source: gs3-extensions/solr/trunk/src/src/java/org/greenstone/gsdl3/service/GS2SolrSearch.java@ 30050

Last change on this file since 30050 was 30050, checked in by Georgiy Litvinov, 9 years ago

Solr repo modifications for Solr side highlighing and snippets
  • Property svn:executable set to *
File size: 14.4 KB
Line 
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
19package org.greenstone.gsdl3.service;
20
21// Greenstone classes
22import java.util.ArrayList;
23import java.util.HashMap;
24import java.util.Iterator;
25import java.util.List;
26import java.util.Map;
27import java.util.Properties;
28import java.util.Set;
29import java.util.Vector;
30
31import org.apache.log4j.Logger;
32import org.apache.solr.client.solrj.SolrServer;
33import org.apache.solr.client.solrj.impl.HttpSolrServer;
34import org.apache.solr.client.solrj.response.FacetField;
35import org.greenstone.LuceneWrapper4.SharedSoleneQueryResult;
36import org.greenstone.gsdl3.util.FacetWrapper;
37import org.greenstone.gsdl3.util.GSFile;
38import org.greenstone.gsdl3.util.GSXML;
39import org.greenstone.gsdl3.util.SolrFacetWrapper;
40import org.greenstone.gsdl3.util.SolrQueryResult;
41import org.greenstone.gsdl3.util.SolrQueryWrapper;
42import org.greenstone.util.GlobalProperties;
43import org.w3c.dom.Document;
44import org.w3c.dom.Element;
45import org.w3c.dom.NodeList;
46
47public class GS2SolrSearch extends SharedSoleneGS2FieldSearch
48{
49
50 public static final String SOLR_SERVLET_SUFFIX = "/solr";
51 protected static final String SORT_ORDER_PARAM = "sortOrder";
52 protected static final String SORT_ORDER_DESCENDING = "1";
53 protected static final String SORT_ORDER_ASCENDING = "0";
54
55 static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.GS2SolrSearch.class.getName());
56
57 protected String solr_servlet_base_url;
58 protected HashMap<String, SolrServer> solr_core_cache;
59 protected SolrQueryWrapper solr_src = null;
60
61 protected ArrayList<String> _facets = new ArrayList<String>();
62
63 public GS2SolrSearch()
64 {
65 paramDefaults.put(SORT_ORDER_PARAM, SORT_ORDER_DESCENDING);
66 does_faceting = true;
67 does_highlight_snippets = true;
68 does_full_field_highlighting = true;
69 // Used to store the solr cores that match the required 'level'
70 // of search (e.g. either document-level=>didx, or
71 // section-level=>sidx. The hashmap is filled out on demand
72 // based on 'level' parameter passed in to 'setUpQueryer()'
73
74 solr_core_cache = new HashMap<String, SolrServer>();
75
76 this.solr_src = new SolrQueryWrapper();
77
78 // Create the solr servlet url on GS3's tomcat. By default it's "http://localhost:8383/solr"
79 // Don't do this in configure(), since the tomcat url will remain unchanged while tomcat is running
80 try {
81 Properties globalProperties = new Properties();
82 globalProperties.load(Class.forName("org.greenstone.util.GlobalProperties").getClassLoader().getResourceAsStream("global.properties"));
83 String host = globalProperties.getProperty("tomcat.server", "localhost");
84 String port = globalProperties.getProperty("tomcat.port", "8383");
85 String protocol = globalProperties.getProperty("tomcat.protocol", "http");
86
87 String portStr = port.equals("80") ? "" : ":"+port;
88 solr_servlet_base_url = protocol+"://"+host+portStr+SOLR_SERVLET_SUFFIX;
89 } catch(Exception e) {
90 logger.error("Error reading greenstone's tomcat solr server properties from global.properties", e);
91 }
92 }
93
94 /** configure this service */
95 public boolean configure(Element info, Element extra_info)
96 {
97 boolean success = super.configure(info, extra_info);
98
99 // clear the map of solr cores for this collection added to the map upon querying
100 solr_core_cache.clear();
101
102 if(!success) {
103 return false;
