source: trunk/gsdl3/src/java/org/greenstone/gsdl3/service/CrossCollectionSearch.java@ 13576

Last change on this file since 13576 was 13270, checked in by shaoqun, 17 years ago

replace Category class which is deprecated with Logger class

  • Property svn:keywords set to Author Date Id Revision
File size: 16.3 KB
Line 
1/*
2 * CrossCollectionSearch.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.GSXML;
22import org.greenstone.gsdl3.util.GSPath;
23
24import org.w3c.dom.Element;
25import org.w3c.dom.NodeList;
26
27import java.util.ArrayList;
28import java.util.Map;
29import java.util.Iterator;
30import java.util.Set;
31import java.util.HashMap;
32
33import org.apache.log4j.*;
34
35/**
36 *
37 * @author <a href="mailto:[email protected]">Katherine Don</a>
38 */
39
40public class CrossCollectionSearch
41 extends ServiceRack {
42
43 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.CrossCollectionSearch.class.getName());
44 protected static final String QUERY_PARAM = "query";
45 protected static final String COLLECTION_PARAM = "collection";
46
47 // the services on offer - these proxy the actual collection ones
48 protected static final String TEXT_QUERY_SERVICE = "TextQuery";
49 protected static final String ADV_QUERY_SERVICE = "AdvTextQuery";
50 protected static final String DOCUMENT_METADATA_RETRIEVE_SERVICE = "DocumentMetadataRetrieve";
51
52 protected String [] coll_ids_list = null;
53 protected String [] coll_names_list = null;
54
55
56 /** constructor */
57 public CrossCollectionSearch()
58 {
59 }
60
61 public boolean configure(Element info, Element extra_info)
62 {
63 // any parameters? colls to include??
64 logger.info("Configuring CrossCollectionSearch...");
65 // query service
66 Element ccs_service = this.doc.createElement(GSXML.SERVICE_ELEM);
67 ccs_service.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_QUERY);
68 ccs_service.setAttribute(GSXML.NAME_ATT, TEXT_QUERY_SERVICE);
69 this.short_service_info.appendChild(ccs_service);
70
71 // metadata service
72 Element dmr_service = this.doc.createElement(GSXML.SERVICE_ELEM);
73 dmr_service.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_RETRIEVE);
74 dmr_service.setAttribute(GSXML.NAME_ATT, DOCUMENT_METADATA_RETRIEVE_SERVICE);
75 this.short_service_info.appendChild(dmr_service);
76
77 // get any format info
78 Element format_info = (Element)GSXML.getChildByTagName(info, GSXML.FORMAT_ELEM);
79 if (format_info != null) {
80 this.format_info_map.put(TEXT_QUERY_SERVICE, this.doc.importNode(format_info, true));
81 } else {
82 // add in a default format statement
83 String format_string = "<format xmlns:gsf='http://www.greenstone.org/greenstone3/schema/ConfigFormat' xmlns:xsl='http://www.w3.org/1999/XSL/Transform'><gsf:template match='documentNode'><td><a><xsl:attribute name='href'>?a=d&amp;c=<xsl:value-of select='@collection'/>&amp;d=<xsl:value-of select='@nodeID'/><xsl:if test=\"@nodeType='leaf'\">&amp;sib=1</xsl:if>&amp;dt=<xsl:value-of select='@docType'/>&amp;p.a=q&amp;&amp;p.s="+TEXT_QUERY_SERVICE+"&amp;p.c=";
84 if (this.cluster_name!=null) {
85 format_string += this.cluster_name;
86 }
87 format_string += "</xsl:attribute><gsf:icon/></a></td><td><gsf:metadata name='Title'/> (<xsl:value-of select='@collection'/>) </td></gsf:template></format>";
88 this.format_info_map.put(TEXT_QUERY_SERVICE, this.doc.importNode(this.converter.getDOM(format_string).getDocumentElement(), true));
89 }
90 return true;
91 }
92
93 protected Element getServiceDescription(String service, String lang, String subset)
94 {
95 if (service.equals(TEXT_QUERY_SERVICE)) {
96
97 Element ccs_service = this.doc.createElement(GSXML.SERVICE_ELEM);
98 ccs_service.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_QUERY);
99 ccs_service.setAttribute(GSXML.NAME_ATT, TEXT_QUERY_SERVICE);
100
101 // display info
102 if (subset==null || subset.equals(GSXML.DISPLAY_TEXT_ELEM+GSXML.LIST_MODIFIER)) {
