source: greenstone3/trunk/src/java/org/greenstone/gsdl3/service/PhindPhraseBrowse.java@ 14483

Last change on this file since 14483 was 14483, checked in by xiao, 17 years ago

make MGPPRetrieveWrapper and MGPPSearchWrapper static variables; synchronize the method findPhraseNumberFromWord() for search and getPhraseData() for retrieve.
  • Property svn:keywords set to Author Date Id Revision
File size: 16.5 KB
Line 
1/*
2 * PhindServices.java
3 * Copyright (C) 2002 New Zealand Digital Library, http://www.nzdl.org
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 */
19package org.greenstone.gsdl3.service;
20
21import org.greenstone.gsdl3.util.*;
22
23import org.greenstone.mgpp.*;
24import org.w3c.dom.Document;
25import org.w3c.dom.Node;
26import org.w3c.dom.Element;
27import org.w3c.dom.Text;
28
29import java.util.Vector;
30import java.util.HashMap;
31import java.io.File;
32
33import org.apache.log4j.*;
34
35/**
36 * PhindServices - the phind phrase browsing service
37 *
38 * @author <a href="mailto:[email protected]">Katherine Don</a>
39 * @version $Revision: 14483 $
40 */
41public class PhindPhraseBrowse
42 extends ServiceRack {
43
44 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.PhindPhraseBrowse.class.getName());
45
46 // the services on offer
47 private static final String PHIND_SERVICE = "PhindApplet";
48
49 private static MGPPRetrieveWrapper mgpp_retrieve_src=null;
50 private static MGPPSearchWrapper mgpp_search_src=null;
51 private String basepath = null;
52
53 private Element applet_description = null;
54
55 public PhindPhraseBrowse() {
56 if(this.mgpp_retrieve_src == null) {
57 this.mgpp_retrieve_src = new MGPPRetrieveWrapper();
58 }
59 if(this.mgpp_search_src == null) {
60 this.mgpp_search_src = new MGPPSearchWrapper();
61 }
62 // set up the default params
63 this.mgpp_search_src.setQueryLevel("Document");
64 this.mgpp_search_src.setReturnLevel("Document");
65 this.mgpp_search_src.setMaxDocs(5);
66 this.mgpp_search_src.setStem(false);
67 this.mgpp_search_src.setCase(true);
68 }
69
70 public void cleanUp() {
71 super.cleanUp();
72 this.mgpp_retrieve_src.unloadIndexData();
73 this.mgpp_search_src.unloadIndexData();
74 }
75
76 /** configure the service module
77 *
78 * @param info a DOM Element containing any config info for the service
79 * @return true if configured
80 */
81 public boolean configure(Element info, Element extra_info) {
82
83 if (!super.configure(info, extra_info)){
84 return false;
85 }
86
87 logger.info("configuring PhindPhraseBrowse");
88
89 // set up short_service_info_ - for now just has name and type
90 Element e = this.doc.createElement(GSXML.SERVICE_ELEM);
91 e.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_APPLET);
92 e.setAttribute(GSXML.NAME_ATT, PHIND_SERVICE);
93 this.short_service_info.appendChild(e);
94
95 // set up the static applet description
96
97 applet_description = this.doc.createElement(GSXML.SERVICE_ELEM);
98 applet_description.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_APPLET);
99 applet_description.setAttribute(GSXML.NAME_ATT, PHIND_SERVICE);
100
101 // add in the applet info for the phind applet
102 // need to make this dynamic - library names etc
103 // change the applet params - have a single param with the library name
104 // this is left blank at this end, and must be filled in by applet action - if the library name is not needed, this param is left out
105 // phindcgi param now is not complete - library must be prepended to it.