104 }
105
106 // Setting up facets
107 // TODO - get these from build config, in case some haven't built
108 Element searchElem = (Element) GSXML.getChildByTagName(extra_info, GSXML.SEARCH_ELEM);
109 NodeList facet_list = info.getElementsByTagName("facet");
110 for (int i=0; i<facet_list.getLength(); i++) {
111 _facets.add(((Element)facet_list.item(i)).getAttribute(GSXML.SHORTNAME_ATT));
112 }
113 // NodeList configIndexElems = searchElem.getElementsByTagName(GSXML.INDEX_ELEM);
114
115 // ArrayList<String> chosenFacets = new ArrayList<String>();
116 // for (int i = 0; i < configIndexElems.getLength(); i++)
117 // {
118 // Element current = (Element) configIndexElems.item(i);
119 // if (current.getAttribute(GSXML.FACET_ATT).equals("true"))
120 // {
121 // chosenFacets.add(current.getAttribute(GSXML.NAME_ATT));
122 // }
123 // }
124
125 // Element indexListElem = (Element) GSXML.getChildByTagName(info, GSXML.INDEX_ELEM + GSXML.LIST_MODIFIER);
126 // NodeList buildIndexElems = indexListElem.getElementsByTagName(GSXML.INDEX_ELEM);
127
128 // for (int j = 0; j < buildIndexElems.getLength(); j++)
129 // {
130 // Element current = (Element) buildIndexElems.item(j);
131 // for (int i = 0; i < chosenFacets.size(); i++)
132 // {
133 // if (current.getAttribute(GSXML.NAME_ATT).equals(chosenFacets.get(i)))
134 // {
135 // _facets.add(current.getAttribute(GSXML.SHORTNAME_ATT));
136 // }
137 // }
138 // }
139
140 return true;
141 }
142
143 public void cleanUp()
144 {
145 super.cleanUp();
146 this.solr_src.cleanUp();
147
148 // clear the map keeping track of the SolrServers in this collection
149 solr_core_cache.clear();
150 }
151
152 /** add in the SOLR specific params to TextQuery */
153 protected void addCustomQueryParams(Element param_list, String lang)
154 {
155 super.addCustomQueryParams(param_list, lang);
156 /** Add in the sort order asc/desc param */
157 createParameter(SORT_ORDER_PARAM, param_list, lang);
158 }
159 /** add in SOLR specific params for AdvancedFieldQuery */
160 protected void addCustomQueryParamsAdvField(Element param_list, String lang)
161 {
162 super.addCustomQueryParamsAdvField(param_list, lang);
163 createParameter(SORT_ORDER_PARAM, param_list, lang);
164
165 }
166 /** create a param and add to the list */
167 protected void createParameter(String name, Element param_list, String lang)
168 {
169 Document doc = param_list.getOwnerDocument();
170 Element param = null;
171 String param_default = paramDefaults.get(name);
172 if (name.equals(SORT_ORDER_PARAM)) {
173 String[] vals = { SORT_ORDER_ASCENDING, SORT_ORDER_DESCENDING };
174 String[] vals_texts = { getTextString("param." + SORT_ORDER_PARAM + "." + SORT_ORDER_ASCENDING, lang), getTextString("param." + SORT_ORDER_PARAM + "." + SORT_ORDER_DESCENDING, lang) };
175
176 param = GSXML.createParameterDescription(doc, SORT_ORDER_PARAM, getTextString("param." + SORT_ORDER_PARAM, lang), GSXML.PARAM_TYPE_ENUM_SINGLE, param_default, vals, vals_texts);
177 }
178
179 if (param != null)
180 {
181 param_list.appendChild(param);
182 }
183 else
184 {
185 super.createParameter(name, param_list, lang);
186 }
187
188 }
189
190 /** methods to handle actually doing the query */
191
192 /** do any initialisation of the query object */
193 protected boolean setUpQueryer(HashMap params)
194 {
195 this.solr_src.clearFacets();
196 this.solr_src.clearFacetQueries();
197
198 for (int i = 0; i < _facets.size(); i++)
199 {
200 this.solr_src.addFacet(_facets.get(i));
201 }
202
203 String index = "didx";
204 String physical_index_language_name = null;
205 String physical_sub_index_name = null;
206 int maxdocs = 100;
207 int hits_per_page = 20;
208 int start_page = 1;
209 // set up the query params
210 Set entries = params.entrySet();
211 Iterator i = entries.iterator();