103 ccs_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_NAME, getTextString(TEXT_QUERY_SERVICE+".name", lang)));
104 ccs_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_SUBMIT, getTextString(TEXT_QUERY_SERVICE+".submit", lang)));
105 ccs_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_DESCRIPTION, getTextString(TEXT_QUERY_SERVICE+".description", lang)));
106 }
107 // param info
108 if (subset==null || subset.equals(GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER)) {
109 Element param_list = this.doc.createElement(GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
110 // collection list
111 if (coll_ids_list ==null ) {
112 initCollectionList();
113 }
114 Element param = GSXML.createParameterDescription(this.doc, COLLECTION_PARAM, getTextString("param."+COLLECTION_PARAM, lang), GSXML.PARAM_TYPE_ENUM_MULTI, "all", coll_ids_list, coll_names_list);
115 param_list.appendChild(param);
116 // query param
117 param = GSXML.createParameterDescription(this.doc, QUERY_PARAM, getTextString("param."+QUERY_PARAM, lang), GSXML.PARAM_TYPE_STRING, null, null, null);
118 param_list.appendChild(param);
119 ccs_service.appendChild(param_list);
120 }
121
122 logger.debug("service description="+this.converter.getPrettyString(ccs_service));
123 return ccs_service;
124 }
125 // these ones are probably never called, but put them here just in case
126 Element service_elem = this.doc.createElement(GSXML.SERVICE_ELEM);
127 service_elem.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_RETRIEVE);
128 service_elem.setAttribute(GSXML.NAME_ATT, service);
129 return service_elem;
130
131
132 }
133
134
135 protected Element processTextQuery(Element request) {
136 // Create a new (empty) result message
137 Element result = this.doc.createElement(GSXML.RESPONSE_ELEM);
138 result.setAttribute(GSXML.FROM_ATT, TEXT_QUERY_SERVICE);
139 result.setAttribute(GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
140
141 String lang = request.getAttribute(GSXML.LANG_ATT);
142 // Get the parameters of the request
143 Element param_list = (Element) GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
144 if (param_list == null) {
145 logger.error("TextQuery request had no paramList.");
146 return result; // Return the empty result
147 }
148
149 // get the collection list
150 String [] colls_list;
151 Element coll_param = GSXML.getNamedElement(param_list, GSXML.PARAM_ELEM, GSXML.NAME_ATT, COLLECTION_PARAM);
152 if (coll_param == null) {
153 colls_list = coll_ids_list;
154 } else {
155 String coll_list = GSXML.getValue(coll_param);
156 if (coll_list.equals("all") || coll_list.equals("")) {
157 colls_list = coll_ids_list;
158 } else {
159 colls_list = coll_list.split(",");
160 }
161 }
162
163 Element query_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
164 // we are sending the same request to each collection - build up the to
165 // attribute for the request
166 StringBuffer to_att = new StringBuffer();
167 for (int i=0; i<colls_list.length; i++) {
168 if (i>0) {
169 to_att.append(",");
170 }
171 to_att.append(GSPath.appendLink(colls_list[i], "TextQuery"));
172
173 }
174 // send the query to all colls
175 Element query_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_PROCESS, to_att.toString(), lang, "");
176 query_message.appendChild(query_request);
177 // should we add params individually?
178 Element new_param_list = this.doc.createElement(GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
179 query_request.appendChild(new_param_list);
180 new_param_list.appendChild(this.doc.importNode(GSXML.getNamedElement(param_list, GSXML.PARAM_ELEM, GSXML.NAME_ATT, QUERY_PARAM), true));
181 Element query_result = this.router.process(query_message);
182
183 // gather up the data from each response
184 int numDocsMatched = 0;
185 int numDocsReturned = 0;
186
187 //term info??
