1 | /*
|
---|
2 | * GS2SolrSearch.java
|
---|
3 | * Copyright (C) 2006 New Zealand Digital Library, http://www.nzdl.org
|
---|
4 | *
|
---|
5 | * This program is free software; you can redistribute it and/or modify
|
---|
6 | * the Free Software Foundation; either version 2 of the License, or
|
---|
7 | * (at your option) any later version.
|
---|
8 | *
|
---|
9 | * This program is distributed in the hope that it will be useful,
|
---|
10 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
11 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
12 | * GNU General Public License for more details.
|
---|
13 | *
|
---|
14 | * You should have received a copy of the GNU General Public License
|
---|
15 | * along with this program; if not, write to the Free Software
|
---|
16 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
17 | */
|
---|
18 |
|
---|
19 | package org.greenstone.gsdl3.service;
|
---|
20 |
|
---|
21 | // Greenstone classes
|
---|
22 | import java.io.File;
|
---|
23 | import java.util.ArrayList;
|
---|
24 | import java.util.HashMap;
|
---|
25 | import java.util.Iterator;
|
---|
26 | import java.util.List;
|
---|
27 | import java.util.Map;
|
---|
28 | import java.util.Set;
|
---|
29 | import java.util.Vector;
|
---|
30 |
|
---|
31 | import org.apache.log4j.Logger;
|
---|
32 | import org.apache.solr.client.solrj.embedded.EmbeddedSolrServer;
|
---|
33 | import org.apache.solr.client.solrj.response.FacetField;
|
---|
34 | import org.apache.solr.core.CoreContainer;
|
---|
35 | import org.greenstone.LuceneWrapper3.SharedSoleneQueryResult;
|
---|
36 | import org.greenstone.gsdl3.util.FacetWrapper;
|
---|
37 | import org.greenstone.gsdl3.util.GSFile;
|
---|
38 | import org.greenstone.gsdl3.util.GSXML;
|
---|
39 | import org.greenstone.gsdl3.util.SolrFacetWrapper;
|
---|
40 | import org.greenstone.gsdl3.util.SolrQueryResult;
|
---|
41 | import org.greenstone.gsdl3.util.SolrQueryWrapper;
|
---|
42 | import org.greenstone.util.GlobalProperties;
|
---|
43 | import org.w3c.dom.Element;
|
---|
44 | import org.w3c.dom.NodeList;
|
---|
45 |
|
---|
46 | public class GS2SolrSearch extends SharedSoleneGS2FieldSearch
|
---|
47 | {
|
---|
48 | static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.GS2SolrSearch.class.getName());
|
---|
49 |
|
---|
50 | static protected CoreContainer all_solr_cores = null;
|
---|
51 |
|
---|
52 | protected HashMap solr_core_cache;
|
---|
53 | protected SolrQueryWrapper solr_src = null;
|
---|
54 |
|
---|
55 | protected ArrayList<String> _facets = new ArrayList<String>();
|
---|
56 |
|
---|
57 | public GS2SolrSearch()
|
---|
58 | {
|
---|
59 | does_faceting = true;
|
---|
60 | // Used to store the solr cores that match the required 'level'
|
---|
61 | // of search (e.g. either document-level=>didx, or
|
---|
62 | // section-level=>sidx. The hashmap is filled out on demand
|
---|
63 | // based on 'level' parameter passed in to 'setUpQueryer()'
|
---|
64 |
|
---|
65 | solr_core_cache = new HashMap();
|
---|
66 |
|
---|
67 | if (all_solr_cores == null)
|
---|
68 | {
|
---|
69 | // Share one CoreContainer across all sites/collections
|
---|
70 | try
|
---|
71 | {
|
---|
72 | String gsdl3_home = GlobalProperties.getGSDL3Home();
|
---|
73 | String solr_ext_name = GlobalProperties.getProperty("gsdlext.solr.dirname", "solr");
|
---|
74 |
|
---|
75 | String solr_home_str = GSFile.extHome(gsdl3_home, solr_ext_name);
|
---|
76 |
|
---|
77 | all_solr_cores = new CoreContainer(solr_home_str);
|
---|
78 | }
|
---|
79 | catch (Exception e)
|
---|
80 | {
|
---|
81 | e.printStackTrace();
|
---|
82 | }
|
---|
83 | }
|
---|
84 |
|
---|
85 | this.solr_src = new SolrQueryWrapper();
|
---|
86 | }
|
---|
87 |
|
---|
88 | /** configure this service */
|
---|
89 | public boolean configure(Element info, Element extra_info)
|
---|
90 | {
|
---|
91 | boolean success = super.configure(info, extra_info);
|
---|
92 |
|
---|
93 | // 1. Make the CoreContainer reload solr.xml
|
---|
94 | // This is particularly needed for when activate.pl is executed during
|
---|
95 | // a running GS3 server. At that point, the solr collection is reactivated and
|
---|
96 | // we need to tell Greenstone that the solr index has changed. This requires
|
---|
97 | // the CoreContainer to reload the solr.xml file, and it all works again.
