source: greenstone3/trunk/src/java/org/greenstone/gsdl3/action/QueryAction.java@ 19984

Last change on this file since 19984 was 19984, checked in by oranfry, 15 years ago

initial implementation of support for site-level metadata
  • Property svn:keywords set to Author Date Id Revision
File size: 10.6 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.Text;
9import org.w3c.dom.Document;
10import org.w3c.dom.Element;
11
12import java.util.HashMap;
13import java.util.HashSet;
14import java.util.Vector;
15import java.util.Map;
16import java.util.Iterator;
17import java.io.File;
18
19import org.apache.log4j.*;
20
21/** action class for queries */
22public class QueryAction extends Action {
23
24
25 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.action.QueryAction.class.getName());
26
27 /** process - processes a request.
28 */
29 public Node process (Node message_node) {
30
31 Element message = this.converter.nodeToElement(message_node);
32
33 // get the request - assume there is only one
34 Element request = (Element)GSXML.getChildByTagName(message, GSXML.REQUEST_ELEM);
35
36 // create the return message
37 Element result = this.doc.createElement(GSXML.MESSAGE_ELEM);
38 Element response = basicQuery(request);
39 result.appendChild(this.doc.importNode(response, true));
40 return result;
41 }
42
43 /** a generic query handler
44 * this gets the service description, does the query (just passes all the
45 * params to the service, then gets the titles for any results
46 */
47 protected Element basicQuery(Element request) {
48
49 // the result
50 Element page_response = this.doc.createElement(GSXML.RESPONSE_ELEM);
51
52 // extract the params from the cgi-request, and check that we have a coll specified
53 Element cgi_param_list = (Element)GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
54 HashMap params = GSXML.extractParams(cgi_param_list, false);
55
56 String request_type = (String)params.get(GSParams.REQUEST_TYPE);
57 String service_name = (String)params.get(GSParams.SERVICE);
58 String collection = (String)params.get(GSParams.COLLECTION);
59
60 // collection may be null or empty when we are doing cross coll services
61 if (collection == null || collection.equals("")) {
62 collection = null;
63 }
64
65 String lang = request.getAttribute(GSXML.LANG_ATT);
66 String uid = request.getAttribute(GSXML.USER_ID_ATT);
67 String to = service_name;
68 if (collection != null) {
69 to = GSPath.prependLink(to, collection);
70 }
71
72 if (request_type.indexOf("d")!=-1) {
73 // we have been asked for the service description
74 Element mr_info_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
75 Element mr_info_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_DESCRIBE, to, lang, uid);
76 mr_info_message.appendChild(mr_info_request);
77
78 // process the message
79 Element mr_info_response = (Element) this.mr.process(mr_info_message);
80 // the response
81 Element service_response = (Element)GSXML.getChildByTagName(mr_info_response, GSXML.RESPONSE_ELEM);
82
83 Element service_description = (Element)this.doc.importNode(GSXML.getChildByTagName(service_response, GSXML.SERVICE_ELEM), true);
84 page_response.appendChild(service_description);
85 }
86
87 if (request_type.indexOf("r") == -1) {
88 // just a display request, no actual processing to do
89 System.err.println("just a display request, no actual processing to do");
90 //append site metadata
91 addSiteMetadata( page_response, lang, uid);
92 return page_response;
93 }
94
95 // check that we have some service params
96 HashMap service_params = (HashMap)params.get("s1");
97 if (service_params == null) { // no query
98 //append site metadata
99 addSiteMetadata( page_response, lang, uid);
100 return page_response;
101 }
102
103 // create the query request
104 Element mr_query_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
105 Element mr_query_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_PROCESS, to, lang, uid);
106 mr_query_message.appendChild(mr_query_request);
107
108 Element query_param_list = this.doc.createElement(GSXML.PARAM_ELEM + GSXML.LIST_MODIFIER);
109 GSXML.addParametersToList(this.doc, query_param_list, service_params);
110 mr_query_request.appendChild(query_param_list);
111
112 // also get the format stuff now if there is some
113 Element format_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_FORMAT, to, lang, uid);
114 mr_query_message.appendChild(format_request);
115
116 logger.debug(GSXML.xmlNodeToString(mr_query_message));
117
118 // do the query
119 Element mr_query_response = (Element)this.mr.process(mr_query_message);
120
121 // check for errors
122 if (processErrorElements(mr_query_response, page_response)) {
123 //append site metadata
124 addSiteMetadata( page_response, lang, uid);
125 return page_response;
126 }
127
128 NodeList responses = mr_query_response.getElementsByTagName(GSXML.RESPONSE_ELEM);
129 Element query_response = (Element) responses.item(0);
130 Element format_response = (Element) responses.item(1);
131
132 Element query_result_metadata_list = (Element) GSXML.getChildByTagName(query_response, GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER);
133 if (query_result_metadata_list == null) {
134 logger.error("No query result metadata.\n");
135 } else { // add it into the page response
136 page_response.appendChild(this.doc.importNode(query_result_metadata_list, true));
137 }
138
139 Element query_term_info_list = (Element) GSXML.getChildByTagName(query_response, GSXML.TERM_ELEM+GSXML.LIST_MODIFIER);
140 if (query_term_info_list == null) {
141 logger.error("No query term information.\n");
142 } else { // add it into the page response
143 page_response.appendChild(this.doc.importNode(query_term_info_list, true));
