| 1 | /**
|
|---|
| 2 | *#########################################################################
|
|---|
| 3 | *
|
|---|
| 4 | * A component of the Gatherer application, part of the Greenstone digital
|
|---|
| 5 | * library suite from the New Zealand Digital Library Project at the
|
|---|
| 6 | * University of Waikato, New Zealand.
|
|---|
| 7 | *
|
|---|
| 8 | * Author: John Thompson, Greenstone Digital Library, University of Waikato
|
|---|
| 9 | *
|
|---|
| 10 | * Copyright (C) 1999 New Zealand Digital Library Project
|
|---|
| 11 | *
|
|---|
| 12 | * This program is free software; you can redistribute it and/or modify
|
|---|
| 13 | * it under the terms of the GNU General Public License as published by
|
|---|
| 14 | * the Free Software Foundation; either version 2 of the License, or
|
|---|
| 15 | * (at your option) any later version.
|
|---|
| 16 | *
|
|---|
| 17 | * This program is distributed in the hope that it will be useful,
|
|---|
| 18 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 19 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 20 | * GNU General Public License for more details.
|
|---|
| 21 | *
|
|---|
| 22 | * You should have received a copy of the GNU General Public License
|
|---|
| 23 | * along with this program; if not, write to the Free Software
|
|---|
| 24 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 25 | *########################################################################
|
|---|
| 26 | */
|
|---|
| 27 | package org.greenstone.gatherer.gui.metaaudit;
|
|---|
| 28 |
|
|---|
| 29 | /**************************************************************************************
|
|---|
| 30 | * Title: Gatherer
|
|---|
| 31 | * Description: The Gatherer: a tool for gathering and enriching a digital collection.
|
|---|
| 32 | * Copyright: Copyright (c) 2001
|
|---|
| 33 | * Company: The University of Waikato
|
|---|
| 34 | * Written: /05/02
|
|---|
| 35 | * Revised: 22/08/02 Revamped, Optimized and Commented.
|
|---|
| 36 | * 13-08-03 Extended matching functionality to correctly handle the wildcard character
|
|---|
| 37 | **************************************************************************************/
|
|---|
| 38 |
|
|---|
| 39 | import java.util.ArrayList;
|
|---|
| 40 | import org.greenstone.gatherer.util.StaticStrings;
|
|---|
| 41 |
|
|---|
| 42 | /** An Autofilter object stores the filters set on a single column, and provides a method for determining if a certain value passes the filter.
|
|---|
| 43 | * @author John Thompson
|
|---|
| 44 | * @version 2.3
|
|---|
| 45 | */
|
|---|
| 46 | public class Autofilter {
|
|---|
| 47 |
|
|---|
| 48 | static public void main(String[] args) {
|
|---|
| 49 | if(args.length != 2) {
|
|---|
| 50 | System.err.println("Usage: java Autofilter <string> <pattern>");
|
|---|
| 51 | }
|
|---|
| 52 | else {
|
|---|
| 53 | switch(compareToPattern(args[0], args[1])) {
|
|---|
| 54 | case LESS_THAN:
|
|---|
| 55 | System.err.print(args[0] + " is less than");
|
|---|
| 56 | break;
|
|---|
| 57 | case GREATER_THAN:
|
|---|
| 58 | System.err.print(args[0] + " is greater than");
|
|---|
| 59 | break;
|
|---|
| 60 | case EQUAL:
|
|---|
| 61 | System.err.print(args[0] + " is equal to");
|
|---|
| 62 | break;
|
|---|
| 63 | default:
|
|---|
| 64 | System.err.print(args[0] + " is something to");
|
|---|
| 65 | }
|
|---|
| 66 | System.err.println(" " + args[1]);
|
|---|
| 67 | }
|
|---|
| 68 | }
|
|---|
| 69 |
|
|---|
| 70 | /** A String comparison which also recognises the WildCard character. The basic idea is that '*' matches zero or more of any characters, and if we limited patterns to be like "a*" then as soon as we hit the star we could return a match. The problem is that we want patterns like "b*y" which would match "boy" but not "bee" nor "bus" (but should return less than or greater than for each of these respectively). Thus our testing has to become non-deterministic so the matches are conducted like so:
|
|---|
| 71 | * Given p = [ b * y ]
|
|---|
| 72 | * Test s = [ b e e ] Result: p0=s0,{(p2>s1=>ERROR),(p1=s1,{(p1>s2=>ERROR),(p2>s2=>LESS_THAN)})} => LESS_THAN
|
|---|
| 73 | * Test s = [ b o y ] Result: p0=s0,{(p2>s1=>ERROR),(p1=s1,{(p1>s2=>ERROR),(p2=s2=>EQUAL)})} => EQUAL
|
|---|
| 74 | * Test s = [ b u s ] Result: p0=s0,{(p2>s1=>ERROR),(p1=s1,{(p1>s2=>ERROR),(p2<s2=>GREATER_THAN)})} => GREATER_THAN
|
|---|
| 75 | * where the two () phrases within a {} represent two different 'threads' of processing. Errors are generated when one string runs out of characters before the other. Only non-error results are propogated. I've decided to implement this non-deterministic matching as a recursive function call.
