//break; //<- stops when the first set of coordinates is found.
242
currentReadLine = this.input.readLine();
248
* Reads a set of coordinates from the "file system" file through the use of
249
* the "input" field, scales coordinate to angstr???m unit, builds each atom with
250
* the right associated coordinates, builds a new molecule with these atoms
251
* and returns the complete molecule.
253
* <p><b>Implementation</b>:
254
* <br>Dummy atoms are ignored.
256
* @param coordinatesUnits The unit in which coordinates are given.
258
* @throws IOException may be thrown by the "input" object.
260
* @see org.openscience.cdk.io.GamessReader#input
262
//TODO Update method comments with appropriate information.
263
private IMolecule readCoordinates(IMolecule molecule, boolean coordinatesUnits) throws IOException {
266
* Coordinates must all be given in angstr???ms.
268
double unitScaling = GamessReader.scalesCoordinatesUnits(coordinatesUnits);
270
String retrievedLineFromFile;
272
while (this.input.ready() == true) {
273
retrievedLineFromFile = this.input.readLine();
275
* A coordinate set is followed by an empty line, so when this line
276
* is reached, there are no more coordinates to add to the current set.
278
if ((retrievedLineFromFile == null) || (retrievedLineFromFile.trim().length() == 0)) {
285
//StringReader sr = new StringReader(retrievedLineFromFile);
286
StreamTokenizer token = new StreamTokenizer(new StringReader(retrievedLineFromFile));
289
* The first token is ignored. It contains the atomic symbol and may
290
* be concatenated with a number.
294
if (token.nextToken() == StreamTokenizer.TT_NUMBER) {
295
atomicNumber = (int) token.nval;
296
atomicSymbol = this.identifyAtomicSymbol(atomicNumber);
298
* Dummy atoms are assumed to be given with an atomic number set
299
* to zero. We will do not add them to the molecule.
301
if (atomicNumber == 0) {
305
throw new IOException("Error reading coordinates");
309
* Atom's coordinates are stored in an array.
311
double[] coordinates = new double[3];
312
for (int i = 0; i < coordinates.length; i++) {
313
if (token.nextToken() == StreamTokenizer.TT_NUMBER) {
314
coordinates[i] = token.nval * unitScaling;
316
throw new IOException("Error reading coordinates");
319
IAtom atom = molecule.getBuilder().newAtom(atomicSymbol, new Point3d(coordinates[0],coordinates[1],coordinates[2]));
320
molecule.addAtom(atom);
326
* Identifies the atomic symbol of an atom given its default atomic number.
328
* <p><b>Implementation</b>:
329
* <br>This is not a definitive method. It will probably be replaced with a
330
* more appropriate one. Be advised that as it is not a definitive version,
331
* it only recognise atoms from Hydrogen (1) to Argon (18).
333
* @param atomicNumber The atomic number of an atom.
335
* @return The Symbol corresponding to the atom or "null" is the atom was not recognised.
337
//TODO Update method comments with appropriate information.
338
private String identifyAtomicSymbol(int atomicNumber) {
340
switch (atomicNumber) {
403
* Scales coordinates to Ångström unit if they are given in Bohr unit.
404
* If coordinates are already given in Ångström unit, then no modifications
407
* @param coordinatesUnits <code>BOHR_UNIT</code> if coordinates are given in Bohr unit and <code>ANGSTROM_UNIT</code>
408
* if they are given in Ångström unit.
410
* @return The scaling convertion factor: 1 if no scaling is needed and <code>BOHR_TO_ANGSTROM</code> if scaling has to be performed.
412
* @see org.openscience.cdk.PhysicalConstants#BOHR_TO_ANGSTROM
413
* @see org.openscience.cdk.io.GamessReader#BOHR_UNIT
414
* @see org.openscience.cdk.io.GamessReader#ANGSTROM_UNIT
416
//TODO Update method comments with appropriate information.
417
private static double scalesCoordinatesUnits(boolean coordinatesUnits) {
418
if (coordinatesUnits == GamessReader.BOHR_UNIT) {
419
return PhysicalConstants.BOHR_TO_ANGSTROM;
420
} else { //condition is: (coordinatesUnits == GamessReader.ANGTROM_UNIT)
425
/* (non-Javadoc) (Javadoc is automaticly inherited from the link below)
426
* @see org.openscience.cdk.io.ChemObjectIO#close()
428
//TODO Answer the question : What are all concerned ressources ?
429
public void close() throws IOException {
431
* Closes the BufferedReader used to read the file content.
b'\\ No newline at end of file'
added
removed
src/org/openscience/cdk/io/GamessReader.java
