1 | /*
|
---|
2 | * KEAModelBuilder.java
|
---|
3 | * Copyright (C) 2001 Eibe Frank
|
---|
4 | *
|
---|
5 | * This program is free software; you can redistribute it and/or modify
|
---|
6 | * it under the terms of the GNU General Public License as published by
|
---|
7 | * the Free Software Foundation; either version 2 of the License, or
|
---|
8 | * (at your option) any later version.
|
---|
9 | *
|
---|
10 | * This program is distributed in the hope that it will be useful,
|
---|
11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
13 | * GNU General Public License for more details.
|
---|
14 | *
|
---|
15 | * You should have received a copy of the GNU General Public License
|
---|
16 | * along with this program; if not, write to the Free Software
|
---|
17 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
18 | */
|
---|
19 |
|
---|
20 | import java.io.*;
|
---|
21 | import java.util.*;
|
---|
22 | import weka.core.*;
|
---|
23 | import weka.filters.*;
|
---|
24 |
|
---|
25 | /**
|
---|
26 | * Builds a keyphrase extraction model from the documents in a given
|
---|
27 | * directory. Assumes that the file names for the documents end with
|
---|
28 | * ".txt". Assumes that files containing corresponding
|
---|
29 | * author-assigned keyphrases end with ".key". Optionally an encoding
|
---|
30 | * for the documents/keyphrases can be defined (e.g. for Chinese
|
---|
31 | * text).
|
---|
32 | *
|
---|
33 | * Valid options are:<p>
|
---|
34 | *
|
---|
35 | * -l "directory name"<br>
|
---|
36 | * Specifies name of directory.<p>
|
---|
37 | *
|
---|
38 | * -m "model name"<br>
|
---|
39 | * Specifies name of model.<p>
|
---|
40 | *
|
---|
41 | * -e "encoding"<br>
|
---|
42 | * Specifies encoding.<p>
|
---|
43 | *
|
---|
44 | * -d<br>
|
---|
45 | * Turns debugging mode on.<p>
|
---|
46 | *
|
---|
47 | * -k<br>
|
---|
48 | * Use keyphrase frequency statistic.<p>
|
---|
49 | *
|
---|
50 | * -p<br>
|
---|
51 | * Disallow internal periods.<p>
|
---|
52 | *
|
---|
53 | * -x "length"<br>
|
---|
54 | * Sets maximum phrase length (default: 3).<p>
|
---|
55 | *
|
---|
56 | * -y "length"<br>
|
---|
57 | * Sets minimum phrase length (default: 1).<p>
|
---|
58 | *
|
---|
59 | * -o "number"<br>
|
---|
60 | * The minimum number of times a phrase needs to occur (default: 2). <p>
|
---|
61 | *
|
---|
62 | * -s "name of class implementing list of stop words"<br>
|
---|
63 | * Sets list of stop words to used (default: StopwordsEnglish).<p>
|
---|
64 | *
|
---|
65 | * -t "name of class implementing stemmer"<br>
|
---|
66 | * Sets stemmer to use (default: IteratedLovinsStemmer). <p>
|
---|
67 | *
|
---|
68 | * -n<br>
|
---|
69 | * Do not check for proper nouns. <p>
|
---|
70 | *
|
---|
71 | * @author Eibe Frank ([email protected])
|
---|
72 | * @version 1.0
|
---|
73 | */
|
---|
74 | public class KEAModelBuilder implements OptionHandler {
|
---|
75 |
|
---|
76 | /** Name of directory */
|
---|
77 | String m_dirName = null;
|
---|
78 |
|
---|
79 | /** Name of model */
|
---|
80 | String m_modelName = null;
|
---|
81 |
|
---|
82 | /** Encoding */
|
---|
83 | String m_encoding = "default";
