root / trunk / extensions / extNormalization / src / org / gvsig / normalization / operations / NormAlgorithm.java @ 22823
History | View | Annotate | Download (14.3 KB)
| 1 |
/* gvSIG. Geographic Information System of the Valencian Government
|
|---|---|
| 2 |
*
|
| 3 |
* Copyright (C) 2007-2008 Infrastructures and Transports Department
|
| 4 |
* of the Valencian Government (CIT)
|
| 5 |
*
|
| 6 |
* This program is free software; you can redistribute it and/or
|
| 7 |
* modify it under the terms of the GNU General Public License
|
| 8 |
* as published by the Free Software Foundation; either version 2
|
| 9 |
* of the License, or (at your option) any later version.
|
| 10 |
*
|
| 11 |
* This program is distributed in the hope that it will be useful,
|
| 12 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
| 13 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
| 14 |
* GNU General Public License for more details.
|
| 15 |
*
|
| 16 |
* You should have received a copy of the GNU General Public License
|
| 17 |
* along with this program; if not, write to the Free Software
|
| 18 |
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
|
| 19 |
* MA 02110-1301, USA.
|
| 20 |
*
|
| 21 |
*/
|
| 22 |
|
| 23 |
/*
|
| 24 |
* AUTHORS (In addition to CIT):
|
| 25 |
* 2008 Prodevelop S.L. main development
|
| 26 |
*/
|
| 27 |
|
| 28 |
package org.gvsig.normalization.operations; |
| 29 |
|
| 30 |
import java.text.DecimalFormat; |
| 31 |
import java.text.DecimalFormatSymbols; |
| 32 |
import java.util.ArrayList; |
| 33 |
import java.util.List; |
| 34 |
import java.util.Locale; |
| 35 |
|
| 36 |
import javax.swing.event.ChangeEvent; |
| 37 |
import javax.swing.event.ChangeListener; |
| 38 |
|
| 39 |
import org.apache.log4j.Logger; |
| 40 |
import org.gvsig.geocoding.patterns.normalization.Addresselement; |
| 41 |
import org.gvsig.geocoding.patterns.normalization.Datevalue; |
| 42 |
import org.gvsig.geocoding.patterns.normalization.Decimalvalue; |
| 43 |
import org.gvsig.geocoding.patterns.normalization.Integervalue; |
| 44 |
import org.gvsig.geocoding.patterns.normalization.Newfieldtype; |
| 45 |
import org.gvsig.geocoding.patterns.normalization.Patternnormalization; |
| 46 |
import org.gvsig.geocoding.patterns.normalization.Stringvalue; |
| 47 |
|
| 48 |
/**
|
| 49 |
* This class tokens strings
|
| 50 |
*
|
| 51 |
* @author <a href="mailto:jsanz@prodevelop.es"> Jorge Gaspar Sanz Salinas</a>
|
| 52 |
* @author <a href="mailto:vsanjaime@prodevelop.es"> Vicente Sanjaime Calvet</a>
|
| 53 |
*
|
| 54 |
*/
|
| 55 |
|
| 56 |
public class NormAlgorithm { |
| 57 |
|
| 58 |
// ATTRIBUTES
|
| 59 |
|
| 60 |
private static final Logger log = Logger.getLogger(NormAlgorithm.class); |
| 61 |
|
| 62 |
private Patternnormalization pat;
|
| 63 |
|
| 64 |
private int numFields; |
| 65 |
|
| 66 |
private int row; |
| 67 |
|
| 68 |
private Addresselement[] aad; |
| 69 |
|
| 70 |