|
---|
98 |
|
---|
99 | solr_core_cache.clear(); // clear the map of existing solr cores
|
---|
100 |
|
---|
101 | // Reload the updated solr.xml into the CoreContainer
|
---|
102 | // (Doing an all_solr_cores.shutdown() first doesn't seem to be required)
|
---|
103 | try {
|
---|
104 | String solr_home_str = all_solr_cores.getSolrHome();
|
---|
105 | File solr_home = new File(solr_home_str);
|
---|
106 | File solr_xml = new File( solr_home,"solr.xml" );
|
---|
107 |
|
---|
108 | all_solr_cores.load(solr_home_str,solr_xml);
|
---|
109 | } catch (Exception e) {
|
---|
110 | logger.error("Exception in GS2SolrSearch.configure(): " + e.getMessage());
|
---|
111 | e.printStackTrace();
|
---|
112 | return false;
|
---|
113 | }
|
---|
114 |
|
---|
115 | if(!success) {
|
---|
116 | return false;
|
---|
117 | }
|
---|
118 |
|
---|
119 | // 2. Setting up facets
|
---|
120 | Element searchElem = (Element) GSXML.getChildByTagName(extra_info, GSXML.SEARCH_ELEM);
|
---|
121 | NodeList configIndexElems = searchElem.getElementsByTagName(GSXML.INDEX_ELEM);
|
---|
122 |
|
---|
123 | ArrayList<String> chosenFacets = new ArrayList<String>();
|
---|
124 | for (int i = 0; i < configIndexElems.getLength(); i++)
|
---|
125 | {
|
---|
126 | Element current = (Element) configIndexElems.item(i);
|
---|
127 | if (current.getAttribute(GSXML.FACET_ATT).equals("true"))
|
---|
128 | {
|
---|
129 | chosenFacets.add(current.getAttribute(GSXML.NAME_ATT));
|
---|
130 | }
|
---|
131 | }
|
---|
132 |
|
---|
133 | Element indexListElem = (Element) GSXML.getChildByTagName(info, GSXML.INDEX_ELEM + GSXML.LIST_MODIFIER);
|
---|
134 | NodeList buildIndexElems = indexListElem.getElementsByTagName(GSXML.INDEX_ELEM);
|
---|
135 |
|
---|
136 | for (int j = 0; j < buildIndexElems.getLength(); j++)
|
---|
137 | {
|
---|
138 | Element current = (Element) buildIndexElems.item(j);
|
---|
139 | for (int i = 0; i < chosenFacets.size(); i++)
|
---|
140 | {
|
---|
141 | if (current.getAttribute(GSXML.NAME_ATT).equals(chosenFacets.get(i)))
|
---|
142 | {
|
---|
143 | _facets.add(current.getAttribute(GSXML.SHORTNAME_ATT));
|
---|
144 | }
|
---|
145 | }
|
---|
146 | }
|
---|
147 |
|
---|
148 | return true;
|
---|
149 | }
|
---|
150 |
|
---|
151 | public void cleanUp()
|
---|
152 | {
|
---|
153 | super.cleanUp();
|
---|
154 | this.solr_src.cleanUp();
|
---|
155 | all_solr_cores.shutdown();
|
---|
156 | }
|
---|
157 |
|
---|
158 | /** methods to handle actually doing the query */
|
---|
159 |
|
---|
160 | /** do any initialisation of the query object */
|
---|
161 | protected boolean setUpQueryer(HashMap params)
|
---|
162 | {
|
---|
163 | this.solr_src.clearFacets();
|
---|
164 | this.solr_src.clearFacetQueries();
|
---|
165 |
|
---|
166 | for (int i = 0; i < _facets.size(); i++)
|
---|
167 | {
|
---|
168 | this.solr_src.addFacet(_facets.get(i));
