source: trunk/gsdl3/src/java/org/greenstone/gsdl3/service/PhindPhraseBrowse.java@ 6273

Last change on this file since 6273 was 6273, checked in by kjdon, 20 years ago

was using old request types
  • Property svn:keywords set to Author Date Id Revision
File size: 14.9 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/**
33 * PhindServices - the phind phrase browsing service
34 *
35 * @author <a href="mailto:[email protected]">Katherine Don</a>
36 * @version $Revision: 6273 $
37 */
38public class PhindPhraseBrowse
39 extends ServiceRack {
40
41 // the services on offer
42 private static final String PHIND_SERVICE = "PhindApplet";
43
44 private MGPPWrapper mgpp_src=null;
45 private String basepath = null;
46
47 private Element applet_description = null;
48
49 public PhindPhraseBrowse() {
50 this.mgpp_src = new MGPPWrapper();
51 // set up the default params
52 this.mgpp_src.setQueryLevel("Document");
53 this.mgpp_src.setReturnLevel("Document");
54 this.mgpp_src.setMaxDocs(5);
55 this.mgpp_src.setStem(false);
56 this.mgpp_src.setCase(true);
57 }
58 /** configure the service module
59 *
60 * @param info a DOM Element containing any config info for the service
61 * @return true if configured
62 */
63 public boolean configure(Element info, Element extra_info) {
64
65 System.out.println("configuring PhindPhraseBrowse");
66
67 // set up short_service_info_ - for now just has name and type
68 Element e = this.doc.createElement(GSXML.SERVICE_ELEM);
69 e.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_APPLET);
70 e.setAttribute(GSXML.NAME_ATT, PHIND_SERVICE);
71 this.short_service_info.appendChild(e);
72
73 // set up the static applet description
74
75 applet_description = this.doc.createElement(GSXML.SERVICE_ELEM);
76 applet_description.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_APPLET);
77 applet_description.setAttribute(GSXML.NAME_ATT, PHIND_SERVICE);
78
79 // add in the applet info for the phind applet
80 // need to make this dynamic - library names etc
81 // change the applet params - have a single param with the library name
82 // 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
83 // phindcgi param now is not complete - library must be prepended to it.
84 String app_info = "<"+GSXML.APPLET_ELEM+" CODEBASE='lib/java' CODE='org.greenstone.applet.phind.Phind.class' ARCHIVE='phind.jar, xercesImpl.jar, gsdl3.jar, jaxp.jar, xml-apis.jar' WIDTH='500' HEIGHT='400'><PARAM NAME='library' VALUE=''/> <PARAM NAME='phindcgi' VALUE='?";
85 app_info += GSParams.ACTION +"=a&amp;"+GSParams.REQUEST_TYPE +"=r&amp;"+GSParams.SERVICE+"="+PHIND_SERVICE+"&amp;"+GSParams.OUTPUT+"=xml&amp;"+GSParams.RESPONSE_ONLY+"=1'/>";
86 app_info +="<PARAM NAME='collection' VALUE='";
87 app_info += this.cluster_name;
88 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+">";
89
90 Document dom = this.converter.getDOM(app_info);
91 Element app_elem = dom.getDocumentElement();
92 applet_description.appendChild(this.doc.importNode(app_elem, true));
93
94 return true;
95 }
96
97 protected Element getServiceDescription(String service, String lang, String subset) {
98 if (!service.equals(PHIND_SERVICE)) {
99 return null;
100 }
101 Element describe = (Element) applet_description.cloneNode(true);
102 describe.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_NAME, getTextString(PHIND_SERVICE+".name", lang)));
103 describe.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_DESCRIPTION, getTextString(PHIND_SERVICE+".description", lang)));
104 return describe;
105 }
106
107 protected Element processPhindApplet(Element request) {
108
109 Element param_elem = (Element)GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