106 String app_info = "<"+GSXML.APPLET_ELEM+" CODEBASE='applet' CODE='org.greenstone.applet.phind.Phind.class' ARCHIVE='phind.jar, xercesImpl.jar, xml-apis.jar' WIDTH='500' HEIGHT='400'><PARAM NAME='library' VALUE=''/> <PARAM NAME='phindcgi' VALUE='?";
107 app_info += GSParams.ACTION +"=a&amp;"+GSParams.REQUEST_TYPE +"=r&amp;"+GSParams.SERVICE+"="+PHIND_SERVICE+"&amp;"+GSParams.OUTPUT+"=xml&amp;"+GSParams.RESPONSE_ONLY+"=1'/>";
108 app_info +="<PARAM NAME='collection' VALUE='";
109 app_info += this.cluster_name;
110 app_info += "'/> <PARAM NAME='classifier' VALUE='1'/> <PARAM NAME='orientation' VALUE='vertical'/> <PARAM NAME='depth' VALUE='2'/> <PARAM NAME='resultorder' VALUE='L,l,E,e,D,d'/> <PARAM NAME='backdrop' VALUE='interfaces/default/images/phindbg1.jpg'/><PARAM NAME='fontsize' VALUE='10'/> <PARAM NAME='blocksize' VALUE='10'/>The Phind java applet.</"+GSXML.APPLET_ELEM+">";
111
112 Document dom = this.converter.getDOM(app_info);
113 if (dom==null) {
114 logger.error("Couldn't parse applet info");
115 return false;
116 }
117 Element app_elem = dom.getDocumentElement();
118 applet_description.appendChild(this.doc.importNode(app_elem, true));
119
120 return true;
121 }
122
123 protected Element getServiceDescription(String service, String lang, String subset) {
124 if (!service.equals(PHIND_SERVICE)) {
125 return null;
126 }
127 Element describe = (Element) applet_description.cloneNode(true);
128 describe.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_NAME, getTextString(PHIND_SERVICE+".name", lang)));
129 describe.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_DESCRIPTION, getTextString(PHIND_SERVICE+".description", lang)));
130 return describe;
131 }
132
133 protected Element processPhindApplet(Element request) {
134
135 Element param_elem = (Element)GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
136 HashMap params = GSXML.extractParams(param_elem, false);
137
138 long first_e = Long.parseLong((String)params.get("pfe"));
139 long last_e = Long.parseLong((String)params.get("ple"));
140 long first_l = Long.parseLong((String)params.get("pfl"));
141 long last_l = Long.parseLong((String)params.get("pll"));
142 long first_d = Long.parseLong((String)params.get("pfd"));
143 long last_d = Long.parseLong((String)params.get("pld"));
144
145 long phrase;
146 String phrase_str = (String)params.get("ppnum");
147 if (phrase_str == null || phrase_str.equals("")) {
148 phrase=0;
149 } else {
150 phrase = Long.parseLong(phrase_str);
151 }
152 String word = (String)params.get("pptext");
153 String phind_index = (String)params.get("pc");
154 // the location of the mgpp database files
155 this.basepath = GSFile.phindBaseDir(this.site_home, this.cluster_name, phind_index);
156
157 // the result element
158 Element result = this.doc.createElement(GSXML.RESPONSE_ELEM);
159 result.setAttribute(GSXML.FROM_ATT, PHIND_SERVICE);
160 result.setAttribute(GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
161
162 // applet result info must be in appletInfo element
163 Element applet_data = this.doc.createElement(GSXML.APPLET_DATA_ELEM);
164 result.appendChild(applet_data);
165 Element phind_data = this.doc.createElement("phindData");
166 applet_data.appendChild(phind_data);
167
168
169 // if we dont know the phrase number, look it up
170 if (phrase == 0) {
171 if (word==null || word.equals("")) {
172 Element error = phindError("no word or phrase");
173 phind_data.appendChild(error);
174 return result;
175 }
176 phrase = findPhraseNumberFromWord( word);
177 }
178 if (phrase==0) {
179 // the word is not in the collection
180 // return a phind error string
181 Element error = phindError("the term "+word+" is not in the collection");
182 phind_data.appendChild(error);
183 return result;
184 }
185
186 // get the phrase data into the phind_data node
187 getPhraseData(phind_data, phrase, first_l, last_l,
188 first_e, last_e, first_d, last_d);
189 return result;
190
191
192 }// processPhindApplet
193
194 protected long findPhraseNumberFromWord(String word) {
195 synchronized (mgpp_search_src) {
196 // set the mgpp index data - we are looking up pword
197 mgpp_search_src.loadIndexData(this.basepath+File.separatorChar+"pword");
198
199 mgpp_search_src.runQuery(word);
200
201 MGPPQueryResult res = mgpp_search_src.getQueryResult();
202 Vector docs = res.getDocs();