|
|---|
| 76 | * If you analyse it you can see there are three possible actions when a '*' is detected.
|
|---|
| 77 | * 1. * is an empty string, ignore it in pattern and try to compare the next character in p with the current character in s. This case only occurs if there are more characters in p (otherwise apple would be greater than a*)
|
|---|
| 78 | * 2. * matches exactly the current character in s, try to match the next character in p with the next character in s
|
|---|
| 79 | * 3. * matches some string of characters including the current character in s, try to match '*' against the next character in s
|
|---|
| 80 | * @param s the String being matched
|
|---|
| 81 | * @param p the Pattern to match it against
|
|---|
| 82 | * @return -1 if s is less than p, 0 if they are equal, 1 if s is greater than p
|
|---|
| 83 | */
|
|---|
| 84 | static public int compareToPattern(String s, String p) {
|
|---|
| 85 | if(p.indexOf(StaticStrings.STAR_CHAR) == -1) {
|
|---|
| 86 | return compareToPattern(s, p, 0, 0, false);
|
|---|
| 87 | }
|
|---|
| 88 | else {
|
|---|
| 89 | return compareToPattern(s, p, 0, 0, true);
|
|---|
| 90 | }
|
|---|
| 91 | }
|
|---|
| 92 |
|
|---|
| 93 | static final private int LESS_THAN = -1;
|
|---|
| 94 | static final private int EQUAL = 0;
|
|---|
| 95 | static final private int GREATER_THAN = 1;
|
|---|
| 96 | static final private int ERROR = 42;
|
|---|
| 97 |
|
|---|
| 98 | static private int compareToPattern(String s, String p, int s_index, int p_index, boolean wildcard_matching) {
|
|---|
| 99 | // Lets do the simple cases first.
|
|---|
| 100 | // Both out of characters at the same time
|
|---|
| 101 | if(s_index >= s.length() && p_index >= p.length()) {
|
|---|
| 102 | ///ystem.err.println("Both out of characters. Equal.");
|
|---|
| 103 | return EQUAL;
|
|---|
| 104 | }
|
|---|
| 105 | // s out of characters first
|
|---|
| 106 | if(s_index >= s.length()) {
|
|---|
| 107 | // Not allowed to run out if matching wildcard
|
|---|
| 108 | if(wildcard_matching) {
|
|---|
| 109 | return ERROR;
|
|---|
| 110 | }
|
|---|
| 111 | // Remember that wildcard matches the empty string too, so as long as there are only '*' characters left in the pattern we can say they are equal as well.