212 while (i.hasNext())
213 {
214 Map.Entry m = (Map.Entry) i.next();
215 String name = (String) m.getKey();
216 String value = (String) m.getValue();
217
218 ///System.err.println("### GS2SolrSearch.java: name " + name + " - value " + value);
219
220 if (name.equals(MAXDOCS_PARAM) && !value.equals(""))
221 {
222 maxdocs = Integer.parseInt(value);
223 }
224 else if (name.equals(HITS_PER_PAGE_PARAM))
225 {
226 hits_per_page = Integer.parseInt(value);
227 }
228 else if (name.equals(START_PAGE_PARAM))
229 {
230 start_page = Integer.parseInt(value);
231 }
232 else if (name.equals(MATCH_PARAM))
233 {
234 if (value.equals(MATCH_PARAM_ALL))
235 {
236 this.solr_src.setDefaultConjunctionOperator("AND");
237 }
238 else
239 {
240 this.solr_src.setDefaultConjunctionOperator("OR");
241 }
242 }
243 else if (name.equals(RANK_PARAM))
244 {
245 if (value.equals(RANK_PARAM_RANK))
246 {
247 value = SolrQueryWrapper.SORT_BY_RANK;
248 } else if (value.equals(RANK_PARAM_NONE)) {
249 value = SolrQueryWrapper.SORT_BY_INDEX_ORDER;
250 }
251
252 this.solr_src.setSortField(value);
253 }
254 else if (name.equals(SORT_ORDER_PARAM)) {
255 if (value.equals(SORT_ORDER_DESCENDING)) {
256 this.solr_src.setSortOrder(SolrQueryWrapper.SORT_DESCENDING);
257 } else {
258 this.solr_src.setSortOrder(SolrQueryWrapper.SORT_ASCENDING);
259 }
260 }
261 else if (name.equals(LEVEL_PARAM))
262 {
263 if (value.toUpperCase().equals("SEC"))
264 {
265 index = "sidx";
266 }
267 else
268 {
269 index = "didx";
270 }
271 }
272 // Would facets ever come in through params???
273 else if (name.equals("facets") && value.length() > 0)
274 {
275 String[] facets = value.split(",");
276
277 for (String facet : facets)
278 {
279 this.solr_src.addFacet(facet);
280 }
281 }
282 else if (name.equals("facetQueries") && value.length() > 0)
283 {
284 this.solr_src.addFacetQuery(value);
285 }
286 else if (name.equals(INDEX_SUBCOLLECTION_PARAM))
287 {
288 physical_sub_index_name = value;
289 }
290 else if (name.equals(INDEX_LANGUAGE_PARAM))
291 {
292 physical_index_language_name = value;
293 } // ignore any others
294 }
295 // set up start and end results if necessary
296 int start_results = 1;
297 if (start_page != 1)
298 {
299 start_results = ((start_page - 1) * hits_per_page) + 1;
300 }
301 int end_results = hits_per_page * start_page;
302 this.solr_src.setStartResults(start_results);
303 this.solr_src.setEndResults(end_results);
304 this.solr_src.setMaxDocs(maxdocs);
305
306 if (index.equals("sidx") || index.equals("didx"))
307 {
308 if (physical_sub_index_name != null)
309 {
310 index += physical_sub_index_name;
311 }
312 if (physical_index_language_name != null)
313 {
314 index += physical_index_language_name;
315 }
316 }
317
318 // now we know the index level, we can dig out the required
319 // solr-core, (caching the result in 'solr_core_cache')
320 String core_name = getCollectionCoreNamePrefix() + "-" + index;
321
322 SolrServer solr_core = null;
323
324 if (!solr_core_cache.containsKey(core_name))
325 {
326 solr_core = new HttpSolrServer(this.solr_servlet_base_url+"/"+core_name);
327 solr_core_cache.put(core_name, solr_core);
328 }
329 else
330 {
331 solr_core = solr_core_cache.get(core_name);
332 }
333
334 this.solr_src.setSolrCore(solr_core);
335 this.solr_src.setCollectionCoreNamePrefix(getCollectionCoreNamePrefix());
336 this.solr_src.initialise();
337 return true;
338 }
339
340 /** do the query */
341 protected Object runQuery(String query)
342 {
343 try
344 {
345 //if it is a Highlighting Query - execute it
346 this.solr_src.setHighlightField(indexField);
347 if(hldocOID != null)
348 {
349 String rslt = this.solr_src.runHighlightingQuery(query,hldocOID);
350 return rslt;
351 }
352 SharedSoleneQueryResult sqr = this.solr_src.runQuery(query);