188
189 NodeList metadata = query_result.getElementsByTagName(GSXML.METADATA_ELEM);
190 for (int j=0; j<metadata.getLength(); j++) {
191 Element meta = (Element)metadata.item(j);
192 if (meta.getAttribute(GSXML.NAME_ATT).equals("numDocsReturned")) {
193 numDocsReturned += Integer.parseInt(GSXML.getValue(meta));
194 } else if (meta.getAttribute(GSXML.NAME_ATT).equals("numDocsMatched")) {
195 numDocsMatched += Integer.parseInt(GSXML.getValue(meta));
196 }
197 }
198
199 Element metadata_list = this.doc.createElement(GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER);
200 result.appendChild(metadata_list);
201 GSXML.addMetadata(this.doc, metadata_list, "numDocsReturned", ""+numDocsReturned);
202 //GSXML.addMetadata(this.doc, metadata_list, "numDocsMatched", ""+numDocsMatched);
203
204 Element doc_node_list = this.doc.createElement(GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
205 result.appendChild(doc_node_list);
206
207 NodeList responses = query_result.getElementsByTagName(GSXML.RESPONSE_ELEM);
208
209 for (int k=0; k<responses.getLength(); k++) {
210 String coll_name = GSPath.getFirstLink(((Element)responses.item(k)).getAttribute(GSXML.FROM_ATT));
211 NodeList nodes = ((Element)responses.item(k)).getElementsByTagName(GSXML.DOC_NODE_ELEM);
212 if (nodes==null || nodes.getLength()==0) continue;
213 Element last_node = null;
214 Element this_node = null;
215 for (int n=0; n<nodes.getLength(); n++) {
216 this_node = (Element)nodes.item(n);
217 this_node.setAttribute("collection", coll_name);
218 if (k==0) {
219
220 doc_node_list.appendChild(this.doc.importNode(this_node, true));
221 } else {
222 if (last_node==null) {
223 last_node = (Element)GSXML.getChildByTagName(doc_node_list, GSXML.DOC_NODE_ELEM);
224 }
225 last_node = GSXML.insertIntoOrderedList(doc_node_list, GSXML.DOC_NODE_ELEM, last_node, this_node, "rank", true);
226 }
227
228 }
229 }
230 return result;
231 }
232
233// protected Element processAdvTextQuery(Element request)
234// {
235
236// }
237 protected boolean initCollectionList() {
238 String lang="en";
239 String uid = "";
240 // first, get the message router info
241 Element coll_list_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
242 Element coll_list_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_DESCRIBE, "", lang, ""); // uid
243 coll_list_message.appendChild(coll_list_request);
244 logger.debug("coll list request = "+this.converter.getPrettyString(coll_list_request));
245 Element coll_list_response = (Element)this.router.process(coll_list_message);
246 if (coll_list_response==null) {
247 logger.error("couldn't query the message router!");
248 return false;
249 }
250 logger.debug("coll list response = "+this.converter.getPrettyString(coll_list_response));
251 // second, get some info from each collection. we want the coll name
252 // and whether its got a text query service
253
254 NodeList colls = coll_list_response.getElementsByTagName(GSXML.COLLECTION_ELEM);
255 // we will send all the requests in a single message
256 Element metadata_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
257 for (int i=0; i<colls.getLength(); i++) {
258 Element c = (Element)colls.item(i);
259 String name = c.getAttribute(GSXML.NAME_ATT);
260 Element metadata_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_DESCRIBE, name, lang, uid);
261 metadata_message.appendChild(metadata_request);
262 }
263 logger.debug("metadata request = "+this.converter.getPrettyString(metadata_message));
264 Element metadata_response = (Element)this.router.process(metadata_message);
265 logger.debug("metadata response = "+this.converter.getPrettyString(metadata_response));
266
267 NodeList coll_responses = metadata_response.getElementsByTagName(GSXML.COLLECTION_ELEM);
268 ArrayList valid_colls = new ArrayList();
269 ArrayList valid_coll_names = new ArrayList();
270 for (int i=0; i<coll_responses.getLength(); i++) {
271 Element coll = (Element)coll_responses.item(i);
272 Element service_list = (Element)GSXML.getChildByTagName(coll, GSXML.SERVICE_ELEM+GSXML.LIST_MODIFIER);
273 if (service_list==null) continue;
274 Element query_service = GSXML.getNamedElement(service_list, GSXML.SERVICE_ELEM, GSXML.NAME_ATT, TEXT_QUERY_SERVICE); // should be AbstractSearch.TEXT_QUERY_SERVICE
275 if (query_service == null) continue;
276 String coll_id = coll.getAttribute(GSXML.NAME_ATT);
277 String coll_name = coll_id+": "+GSXML.getDisplayText(coll, GSXML.DISPLAY_TEXT_NAME, "en", "en"); // just use english for now until we do some caching or something