|
|---|
| 112 | else {
|
|---|
| 113 | while(p_index < p.length()) {
|
|---|
| 114 | if(p.charAt(p_index) == StaticStrings.STAR_CHAR) {
|
|---|
| 115 | p_index++;
|
|---|
| 116 | }
|
|---|
| 117 | else {
|
|---|
| 118 | ///ystem.err.println("S out of characters. Less Than.");
|
|---|
| 119 | return LESS_THAN;
|
|---|
| 120 | }
|
|---|
| 121 | }
|
|---|
| 122 | // We've run out of pattern and it only contained '*' so we're still equal
|
|---|
| 123 | ///ystem.err.println("Both out of characters. Equal.");
|
|---|
| 124 | return EQUAL;
|
|---|
| 125 | }
|
|---|
| 126 | }
|
|---|
| 127 | // p out of characters first
|
|---|
| 128 | if(p_index >= p.length()) {
|
|---|
| 129 | // Not allowed to run out if matching wildcard
|
|---|
| 130 | if(wildcard_matching) {
|
|---|
| 131 | return ERROR;
|
|---|
| 132 | }
|
|---|
| 133 | // so it is greater than
|
|---|
| 134 | else {
|
|---|
| 135 | ///ystem.err.println("P out of characters. Greater Than.");
|
|---|
| 136 | return GREATER_THAN;
|
|---|
| 137 | }
|
|---|
| 138 | }
|
|---|
| 139 | char s_char = s.charAt(s_index);
|
|---|
| 140 | char p_char = p.charAt(p_index);
|
|---|
| 141 | ///ystem.err.println("Comparing " + s_char + " against pattern character " + p_char);
|
|---|
| 142 | // Equivelent characters
|
|---|
| 143 | // Now the tricky-dicky bit - WildCard matching.
|
|---|
| 144 | if(p_char == StaticStrings.STAR_CHAR) {
|
|---|
| 145 | int result;
|
|---|
| 146 | // Case 1
|
|---|
| 147 | ///ystem.err.println("WildCard Case 1");
|
|---|
| 148 | if(p_index + 1 < p.length()) {
|
|---|
| 149 | if((result = compareToPattern(s, p, s_index, p_index + 1, wildcard_matching)) != ERROR) {
|
|---|
| 150 | return result;
|
|---|
| 151 | }
|
|---|
| 152 | }
|
|---|
| 153 | // Case 2
|
|---|
| 154 | ///ystem.err.println("WildCard Case 2");
|
|---|
| 155 | if((result = compareToPattern(s, p, s_index + 1, p_index + 1, wildcard_matching)) != ERROR) {
|
|---|
| 156 | return result;
|
|---|
| 157 | }
|
|---|
| 158 | // Case 3
|
|---|
| 159 | ///ystem.err.println("WildCard Case 3");
|
|---|
| 160 | if((result = compareToPattern(s, p, s_index + 1, p_index, wildcard_matching)) != ERROR) {
|
|---|
| 161 | return result;
|
|---|
| 162 | }
|
|---|
| 163 | }
|
|---|
| 164 | if(s_char == p_char) {
|
|---|
| 165 | return compareToPattern(s, p, s_index + 1, p_index + 1, wildcard_matching);
|
|---|
| 166 | }
|
|---|
| 167 | // A preceeding character
|
|---|
| 168 | if(s_char < p_char) {
|
|---|
| 169 | ///ystem.err.println("s char is less than p char");
|
|---|
| 170 | if(wildcard_matching) {
|
|---|
| 171 | ///ystem.err.println("Because we are wildcard matching we still have to match the rest of s incase of errors.");
|
|---|
| 172 | if(compareToPattern(s, p, s_index + 1, p_index + 1, wildcard_matching) != ERROR) {
|
|---|
| 173 | ///ystem.err.println("No error. Less than.");
|
|---|
| 174 | return LESS_THAN;
|
|---|
| 175 | }
|
|---|
| 176 | else {
|
|---|
| 177 | ///ystem.err.println("Error detected.");
|
|---|
| 178 | return ERROR;
|
|---|
| 179 | }
|
|---|
| 180 | }
|
|---|
| 181 | else {
|
|---|
| 182 | ///ystem.err.println("Less than.");
|
|---|
| 183 | return LESS_THAN;
|
|---|
| 184 | }
|
|---|
| 185 | }
|
|---|
| 186 | // A succeeding character
|
|---|
| 187 | if(s_char > p_char) {
|
|---|
| 188 | ///ystem.err.println("s char is greater than p char");
|
|---|
| 189 | if(wildcard_matching) {
|
|---|
| 190 | ///ystem.err.println("Because we are wildcard matching we still have to match the rest of s incase of errors.");
|
|---|
| 191 | if(compareToPattern(s, p, s_index + 1, p_index + 1, wildcard_matching) != ERROR) {