|
---|
169 | }
|
---|
170 |
|
---|
171 | String index = "didx";
|
---|
172 | String physical_index_language_name = null;
|
---|
173 | String physical_sub_index_name = null;
|
---|
174 | int maxdocs = 100;
|
---|
175 | int hits_per_page = 20;
|
---|
176 | int start_page = 1;
|
---|
177 | // set up the query params
|
---|
178 | Set entries = params.entrySet();
|
---|
179 | Iterator i = entries.iterator();
|
---|
180 | while (i.hasNext())
|
---|
181 | {
|
---|
182 | Map.Entry m = (Map.Entry) i.next();
|
---|
183 | String name = (String) m.getKey();
|
---|
184 | String value = (String) m.getValue();
|
---|
185 |
|
---|
186 | if (name.equals(MAXDOCS_PARAM) && !value.equals(""))
|
---|
187 | {
|
---|
188 | maxdocs = Integer.parseInt(value);
|
---|
189 | }
|
---|
190 | else if (name.equals(HITS_PER_PAGE_PARAM))
|
---|
191 | {
|
---|
192 | hits_per_page = Integer.parseInt(value);
|
---|
193 | }
|
---|
194 | else if (name.equals(START_PAGE_PARAM))
|
---|
195 | {
|
---|
196 | start_page = Integer.parseInt(value);
|
---|
197 | }
|
---|
198 | else if (name.equals(MATCH_PARAM))
|
---|
199 | {
|
---|
200 | if (value.equals(MATCH_PARAM_ALL))
|
---|
201 | {
|
---|
202 | this.solr_src.setDefaultConjunctionOperator("AND");
|
---|
203 | }
|
---|
204 | else
|
---|
205 | {
|
---|
206 | this.solr_src.setDefaultConjunctionOperator("OR");
|
---|
207 | }
|
---|
208 | }
|
---|
209 | else if (name.equals(RANK_PARAM))
|
---|
210 | {
|
---|
211 | if (value.equals(RANK_PARAM_RANK_VALUE))
|
---|
212 | {
|
---|
213 | value = null;
|
---|
214 | }
|
---|
215 | this.solr_src.setSortField(value);
|
---|
216 | }
|
---|
217 | else if (name.equals(LEVEL_PARAM))
|
---|
218 | {
|
---|
219 | if (value.toUpperCase().equals("SEC"))
|
---|
220 | {
|
---|
221 | index = "sidx";
|
---|
222 | }
|
---|
223 | else
|
---|
224 | {
|
---|
225 | index = "didx";
|
---|
226 | }
|
---|
227 | }
|
---|
228 | else if (name.equals("facets") && value.length() > 0)
|
---|
229 | {
|
---|
230 | String[] facets = value.split(",");
|
---|
231 |
|
---|
232 | for (String facet : facets)
|
---|
233 | {
|
---|
234 | this.solr_src.addFacet(facet);
|
---|
235 | }
|
---|
236 | }
|
---|
237 | else if (name.equals("facetQueries") && value.length() > 0)
|
---|
238 | {
|
---|
239 | this.solr_src.addFacetQuery(value);
|
---|
240 | }
|
---|
241 | else if (name.equals(INDEX_SUBCOLLECTION_PARAM))
|
---|
242 | {
|
---|
243 | physical_sub_index_name = value;
|
---|
244 | }
|
---|
245 | else if (name.equals(INDEX_LANGUAGE_PARAM))
|
---|
246 | {
|
---|
247 | physical_index_language_name = value;
|
---|
248 | } // ignore any others
|
---|
249 | }
|
---|
250 | // set up start and end results if necessary
|
---|
251 | int start_results = 1;
|
---|
252 | if (start_page != 1)
|
---|
253 | {
|
---|
254 | start_results = ((start_page - 1) * hits_per_page) + 1;
|
---|