110 HashMap params = GSXML.extractParams(param_elem, false);
111
112 long first_e = Long.parseLong((String)params.get("pfe"));
113 long last_e = Long.parseLong((String)params.get("ple"));
114 long first_l = Long.parseLong((String)params.get("pfl"));
115 long last_l = Long.parseLong((String)params.get("pll"));
116 long first_d = Long.parseLong((String)params.get("pfd"));
117 long last_d = Long.parseLong((String)params.get("pld"));
118
119 long phrase;
120 String phrase_str = (String)params.get("ppnum");
121 if (phrase_str == null || phrase_str.equals("")) {
122 phrase=0;
123 } else {
124 phrase = Long.parseLong(phrase_str);
125 }
126 String word = (String)params.get("pptext");
127 String phind_index = (String)params.get("pc");
128 // the location of the mgpp database files
129 this.basepath = GSFile.phindBaseDir(this.site_home, this.cluster_name, phind_index);
130
131 // the result element
132 Element result = this.doc.createElement(GSXML.RESPONSE_ELEM);
133 result.setAttribute(GSXML.FROM_ATT, PHIND_SERVICE);
134 result.setAttribute(GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
135
136 // applet result info must be in appletInfo element
137 Element applet_data = this.doc.createElement(GSXML.APPLET_DATA_ELEM);
138 result.appendChild(applet_data);
139 Element phind_data = this.doc.createElement("phindData");
140 applet_data.appendChild(phind_data);
141
142
143 // if we dont know the phrase number, look it up
144 if (phrase == 0) {
145 if (word==null || word.equals("")) {
146 Element error = phindError("no word or phrase");
147 phind_data.appendChild(error);
148 return result;
149 }
150 phrase = findPhraseNumberFromWord( word);
151 }
152 if (phrase==0) {
153 // the word is not in the collection
154 // return a phind error string
155 Element error = phindError("the term "+word+" is not in the collection");
156 phind_data.appendChild(error);
157 return result;
158 }
159
160 // get the phrase data into the phind_data node
161 getPhraseData(phind_data, phrase, first_l, last_l,
162 first_e, last_e, first_d, last_d);
163 return result;
164
165
166 }// processPhindApplet
167
168 protected long findPhraseNumberFromWord(String word) {
169
170 // set the mgpp index data - we are looking up pword
171 this.mgpp_src.loadIndexData(this.basepath+File.separatorChar+"pword");
172
173 this.mgpp_src.runQuery(word);
174
175 MGPPQueryResult res = this.mgpp_src.getQueryResult();
176 Vector docs = res.getDocs();
177 if (docs.size()==0) {
178 // phrase not found
179 return 0;
180 }
181 MGPPDocInfo doc = (MGPPDocInfo)docs.firstElement();
182 return doc.num_;
183 }
184
185 protected boolean getPhraseData(Element phind_data,
186 long phrase, long first_l, long last_l,
187 long first_e, long last_e, long first_d,
188 long last_d) {
189
190 String record = this.mgpp_src.getDocument(this.basepath+File.separatorChar+"pdata", "Document",
191 phrase);
192 if (record.equals("")) {
193 Element error = phindError("somethings gone wrong - we haven't got a record for phrase number "+phrase);
194 phind_data.appendChild(error);
195 return false;
196 }
197
198 // parse the record - its in gordons cryptic form
199 // ":word:tf:ef:df:el:dl:lf:ll"
200 // el: e,e,e
201 // dl: d;f,d;f,
202 // lf and ll may be null
203 // l: type,dest, dest; type,dest,dest
204
205 // ignore everything up to and including first colon (has
206 // <Document>3505: at the start)
207 record = record.substring(record.indexOf(':')+1);
208
209 // split on ':'
210 String [] fields = record.split(":");
211 String word = fields[0];
212 String tf = fields[1];
213 String ef = fields[2];
214 String df = fields[3];
215
216
217 String expansions = fields[4];