|
---|
84 |
|
---|
85 | /** Debugging mode? */
|
---|
86 | boolean m_debug = false;
|
---|
87 |
|
---|
88 | /** Use keyphrase frequency attribute? */
|
---|
89 | boolean m_useKFrequency = false;
|
---|
90 |
|
---|
91 | /** Disallow internal periods? */
|
---|
92 | boolean m_disallowIPeriods = false;
|
---|
93 |
|
---|
94 | /** The maximum length of phrases */
|
---|
95 | private int m_MaxPhraseLength = 3;
|
---|
96 |
|
---|
97 | /** The minimum length of phrases */
|
---|
98 | private int m_MinPhraseLength = 1;
|
---|
99 |
|
---|
100 | /** The minimum number of occurences of a phrase */
|
---|
101 | private int m_MinNumOccur = 2;
|
---|
102 |
|
---|
103 | /** The KEA filter object */
|
---|
104 | KEAFilter m_KEAFilter = null;
|
---|
105 |
|
---|
106 | /** The stemmer to be used */
|
---|
107 | private Stemmer m_Stemmer = new IteratedLovinsStemmer();
|
---|
108 |
|
---|
109 | /** The list of stop words to be used */
|
---|
110 | private Stopwords m_Stopwords = new StopwordsEnglish();
|
---|
111 |
|
---|
112 | /** Determines whether check for proper nouns is performed */
|
---|
113 | private boolean m_CheckForProperNouns = true;
|
---|
114 |
|
---|
115 | /**
|
---|
116 | * Get the M_CheckProperNouns value.
|
---|
117 | * @return the M_CheckProperNouns value.
|
---|
118 | */
|
---|
119 | public boolean getCheckForProperNouns() {
|
---|
120 | return m_CheckForProperNouns;
|
---|
121 | }
|
---|
122 |
|
---|
123 | /**
|
---|
124 | * Set the M_CheckProperNouns value.
|
---|
125 | * @param newM_CheckProperNouns The new M_CheckProperNouns value.
|
---|
126 | */
|
---|
127 | public void setCheckForProperNouns(boolean newM_CheckProperNouns) {
|
---|
128 | this.m_CheckForProperNouns = newM_CheckProperNouns;
|
---|
129 | }
|
---|
130 |
|
---|
131 | /**
|
---|
132 | * Get the M_Stopwords value.
|
---|
133 | * @return the M_Stopwords value.
|
---|
134 | */
|
---|
135 | public Stopwords getStopwords() {
|
---|
136 |
|
---|
137 | return m_Stopwords;
|
---|
138 | }
|
---|
139 |
|
---|
140 | /**
|
---|
141 | * Set the M_Stopwords value.
|
---|
142 | * @param newM_Stopwords The new M_Stopwords value.
|
---|
143 | */
|
---|
144 | public void setStopwords(Stopwords newM_Stopwords) {
|
---|
145 |
|
---|
146 | this.m_Stopwords = newM_Stopwords;
|
---|
147 | }
|
---|
148 |
|
---|
149 |
|
---|
150 | /**
|
---|
151 | * Get the Stemmer value.
|
---|
152 | * @return the Stemmer value.
|
---|
153 | */
|
---|
154 | public Stemmer getStemmer() {
|
---|
155 |
|
---|
156 | return m_Stemmer;
|
---|
157 | }
|
---|
158 |
|
---|
159 | /**
|
---|
160 | * Set the Stemmer value.
|
---|
161 | * @param newStemmer The new Stemmer value.
|
---|
162 | */
|
---|
163 | public void setStemmer(Stemmer newStemmer) {
|
---|
164 |
|
---|
165 | this.m_Stemmer = newStemmer;
|
---|
166 | }
|
---|
167 |
|
---|
168 | /**
|
---|
169 | * Get the value of MinNumOccur.
|
---|
170 | *
|
---|
171 | * @return Value of MinNumOccur.
|
---|
172 | */
|
---|
173 | public int getMinNumOccur() {
|
---|
174 |
|
---|
175 | return m_MinNumOccur;
|
---|
176 | }
|
---|
177 |
|
---|
178 | /**
|
---|
179 | * Set the value of MinNumOccur.