255 | }
|
---|
256 | int end_results = hits_per_page * start_page;
|
---|
257 | this.solr_src.setStartResults(start_results);
|
---|
258 | this.solr_src.setEndResults(end_results);
|
---|
259 | this.solr_src.setMaxDocs(maxdocs);
|
---|
260 |
|
---|
261 | if (index.equals("sidx") || index.equals("didx"))
|
---|
262 | {
|
---|
263 | if (physical_sub_index_name != null)
|
---|
264 | {
|
---|
265 | index += physical_sub_index_name;
|
---|
266 | }
|
---|
267 | if (physical_index_language_name != null)
|
---|
268 | {
|
---|
269 | index += physical_index_language_name;
|
---|
270 | }
|
---|
271 | }
|
---|
272 |
|
---|
273 | // now we know the index level, we can dig out the required
|
---|
274 | // solr-core, (caching the result in 'solr_core_cache')
|
---|
275 |
|
---|
276 | String site_name = this.router.getSiteName();
|
---|
277 | String coll_name = this.cluster_name;
|
---|
278 |
|
---|
279 | String core_name = site_name + "-" + coll_name + "-" + index;
|
---|
280 |
|
---|
281 | EmbeddedSolrServer solr_core = null;
|
---|
282 |
|
---|
283 | if (!solr_core_cache.containsKey(core_name))
|
---|
284 | {
|
---|
285 | solr_core = new EmbeddedSolrServer(all_solr_cores, core_name);
|
---|
286 |
|
---|
287 | solr_core_cache.put(core_name, solr_core);
|
---|
288 | }
|
---|
289 | else
|
---|
290 | {
|
---|
291 | solr_core = (EmbeddedSolrServer) solr_core_cache.get(core_name);
|
---|
292 | }
|
---|
293 |
|
---|
294 | this.solr_src.setSolrCore(solr_core);
|
---|
295 | this.solr_src.initialise();
|
---|
296 | return true;
|
---|
297 | }
|
---|
298 |
|
---|
299 | /** do the query */
|
---|
300 | protected Object runQuery(String query)
|
---|
301 | {
|
---|
302 | try
|
---|
303 | {
|
---|
304 | //SharedSoleneQueryResult sqr = this.solr_src.runQuery(query);
|
---|
305 | SharedSoleneQueryResult sqr = this.solr_src.runQuery(query);
|
---|
306 |
|
---|
307 | return sqr;
|
---|
308 | }
|
---|
309 | catch (Exception e)
|
---|
310 | {
|
---|
311 | logger.error("Exception happened in run query: ", e);
|
---|
312 | }
|
---|
313 |
|
---|
314 | return null;
|
---|
315 | }
|
---|
316 |
|
---|
317 | /** get the total number of docs that match */
|
---|
318 | protected long numDocsMatched(Object query_result)
|
---|
319 | {
|
---|
320 | return ((SharedSoleneQueryResult) query_result).getTotalDocs();
|
---|
321 |
|
---|
322 | }
|
---|
323 |
|
---|
324 | /** get the list of doc ids */
|
---|
325 | protected String[] getDocIDs(Object query_result)
|
---|
326 | {
|
---|
327 | Vector docs = ((SharedSoleneQueryResult) query_result).getDocs();
|
---|
328 | String[] doc_nums = new String[docs.size()];
|
---|
329 | for (int d = 0; d < docs.size(); d++)
|
---|
330 | {
|
---|
331 | String doc_num = ((SharedSoleneQueryResult.DocInfo) docs.elementAt(d)).id_;
|
---|
332 | doc_nums[d] = doc_num;
|
---|
333 | }
|
---|
334 | return doc_nums;