@SuppressWarnings("unchecked") |
| 71 |
private ArrayList listeners = new ArrayList(); |
| 72 |
|
| 73 |
/**
|
| 74 |
* Builder
|
| 75 |
*
|
| 76 |
* @param _pat
|
| 77 |
* Normalization pattern
|
| 78 |
*/
|
| 79 |
public NormAlgorithm(Patternnormalization _pat) {
|
| 80 |
pat = _pat; |
| 81 |
aad = pat.getAddresselements().getAddresselement(); |
| 82 |
numFields = aad.length; |
| 83 |
|
| 84 |
} |
| 85 |
|
| 86 |
// METHODS
|
| 87 |
|
| 88 |
/**
|
| 89 |
* This method cuts a chain in several parts and they are returned in a List
|
| 90 |
*
|
| 91 |
* @param _chain
|
| 92 |
* strings
|
| 93 |
* @return list with the strings tokenized
|
| 94 |
*/
|
| 95 |
@SuppressWarnings("unchecked") |
| 96 |
public List splitChain(String _chain) { |
| 97 |
String preChain = _chain;
|
| 98 |
String postChain = ""; |
| 99 |
List subStrings = new ArrayList(); |
| 100 |
|
| 101 |
// EXCEPTIONAL CASES
|
| 102 |
if(_chain == null){ |
| 103 |
return subStrings;
|
| 104 |
} |
| 105 |
|
| 106 |
else if (_chain.compareToIgnoreCase("") == 0) { |
| 107 |
subStrings.add(_chain); |
| 108 |
return subStrings;
|
| 109 |
} |
| 110 |
|
| 111 |
|
| 112 |
// NORMAL CASE
|
| 113 |
else {
|
| 114 |
int fw = 0; |
| 115 |
int init = 0; |
| 116 |
String subChain = ""; |
| 117 |
|
| 118 |
for (int i = 0; i < numFields; i++) { |
| 119 |
|
| 120 |
if (preChain.length() == 0) { |
| 121 |
return subStrings;
|
| 122 |
} else {
|
| 123 |
|
| 124 |
fw = aad[i].getFieldwidth(); |
| 125 |
// Cut chain with fixed width
|
| 126 |
if (fw > 0 && preChain.length() > fw) { |
| 127 |
subChain = preChain.substring(init, fw); |
| 128 |
subStrings.add(subChain); |
| 129 |
postChain = preChain.substring(fw); |
| 130 |
preChain = postChain; |
| 131 |
|
| 132 |
} else if (fw > 0 && preChain.length() <= fw) { |
| 133 |
subStrings.add(preChain); |
| 134 |
return subStrings;
|
| 135 |
} |
| 136 |
// Cut chain with separators
|
| 137 |
else {
|
| 138 |
// Load specific separators
|
| 139 |
List separators = loadSpecificSeparators(aad[i]);
|
| 140 |
boolean join = withJoinSeparators(aad[i]);
|
| 141 |
// Search the first delimiter in the chain
|
| 142 |
int posi = calculatePosition(separators, preChain);
|
| 143 |
int tamSep = calculateSizeSep(separators, preChain);
|
| 144 |
// Firsts elements
|
| 145 |
if (i < numFields - 1) { |
| 146 |
|
| 147 |
if (join) {
|
| 148 |
while (posi == 0) { |
| 149 |
preChain = deleteFirst(preChain); |
| 150 |
posi = calculatePosition(separators, |
| 151 |
preChain); |
| 152 |
if(preChain.length() == 0){ |
| 153 |
break;
|
| 154 |
} |
| 155 |
} |
| 156 |
subChain = preChain.substring(0, posi);
|
| 157 |
try {
|
| 158 |
postChain = preChain.substring(posi |
| 159 |
+ tamSep); |
| 160 |
} catch (Exception e) { |
| 161 |
postChain = "";
|
| 162 |
} |
| 163 |
subStrings.add(subChain); |
| 164 |
preChain = postChain; |
| 165 |
} else {
|
| 166 |
subChain = preChain.substring(0, posi);
|
| 167 |
