source: main/trunk/greenstone3/src/java/org/greenstone/gsdl3/action/GS2BrowseAction.java@ 24254

Last change on this file since 24254 was 24254, checked in by ak19, 13 years ago

Commits for ticket 770 concerning the display of multiple values for a metadata (like dc.Title) when classified by that metadata. So when the user browses by dc.Title, they no longer merely see a doc listed once for each dc.Title assigned but under the same (first retrieved) dc.Title, but they should now see the doc listed once for each dc.Title assigned to it with a different dc.Title value each time.
  • Property svn:keywords set to Author Date Id Revision
File size: 14.3 KB
Line 
1package org.greenstone.gsdl3.action;
2
3import org.greenstone.gsdl3.core.ModuleInterface;
4import org.greenstone.gsdl3.util.*;
5// XML classes
6import org.w3c.dom.Node;
7import org.w3c.dom.NodeList;
8import org.w3c.dom.Document;
9import org.w3c.dom.Element;
10
11import java.util.HashMap;
12import java.util.HashSet;
13import java.util.Vector;
14import java.io.File;
15
16import org.apache.log4j.*;
17
18/** action for GS2 style classifier browsing */
19public class GS2BrowseAction extends Action {
20
21 public static final String CLASSIFIER_ARG = "cl";
22
23 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.action.GS2BrowseAction.class.getName());
24
25 /** process the request */
26 public Node process (Node message_node) {
27
28 Element message = this.converter.nodeToElement(message_node);
29
30 // get the request - assume only one
31 Element request = (Element)GSXML.getChildByTagName(message, GSXML.REQUEST_ELEM);
32
33 // the result
34 Element result = this.doc.createElement(GSXML.MESSAGE_ELEM);
35 Element response = classifierBrowse(request);
36 result.appendChild(response);
37 return result;
38 }
39
40
41 protected Element classifierBrowse(Element request) {
42
43 Element page_response = this.doc.createElement(GSXML.RESPONSE_ELEM);
44
45 // extract the params from the cgi-request, and check that we have a coll specified
46 Element cgi_paramList = (Element)GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
47 HashMap params = GSXML.extractParams(cgi_paramList, false);
48
49 String service_name = (String)params.get(GSParams.SERVICE);
50 String collection = (String)params.get(GSParams.COLLECTION);
51 if (collection == null || collection.equals("")) {
52 logger.error("classifierBrowse, need to specify a collection!");
53 return page_response;
54
55 }
56
57 String lang = request.getAttribute(GSXML.LANG_ATT);
58 String uid = request.getAttribute(GSXML.USER_ID_ATT);
59 String to = GSPath.appendLink(collection, service_name);
60
61 // the first part of the response is the service description
62 // for now get this again from the service.
63 // this should be cached somehow later on.
64
65 Element info_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
66 Element info_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_DESCRIBE, to, lang, uid);
67 info_message.appendChild(info_request);
68
69 // also get the format stuff now if there is some
70 Element format_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_FORMAT, to, lang, uid);
71 info_message.appendChild(format_request);
72 // process the requests
73
74 Element info_response = (Element) this.mr.process(info_message);
75
76 // the two responses
77 NodeList responses = info_response.getElementsByTagName(GSXML.RESPONSE_ELEM);
78 Element service_response = (Element)responses.item(0);
79 Element format_response = (Element)responses.item(1);
80
81 Element service_description = (Element)GSXML.getChildByTagName(service_response, GSXML.SERVICE_ELEM);
82 page_response.appendChild(this.doc.importNode(service_description, true));
83
84 //append site metadata
85 addSiteMetadata( page_response, lang, uid);
86
87 // if rt=d, then we are just displaying the service
88 String request_type = (String)params.get(GSParams.REQUEST_TYPE);