|
---|
335 | }
|
---|
336 |
|
---|
337 | /** get the list of doc ranks */
|
---|
338 | protected String[] getDocRanks(Object query_result)
|
---|
339 | {
|
---|
340 | Vector docs = ((SharedSoleneQueryResult) query_result).getDocs();
|
---|
341 | String[] doc_ranks = new String[docs.size()];
|
---|
342 | for (int d = 0; d < docs.size(); d++)
|
---|
343 | {
|
---|
344 | doc_ranks[d] = Float.toString(((SharedSoleneQueryResult.DocInfo) docs.elementAt(d)).rank_);
|
---|
345 | }
|
---|
346 | return doc_ranks;
|
---|
347 | }
|
---|
348 |
|
---|
349 | /** add in term info if available */
|
---|
350 | protected boolean addTermInfo(Element term_list, HashMap params, Object query_result)
|
---|
351 | {
|
---|
352 | String query_level = (String) params.get(LEVEL_PARAM); // the current query level
|
---|
353 |
|
---|
354 | Vector terms = ((SharedSoleneQueryResult) query_result).getTerms();
|
---|
355 | for (int t = 0; t < terms.size(); t++)
|
---|
356 | {
|
---|
357 | SharedSoleneQueryResult.TermInfo term_info = (SharedSoleneQueryResult.TermInfo) terms.get(t);
|
---|
358 |
|
---|
359 | Element term_elem = this.doc.createElement(GSXML.TERM_ELEM);
|
---|
360 | term_elem.setAttribute(GSXML.NAME_ATT, term_info.term_);
|
---|
361 | term_elem.setAttribute(FREQ_ATT, "" + term_info.term_freq_);
|
---|
362 | term_elem.setAttribute(NUM_DOCS_MATCH_ATT, "" + term_info.match_docs_);
|
---|
363 | term_elem.setAttribute(FIELD_ATT, term_info.field_);
|
---|
364 | term_list.appendChild(term_elem);
|
---|
365 | }
|
---|
366 |
|
---|
367 | Vector stopwords = ((SharedSoleneQueryResult) query_result).getStopWords();
|
---|
368 | for (int t = 0; t < stopwords.size(); t++)
|
---|
369 | {
|
---|
370 | String stopword = (String) stopwords.get(t);
|
---|
371 |
|
---|
372 | Element stopword_elem = this.doc.createElement(GSXML.STOPWORD_ELEM);
|
---|
373 | stopword_elem.setAttribute(GSXML.NAME_ATT, stopword);
|
---|
374 | term_list.appendChild(stopword_elem);
|
---|
375 | }
|
---|
376 |
|
---|
377 | return true;
|
---|
378 | }
|
---|
379 |
|
---|
380 | protected ArrayList<FacetWrapper> getFacets(Object query_result)
|
---|
381 | {
|
---|
382 | if (!(query_result instanceof SolrQueryResult))
|
---|
383 | {
|
---|
384 | return null;
|
---|
385 | }
|
---|
386 |
|
---|
387 | SolrQueryResult result = (SolrQueryResult) query_result;
|
---|
388 | List<FacetField> facets = result.getFacetResults();
|
---|
389 |
|
---|
390 | if (facets == null)
|
---|
391 | {
|
---|
392 | return null;
|
---|
393 | }
|
---|
394 |
|
---|
395 | ArrayList<FacetWrapper> newFacetList = new ArrayList<FacetWrapper>();
|
---|
396 |
|
---|
397 | for (FacetField facet : facets)
|
---|
398 | {
|
---|
399 | newFacetList.add(new SolrFacetWrapper(facet));
|
---|
400 | }
|
---|
401 |
|
---|
402 | return newFacetList;
|
---|
403 | }
|
---|
404 | }
|
---|