218 String documents = fields[5];
219 String lf = "0";
220 String linklist = "";
221 if (fields.length > 7) {// have thesaurus stuff
222 lf =fields[6];
223 linklist = fields[7];
224 }
225
226 // the phindData attributes and phrase
227 phind_data.setAttribute("id", Long.toString(phrase));
228 phind_data.setAttribute("df", df);
229 phind_data.setAttribute("ef", ef);
230 phind_data.setAttribute("lf", lf);
231 phind_data.setAttribute("tf", tf);
232 GSXML.createTextElement(this.doc, "phrase", word);
233
234 addExpansionList(phind_data, expansions, word, ef, first_e, last_e);
235 addDocumentList(phind_data, documents, word, df, first_d, last_d);
236 if (!lf.equals("0")) {
237 addThesaurusList(phind_data, linklist, word, lf, first_l, last_l);
238 }
239 return true;
240 }
241
242 protected boolean addExpansionList( Element phind_data, String record,
243 String word,
244 String freq,
245 long first, long last) {
246
247 Element expansion_list = this.doc.createElement("expansionList");
248 phind_data.appendChild(expansion_list);
249 expansion_list.setAttribute("length", freq);
250 expansion_list.setAttribute("start", Long.toString(first));
251 expansion_list.setAttribute("end", Long.toString(last));
252
253 // get the list of strings
254 String [] expansions = record.split(",");
255 int length = expansions.length;
256 if (length < last) last = length;
257 for (long i = first; i < last; i++) {
258 long num = Long.parseLong(expansions[(int)i]);
259 Element expansion = getExpansion( num, word);
260 expansion.setAttribute("num", Long.toString(i));
261 expansion_list.appendChild(expansion);
262 }
263 return true;
264 }
265
266 protected Element getExpansion(long phrase_num,
267 String orig_phrase) {
268
269 // look up the phrase in the pdata thingy
270 String record = this.mgpp_src.getDocument(this.basepath+File.separatorChar+"pdata", "Document",
271 phrase_num);
272
273 if (record ==null || record.equals("")) return null;
274
275 // ignore everything up to and including first colon
276 record = record.substring(record.indexOf(':')+1);
277
278 String [] fields = record.split(":");
279 String phrase = fields[0];
280 String tf = fields[1];
281 //String ef = fields[2]; dont use this
282 String df = fields[3];
283
284 Element expansion = this.doc.createElement("expansion");
285 expansion.setAttribute("tf", tf);
286 expansion.setAttribute("df", df);
287 expansion.setAttribute("id", Long.toString(phrase_num));
288
289 // get teh suffix and prefix
290 String [] ends = splitPhraseOnWord(phrase, orig_phrase);
291 if (!ends[0].equals("")) {
292 expansion.appendChild(GSXML.createTextElement(this.doc, "prefix", ends[0]));
293 }
294 if (!ends[1].equals("")) {
295 expansion.appendChild(GSXML.createTextElement(this.doc, "suffix", ends[1]));
296 }
297
298 return expansion;
299
300 }
301
302 protected boolean addDocumentList(Element phind_data, String record,
303 String word,
304 String freq,
305 long first, long last) {
306
307 Element document_list = this.doc.createElement("documentList");
308 phind_data.appendChild(document_list);
309 document_list.setAttribute("length", freq);
310 document_list.setAttribute("start", Long.toString(first));
311 document_list.setAttribute("end", Long.toString(last));
312
313 // get the list of doc,freq
314 String [] doc_freqs = record.split(";");
315 int length = doc_freqs.length;
316 if (length<last) last=length;
317
318 for (long i = first; i < last; i++) {
319 String doc_elem = doc_freqs[(int)i];
320 int p = doc_elem.indexOf(',');
321 long doc_num;
322 String doc_freq;
323 if (p == -1) { // there is no freq in the record
324 doc_num =Long.parseLong(doc_elem);
325 doc_freq = "1";
326 } else {