89 if (request_type.equals("d")) {
90 //return the page that we have so far
91 return page_response;
92 }
93
94 // get the node that the user has clicked on
95 String classifier_node = (String)params.get(CLASSIFIER_ARG);
96
97 // if the node is not defined, return the page that we have so far
98 if (classifier_node ==null || classifier_node.equals("")) {
99 return page_response;
100 }
101
102 // the id of the classifier is the top id of the selected node
103 String top_id = OID.getTop(classifier_node);
104 HashSet doc_meta_names = new HashSet();
105 HashSet class_meta_names = new HashSet();
106 // add in the defaults
107 doc_meta_names.add("Title");
108 class_meta_names.add("Title");
109
110 // add the format info into the response
111 Element format_elem = (Element)GSXML.getChildByTagName(format_response, GSXML.FORMAT_ELEM);
112 if (format_elem != null) {
113
114 // find the one for the classifier we are in
115 Element this_format = GSXML.getNamedElement(format_elem, GSXML.CLASSIFIER_ELEM, GSXML.NAME_ATT, top_id);
116 if (this_format == null) {
117 this_format = (Element)GSXML.getChildByTagName(format_elem, GSXML.DEFAULT_ELEM);
118 }
119
120 if (this_format != null) {
121 Element new_format = GSXML.duplicateWithNewName(this.doc, this_format, GSXML.FORMAT_ELEM, false);
122 extractMetadataNames(new_format, doc_meta_names, class_meta_names);
123 // set the format type
124 new_format.setAttribute(GSXML.TYPE_ATT, "browse");
125
126 page_response.appendChild(new_format);
127 }
128 }
129
130 // find out if this classifier is horizontal at top
131 Element class_list = (Element)GSXML.getChildByTagName(service_description, GSXML.CLASSIFIER_ELEM+GSXML.LIST_MODIFIER);
132 Element this_classifier = GSXML.getNamedElement(class_list, GSXML.CLASSIFIER_ELEM, GSXML.NAME_ATT, top_id);
133 boolean horizontal_at_top = false;
134 if (!this_classifier.getAttribute("horizontalAtTop").equals("")) {
135 horizontal_at_top = true;
136 }
137 if (top_id.equals(classifier_node) && horizontal_at_top) {
138 // we have asked for a top node - if the first list is horizontal, we will select the first element of that list
139 // this is a hack. also it craps out if the classifier really isn't horizontalAtTop. -
140 classifier_node = classifier_node+".1";
141
142 }
143
144 // get the browse structure for the selected node
145 Element classify_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
146 Element classify_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_PROCESS, to, lang, uid);
147 classify_message.appendChild(classify_request);
148
149 //Create a parameter list to specify the required structure information
150 // for now, always get ancestors and children
151 Element param_list = this.doc.createElement(GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
152 classify_request.appendChild(param_list);
153 Element param = this.doc.createElement(GSXML.PARAM_ELEM);
154 param_list.appendChild(param);
155 param.setAttribute(GSXML.NAME_ATT, "structure");
156 param.setAttribute(GSXML.VALUE_ATT, "ancestors");
157 param = this.doc.createElement(GSXML.PARAM_ELEM);
158 param_list.appendChild(param);
159 param.setAttribute(GSXML.NAME_ATT, "structure");
160 param.setAttribute(GSXML.VALUE_ATT, "children");
161
162 // put the classifier node into a classifier node list
163 Element classifier_list = this.doc.createElement(GSXML.CLASS_NODE_ELEM+GSXML.LIST_MODIFIER);
164 Element classifier = this.doc.createElement(GSXML.CLASS_NODE_ELEM);
165 classifier.setAttribute(GSXML.NODE_ID_ATT, classifier_node);
166 classifier_list.appendChild(classifier);
167 classify_request.appendChild(classifier_list);
168
169 if (horizontal_at_top && !classifier_node.equals(top_id)) {
170 // also put the top id in, to get the persistant horizontal info
171 classifier = this.doc.createElement(GSXML.CLASS_NODE_ELEM);
172 classifier.setAttribute(GSXML.NODE_ID_ATT, top_id);
173 classifier_list.appendChild(classifier);
