FESCnmSnmReader.java
/* Copyright 2002-2019 CS Systèmes d'Information
* Licensed to CS Systèmes d'Information (CS) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* CS licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.orekit.forces.gravity.potential;
import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import org.orekit.errors.OrekitException;
import org.orekit.errors.OrekitMessages;
/** Reader for ocean tides files following the fes2004_Cnm-Snm.dat format.
* @since 6.1
* @author Luc Maisonobe
*/
public class FESCnmSnmReader extends OceanTidesReader {
/** Default pattern for fields with unknown type (non-space characters). */
private static final String UNKNOWN_TYPE_PATTERN = "\\S+";
/** Pattern for fields with integer type. */
private static final String INTEGER_TYPE_PATTERN = "[-+]?\\p{Digit}+";
/** Pattern for fields with real type. */
private static final String REAL_TYPE_PATTERN = "[-+]?(?:(?:\\p{Digit}+(?:\\.\\p{Digit}*)?)|(?:\\.\\p{Digit}+))(?:[eE][-+]?\\p{Digit}+)?";
/** Pattern for fields with Doodson number. */
private static final String DOODSON_TYPE_PATTERN = "\\p{Digit}{2,3}[.,]\\p{Digit}{3}";
/** Scale of the Cnm, Snm parameters. */
private final double scale;
/** Simple constructor.
* @param supportedNames regular expression for supported files names
* @param scale scale of the Cnm, Snm parameters
*/
public FESCnmSnmReader(final String supportedNames, final double scale) {
super(supportedNames);
this.scale = scale;
}
/** {@inheritDoc} */
@Override
public void loadData(final InputStream input, final String name)
throws IOException {
// FES ocean tides models have the following form:
// Coefficients to compute variations in normalized Stokes coefficients (unit = 10^-12)
// Ocean tide model: FES2004 normalized model (fev. 2004) up to (100,100)
// (long period from FES2002 up to (50,50) + equilibrium Om1/Om2, atmospheric tide NOT included)
// Doodson Darw l m DelC+ DelS+ DelC- DelS-
// 55.565 Om1 2 0 6.58128 -0.00000 -0.00000 -0.00000
// 55.575 Om2 2 0 -0.06330 0.00000 0.00000 0.00000
// 56.554 Sa 1 0 -0.00000 -0.00000 -0.00000 -0.00000
// 56.554 Sa 2 0 0.56720 0.01099 -0.00000 -0.00000
// 56.554 Sa 3 0 0.00908 -0.00050 -0.00000 -0.00000
final String[] fieldsPatterns = new String[] {
DOODSON_TYPE_PATTERN,
UNKNOWN_TYPE_PATTERN,
INTEGER_TYPE_PATTERN,
INTEGER_TYPE_PATTERN,
REAL_TYPE_PATTERN,
REAL_TYPE_PATTERN,
REAL_TYPE_PATTERN,
REAL_TYPE_PATTERN
};
final StringBuilder builder = new StringBuilder("^\\p{Space}*");
for (int i = 0; i < fieldsPatterns.length; ++i) {
builder.append("(");
builder.append(fieldsPatterns[i]);
builder.append(")");
builder.append((i < fieldsPatterns.length - 1) ? "\\p{Space}+" : "\\p{Space}*$");
}
final Pattern regularLinePattern = Pattern.compile(builder.toString());
// parse the file
startParse(name);
final BufferedReader r = new BufferedReader(new InputStreamReader(input, "UTF-8"));
int lineNumber = 0;
boolean dataStarted = false;
for (String line = r.readLine(); line != null; line = r.readLine()) {
++lineNumber;
final Matcher regularMatcher = regularLinePattern.matcher(line);
if (regularMatcher.matches()) {
// we have found a regular data line
// parse Doodson, degree and order fields
final int doodson = Integer.parseInt(regularMatcher.group(1).replaceAll("[.,]", ""));
final int n = Integer.parseInt(regularMatcher.group(3));
final int m = Integer.parseInt(regularMatcher.group(4));
if (canAdd(n, m)) {
// parse coefficients
final double cPlus = scale * Double.parseDouble(regularMatcher.group(5));
final double sPlus = scale * Double.parseDouble(regularMatcher.group(6));
final double cMinus = scale * Double.parseDouble(regularMatcher.group(7));
final double sMinus = scale * Double.parseDouble(regularMatcher.group(8));
// store parsed fields
addWaveCoefficients(doodson, n, m, cPlus, sPlus, cMinus, sMinus, lineNumber, line);
dataStarted = true;
}
} else if (dataStarted) {
// once the first data line has been encountered,
// all remaining lines should also be data lines
throw new OrekitException(OrekitMessages.UNABLE_TO_PARSE_LINE_IN_FILE,
lineNumber, name, line);
}
}
endParse();
}
}