203 if (docs.size()==0) {
204 // phrase not found
205 return 0;
206 }
207 MGPPDocInfo doc = (MGPPDocInfo)docs.firstElement();
208 return doc.num_;
209 }
210 }
211
212 protected boolean getPhraseData(Element phind_data,
213 long phrase, long first_l, long last_l,
214 long first_e, long last_e, long first_d,
215 long last_d) {
216
217 synchronized (mgpp_retrieve_src) {
218 String record = this.mgpp_retrieve_src.getDocument(this.basepath+File.separatorChar+"pdata", "Document",
219 phrase);
220 if (record.equals("")) {
221 Element error = phindError("somethings gone wrong - we haven't got a record for phrase number "+phrase);
222 phind_data.appendChild(error);
223 return false;
224 }
225
226 // parse the record - its in gordons cryptic form
227 // ":word:tf:ef:df:el:dl:lf:ll"
228 // el: e,e,e
229 // dl: d;f,d;f,
230 // lf and ll may be null
231 // l: type,dest, dest; type,dest,dest
232
233 // ignore everything up to and including first colon (has
234 // <Document>3505: at the start)
235 record = record.substring(record.indexOf(':')+1);
236
237 // split on ':'
238 String [] fields = record.split(":");
239 String word = fields[0];
240 String tf = fields[1];
241 String ef = fields[2];
242 String df = fields[3];
243
244
245 String expansions = fields[4];
246 String documents = fields[5];
247 String lf = "0";
248 String linklist = "";
249 if (fields.length > 7) {// have thesaurus stuff
250 lf =fields[6];
251 linklist = fields[7];
252 }
253
254 // the phindData attributes and phrase
255 phind_data.setAttribute("id", Long.toString(phrase));
256 phind_data.setAttribute("df", df);
257 phind_data.setAttribute("ef", ef);
258 phind_data.setAttribute("lf", lf);
259 phind_data.setAttribute("tf", tf);
260 GSXML.createTextElement(this.doc, "phrase", word);
261
262 addExpansionList(phind_data, expansions, word, ef, first_e, last_e);
263 addDocumentList(phind_data, documents, word, df, first_d, last_d);
264 if (!lf.equals("0")) {
265 addThesaurusList(phind_data, linklist, word, lf, first_l, last_l);
266 }
267 return true;
268 }
269 }
270
271 protected boolean addExpansionList( Element phind_data, String record,
272 String word,
273 String freq,
274 long first, long last) {
275
276 Element expansion_list = this.doc.createElement("expansionList");
277 phind_data.appendChild(expansion_list);
278 expansion_list.setAttribute("length", freq);
279 expansion_list.setAttribute("start", Long.toString(first));
280 expansion_list.setAttribute("end", Long.toString(last));
281
282 // get the list of strings
283 String [] expansions = record.split(",");
284 int length = expansions.length;
285 if (length < last) last = length;
286 for (long i = first; i < last; i++) {
287 long num = Long.parseLong(expansions[(int)i]);
288 Element expansion = getExpansion( num, word);
289 expansion.setAttribute("num", Long.toString(i));
290 expansion_list.appendChild(expansion);
291 }
292 return true;
293 }
294
295 protected Element getExpansion(long phrase_num,
296 String orig_phrase) {
297
298 // look up the phrase in the pdata thingy
299 String record = this.mgpp_retrieve_src.getDocument(this.basepath+File.separatorChar+"pdata", "Document",
300 phrase_num);
301
302 if (record ==null || record.equals("")) return null;
303
304 // ignore everything up to and including first colon
305 record = record.substring(record.indexOf(':')+1);
306
307 String [] fields = record.split(":");
308 String phrase = fields[0];
309 String tf = fields[1];
310 //String ef = fields[2]; dont use this
311 String df = fields[3];
312
313 Element expansion = this.doc.createElement("expansion");
314 expansion.setAttribute("tf", tf);
315 expansion.setAttribute("df", df);
316 expansion.setAttribute("id", Long.toString(phrase_num));
317
318 // get teh suffix and prefix
319 String [] ends = splitPhraseOnWord(phrase, orig_phrase);
320 if (!ends[0].equals("")) {
321 expansion.appendChild(GSXML.createTextElement(this.doc, "prefix", ends[0]));
322 }
323 if (!ends[1].equals("")) {
324 expansion.appendChild(GSXML.createTextElement(this.doc, "suffix", ends[1]));
325 }
326
327 return expansion;
328
329 }
330
331 protected boolean addDocumentList(Element phind_data, String record,
332 String word,
333 String freq,
334 long first, long last) {
335
336 Element document_list = this.doc.createElement("documentList");
337 phind_data.appendChild(document_list);