|
|---|
| 192 | ///ystem.err.println("No error. Greater than.");
|
|---|
| 193 | return GREATER_THAN;
|
|---|
| 194 | }
|
|---|
| 195 | else {
|
|---|
| 196 | ///ystem.err.println("Error detected.");
|
|---|
| 197 | return ERROR;
|
|---|
| 198 | }
|
|---|
| 199 | }
|
|---|
| 200 | else {
|
|---|
| 201 | ///ystem.err.println("Greater than.");
|
|---|
| 202 | return GREATER_THAN;
|
|---|
| 203 | }
|
|---|
| 204 | }
|
|---|
| 205 | // Oh-No. Well, we'll just assume that string is less than pattern
|
|---|
| 206 | return LESS_THAN;
|
|---|
| 207 | }
|
|---|
| 208 |
|
|---|
| 209 | /** <i>true</i> if the filter should be applied, <i>false</i> to indicate the filter is turned off. */
|
|---|
| 210 | public boolean active;
|
|---|
| 211 | /** <i>true</i> if the matching for the first expression should be case sensitive, <i>false</i> otherwise. */
|
|---|
| 212 | public boolean casesense_one;
|
|---|
| 213 | /** <i>true</i> if the matching for the second expression should be case sensitive, <i>false</i> otherwise. */
|
|---|
| 214 | public boolean casesense_two;
|
|---|
| 215 | /** Used to determine the operation intended when applying two filters, and set using the values of the OPERATION_TYPE enumeration. */
|
|---|
| 216 | public boolean operation;
|
|---|
| 217 | /** Used to determine how the column this filter is applied to should be sorted, and set using the values of the SORT_TYPE enumeration. */
|
|---|
| 218 | public boolean sort;
|
|---|
| 219 | /** The method to be used for the first filter expression, set from the values of the METHOD_LIST enumeration. */
|
|---|
| 220 | public int method_one;
|
|---|
| 221 | /** The method to be used for the second filter expression, set from the values of the METHOD_LIST enumeration. */
|
|---|
| 222 | public int method_two;
|
|---|
| 223 | /** The value to be matched against for the first expression. */
|
|---|
| 224 | public String value_one;
|
|---|
| 225 | /** The value to be matched against for the second expression. */
|
|---|
| 226 | public String value_two;
|
|---|
| 227 | /** An element of the SORT_TYPE enumeration, indicates lowest to highest value column ordering. */
|
|---|
| 228 | public static final boolean ASCENDING = true;
|
|---|
| 229 | /** An element of the OPERATION_TYPE enumeration, indicates that both filter expressions must be met (conjunction). */
|
|---|
| 230 | public static final boolean AND = true;
|
|---|
| 231 | /** An element of the SORT_TYPE enumeration, indicates highest to lowest value column ordering. */
|
|---|
| 232 | public static final boolean DESCENDING = false;
|
|---|
| 233 | /** An element of the OPERATION_TYPE enumeration, indicates that either (or both) filter expressions must be met (disjunction). */
|
|---|
| 234 | public static final boolean OR = false;
|
|---|
| 235 | /** An enumeration of symbolic names of various matching methods. */
|
|---|
| 236 | public static final String METHOD_LIST[] = {"eqeq", "!eq", "<", "<eq", ">", ">eq", "^", "!^", "$", "!$", "?", "!?"};
|
|---|
| 237 | /** Default Constructor. */
|
|---|
| 238 | public Autofilter() {
|
|---|
| 239 | operation = OR;
|
|---|
| 240 | sort = ASCENDING;
|
|---|
| 241 | }
|
|---|
| 242 | /** Determine if this filter is currently active.
|
|---|
| 243 | * @return <i>true</i> if it is active, <i>false</i> otherwise.
|
|---|
| 244 | */
|
|---|
| 245 | public boolean active() {
|
|---|
| 246 | return active;
|
|---|
| 247 | }
|
|---|
| 248 | /** Determine if this list of values (for a certain cell) passes the filter.