174 }
175 // process the request
176 Element classify_response = (Element)this.mr.process(classify_message);
177 String path = GSPath.appendLink(GSXML.RESPONSE_ELEM, GSXML.CLASS_NODE_ELEM+GSXML.LIST_MODIFIER);
178 Element class_node_list = (Element)GSXML.getNodeByPath(classify_response, path);
179
180 path = GSPath.appendLink(GSXML.CLASS_NODE_ELEM, GSXML.NODE_STRUCTURE_ELEM);
181 // assume that we always get back the top level CL1 node - this becomes the page_classifier node
182 path = GSPath.appendLink(path, GSXML.CLASS_NODE_ELEM);
183 Element cl_structure = (Element)GSXML.getNodeByPath(class_node_list,
184 path);
185 if (cl_structure ==null) {
186 logger.error("classifier structure request returned no structure");
187 return page_response;
188 }
189
190 Element page_classifier = null;
191 if (horizontal_at_top && !classifier_node.equals(top_id)) {
192 // get the info for the top node
193 Element top_node = GSXML.getNamedElement(class_node_list, GSXML.CLASS_NODE_ELEM, GSXML.NODE_ID_ATT, top_id);
194 if (top_node !=null) {
195 path = GSPath.appendLink(GSXML.NODE_STRUCTURE_ELEM, GSXML.CLASS_NODE_ELEM);
196 Element top_structure = (Element)GSXML.getNodeByPath(top_node, path);
197 // add this as the classifier elem
198 page_classifier = GSXML.duplicateWithNewName(this.doc, top_structure, GSXML.CLASSIFIER_ELEM, true);
199 page_response.appendChild(page_classifier);
200 // now replace the child with the structure from the other request
201 Element new_classifier = (Element)GSXML.getChildByTagName(cl_structure, GSXML.CLASS_NODE_ELEM);
202 String replace_name = new_classifier.getAttribute(GSXML.NODE_ID_ATT);
203
204 // find the appropriate child
205 Element old_classifier = GSXML.getNamedElement(page_classifier, GSXML.CLASS_NODE_ELEM, GSXML.NODE_ID_ATT, replace_name);
206 page_classifier.replaceChild(this.doc.importNode(new_classifier, true), old_classifier);
207 page_classifier.setAttribute(GSXML.NAME_ATT, top_id);
208 } else {
209 // add the single classifier node as the page classifier
210 page_classifier = GSXML.duplicateWithNewName(this.doc, cl_structure, GSXML.CLASSIFIER_ELEM, true);
211 page_response.appendChild(page_classifier);
212 page_classifier.setAttribute(GSXML.NAME_ATT, top_id);
213 }
214
215 } else {
216 // add the single classifier node as the page classifier
217 page_classifier = GSXML.duplicateWithNewName(this.doc, cl_structure, GSXML.CLASSIFIER_ELEM, true);
218 page_response.appendChild(page_classifier);
219 page_classifier.setAttribute(GSXML.NAME_ATT, top_id);
220 }
221 // get the metadata for each classifier node,
222 // then for each document node
223
224 Element metadata_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
225
226 boolean did_classifier = false;
227 boolean did_documents = false;
228
229
230 // if there are classifier nodes
231 // create a metadata request for the classifier, and add it to
232 // the the message
233 NodeList cl_nodes = page_classifier.getElementsByTagName(GSXML.CLASS_NODE_ELEM);
234
235 if (cl_nodes.getLength() > 0) {
236 did_classifier = true;
237 Element cl_meta_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_PROCESS, to+"MetadataRetrieve", lang, uid);
238 metadata_message.appendChild(cl_meta_request);
239
240 Element new_cl_nodes_list = this.doc.createElement(GSXML.CLASS_NODE_ELEM+GSXML.LIST_MODIFIER);
241 cl_meta_request.appendChild(new_cl_nodes_list);
242
243 for (int c=0; c<cl_nodes.getLength(); c++) {
244
245 Element cl = this.doc.createElement(GSXML.CLASS_NODE_ELEM);
246 cl.setAttribute(GSXML.NODE_ID_ATT, ((Element)cl_nodes.item(c)).getAttribute(GSXML.NODE_ID_ATT));
247 new_cl_nodes_list.appendChild(cl);
248 }
249
250 // create and add in the param list - for now get all the metadata
251 // should be based on info sent in from the recept, and the
252 // format stuff
253 Element cl_param_list = createMetadataParamList(class_meta_names);
254 cl_meta_request.appendChild(cl_param_list);