338 document_list.setAttribute("length", freq);
339 document_list.setAttribute("start", Long.toString(first));
340 document_list.setAttribute("end", Long.toString(last));
341
342 // get the list of doc,freq
343 String [] doc_freqs = record.split(";");
344 int length = doc_freqs.length;
345 if (length<last) last=length;
346
347 for (long i = first; i < last; i++) {
348 String doc_elem = doc_freqs[(int)i];
349 int p = doc_elem.indexOf(',');
350 long doc_num;
351 String doc_freq;
352 if (p == -1) { // there is no freq in the record
353 doc_num =Long.parseLong(doc_elem);
354 doc_freq = "1";
355 } else {
356 doc_num = Long.parseLong(doc_elem.substring(0,p));
357 doc_freq = doc_elem.substring(p+1);
358 }
359 Element document = getDocument( doc_num);
360 document.setAttribute("freq", doc_freq);
361 document.setAttribute("num", Long.toString(i));
362 document_list.appendChild(document);
363 }
364
365
366 return true;
367 }
368
369
370 protected Element getDocument(long doc_num) {
371
372 // look up the phrase in the docs thingy
373 String record = this.mgpp_retrieve_src.getDocument(this.basepath+File.separatorChar+"docs", "Document",
374 doc_num);
375
376 if (record ==null || record.equals("")) return null;
377
378 // ignore everything up to and including first \t
379 record = record.substring(record.indexOf('\t')+1);
380
381 String [] fields = record.split("\t");
382 String hash = fields[0];
383 String title = fields[1];
384
385 Element d = this.doc.createElement("document");
386 d.setAttribute("hash", hash);
387 d.appendChild(GSXML.createTextElement(this.doc, "title", title));
388
389 return d;
390
391 }
392 protected boolean addThesaurusList(Element phind_data, String record,
393 String word,
394 String freq,
395 long first, long last) {
396
397
398 Element thesaurus_list = this.doc.createElement("thesaurusList");
399 phind_data.appendChild(thesaurus_list);
400 thesaurus_list.setAttribute("length", freq);
401 thesaurus_list.setAttribute("start", Long.toString(first));
402 thesaurus_list.setAttribute("end", Long.toString(last));
403
404 // get the list of type,dest,dest
405 String [] links = record.split(";");
406 int length = links.length;
407 long index = 0;
408 for (int i = 0; i < length; i++) { // go through the entries
409 String link_info = links[(int)i];
410 String [] items = link_info.split(",");
411 // the first entry is teh type
412 String type = items[0];
413 for (int j = 1; j<items.length; j++, index++) {
414 if (index >= first && index < last) { // only output the ones we want
415 long phrase = Long.parseLong(items[j]);
416 Element t = getThesaurus(phrase);
417 t.setAttribute("type", type);
418 thesaurus_list.appendChild(t);
419 }
420 }
421 }
422
423 return true;
424 }
425
426 protected Element getThesaurus(long phrase_num) {
427
428 // look up the phrase in the pdata thingy
429 String record = this.mgpp_retrieve_src.getDocument(this.basepath+File.separatorChar+"pdata", "Document",
430 phrase_num);
431
432 if (record ==null || record.equals("")) return null;
433
434 // ignore everything up to and including first colon
435 record = record.substring(record.indexOf(':')+1);
436
437 String [] fields = record.split(":");
438 String phrase = fields[0];
439 String tf = fields[1];
440 //String ef = fields[2]; dont use this
441 String df = fields[3];
442
443 Element thesaurus = this.doc.createElement("thesaurus");
444 thesaurus.setAttribute("tf", tf);
445 thesaurus.setAttribute("df", df);
446 thesaurus.setAttribute("id", Long.toString(phrase_num));
447 thesaurus.appendChild(GSXML.createTextElement(this.doc, "phrase", phrase));
448 return thesaurus;
449
450 }
451
452 /** returns an array of two elements - the prefix and the suffix*/
453 protected String [] splitPhraseOnWord(String phrase, String word) {
454
455 if (word.equals("")) {
456
457 String [] res = {phrase, ""};
458 return res;
459 }
460 // use 2 so that we only split on the first occurrance. trailing empty strings should be included
461 String [] result = phrase.split(word, 2);
462 return result;
463
464 }
465
466 protected Element phindError(String message) {
467 Element e = this.doc.createElement("phindError");
468 Text t = this.doc.createTextNode(message);
469 e.appendChild(t);
470 return e;
471 }
472
473}
474
Note: See TracBrowser for help on using the repository browser.