|
|---|
| 249 | * @param values An <strong>ArrayList</strong> of values sourced from a single cell in the associated column.
|
|---|
| 250 | * @return <i>true</i> if the values match and should be displayed, <i>false</i> otherwise.
|
|---|
| 251 | */
|
|---|
| 252 | public boolean filter(ArrayList values) {
|
|---|
| 253 | boolean result = false;
|
|---|
| 254 | if(value_one != null) {
|
|---|
| 255 | result = filter(values, method_one, value_one, casesense_one);
|
|---|
| 256 | if(result) {
|
|---|
| 257 | if(operation == AND && value_two != null) {
|
|---|
| 258 | result = filter(values, method_two, value_two, casesense_two);
|
|---|
| 259 | }
|
|---|
| 260 | }
|
|---|
| 261 | else if(operation == OR && value_two != null) {
|
|---|
| 262 | result = filter(values, method_two, value_two, casesense_two);
|
|---|
| 263 | }
|
|---|
| 264 | }
|
|---|
| 265 | return result;
|
|---|
| 266 | }
|
|---|
| 267 | /** Set the current activity state of this filter.
|
|---|
| 268 | * @param active The new state of this filter, <i>true</i> to activate, <i>false</i> otherwise.
|
|---|
| 269 | */
|
|---|
| 270 | public void setActive(boolean active) {
|
|---|
| 271 | this.active = active;
|
|---|
| 272 | }
|
|---|
| 273 | /** Set one of the filter expressions using the given information.
|
|---|
| 274 | * @param number The number of the filter you wish to set as an <i>int</i>. Either 1 or 2.
|
|---|
| 275 | * @param method An <i>int</i> indicating the method to be used when matching.
|
|---|
| 276 | * @param value The <strong>String</strong> to be matched against.
|
|---|
| 277 | * @param casesense <i>true</i> if this expression should be case sensitive, <i>false</i> otherwise.
|
|---|
| 278 | */
|
|---|
| 279 | public void setFilter(int number, int method, String value, boolean casesense) {
|
|---|
| 280 | if(!casesense && value != null) {
|
|---|
| 281 | value = value.toLowerCase();
|
|---|
| 282 | }
|
|---|
| 283 | if(number == 1) {
|
|---|
| 284 | casesense_one = casesense;
|
|---|
| 285 | method_one = method;
|
|---|
| 286 | value_one = value;
|
|---|
| 287 | }
|
|---|
| 288 | else {
|
|---|
| 289 | casesense_two = casesense;
|
|---|
| 290 | method_two = method;
|
|---|
| 291 | value_two = value;
|
|---|
| 292 | }
|
|---|
| 293 | }
|
|---|
| 294 | /** Set the operation to be used to join the two filters (if a second filter is set).
|
|---|
| 295 | * @param operation <i>true</i> for conjunct filters, <i>false</i> for disjunct.
|
|---|
| 296 | */
|
|---|
| 297 | public void setOperation(boolean operation) {
|
|---|
| 298 | this.operation = operation;
|
|---|
| 299 | }
|
|---|
| 300 | /** Set the sort order of this column.
|
|---|
| 301 | * @param sort <i>true</i> for ascending sort, <i>false</i> for descending.
|
|---|
| 302 | */
|
|---|
| 303 | public void setSort(boolean sort) {
|
|---|
| 304 | this.sort = sort;
|
|---|
| 305 | }
|
|---|
| 306 | /** Decide whether a row should be displayed or filtered. The result depends on the selector set with setFilterType.
|
|---|
| 307 | * @param values An <strong>ArrayList</strong> of values to be checked against the filter.
|
|---|
| 308 | * @param method The method of matching to be used, as an <i>int</i>.
|
|---|
| 309 | * @param target The <strong>String</Strong> to match against.
|
|---|
| 310 | * @param casesense <i>true</i> if the match is to be case sensitive, <i>false</i> otherwise.
|
|---|
| 311 | * @return <i>true</i> to display the row, <i>false</i> to hide it.