postChain = preChain.substring(posi + tamSep); |
| 168 |
subStrings.add(subChain); |
| 169 |
preChain = postChain; |
| 170 |
} |
| 171 |
|
| 172 |
} |
| 173 |
// Last element
|
| 174 |
else {
|
| 175 |
subStrings.add(preChain); |
| 176 |
} |
| 177 |
} |
| 178 |
} |
| 179 |
} |
| 180 |
} |
| 181 |
return subStrings;
|
| 182 |
} |
| 183 |
|
| 184 |
/**
|
| 185 |
* This method cuts a chain in several parts from separators
|
| 186 |
*
|
| 187 |
* @param chain
|
| 188 |
* string
|
| 189 |
* @param fields
|
| 190 |
* fields number
|
| 191 |
* @param separators
|
| 192 |
* array of characters
|
| 193 |
* @param joinDelimiters
|
| 194 |
* with or without joinDelimiters
|
| 195 |
* @return list with the strings tokenized
|
| 196 |
*/
|
| 197 |
@SuppressWarnings("unchecked") |
| 198 |
public static List splitChainBySeparators(String chain, int fields, |
| 199 |
String[] separators, boolean joinDelimiters) { |
| 200 |
|
| 201 |
List subStrings = new ArrayList(); |
| 202 |
int posTemp = -1; |
| 203 |
String separator;
|
| 204 |
String chain2 = chain;
|
| 205 |
int campos = fields;
|
| 206 |
|
| 207 |
// EXCEPTIONAL CASES
|
| 208 |
if (chain.compareToIgnoreCase("") == 0 || campos == 0 |
| 209 |
|| separators.length < 1) {
|
| 210 |
subStrings.add(chain); |
| 211 |
return subStrings;
|
| 212 |
} |
| 213 |
|
| 214 |
// NORMAL CASE
|
| 215 |
else {
|
| 216 |
|
| 217 |
// Only (parts-1) loops
|
| 218 |
for (int i = 0; i < (campos - 1); i++) { |
| 219 |
int posi = Integer.MAX_VALUE; |
| 220 |
String firstChain;
|
| 221 |
for (int j = 0; j < separators.length; j++) { |
| 222 |
separator = separators[j]; |
| 223 |
posTemp = chain2.indexOf(separator); |
| 224 |
if (posTemp != -1 && posTemp < posi) { |
| 225 |
posi = posTemp; |
| 226 |
} |
| 227 |
posTemp = -1;
|
| 228 |
} |
| 229 |
if (posi == 0 && joinDelimiters) { |
| 230 |
campos++; |
| 231 |
chain2 = chain2.substring(posi + 1);
|
| 232 |
} else {
|
| 233 |
firstChain = chain2.substring(0, posi);
|
| 234 |
chain2 = chain2.substring(posi + 1);
|
| 235 |
subStrings.add(firstChain); |
| 236 |
// In the last loop add the first chain and the rest of
|
| 237 |
// chain
|
| 238 |
if (i == (campos - 2)) { |
| 239 |
subStrings.add(chain2); |
| 240 |
} |
| 241 |
} |
| 242 |
} |
| 243 |
return subStrings;
|
| 244 |
} |
| 245 |
} |
| 246 |
|
| 247 |
/**
|
| 248 |
* This method cuts a chain in several parts from fixed width
|
| 249 |
*
|
| 250 |
* @param chain
|
| 251 |
* string
|
| 252 |
* @param fieldWidth
|
| 253 |
* array with fields widths
|
| 254 |
* @return list with the strings tokenized
|
| 255 |
*/
|
| 256 |
@SuppressWarnings("unchecked") |
| 257 |
public static List splitChainByFixedWidth(String chain, int[] fieldWidth) { |
| 258 |
|
| 259 |
List subStrings = new ArrayList(); |
| 260 |
int elements = fieldWidth.length;
|
| 261 |
String subChain;
|
| 262 |
int inicio = 0; |
| 263 |
int fin = 0; |
| 264 |
|
| 265 |
// EXCEPTIONAL CASES
|
| 266 |