144 }
145
146 // check that there are some documents - for now check the list, but later should use a numdocs metadata elem
147 Element document_list = (Element)GSXML.getChildByTagName(query_response, GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
148 // documentList not present if no docs found
149 if (document_list == null) {
150 // add in a dummy doc node list - used by the display. need to think about this
151 page_response.appendChild(this.doc.createElement(GSXML.DOC_NODE_ELEM + GSXML.LIST_MODIFIER));
152 //append site metadata
153 addSiteMetadata( page_response, lang, uid);
154 return page_response;
155 }
156
157 // now we check to see if there is metadata already - some search services return predefined metadata. if there is some, don't do a metadata request
158 NodeList doc_metadata = document_list.getElementsByTagName(GSXML.METADATA_ELEM+GSXML.LIST_MODIFIER);
159 if (doc_metadata.getLength()>0) {
160 logger.error("have already found metadata!");
161 // append the doc list to the result
162 page_response.appendChild(this.doc.importNode(document_list, true));
163 //append site metadata
164 addSiteMetadata( page_response, lang, uid);
165 return page_response;
166 }
167
168 // get the metadata elements needed from the format statement if any
169 HashSet metadata_names = new HashSet();
170 metadata_names.add("Title");
171 // add in the format info to the stylesheet if there is any
172 Element format_elem = (Element)GSXML.getChildByTagName(format_response, GSXML.FORMAT_ELEM);
173 if (format_elem != null) {
174 // set the format type
175 format_elem.setAttribute(GSXML.TYPE_ATT, "search");
176 // for now just add to the response
177 page_response.appendChild(this.doc.importNode(format_elem, true));
178 extractMetadataNames(format_elem, metadata_names);
179 }
180
181 // paging of the results is done here - we filter the list to remove unwanted entries before retrieving metadata
182 Element filtered_doc_list = filterDocList(params, service_params, document_list);
183
184 // do the metadata request on the filtered list
185 Element mr_metadata_message = this.doc.createElement(GSXML.MESSAGE_ELEM);
186 to = "DocumentMetadataRetrieve";
187 if (collection != null) {
188 to = GSPath.prependLink(to, collection);
189 }
190 Element mr_metadata_request = GSXML.createBasicRequest(this.doc, GSXML.REQUEST_TYPE_PROCESS, to, lang, uid);
191 mr_metadata_message.appendChild(mr_metadata_request);
192
193 // just get all for now - the receptionist should perhaps pass in some
194 // metadata that it wants, and QueryAction should look through the format stuff to see if there is any other?
195
196 Element dm_param_list = createMetadataParamList(metadata_names);
197
198 mr_metadata_request.appendChild(dm_param_list);
199
200 // add in the doc node list too
201 mr_metadata_request.appendChild(filtered_doc_list);
202
203 Element mr_metadata_response = (Element) this.mr.process(mr_metadata_message);
204 // check for errors
205 processErrorElements(mr_metadata_response, page_response);
206
207 Element metadata_response = (Element) GSXML.getChildByTagName(mr_metadata_response, GSXML.RESPONSE_ELEM);
208
209 Element query_result_document_list = (Element) GSXML.getChildByTagName(metadata_response, GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
210
211 if (query_result_document_list != null) {
212 page_response.appendChild(this.doc.importNode(query_result_document_list, true));
213 }
214
215 logger.debug("Query page:\n" + this.converter.getPrettyString(page_response));
216 //append site metadata
217 addSiteMetadata( page_response, lang, uid);
218 return page_response;
219 }
220
221 /** this filters out some of the doc results for result paging */
222 protected Element filterDocList(HashMap params, HashMap service_params, Element orig_doc_list) {
223
224 // check the hits_per_page param - is it a service param??
225 String hits_pp = (String) service_params.get("hitsPerPage");
226 if (hits_pp == null) {
227 // the service is doing the paging, so we want to display all of the returned docs(???)
228 // return (Element)this.doc.importNode(orig_doc_list, true);
229 // try hitsPerPage in the globle param
230 hits_pp = (String)params.get("hitsPerPage");
231 }
232
233 int hits = 20;
234 if (hits_pp != null && !hits_pp.equals("")) {
235 try {
236 hits = Integer.parseInt(hits_pp);
237 } catch (Exception e) {
238 hits=20;
239 }
240 }
241
242 if (hits == -1) { // all
243 return (Element)this.doc.importNode(orig_doc_list, true);
244 }
245 NodeList result_docs = orig_doc_list.getElementsByTagName(GSXML.DOC_NODE_ELEM);
246
247 int num_docs = result_docs.getLength();
248 if (num_docs <= hits) {
249 // too few docs to do paging
250 return (Element)this.doc.importNode(orig_doc_list, true);
251 }
252
253 // now we need our own doc list
254 Element result_list = this.doc.createElement(GSXML.DOC_NODE_ELEM + GSXML.LIST_MODIFIER);
255
256 String start_p = (String) service_params.get("startPage");
257 if(start_p == null){
258 start_p = (String)params.get("startPage");
259 }
260
261 int start = 1;
262 if (start_p != null && !start_p.equals("")) {
263 try {
264 start = Integer.parseInt(start_p);
265 } catch (Exception e) {
266 start = 1;
267 }
268 }
269
270 int start_from = (start-1)*hits;
271 int end_at = (start*hits)-1;
272
273 if (start_from > num_docs) {
274 // something has gone wrong
275 return result_list;
276 }
277
278 if (end_at > num_docs) {
279 end_at = num_docs-1;
280 }
281
282 // now we finally have the docs numbers to use
283 for (int i=start_from; i<=end_at; i++) {
284 result_list.appendChild(this.doc.importNode(result_docs.item(i), true));
285 }
286
287 return result_list;
288 }
289
290}
Note: See TracBrowser for help on using the repository browser.