source: main/trunk/greenstone3/src/java/org/greenstone/gsdl3/service/AbstractSearch.java@ 25885

Last change on this file since 25885 was 25885, checked in by sjm84, 12 years ago

Removed the Solr code from the main greenstone code
  • Property svn:keywords set to Author Date Id Revision
File size: 12.6 KB
Line 
1/*
2 * AbstractSearch.java
3 * Copyright (C) 2005 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
21// Greenstone classes
22import java.util.ArrayList;
23
24import org.apache.log4j.Logger;
25import org.greenstone.gsdl3.util.AbstractSimpleDocument;
26import org.greenstone.gsdl3.util.GSPath;
27import org.greenstone.gsdl3.util.GSXML;
28import org.greenstone.gsdl3.util.SimpleDocument;
29import org.w3c.dom.Element;
30import org.w3c.dom.NodeList;
31
32/**
33 * Partially implements a generic search service
34 *
35 * @author Katherine Don
36 */
37
38public abstract class AbstractSearch extends ServiceRack
39{
40
41 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.AbstractSearch.class.getName());
42
43 // the search service
44 protected String QUERY_SERVICE = null; // set by subclass
45
46 // compulsory params
47 protected static final String INDEX_PARAM = "index";
48 protected static final String QUERY_PARAM = "query";
49 protected static final String RAW_PARAM = "rawquery";
50
51 // optional standard params - some of these have to be implemented
52 protected static final String MAXDOCS_PARAM = "maxDocs";
53 protected static final String HITS_PER_PAGE_PARAM = "hitsPerPage";
54 protected static final String START_PAGE_PARAM = "startPage";
55
56 protected AbstractSimpleDocument gs_doc = null;
57
58 /** can more than one index be searched at the same time? */
59 protected boolean does_multi_index_search = false;
60 /** does this service support paging of results? */
61 protected boolean does_paging = false;
62 /** does this service support asking for a subset of results? */
63 protected boolean does_chunking = false;
64 /** does this service support faceting search results */
65 protected boolean does_faceting = false;
66 /**
67 * the default document type - use if all documents are the same type
68 */
69 protected String default_document_type = null;
70 /**
71 * the default index, or comma separated list if more than one is the
72 * default (with start and end commas, eg ,TI,SU,). Should be set by
73 * configure()
74 */
75 protected String default_index = "";
76
77 protected String default_max_docs = "100";
78
79 protected String default_hits_per_page = "10";
80
81 public AbstractSearch()
82 {
83 }
84
85 /**
86 * Sets up the short service info for service by QUERY_SERVICE (e.g.
87 * TextQuery or AudioQuery) If other services will be provided, should be
88 * added in the subclass configure also looks for search format info, and
89 * document format info
90 */
91 public boolean configure(Element info, Element extra_info)
92 {
93 if (!super.configure(info, extra_info))
94 {
95 return false;
96 }
97
98 logger.info("Configuring AbstractSearch...");
99
100 this.config_info = info;
101
102 // set up short_service_info_
103 // => for now just has id and type. the name (lang dependent)
104 // will be added in if the list is requested.
105
106 Element tq_service = this.doc.createElement(GSXML.SERVICE_ELEM);
107 tq_service.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_QUERY);
108 tq_service.setAttribute(GSXML.NAME_ATT, QUERY_SERVICE);
109 this.short_service_info.appendChild(tq_service);
110
111 // add some format info to service map if there is any
112 // => lookin extra info first look in buildConfig
113
114 Element format = (Element) GSXML.getChildByTagName(info, GSXML.FORMAT_ELEM);
115
116 if (format == null)
117 {
118 String path = GSPath.appendLink(GSXML.SEARCH_ELEM, GSXML.FORMAT_ELEM);
119
120 // Note by xiao: instead of retrieving the first 'format'
121 // element inside the 'search' element, we are trying to
122 // find the real format element which has at least one
123 // 'gsf:template' child element. (extra_info is
124 // collectionConfig.xml)
125
126 //format = (Element) GSXML.getNodeByPath(extra_info, path);
127
128 Element search_elem = (Element) GSXML.getChildByTagName(extra_info, GSXML.SEARCH_ELEM);
129 NodeList format_elems = null;
130 if (search_elem != null)
131 {
132 format_elems = search_elem.getElementsByTagName(GSXML.FORMAT_ELEM);
133 }
134 for (int i = 0; i < format_elems.getLength(); i++)
135 {
136 format = (Element) format_elems.item(i);
137 if (format.getElementsByTagName("gsf:template").getLength() != 0)
138 {
139 break;
140 }
141 }
142 }//end of if(format==null)
143 //
144 if (format != null)
145 {
146 this.format_info_map.put(QUERY_SERVICE, this.doc.importNode(format, true));
147 }
148
149 // look for document display format - for documentType
150 String path = GSPath.appendLink(GSXML.DISPLAY_ELEM, GSXML.FORMAT_ELEM);
151 Element display_format = (Element) GSXML.getNodeByPath(extra_info, path);
152 if (display_format != null)
153 {
154 // check for docType option.