278 valid_colls.add(coll_id);
279 valid_coll_names.add(coll_name);
280 }
281
282 if (valid_colls.size()==0) {
283 return false;
284 }
285 valid_colls.add(0, "all");
286 valid_coll_names.add(0, getTextString("param."+COLLECTION_PARAM+".all", "en" ));
287 this.coll_ids_list = new String[1];
288 this.coll_names_list = new String[1];
289 this.coll_ids_list = (String []) valid_colls.toArray(coll_ids_list);
290 this.coll_names_list = (String []) valid_coll_names.toArray(coll_names_list);
291 return true;
292 }
293
294 protected Element processDocumentMetadataRetrieve(Element request) {
295 // Create a new (empty) result message
296 Element result = this.doc.createElement(GSXML.RESPONSE_ELEM);
297 result.setAttribute(GSXML.FROM_ATT, DOCUMENT_METADATA_RETRIEVE_SERVICE);
298 result.setAttribute(GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
299
300 String lang = request.getAttribute(GSXML.LANG_ATT);
301 // Get the parameters of the request
302 Element param_list = (Element) GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
303 if (param_list == null) {
304 logger.error("DocumentMetadataRetrieve request had no paramList.");
305 return result; // Return the empty result
306 }
307
308 NodeList query_doc_list = request.getElementsByTagName(GSXML.DOC_NODE_ELEM);
309 if (query_doc_list.getLength()==0) {
310 logger.error("DocumentMetadataRetrieve request had no documentNodes.");
311 return result; // Return the empty result
312 }
313
314 // the resulting doc node list
315 Element result_node_list = this.doc.createElement(GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
316 result.appendChild(result_node_list);
317
318 // get all the metadata params
319 Element new_param_list = this.doc.createElement(GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
320 Element param = GSXML.createParameter(this.doc, "metadata", "Title");
321 new_param_list.appendChild(param);
322
323 // organise the nodes into collection lists
324 HashMap coll_map = new HashMap();
325
326 for (int i=0; i<query_doc_list.getLength(); i++) {
327 Element doc_node = (Element)query_doc_list.item(i);
328 String coll_name = doc_node.getAttribute("collection");
329 Element coll_items = (Element)coll_map.get(coll_name);
330 if (coll_items==null) {
331 coll_items=this.doc.createElement(GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
332 coll_map.put(coll_name, coll_items);
333 }
334 coll_items.appendChild(this.doc.importNode(doc_node, true));
335 }
336
337 // create teh individual requests
338 Element meta_request_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
339 Set mapping_set = coll_map.entrySet();
340 Iterator iter = mapping_set.iterator();
341
342 while (iter.hasNext()) {
343 Map.Entry e = (Map.Entry)iter.next();
344 String cname = (String)e.getKey();
345 Element doc_nodes = (Element)e.getValue();
346 Element meta_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_PROCESS, GSPath.appendLink(cname, DOCUMENT_METADATA_RETRIEVE_SERVICE), lang, "");
347 meta_request.appendChild(doc_nodes);
348 meta_request.appendChild(new_param_list.cloneNode(true));
349 meta_request_message.appendChild(meta_request);
350
351 }
352 Element meta_result = this.router.process(meta_request_message);
353
354 // now need to put the doc nodes back in the right order
355 // go through the original list again. keep an element pointer to
356 // the next element in each collections list
357 NodeList meta_responses = meta_result.getElementsByTagName(GSXML.RESPONSE_ELEM);
358 for (int i=0; i<meta_responses.getLength(); i++) {
359 String collname = GSPath.getFirstLink(((Element)meta_responses.item(i)).getAttribute(GSXML.FROM_ATT));
360 Element first_elem = (Element)GSXML.getNodeByPath(meta_responses.item(i), "documentNodeList/documentNode");
361 coll_map.put(collname, first_elem);
362 }
363
364 for (int i=0; i<query_doc_list.getLength(); i++) {
365 Element doc_node = (Element)query_doc_list.item(i);
366 Element new_node = (Element)this.doc.importNode(doc_node, false);
367 result_node_list.appendChild(new_node);
368 String coll_name = doc_node.getAttribute("collection");
369
370 Element meta_elem = (Element)coll_map.get(coll_name);
371 GSXML.mergeMetadataLists(new_node, meta_elem);
372 coll_map.put(coll_name, meta_elem.getNextSibling());
373 }
374 return result;
375 }
376}
377
Note: See TracBrowser for help on using the repository browser.