327 doc_num = Long.parseLong(doc_elem.substring(0,p));
328 doc_freq = doc_elem.substring(p+1);
329 }
330 Element document = getDocument( doc_num);
331 document.setAttribute("freq", doc_freq);
332 document.setAttribute("num", Long.toString(i));
333 document_list.appendChild(document);
334 }
335
336
337 return true;
338 }
339
340
341 protected Element getDocument(long doc_num) {
342
343 // look up the phrase in the docs thingy
344 String record = this.mgpp_src.getDocument(this.basepath+File.separatorChar+"docs", "Document",
345 doc_num);
346
347 if (record ==null || record.equals("")) return null;
348
349 // ignore everything up to and including first \t
350 record = record.substring(record.indexOf('\t')+1);
351
352 String [] fields = record.split("\t");
353 String hash = fields[0];
354 String title = fields[1];
355
356 Element d = this.doc.createElement("document");
357 d.setAttribute("hash", hash);
358 d.appendChild(GSXML.createTextElement(this.doc, "title", title));
359
360 return d;
361
362 }
363 protected boolean addThesaurusList(Element phind_data, String record,
364 String word,
365 String freq,
366 long first, long last) {
367
368
369 Element thesaurus_list = this.doc.createElement("thesaurusList");
370 phind_data.appendChild(thesaurus_list);
371 thesaurus_list.setAttribute("length", freq);
372 thesaurus_list.setAttribute("start", Long.toString(first));
373 thesaurus_list.setAttribute("end", Long.toString(last));
374
375 // get the list of type,dest,dest
376 String [] links = record.split(";");
377 int length = links.length;
378 long index = 0;
379 for (int i = 0; i < length; i++) { // go through the entries
380 String link_info = links[(int)i];
381 String [] items = link_info.split(",");
382 // the first entry is teh type
383 String type = items[0];
384 for (int j = 1; j<items.length; j++, index++) {
385 if (index >= first && index < last) { // only output the ones we want
386 long phrase = Long.parseLong(items[j]);
387 Element t = getThesaurus(phrase);
388 t.setAttribute("type", type);
389 thesaurus_list.appendChild(t);
390 }
391 }
392 }
393
394 return true;
395 }
396
397 protected Element getThesaurus(long phrase_num) {
398
399 // look up the phrase in the pdata thingy
400 String record = this.mgpp_src.getDocument(this.basepath+File.separatorChar+"pdata", "Document",
401 phrase_num);
402
403 if (record ==null || record.equals("")) return null;
404
405 // ignore everything up to and including first colon
406 record = record.substring(record.indexOf(':')+1);
407
408 String [] fields = record.split(":");
409 String phrase = fields[0];
410 String tf = fields[1];
411 //String ef = fields[2]; dont use this
412 String df = fields[3];
413
414 Element thesaurus = this.doc.createElement("thesaurus");
415 thesaurus.setAttribute("tf", tf);
416 thesaurus.setAttribute("df", df);
417 thesaurus.setAttribute("id", Long.toString(phrase_num));
418 thesaurus.appendChild(GSXML.createTextElement(this.doc, "phrase", phrase));
419 return thesaurus;
420
421 }
422
423 /** returns an array of two elements - the prefix and the suffix*/
424 protected String [] splitPhraseOnWord(String phrase, String word) {
425
426 if (word.equals("")) {
427
428 String [] res = {phrase, ""};
429 return res;
430 }
431 // use 2 so that we only split on the first occurrance. trailing empty strings should be included
432 String [] result = phrase.split(word, 2);
433 return result;
434
435 }
436
437 protected Element phindError(String message) {
438 Element e = this.doc.createElement("phindError");
439 Text t = this.doc.createTextNode(message);
440 e.appendChild(t);
441 return e;
442 }
443
444}
445
Note: See TracBrowser for help on using the repository browser.