|
---|
180 | *
|
---|
181 | * @param newMinNumOccur Value to assign to MinNumOccur.
|
---|
182 | */
|
---|
183 | public void setMinNumOccur(int newMinNumOccur) {
|
---|
184 |
|
---|
185 | m_MinNumOccur = newMinNumOccur;
|
---|
186 | }
|
---|
187 |
|
---|
188 | /**
|
---|
189 | * Get the value of MaxPhraseLength.
|
---|
190 | *
|
---|
191 | * @return Value of MaxPhraseLength.
|
---|
192 | */
|
---|
193 | public int getMaxPhraseLength() {
|
---|
194 |
|
---|
195 | return m_MaxPhraseLength;
|
---|
196 | }
|
---|
197 |
|
---|
198 | /**
|
---|
199 | * Set the value of MaxPhraseLength.
|
---|
200 | *
|
---|
201 | * @param newMaxPhraseLength Value to assign to MaxPhraseLength.
|
---|
202 | */
|
---|
203 | public void setMaxPhraseLength(int newMaxPhraseLength) {
|
---|
204 |
|
---|
205 | m_MaxPhraseLength = newMaxPhraseLength;
|
---|
206 | }
|
---|
207 |
|
---|
208 | /**
|
---|
209 | * Get the value of MinPhraseLength.
|
---|
210 | *
|
---|
211 | * @return Value of MinPhraseLength.
|
---|
212 | */
|
---|
213 | public int getMinPhraseLength() {
|
---|
214 |
|
---|
215 | return m_MinPhraseLength;
|
---|
216 | }
|
---|
217 |
|
---|
218 | /**
|
---|
219 | * Set the value of MinPhraseLength.
|
---|
220 | *
|
---|
221 | * @param newMinPhraseLength Value to assign to MinPhraseLength.
|
---|
222 | */
|
---|
223 | public void setMinPhraseLength(int newMinPhraseLength) {
|
---|
224 |
|
---|
225 | m_MinPhraseLength = newMinPhraseLength;
|
---|
226 | }
|
---|
227 |
|
---|
228 | /**
|
---|
229 | * Get the value of disallowIPeriods.
|
---|
230 | *
|
---|
231 | * @return Value of disallowIPeriods.
|
---|
232 | */
|
---|
233 | public boolean getDisallowIPeriods() {
|
---|
234 |
|
---|
235 | return m_disallowIPeriods;
|
---|
236 | }
|
---|
237 |
|
---|
238 | /**
|
---|
239 | * Set the value of disallowIPeriods.
|
---|
240 | *
|
---|
241 | * @param newdisallowIPeriods Value to assign to disallowIPeriods.
|
---|
242 | */
|
---|
243 | public void setDisallowIPeriods(boolean newdisallowIPeriods) {
|
---|
244 |
|
---|
245 | m_disallowIPeriods = newdisallowIPeriods;
|
---|
246 | }
|
---|
247 |
|
---|
248 | /**
|
---|
249 | * Get the value of useKFrequency.
|
---|
250 | *
|
---|
251 | * @return Value of useKFrequency.
|
---|
252 | */
|
---|
253 | public boolean getUseKFrequency() {
|
---|
254 |
|
---|
255 | return m_useKFrequency;
|
---|
256 | }
|
---|
257 |
|
---|
258 | /**
|
---|
259 | * Set the value of useKFrequency.
|
---|
260 | *
|
---|
261 | * @param newuseKFrequency Value to assign to useKFrequency.
|
---|
262 | */
|
---|
263 | public void setUseKFrequency(boolean newuseKFrequency) {
|
---|
264 |
|
---|
265 | m_useKFrequency = newuseKFrequency;
|
---|
266 | }
|
---|
267 |
|
---|
268 | /**
|
---|
269 | * Get the value of debug.
|
---|
270 | *
|
---|
271 | * @return Value of debug.
|
---|
272 | */
|
---|
273 | public boolean getDebug() {
|
---|
274 |
|
---|
275 | return m_debug;
|
---|
276 | }
|
---|
277 |
|
---|
278 | /**
|
---|
279 | * Set the value of debug.