353
354 return sqr;
355 }
356 catch (Exception e)
357 {
358 logger.error("Exception happened in run query: ", e);
359 }
360
361 return null;
362 }
363
364
365 /** get the total number of docs that match */
366 protected long numDocsMatched(Object query_result)
367 {
368 return ((SharedSoleneQueryResult) query_result).getTotalDocs();
369
370 }
371
372 /** get the list of doc ids */
373 protected String[] getDocIDs(Object query_result)
374 {
375 Vector docs = ((SharedSoleneQueryResult) query_result).getDocs();
376 String[] doc_nums = new String[docs.size()];
377 for (int d = 0; d < docs.size(); d++)
378 {
379 String doc_num = ((SharedSoleneQueryResult.DocInfo) docs.elementAt(d)).id_;
380 doc_nums[d] = doc_num;
381 }
382 return doc_nums;
383 }
384
385 /** get the list of doc ranks */
386 protected String[] getDocRanks(Object query_result)
387 {
388 Vector docs = ((SharedSoleneQueryResult) query_result).getDocs();
389 String[] doc_ranks = new String[docs.size()];
390 for (int d = 0; d < docs.size(); d++)
391 {
392 doc_ranks[d] = Float.toString(((SharedSoleneQueryResult.DocInfo) docs.elementAt(d)).rank_);
393 }
394 return doc_ranks;
395 }
396
397 /** add in term info if available */
398 protected boolean addTermInfo(Element term_list, HashMap params, Object query_result)
399 {
400 Document doc = term_list.getOwnerDocument();
401 String query_level = (String) params.get(LEVEL_PARAM); // the current query level
402
403 Vector terms = ((SharedSoleneQueryResult) query_result).getTerms();
404 for (int t = 0; t < terms.size(); t++)
405 {
406 SharedSoleneQueryResult.TermInfo term_info = (SharedSoleneQueryResult.TermInfo) terms.get(t);
407
408 Element term_elem = doc.createElement(GSXML.TERM_ELEM);
409 term_elem.setAttribute(GSXML.NAME_ATT, term_info.term_);
410 term_elem.setAttribute(FREQ_ATT, "" + term_info.term_freq_);
411 term_elem.setAttribute(NUM_DOCS_MATCH_ATT, "" + term_info.match_docs_);
412 term_elem.setAttribute(FIELD_ATT, term_info.field_);
413 term_list.appendChild(term_elem);
414 }
415
416 Vector stopwords = ((SharedSoleneQueryResult) query_result).getStopWords();
417 for (int t = 0; t < stopwords.size(); t++)
418 {
419 String stopword = (String) stopwords.get(t);
420
421 Element stopword_elem = doc.createElement(GSXML.STOPWORD_ELEM);
422 stopword_elem.setAttribute(GSXML.NAME_ATT, stopword);
423 term_list.appendChild(stopword_elem);
424 }
425
426 return true;
427 }
428
429 protected ArrayList<FacetWrapper> getFacets(Object query_result)
430 {
431 if (!(query_result instanceof SolrQueryResult))
432 {
433 return null;
434 }
435
436 SolrQueryResult result = (SolrQueryResult) query_result;
437 List<FacetField> facets = result.getFacetResults();
438
439 if (facets == null)
440 {
441 return null;
442 }
443
444 ArrayList<FacetWrapper> newFacetList = new ArrayList<FacetWrapper>();
445
446 for (FacetField facet : facets)
447 {
448 SolrFacetWrapper wrap = new SolrFacetWrapper(facet);
449 // String name = wrap.getName();
450 // String display_name = "Poo";
451 // wrap.setDisplayName(display_name);
452
453 newFacetList.add(wrap);
454 }
455
456 return newFacetList;
457 }
458 @Override
459 protected Map<String, Map<String, List<String>>> getHighlightSnippets(Object query_result)
460 {
461 if (!(query_result instanceof SolrQueryResult))
462 {
463 return null;
464 }
465
466 SolrQueryResult result = (SolrQueryResult) query_result;
467
468 return result.getHighlightResults();
469 }
470
471
472 protected String getCollectionCoreNamePrefix() {
473 String site_name = this.router.getSiteName();
474 String coll_name = this.cluster_name;
475 String collection_core_name_prefix = site_name + "-" + coll_name;
476 return collection_core_name_prefix;
477 }
478}
Note: See TracBrowser for help on using the repository browser.