if (chain.compareToIgnoreCase("") == 0 || fieldWidth.length < 1) { |
| 267 |
subStrings.add(chain); |
| 268 |
return subStrings;
|
| 269 |
} |
| 270 |
|
| 271 |
// NORMAL CASE
|
| 272 |
else {
|
| 273 |
for (int i = 0; i < elements; i++) { |
| 274 |
fin = fin + fieldWidth[i]; |
| 275 |
subChain = chain.substring(inicio, fin); |
| 276 |
subStrings.add(subChain); |
| 277 |
inicio = fin; |
| 278 |
} |
| 279 |
return subStrings;
|
| 280 |
} |
| 281 |
} |
| 282 |
|
| 283 |
/**
|
| 284 |
* This method filters the split chains with the in-separators
|
| 285 |
*
|
| 286 |
* @param chains
|
| 287 |
* strings of the one row
|
| 288 |
* @return strings filtered by type
|
| 289 |
*/
|
| 290 |
@SuppressWarnings("unchecked") |
| 291 |
public List filterSplitChains(List chains) { |
| 292 |
|
| 293 |
DecimalFormat numForm = (DecimalFormat) DecimalFormat |
| 294 |
.getInstance(Locale.getDefault());
|
| 295 |
DecimalFormatSymbols simb = numForm.getDecimalFormatSymbols();
|
| 296 |
|
| 297 |
Newfieldtype nft = null;
|
| 298 |
|
| 299 |
String decsep;
|
| 300 |
char cdecsep;
|
| 301 |
String thosep;
|
| 302 |
char cthosep;
|
| 303 |
String txsep;
|
| 304 |
|
| 305 |
List postChain = new ArrayList(); |
| 306 |
String aux = ""; |
| 307 |
|
| 308 |
for (int i = 0; i < numFields; i++) { |
| 309 |
|
| 310 |
nft = aad[i].getNewfieldtype(); |
| 311 |
|
| 312 |
decsep = aad[i].getInfieldseparators().getDecimalseparator() |
| 313 |
.getValue().trim(); |
| 314 |
cdecsep = decsep.compareTo("") == 0 ? (char) 0x20 : decsep |
| 315 |
.charAt(0);
|
| 316 |
thosep = aad[i].getInfieldseparators().getThousandseparator() |
| 317 |
.getValue().trim(); |
| 318 |
cthosep = thosep.compareTo("") == 0 ? (char) 0x20 : thosep |
| 319 |
.charAt(0);
|
| 320 |
|
| 321 |
txsep = aad[i].getInfieldseparators().getTextseparator().getValue() |
| 322 |
.trim(); |
| 323 |
|
| 324 |
if (txsep.compareToIgnoreCase("\"") == 0) { |
| 325 |
txsep = "\"";
|
| 326 |
} |
| 327 |
|
| 328 |
simb.setDecimalSeparator(cdecsep); |
| 329 |
try {
|
| 330 |
simb.setGroupingSeparator(cthosep); |
| 331 |
} catch (RuntimeException e1) { |
| 332 |
log.error("Error setting the group separator", e1);
|
| 333 |
} |
| 334 |
// Fill fields empties
|
| 335 |
if (chains.size() < numFields) {
|
| 336 |
|
| 337 |
for (int j = chains.size(); j < numFields; j++) { |
| 338 |
chains.add(j, "");
|
| 339 |
} |
| 340 |
} |
| 341 |
|
| 342 |
if (((Integervalue) nft.getIntegervalue()) != null |
| 343 |
|| ((Decimalvalue) nft.getDecimalvalue()) != null) {
|
| 344 |
numForm.setDecimalFormatSymbols(simb); |
| 345 |
try {
|
| 346 |
String cadena = (String) chains.get(i); |
| 347 |
int num = cadena.length();
|
| 348 |
boolean comproba = test(cadena, cdecsep, cthosep);
|
| 349 |
if (num > 0 && comproba) { |
| 350 |
aux = (numForm.parse(cadena)).toString().trim(); |
| 351 |
} else {
|
| 352 |
aux = "";
|
| 353 |
} |
| 354 |
} catch (Exception e) { |
| 355 |
aux = "";