|
---|
280 | *
|
---|
281 | * @param newdebug Value to assign to debug.
|
---|
282 | */
|
---|
283 | public void setDebug(boolean newdebug) {
|
---|
284 |
|
---|
285 | m_debug = newdebug;
|
---|
286 | }
|
---|
287 |
|
---|
288 | /**
|
---|
289 | * Get the value of encoding.
|
---|
290 | *
|
---|
291 | * @return Value of encoding.
|
---|
292 | */
|
---|
293 | public String getEncoding() {
|
---|
294 |
|
---|
295 | return m_encoding;
|
---|
296 | }
|
---|
297 |
|
---|
298 | /**
|
---|
299 | * Set the value of encoding.
|
---|
300 | *
|
---|
301 | * @param newencoding Value to assign to encoding.
|
---|
302 | */
|
---|
303 | public void setEncoding(String newencoding) {
|
---|
304 |
|
---|
305 | m_encoding = newencoding;
|
---|
306 | }
|
---|
307 |
|
---|
308 | /**
|
---|
309 | * Get the value of modelName.
|
---|
310 | *
|
---|
311 | * @return Value of modelName.
|
---|
312 | */
|
---|
313 | public String getModelName() {
|
---|
314 |
|
---|
315 | return m_modelName;
|
---|
316 | }
|
---|
317 |
|
---|
318 | /**
|
---|
319 | * Set the value of modelName.
|
---|
320 | *
|
---|
321 | * @param newmodelName Value to assign to modelName.
|
---|
322 | */
|
---|
323 | public void setModelName(String newmodelName) {
|
---|
324 |
|
---|
325 | m_modelName = newmodelName;
|
---|
326 | }
|
---|
327 |
|
---|
328 | /**
|
---|
329 | * Get the value of dirName.
|
---|
330 | *
|
---|
331 | * @return Value of dirName.
|
---|
332 | */
|
---|
333 | public String getDirName() {
|
---|
334 |
|
---|
335 | return m_dirName;
|
---|
336 | }
|
---|
337 |
|
---|
338 | /**
|
---|
339 | * Set the value of dirName.
|
---|
340 | *
|
---|
341 | * @param newdirName Value to assign to dirName.
|
---|
342 | */
|
---|
343 | public void setDirName(String newdirName) {
|
---|
344 |
|
---|
345 | m_dirName = newdirName;
|
---|
346 | }
|
---|
347 |
|
---|
348 | /**
|
---|
349 | * Parses a given list of options controlling the behaviour of this object.
|
---|
350 | * Valid options are:<p>
|
---|
351 | *
|
---|
352 | * -l "directory name" <br>
|
---|
353 | * Specifies name of directory.<p>
|
---|
354 | *
|
---|
355 | * -m "model name" <br>
|
---|
356 | * Specifies name of model.<p>
|
---|
357 | *
|
---|
358 | * -e "encoding" <br>
|
---|
359 | * Specifies encoding.<p>
|
---|
360 | *
|
---|
361 | * -d<br>
|
---|
362 | * Turns debugging mode on.<p>
|
---|
363 | *
|
---|
364 | * -k<br>
|
---|
365 | * Use keyphrase frequency statistic.<p>
|
---|
366 | *
|
---|
367 | * -p<br>
|
---|
368 | * Disallow internal periods. <p>
|
---|
369 | *
|
---|
370 | * -x "length"<br>
|
---|
371 | * Sets maximum phrase length (default: 3).<p>
|
---|
372 | *
|
---|
373 | * -y "length"<br>
|
---|
374 | * Sets minimum phrase length (default: 3).<p>
|
---|
375 | *
|
---|
376 | * -o "number"<br>
|
---|
377 | * The minimum number of times a phrase needs to occur (default: 2). <p>
|
---|
378 | *
|
---|
379 | * -s "name of class implementing list of stop words"<br>
|
---|