155 Element doc_type_opt = GSXML.getNamedElement(display_format, "gsf:option", GSXML.NAME_ATT, "documentType");
156 if (doc_type_opt != null)
157 {
158 String value = doc_type_opt.getAttribute(GSXML.VALUE_ATT);
159 if (!value.equals(""))
160 {
161 this.default_document_type = value;
162 }
163 }
164 }
165
166 // Base line for document (might be overriden by sub-classes)
167 gs_doc = new SimpleDocument(this.doc, this.default_document_type);
168
169 return true;
170 }
171
172 /**
173 * returns a basic description for QUERY_SERVICE. If a subclass provides
174 * other services they need to provide their own descriptions
175 */
176 protected Element getServiceDescription(String service, String lang, String subset)
177 {
178 if (!service.equals(QUERY_SERVICE))
179 {
180 return null;
181 }
182
183 Element tq_service = this.doc.createElement(GSXML.SERVICE_ELEM);
184 tq_service.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_QUERY);
185 tq_service.setAttribute(GSXML.NAME_ATT, QUERY_SERVICE);
186 if (subset == null || subset.equals(GSXML.DISPLAY_TEXT_ELEM + GSXML.LIST_MODIFIER))
187 {
188 tq_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_NAME, getServiceName(QUERY_SERVICE, lang)));
189 tq_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_SUBMIT, getServiceSubmit(QUERY_SERVICE, lang)));
190 tq_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_DESCRIPTION, getServiceDescription(QUERY_SERVICE, lang)));
191 }
192 if (subset == null || subset.equals(GSXML.PARAM_ELEM + GSXML.LIST_MODIFIER))
193 {
194 Element param_list = this.doc.createElement(GSXML.PARAM_ELEM + GSXML.LIST_MODIFIER);
195 addCustomQueryParams(param_list, lang);
196 addStandardQueryParams(param_list, lang);
197 tq_service.appendChild(param_list);
198 }
199 return tq_service;
200
201 }
202
203 // perhaps these should be changed to search down the class hierarchy for
204 // values - do can just put the info in the resource bundle to use it
205 /** returns the default name for the TextQuery service */
206 protected String getServiceName(String service_id, String lang)
207 {
208 return getTextString(service_id + ".name", lang);
209 }
210
211 /** returns the default description for the TextQuery service */
212 protected String getServiceDescription(String service_id, String lang)
213 {
214 return getTextString(service_id + ".description", lang);
215 }
216
217 /** returns the default submit button text for the TextQuery service */
218 protected String getServiceSubmit(String service_id, String lang)
219 {
220 return getTextString(service_id + ".submit", lang);
221
222 }
223
224 /** adds the standard query params into the service description */
225 protected void addStandardQueryParams(Element param_list, String lang)
226 {
227 // this test is not so good. here we are using absence of default index
228 // to determine whether we have indexes or not. But in other places,
229 // absence of default index just means to use the first one as default.
230 if (!default_index.equals(""))
231 {
232 createParameter(INDEX_PARAM, param_list, lang);
233 }
234 if (does_chunking)
235 {
236 createParameter(MAXDOCS_PARAM, param_list, lang);
237 }
238 if (does_paging)
239 {
240 createParameter(HITS_PER_PAGE_PARAM, param_list, lang);
241 createParameter(START_PAGE_PARAM, param_list, lang);
242 }
243 createParameter(QUERY_PARAM, param_list, lang);
244 }
245
246 /**
247 * adds any service specific query params into the service default
248 * implementation: add nothing. subclasses may need to override this to add
249 * in their specific parameters
250 */
251 protected void addCustomQueryParams(Element param_list, String lang)
252 {
253 // default behaviour, do nothing
254 }
255
256 protected void createParameter(String name, Element param_list, String lang)
257 {
258 createParameter(name, param_list, lang, null);
259 }
260
261 protected void createParameter(String name, Element param_list, String lang, String default_value)
262 {
263 // at this level, not interested in boolean return type
264 createParameterChain(name, param_list, lang, default_value);
265 }
266
267 /**
268 * default implementations for the standard parameters plus some other
269 * common ones index, maxDocs, hitsPerPage, startPage
270 */
271
272 protected boolean createParameterChain(String name, Element param_list, String lang, String default_value)
273 {
274 Element param = null;
275 String param_default = default_value;
276
277 if (name.equals(QUERY_PARAM) || name.equals(RAW_PARAM))
278 {
279 param = GSXML.createParameterDescription(this.doc, name, getTextString("param." + name, lang), GSXML.PARAM_TYPE_STRING, param_default, null, null);
280 param_list.appendChild(param);
281 return true;
282 }
283 else if (name.equals(INDEX_PARAM))
284 {
285 // should we make these class fields?