|
|---|
| 312 | */
|
|---|
| 313 | public boolean filter(ArrayList values, int method, String target, boolean casesense) {
|
|---|
| 314 | boolean result = false;
|
|---|
| 315 | // There are several special cases when the filter always returns turn, such as when the target is the wildcard character.
|
|---|
| 316 | if(target == null || target.length() == 0 || target.equals("*")) {
|
|---|
| 317 | result = true;
|
|---|
| 318 | }
|
|---|
| 319 | else {
|
|---|
| 320 | // For each value in the list...
|
|---|
| 321 | for(int i = 0; i < values.size(); i++) {
|
|---|
| 322 | boolean pass;
|
|---|
| 323 | String source;
|
|---|
| 324 | // Account for case sensitivity.
|
|---|
| 325 | if(casesense) {
|
|---|
| 326 | source = values.get(i).toString();
|
|---|
| 327 | }
|
|---|
| 328 | else {
|
|---|
| 329 | source = values.get(i).toString().toLowerCase();
|
|---|
| 330 | }
|
|---|
| 331 | ///ystem.err.println("Testing " + source + " against pattern " + target);
|
|---|
| 332 | // Perform the match, based on the selected method.
|
|---|
| 333 | switch(method) {
|
|---|
| 334 | case 1: // !EQ
|
|---|
| 335 | pass = (compareToPattern(source, target) != 0);
|
|---|
| 336 | break;
|
|---|
| 337 | case 2: // <
|
|---|
| 338 | pass = (compareToPattern(source, target) < 0);
|
|---|
| 339 | break;
|
|---|
| 340 | case 3: // <eq
|
|---|
| 341 | pass = (compareToPattern(source, target) <= 0);
|
|---|
| 342 | break;
|
|---|
| 343 | case 4: // >
|
|---|
| 344 | pass = (compareToPattern(source, target) > 0);
|
|---|
| 345 | break;
|
|---|
| 346 | case 5: // >eq
|
|---|
| 347 | pass = (compareToPattern(source, target) >= 0);
|
|---|
| 348 | break;
|
|---|
| 349 | case 6: // ^
|
|---|
| 350 | pass = source.startsWith(target);
|
|---|
| 351 | break;
|
|---|
| 352 | case 7: // !^
|
|---|
| 353 | pass = !(source.startsWith(target));
|
|---|
| 354 | break;
|
|---|
| 355 | case 8: // $
|
|---|
| 356 | pass = source.endsWith(target);
|
|---|
| 357 | break;
|
|---|
| 358 | case 9: // !$
|
|---|
| 359 | pass = !(source.endsWith(target));
|
|---|
| 360 | break;
|
|---|
| 361 | case 10: // ?
|
|---|
| 362 | pass = (source.indexOf(target) != -1);
|
|---|
| 363 | break;
|
|---|
| 364 | case 11: // !?
|
|---|
| 365 | pass = (source.indexOf(target) == -1);
|
|---|
| 366 | break;
|
|---|
| 367 | default: // EQEQ
|
|---|
| 368 | pass = (compareToPattern(source, target) == 0);
|
|---|
| 369 | break;
|
|---|
| 370 | }
|
|---|
| 371 | result = result || pass;
|
|---|
| 372 | }
|
|---|
| 373 | }
|
|---|
| 374 | return result;
|
|---|
| 375 | }
|
|---|
| 376 | /** Produce a textual representation of this autofilter.
|
|---|
| 377 | * @return A <strong>String</strong> displaying details of this autofilter.
|
|---|
| 378 | */
|
|---|
| 379 | public String toString() {
|
|---|
| 380 | String result = "One: " + method_one + " - " + value_one + " - " + casesense_one;
|
|---|
| 381 | if(value_two != null) {
|
|---|
| 382 | result = result + "\n" + "Operation: " + (operation?"AND":"OR") + "\n";
|
|---|
| 383 | result = result + "Two: " + method_two + " - " + value_two + " - " + casesense_two;
|
|---|
| 384 | }
|
|---|
| 385 | return result;
|
|---|
| 386 | }
|
|---|
| 387 | }
|
|---|