380 | * Sets list of stop words to used (default: StopwordsEnglish).<p>
|
---|
381 | *
|
---|
382 | * -t "name of class implementing stemmer"<br>
|
---|
383 | * Sets stemmer to use (default: IteratedLovinsStemmer). <p>
|
---|
384 | *
|
---|
385 | * -n<br>
|
---|
386 | * Do not check for proper nouns. <p>
|
---|
387 | *
|
---|
388 | * @param options the list of options as an array of strings
|
---|
389 | * @exception Exception if an option is not supported
|
---|
390 | */
|
---|
391 | public void setOptions(String[] options) throws Exception {
|
---|
392 |
|
---|
393 | String dirName = Utils.getOption('l', options);
|
---|
394 | if (dirName.length() > 0) {
|
---|
395 | setDirName(dirName);
|
---|
396 | } else {
|
---|
397 | setDirName(null);
|
---|
398 | throw new Exception("Name of directory required argument.");
|
---|
399 | }
|
---|
400 | String modelName = Utils.getOption('m', options);
|
---|
401 | if (modelName.length() > 0) {
|
---|
402 | setModelName(modelName);
|
---|
403 | } else {
|
---|
404 | setModelName(null);
|
---|
405 | throw new Exception("Name of model required argument.");
|
---|
406 | }
|
---|
407 | String encoding = Utils.getOption('e', options);
|
---|
408 | if (encoding.length() > 0) {
|
---|
409 | setEncoding(encoding);
|
---|
410 | } else {
|
---|
411 | setEncoding("default");
|
---|
412 | }
|
---|
413 | String maxPhraseLengthString = Utils.getOption('x', options);
|
---|
414 | if (maxPhraseLengthString.length() > 0) {
|
---|
415 | setMaxPhraseLength(Integer.parseInt(maxPhraseLengthString));
|
---|
416 | } else {
|
---|
417 | setMaxPhraseLength(3);
|
---|
418 | }
|
---|
419 | String minPhraseLengthString = Utils.getOption('y', options);
|
---|
420 | if (minPhraseLengthString.length() > 0) {
|
---|
421 | setMinPhraseLength(Integer.parseInt(minPhraseLengthString));
|
---|
422 | } else {
|
---|
423 | setMinPhraseLength(1);
|
---|
424 | }
|
---|
425 | String minNumOccurString = Utils.getOption('o', options);
|
---|
426 | if (minNumOccurString.length() > 0) {
|
---|
427 | setMinNumOccur(Integer.parseInt(minNumOccurString));
|
---|
428 | } else {
|
---|
429 | setMinNumOccur(2);
|
---|
430 | }
|
---|
431 | String stopwordsString = Utils.getOption('s', options);
|
---|
432 | if (stopwordsString.length() > 0) {
|
---|
433 | setStopwords((Stopwords)Class.forName(stopwordsString).newInstance());
|
---|
434 | }
|
---|
435 | String stemmerString = Utils.getOption('t', options);
|
---|
436 | if (stemmerString.length() > 0) {
|
---|
437 | setStemmer((Stemmer)Class.forName(stemmerString).newInstance());
|
---|
438 | }
|
---|
439 | setDebug(Utils.getFlag('d', options));
|
---|
440 | setUseKFrequency(Utils.getFlag('k', options));
|
---|
441 | setDisallowIPeriods(Utils.getFlag('p', options));
|
---|
442 | setCheckForProperNouns(!Utils.getFlag('n', options));
|
---|
443 | Utils.checkForRemainingOptions(options);
|
---|
444 | }
|
---|
445 |
|
---|
446 | /**
|
---|
447 | * Gets the current option settings.