|
| 356 |
} |
| 357 |
} |
| 358 |
|
| 359 |
// Field type (VARCHAR)
|
| 360 |
if (((Stringvalue) nft.getStringvalue()) != null) { |
| 361 |
String cadena = (String) chains.get(i); |
| 362 |
if (cadena != null) { |
| 363 |
aux = cadena.replace(txsep, "");
|
| 364 |
} else {
|
| 365 |
aux = "";
|
| 366 |
} |
| 367 |
} |
| 368 |
|
| 369 |
// Field type (DATE)
|
| 370 |
if (((Datevalue) nft.getDatevalue()) != null) { |
| 371 |
String cadena = (String) chains.get(i); |
| 372 |
if (cadena.length() > 0 || cadena != null) { |
| 373 |
aux = cadena.replace(txsep, "");
|
| 374 |
} else {
|
| 375 |
aux = "";
|
| 376 |
} |
| 377 |
} |
| 378 |
postChain.add(aux); |
| 379 |
} |
| 380 |
return postChain;
|
| 381 |
|
| 382 |
} |
| 383 |
|
| 384 |
/**
|
| 385 |
* This method registers the listeners
|
| 386 |
*
|
| 387 |
* @param l
|
| 388 |
* listener
|
| 389 |
*/
|
| 390 |
@SuppressWarnings("unchecked") |
| 391 |
public void registerListener(ChangeListener l) { |
| 392 |
this.listeners.add(l);
|
| 393 |
} |
| 394 |
|
| 395 |
/**
|
| 396 |
* This method remove the listener registred
|
| 397 |
*
|
| 398 |
* @param l
|
| 399 |
* listener
|
| 400 |
*/
|
| 401 |
public void removeListener(ChangeListener l) { |
| 402 |
this.listeners.remove(l);
|
| 403 |
} |
| 404 |
|
| 405 |
/**
|
| 406 |
* This method removes all listeners
|
| 407 |
*/
|
| 408 |
public void removeAllListeners() { |
| 409 |
this.listeners.clear();
|
| 410 |
} |
| 411 |
|
| 412 |
/**
|
| 413 |
*
|
| 414 |
* @param evt
|
| 415 |
* event
|
| 416 |
*/
|
| 417 |
public void update(ChangeEvent evt) { |
| 418 |
|
| 419 |
for (int i = 0; i < listeners.size(); i++) { |
| 420 |
((ChangeListener) listeners.get(i)).stateChanged(evt);
|
| 421 |
} |
| 422 |
} |
| 423 |
|
| 424 |
/**
|
| 425 |
* Add message
|
| 426 |
*
|
| 427 |
* @param message
|
| 428 |
*/
|
| 429 |
public void update(String message) { |
| 430 |
ChangeEvent evt = new ChangeEvent(message); |
| 431 |
update(evt); |
| 432 |
} |
| 433 |
|
| 434 |
/**
|
| 435 |
* Set the row
|
| 436 |
*
|
| 437 |
* @param _row
|
| 438 |
*/
|
| 439 |
public void setRow(int _row) { |
| 440 |
row = _row; |
| 441 |
} |
| 442 |
|
| 443 |
/**
|
| 444 |
* This method loads the join separators attribute of one Addresselement
|
| 445 |
*
|
| 446 |
* @param ad
|
| 447 |
* @return with or without joinSeparators
|
| 448 |
*/
|
| 449 |
private boolean withJoinSeparators(Addresselement ad) { |
| 450 |
return ad.getFieldseparator().getJoinsep();
|
| 451 |
} |
| 452 |
|
| 453 |
/**
|
| 454 |
* This method deletes the first element of one substring
|
| 455 |
*
|
| 456 |
* @param chain
|
| 457 |
* initial string
|
| 458 |
* @return string
|
| 459 |
*/
|
| 460 |
private String deleteFirst(String chain) { |
| 461 |
String del = chain.substring(1); |
| 462 |
return del;
|
| 463 |
} |
| 464 |
|
| 465 |
/**
|
| 466 |
* This method gets the first position in the string of the separators
|
| 467 |
* group.