255
256 }
257
258 // if there are document nodes in the classification (happens
259 // sometimes), create a second request for document metadata and
260 // append to the message
261 NodeList doc_nodes = page_classifier.getElementsByTagName(GSXML.DOC_NODE_ELEM);
262 if (doc_nodes.getLength() > 0) {
263 did_documents = true;
264 Element doc_meta_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_PROCESS, GSPath.appendLink(collection, "DocumentMetadataRetrieve"), lang, uid);
265 metadata_message.appendChild(doc_meta_request);
266
267 Element doc_list = this.doc.createElement(GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
268 doc_meta_request.appendChild(doc_list);
269
270 for (int c=0; c<doc_nodes.getLength(); c++) {
271
272 Element d = this.doc.createElement(GSXML.DOC_NODE_ELEM);
273 d.setAttribute(GSXML.NODE_ID_ATT, ((Element)doc_nodes.item(c)).getAttribute(GSXML.NODE_ID_ATT));
274 if(((Element)doc_nodes.item(c)).hasAttribute(GSXML.NODE_MDOFFSET_ATT)) {
275 d.setAttribute(GSXML.NODE_MDOFFSET_ATT, ((Element)doc_nodes.item(c)).getAttribute(GSXML.NODE_MDOFFSET_ATT));
276 }
277 doc_list.appendChild(d);
278 }
279
280 // create and add in the param list - add all for now
281 Element doc_param_list = createMetadataParamList(doc_meta_names);
282 doc_meta_request.appendChild(doc_param_list);
283
284 }
285
286 // process the metadata requests
287 Element metadata_response = (Element)this.mr.process(metadata_message);
288 if (did_classifier) {
289 // the classifier one will be the first response
290 // add the metadata lists for each node back into the
291 // page_classifier nodes
292 path = GSPath.appendLink(GSXML.RESPONSE_ELEM,
293 GSXML.CLASS_NODE_ELEM+GSXML.LIST_MODIFIER);
294 Node pathNode = GSXML.getNodeByPath(metadata_response, path);
295 if (pathNode ==null){
296 return page_response;
297 }
298 //NodeList meta_response_cls = (Element)pathNode.getChildNodes(); // can't handle empty elements from converting formatted strings (with empty newlines) into XML
299 NodeList meta_response_cls = ((Element)pathNode).getElementsByTagName(GSXML.CLASS_NODE_ELEM);
300 for (int i = 0; i < cl_nodes.getLength(); i++) {
301 GSXML.mergeMetadataLists(cl_nodes.item(i), meta_response_cls.item(i));
302 }
303 }
304
305 if (did_documents) {
306 NodeList meta_response_docs = null;
307 if (!did_classifier) {
308 // its the first response
309 path = GSPath.appendLink(GSXML.RESPONSE_ELEM,
310 GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
311 Node pathNode = GSXML.getNodeByPath(metadata_response, path);
312 if (pathNode ==null){
313 return page_response;
314 }
315
316 meta_response_docs = pathNode.getChildNodes();
317
318
319 } else { // its the second response
320 Node nodes = GSXML.getChildByTagName(metadata_response.getElementsByTagName(GSXML.RESPONSE_ELEM).item(1), GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
321 if (nodes ==null){
322 return page_response;
323 }
324 meta_response_docs = nodes.getChildNodes();
325 }
326
327 for (int i = 0; i < doc_nodes.getLength(); i++) {
328 GSXML.mergeMetadataLists(doc_nodes.item(i), meta_response_docs.item(i));
329 }
330 }
331
332
333 logger.debug("(GS2BrowseAction) Page:\n" + this.converter.getPrettyString(page_response));
334 return page_response;
335 }
336
337
338 protected void extractMetadataNames(Element new_format, HashSet doc_meta_names, HashSet class_meta_names) {
339
340 NodeList templates = new_format.getElementsByTagName("gsf:template");
341 for (int i=0; i<templates.getLength(); i++) {
342 Element template = (Element)templates.item(i);
343 String match = template.getAttribute("match");
344 if (match.equals("documentNode")) {
345 extractMetadataNames(template, doc_meta_names);
346 } else if (match.equals("classifierNode")) {
347 extractMetadataNames(template, class_meta_names);
348 }
349 }
350 }
351
352}
353
354
Note: See TracBrowser for help on using the repository browser.