|
---|
448 | *
|
---|
449 | * @return an array of strings suitable for passing to setOptions
|
---|
450 | */
|
---|
451 | public String [] getOptions() {
|
---|
452 |
|
---|
453 | String [] options = new String [20];
|
---|
454 | int current = 0;
|
---|
455 |
|
---|
456 | options[current++] = "-l";
|
---|
457 | options[current++] = "" + (getDirName());
|
---|
458 | options[current++] = "-m";
|
---|
459 | options[current++] = "" + (getModelName());
|
---|
460 | options[current++] = "-e";
|
---|
461 | options[current++] = "" + (getEncoding());
|
---|
462 | if (getUseKFrequency()) {
|
---|
463 | options[current++] = "-k";
|
---|
464 | }
|
---|
465 | if (getDebug()) {
|
---|
466 | options[current++] = "-d";
|
---|
467 | }
|
---|
468 | if (getDisallowIPeriods()) {
|
---|
469 | options[current++] = "-p";
|
---|
470 | }
|
---|
471 | options[current++] = "-x";
|
---|
472 | options[current++] = "" + (getMaxPhraseLength());
|
---|
473 | options[current++] = "-y";
|
---|
474 | options[current++] = "" + (getMinPhraseLength());
|
---|
475 | options[current++] = "-o";
|
---|
476 | options[current++] = "" + (getMinNumOccur());
|
---|
477 | options[current++] = "-s";
|
---|
478 | options[current++] = "" + (getStopwords().getClass().getName());
|
---|
479 | options[current++] = "-t";
|
---|
480 | options[current++] = "" + (getStemmer().getClass().getName());
|
---|
481 | if (getCheckForProperNouns()) {
|
---|
482 | options[current++] = "-n";
|
---|
483 | }
|
---|
484 |
|
---|
485 | while (current < options.length) {
|
---|
486 | options[current++] = "";
|
---|
487 | }
|
---|
488 | return options;
|
---|
489 | }
|
---|
490 |
|
---|
491 | /**
|
---|
492 | * Returns an enumeration describing the available options.
|
---|
493 | *
|
---|
494 | * @return an enumeration of all the available options
|
---|
495 | */
|
---|
496 | public Enumeration listOptions() {
|
---|
497 |
|
---|
498 | Vector newVector = new Vector(12);
|
---|
499 |
|
---|
500 | newVector.addElement(new Option(
|
---|
501 | "\tSpecifies name of directory.",
|
---|
502 | "l", 1, "-l <directory name>"));
|
---|
503 | newVector.addElement(new Option(
|
---|
504 | "\tSpecifies name of model.",
|
---|
505 | "m", 1, "-m <model name>"));
|
---|
506 | newVector.addElement(new Option(
|
---|
507 | "\tSpecifies encoding.",
|
---|
508 | "e", 1, "-e <encoding>"));
|
---|
509 | newVector.addElement(new Option(
|
---|
510 | "\tTurns debugging mode on.",
|
---|
511 | "d", 0, "-d"));
|
---|
512 | newVector.addElement(new Option(
|
---|
513 | "\tUse keyphrase frequency statistic.",
|
---|
514 | "k", 0, "-k"));
|
---|
515 | newVector.addElement(new Option(
|
---|
516 | "\tDisallow internal periods.",
|
---|
517 | "p", 0, "-p"));
|
---|
518 | newVector.addElement(new Option(
|
---|
519 | "\tSets the maximum phrase length (default: 3).",
|
---|
520 | "x", 1, "-x <length>"));
|
---|
521 | newVector.addElement(new Option(
|
---|
522 | "\tSets the minimum phrase length (default: 1).",
|
---|
523 | "y", 1, "-y <length>"));
|
---|
524 | newVector.addElement(new Option(
|
---|
525 | "\tSet the minimum number of occurences (default: 2).",
|
---|
526 | "o", 1, "-o"));
|
---|
527 | newVector.addElement(new Option(
|
---|
528 | "\tSets the list of stopwords to use (default: StopwordsEnglish).",
|
---|
529 | "s", 1, "-s <name of stopwords class>"));
|
---|
530 | newVector.addElement(new Option(
|
---|
531 | "\tSet the stemmer to use (default: IteratedLovinsStemmer).",
|
---|
532 | "t", 1, "-t <name of stemmer class>"));
|
---|
533 | newVector.addElement(new Option(
|
---|
534 | "\tDo not check for proper nouns.",
|
---|
535 | "n", 0, "-n"));
|
---|
536 |
|
---|
537 | return newVector.elements();
|
---|
538 | }
|
---|
539 |
|
---|
540 | /**
|
---|
541 | * Collects the stems of the file names.