|
| 468 |
*
|
| 469 |
* @param separators
|
| 470 |
* separators characters list
|
| 471 |
* @param preChain
|
| 472 |
* initial string
|
| 473 |
* @return
|
| 474 |
*/
|
| 475 |
@SuppressWarnings("unchecked") |
| 476 |
private int calculatePosition(List separators, String preChain) { |
| 477 |
|
| 478 |
String separator;
|
| 479 |
int posTemp = -1; |
| 480 |
int posi = Integer.MAX_VALUE; |
| 481 |
for (int j = 0; j < separators.size(); j++) { |
| 482 |
separator = (String) separators.get(j);
|
| 483 |
posTemp = preChain.indexOf(separator); |
| 484 |
if (posTemp != -1 && posTemp < posi) { |
| 485 |
posi = posTemp; |
| 486 |
} |
| 487 |
posTemp = -1;
|
| 488 |
} |
| 489 |
if (posi > preChain.length()) {
|
| 490 |
posi = preChain.length(); |
| 491 |
} |
| 492 |
|
| 493 |
return posi;
|
| 494 |
} |
| 495 |
|
| 496 |
/**
|
| 497 |
* This method calculates the number of elements of one separators
|
| 498 |
*
|
| 499 |
* @param separators
|
| 500 |
* @param preChain
|
| 501 |
* @return number of elements
|
| 502 |
*/
|
| 503 |
|
| 504 |
@SuppressWarnings("unchecked") |
| 505 |
private int calculateSizeSep(List separators, String preChain) { |
| 506 |
|
| 507 |
String separator;
|
| 508 |
int posTemp = -1; |
| 509 |
int posi = Integer.MAX_VALUE; |
| 510 |
String sep = ""; |
| 511 |
for (int j = 0; j < separators.size(); j++) { |
| 512 |
separator = (String) separators.get(j);
|
| 513 |
posTemp = preChain.indexOf(separator); |
| 514 |
if (posTemp != -1 && posTemp < posi) { |
| 515 |
posi = posTemp; |
| 516 |
sep = separator; |
| 517 |
} |
| 518 |
posTemp = -1;
|
| 519 |
} |
| 520 |
|
| 521 |
return sep.length();
|
| 522 |
} |
| 523 |
|
| 524 |
/**
|
| 525 |
* This method loads the specifics separators of one Addresselement
|
| 526 |
*
|
| 527 |
* @param adrElem
|
| 528 |
* @return separators list
|
| 529 |
*/
|
| 530 |
@SuppressWarnings("unchecked") |
| 531 |
private List loadSpecificSeparators(Addresselement adrElem) { |
| 532 |
List separators = new ArrayList(); |
| 533 |
if (adrElem.getFieldseparator().getColonsep()) {
|
| 534 |
separators.add(",");
|
| 535 |
} |
| 536 |
if (adrElem.getFieldseparator().getSemicolonsep()) {
|
| 537 |
separators.add(";");
|
| 538 |
} |
| 539 |
if (adrElem.getFieldseparator().getTabsep()) {
|
| 540 |
separators.add("\t");
|
| 541 |
} |
| 542 |
if (adrElem.getFieldseparator().getSpacesep()) {
|
| 543 |
separators.add(" ");
|
| 544 |
} |
| 545 |
if (adrElem.getFieldseparator().getOthersep() != null |
| 546 |
&& adrElem.getFieldseparator().getOthersep() |
| 547 |
.compareToIgnoreCase("") != 0) { |
| 548 |
separators.add((String) adrElem.getFieldseparator().getOthersep());
|
| 549 |
} |
| 550 |
|
| 551 |
return separators;
|
| 552 |
} |
| 553 |
|
| 554 |
/**
|
| 555 |
* This method tests the numbers format
|
| 556 |
*
|
| 557 |
* @param str
|
| 558 |
* @param dec
|
| 559 |
* @param sep
|
| 560 |
* @return true if there aren't other characters
|
| 561 |
*/
|
| 562 |
private boolean test(String str, char dec, char sep) { |
| 563 |
String str2 = str.replaceAll("[0-9]", ""); |
| 564 |
str2 = str2.replaceAll("-", ""); |
| 565 |
|
| 566 |
String str3 = str2;
|
| 567 |
if (str2.indexOf(dec) >= 0) { |
| 568 |
int ind = str2.indexOf(String.valueOf(dec)); |
| 569 |
str3 = str2.substring(0, ind)
|
| 570 |
+ str2.substring(ind + 1, str2.length());
|
| 571 |
} |
| 572 |
String str4 = str3;
|
| 573 |
if (str3.indexOf(sep) >= 0) { |
| 574 |
int ind = str3.indexOf(String.valueOf(sep)); |
| 575 |
str4 = str3.substring(0, ind)
|
| 576 |
+ str3.substring(ind + 1, str3.length());
|
| 577 |
} |
| 578 |
|
| 579 |
return str4.length() == 0; |
| 580 |
} |
| 581 |
|
| 582 |
} |