286 ArrayList<String> index_ids = new ArrayList<String>();
287 ArrayList<String> index_names = new ArrayList<String>();
288 getIndexData(index_ids, index_names, lang);
289 String param_type = GSXML.PARAM_TYPE_ENUM_SINGLE;
290 if (does_multi_index_search)
291 {
292 param_type = GSXML.PARAM_TYPE_ENUM_MULTI;
293 }
294 if (param_default == null)
295 {
296 param_default = this.default_index;
297 }
298 param = GSXML.createParameterDescription2(this.doc, INDEX_PARAM, getTextString("param." + INDEX_PARAM, lang), param_type, param_default, index_ids, index_names);
299 param_list.appendChild(param);
300 return true;
301 }
302 else if (name.equals(MAXDOCS_PARAM))
303 {
304 if (param_default == null)
305 {
306 param_default = this.default_max_docs;
307 }
308
309 param = GSXML.createParameterDescription(this.doc, name, getTextString("param." + name, lang), GSXML.PARAM_TYPE_INTEGER, param_default, null, null);
310 param_list.appendChild(param);
311 return true;
312 }
313 else if (name.equals(HITS_PER_PAGE_PARAM))
314 {
315 if (param_default == null)
316 {
317 param_default = this.default_hits_per_page;
318 }
319
320 param = GSXML.createParameterDescription(this.doc, name, getTextString("param." + name, lang), GSXML.PARAM_TYPE_INTEGER, param_default, null, null);
321 param_list.appendChild(param);
322 return true;
323 }
324 else if (name.equals(START_PAGE_PARAM))
325 {
326 if (param_default == null)
327 {
328 param_default = "1";
329 }
330
331 // start page - set to 1 for the search page
332 param = GSXML.createParameterDescription(this.doc, START_PAGE_PARAM, "", GSXML.PARAM_TYPE_INVISIBLE, param_default, null, null);
333 param_list.appendChild(param);
334 return true;
335 }
336
337 // Get to there then none of the above params matched
338 // => return false so the chain can continue
339 return false;
340 }
341
342 /**
343 * create an element to go into the search results list. A node element has
344 * the form <docNode nodeId='xxx' nodeType='leaf' docType='hierarchy'
345 * rank='0.23'/>
346 */
347 protected Element createDocNode(String node_id, String rank)
348 {
349 return this.gs_doc.createDocNode(node_id, rank);
350 }
351
352 /**
353 * returns the document type of the doc that the specified node belongs to.
354 * should be one of GSXML.DOC_TYPE_SIMPLE, GSXML.DOC_TYPE_PAGED,
355 * GSXML.DOC_TYPE_HIERARCHY
356 */
357 protected String getDocType(String node_id)
358 {
359 return this.gs_doc.getDocType(node_id);
360 }
361
362 /**
363 * returns the node type of the specified node. should be one of
364 * GSXML.NODE_TYPE_LEAF, GSXML.NODE_TYPE_INTERNAL, GSXML.NODE_TYPE_ROOT
365 */
366 protected String getNodeType(String node_id, String doc_type)
367 {
368 return this.gs_doc.getNodeType(node_id, doc_type);
369 }
370
371 /** returns true if the node has child nodes */
372 protected boolean hasChildren(String node_id)
373 {
374 return this.gs_doc.hasChildren(node_id);
375 }
376
377 /** returns true if the node has a parent */
378 protected boolean hasParent(String node_id)
379 {
380 return this.gs_doc.hasParent(node_id);
381 }
382
383 /**
384 * get the details about the indexes available must be implemented by
385 * subclass there must be at least one index
386 */
387 abstract protected void getIndexData(ArrayList<String> index_ids, ArrayList<String> index_names, String lang);
388
389}
Note: See TracBrowser for help on using the repository browser.