|
---|
542 | */
|
---|
543 | public Hashtable collectStems() throws Exception {
|
---|
544 |
|
---|
545 | Hashtable stems = new Hashtable();
|
---|
546 |
|
---|
547 | try {
|
---|
548 | File dir = new File(m_dirName);
|
---|
549 | String[] files = dir.list();
|
---|
550 | for (int i = 0; i < files.length; i++) {
|
---|
551 | if (files[i].endsWith(".key") ||
|
---|
552 | files[i].endsWith(".txt")) {
|
---|
553 | String stem = files[i].substring(0, files[i].length() - 4);
|
---|
554 | if (!stems.containsKey(stem)) {
|
---|
555 | stems.put(stem, new Double(0));
|
---|
556 | }
|
---|
557 | }
|
---|
558 | }
|
---|
559 | } catch (Exception e) {
|
---|
560 | throw new Exception("Problem opening directory " + m_dirName);
|
---|
561 | }
|
---|
562 | return stems;
|
---|
563 | }
|
---|
564 |
|
---|
565 | /**
|
---|
566 | * Builds the model from the files
|
---|
567 | */
|
---|
568 | public void buildModel(Hashtable stems) throws Exception {
|
---|
569 |
|
---|
570 | // Check whether there is actually any data
|
---|
571 | if (stems.size() == 0) {
|
---|
572 | throw new Exception("Couldn't find any data!");
|
---|
573 | }
|
---|
574 |
|
---|
575 | FastVector atts = new FastVector(2);
|
---|
576 | atts.addElement(new Attribute("doc", null));
|
---|
577 | atts.addElement(new Attribute("keyphrases", null));
|
---|
578 | Instances data = new Instances("keyphrase_training_data", atts, 0);
|
---|
579 |
|
---|
580 | // Build model
|
---|
581 | m_KEAFilter = new KEAFilter();
|
---|
582 | m_KEAFilter.setDebug(m_debug);
|
---|
583 | m_KEAFilter.setDisallowInternalPeriods(getDisallowIPeriods());
|
---|
584 | m_KEAFilter.setKFused(getUseKFrequency());
|
---|
585 | m_KEAFilter.setMaxPhraseLength(getMaxPhraseLength());
|
---|
586 | m_KEAFilter.setMinPhraseLength(getMinPhraseLength());
|
---|
587 | m_KEAFilter.setMinNumOccur(getMinNumOccur());
|
---|
588 | m_KEAFilter.setInputFormat(data);
|
---|
589 | m_KEAFilter.setStemmer(getStemmer());
|
---|
590 | m_KEAFilter.setStopwords(getStopwords());
|
---|
591 | m_KEAFilter.setCheckForProperNouns(getCheckForProperNouns());
|
---|
592 | Enumeration elem = stems.keys();
|
---|
593 | while (elem.hasMoreElements()) {
|
---|
594 | String str = (String)elem.nextElement();
|
---|
595 | double[] newInst = new double[2];
|
---|
596 | try {
|
---|
597 | File txt = new File(m_dirName + "/" + str + ".txt");
|
---|
598 | InputStreamReader is;
|
---|
599 | if (!m_encoding.equals("default")) {
|
---|
600 | is = new InputStreamReader(new FileInputStream(txt), m_encoding);
|
---|
601 | } else {
|
---|
602 | is = new InputStreamReader(new FileInputStream(txt));
|
---|
603 | }
|
---|
604 | StringBuffer txtStr = new StringBuffer();
|
---|
605 | int c;
|
---|
606 | while ((c = is.read()) != -1) {
|
---|
607 | txtStr.append((char)c);
|
---|
608 | }
|
---|
609 | newInst[0] = (double)data.attribute(0).addStringValue(txtStr.toString());
|
---|
610 | } catch (Exception e) {
|
---|
611 | if (m_debug) {
|
---|
612 | System.err.println("Can't find document for stem " + str + ".");
|
---|
613 | }
|
---|
614 | newInst[0] = Instance.missingValue();
|
---|
615 | }
|
---|
616 | try {
|
---|
617 | File key = new File(m_dirName + "/" + str + ".key");
|
---|
618 | InputStreamReader is;
|
---|
619 | if (!m_encoding.equals("default")) {
|
---|
620 | is = new InputStreamReader(new FileInputStream(key), m_encoding);
|
---|
621 | } else {
|
---|
622 | is = new InputStreamReader(new FileInputStream(key));
|
---|
623 | }
|
---|
624 | StringBuffer keyStr = new StringBuffer();
|
---|
625 | int c;
|
---|
626 | while ((c = is.read()) != -1) {
|
---|
627 | keyStr.append((char)c);
|
---|
628 | }
|
---|
629 | newInst[1] = (double)data.attribute(1).addStringValue(keyStr.toString());
|
---|
630 | } catch (Exception e) {
|
---|
631 | if (m_debug) {
|
---|
632 | System.err.println("Can't find keyphrases for stem " + str + ".");
|
---|
633 | }
|
---|
634 | newInst[1] = Instance.missingValue();
|
---|
635 | }
|
---|
636 | data.add(new Instance(1.0, newInst));
|
---|
637 | m_KEAFilter.input(data.instance(0));
|
---|
638 | data = data.stringFreeStructure();
|
---|
639 | }
|
---|
640 | m_KEAFilter.batchFinished();
|
---|
641 |
|
---|
642 | // Get rid of instances in filter
|
---|
643 | Instance dummy;
|
---|
644 | while ((dummy = m_KEAFilter.output()) != null) {};
|
---|
645 | }
|
---|
646 |
|
---|
647 | /**
|
---|
648 | * Saves the extraction model to the file.
|
---|
649 | */
|
---|
650 | public void saveModel() throws Exception {
|
---|
651 |
|
---|
652 | BufferedOutputStream bufferedOut =
|
---|
653 | new BufferedOutputStream(new FileOutputStream(m_modelName));
|
---|
654 | ObjectOutputStream out = new ObjectOutputStream(bufferedOut);
|
---|
655 | out.writeObject(m_KEAFilter);
|
---|
656 | out.flush();
|
---|
657 | out.close();
|
---|
658 | }
|
---|
659 |
|
---|
660 | /**
|
---|
661 | * The main method.
|
---|
662 | */
|
---|
663 | public static void main(String[] ops) {
|
---|
664 |
|
---|
665 | KEAModelBuilder kmb = new KEAModelBuilder();
|
---|
666 | try {
|
---|
667 | kmb.setOptions(ops);
|
---|
668 | System.err.print("Building model with options: ");
|
---|
669 | String[] optionSettings = kmb.getOptions();
|
---|
670 | for (int i = 0; i < optionSettings.length; i++) {
|
---|
671 | System.err.print(optionSettings[i] + " ");
|
---|
672 | }
|
---|
673 | System.err.println();
|
---|
674 | kmb.buildModel(kmb.collectStems());
|
---|
675 | kmb.saveModel();
|
---|
676 | } catch (Exception e) {
|
---|
677 | e.printStackTrace();
|
---|
678 | System.err.println(e.getMessage());
|
---|
679 | System.err.println("\nOptions:\n");
|
---|
680 | Enumeration enum = kmb.listOptions();
|
---|
681 | while (enum.hasMoreElements()) {
|
---|
682 | Option option = (Option) enum.nextElement();
|
---|
683 | System.err.println(option.synopsis());
|
---|
684 | System.err.println(option.description());
|
---|
685 | }
|
---|
686 | }
|
---|
687 | }
|
---|
688 | }
|
---|
689 |
|
---|