1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * SolrQueryWrapper.java
|
---|
4 | *
|
---|
5 | * Copyright 2004 The New Zealand Digital Library Project
|
---|
6 | *
|
---|
7 | * A component of the Greenstone digital library software
|
---|
8 | * from the New Zealand Digital Library Project at the
|
---|
9 | * University of Waikato, New Zealand.
|
---|
10 | *
|
---|
11 | * This program is free software; you can redistribute it and/or modify
|
---|
12 | * it under the terms of the GNU General Public License as published by
|
---|
13 | * the Free Software Foundation; either version 2 of the License, or
|
---|
14 | * (at your option) any later version.
|
---|
15 | *
|
---|
16 | * This program is distributed in the hope that it will be useful,
|
---|
17 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
18 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
19 | * GNU General Public License for more details.
|
---|
20 | *
|
---|
21 | * You should have received a copy of the GNU General Public License
|
---|
22 | * along with this program; if not, write to the Free Software
|
---|
23 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
24 | *
|
---|
25 | *********************************************************************/
|
---|
26 | package org.greenstone.gsdl3.util;
|
---|
27 |
|
---|
28 | import java.lang.reflect.Type;
|
---|
29 | import java.net.URLDecoder;
|
---|
30 | import java.util.ArrayList;
|
---|
31 | import java.util.Collection;
|
---|
32 | import java.util.HashMap;
|
---|
33 | import java.util.Iterator;
|
---|
34 | import java.util.List;
|
---|
35 | import java.util.Set;
|
---|
36 | import java.util.HashSet;
|
---|
37 |
|
---|
38 | import org.apache.log4j.Logger;
|
---|
39 | import org.apache.solr.client.solrj.SolrQuery; // subclass of ModifiableSolrParams
|
---|
40 | import org.apache.solr.client.solrj.SolrServer;
|
---|
41 | import org.apache.solr.client.solrj.SolrServerException;
|
---|
42 | import org.apache.solr.client.solrj.embedded.EmbeddedSolrServer;
|
---|
43 | import org.apache.solr.client.solrj.response.QueryResponse;
|
---|
44 | import org.apache.solr.client.solrj.response.TermsResponse;
|
---|
45 |
|
---|
46 | import org.apache.solr.core.CoreContainer;
|
---|
47 | import org.apache.solr.core.SolrCore;
|
---|
48 |
|
---|
49 | import org.apache.solr.common.SolrDocument;
|
---|
50 | import org.apache.solr.common.SolrDocumentList;
|
---|
51 | import org.apache.solr.common.params.ModifiableSolrParams;
|
---|
52 | import org.greenstone.LuceneWrapper4.SharedSoleneQuery;
|
---|
53 | import org.greenstone.LuceneWrapper4.SharedSoleneQueryResult;
|
---|
54 |
|
---|
55 | import org.apache.lucene.search.Query; // Query, TermQuery, BooleanQuery, BooleanClause and more
|
---|
56 | import org.apache.lucene.index.IndexReader;
|
---|
57 | import org.apache.lucene.index.Term;
|
---|
58 | import org.apache.solr.search.QParser;
|
---|
59 | import org.apache.solr.search.SolrIndexSearcher;
|
---|
60 | import org.apache.solr.request.LocalSolrQueryRequest;
|
---|
61 |
|
---|
62 | import com.google.gson.Gson;
|
---|
63 | import com.google.gson.reflect.TypeToken;
|
---|
64 |
|
---|
65 | public class SolrQueryWrapper extends SharedSoleneQuery
|
---|
66 | {
|
---|
67 | public static String SORT_ASCENDING = "asc";
|
---|
68 | public static String SORT_DESCENDING = "desc";
|
---|
69 | public static String SORT_BY_RANK = "score";
|
---|
70 | public static String SORT_BY_INDEX_ORDER = "_docid_";
|
---|
71 |
|
---|
72 | static Logger logger = Logger.getLogger(org.greenstone.gsdl3.util.SolrQueryWrapper.class.getName());
|
---|
73 | protected int max_docs = 100;
|
---|
74 | protected String sort_order = SORT_DESCENDING;
|
---|
75 | protected String sort_field = SORT_BY_RANK; // don't want null default for solr
|
---|
76 | protected ArrayList<String> _facets = new ArrayList<String>();
|
---|
77 | protected ArrayList<String> _facetQueries = new ArrayList<String>();
|
---|
78 | SolrServer solr_core = null;
|
---|
79 |
|
---|
80 | public SolrQueryWrapper()
|
---|
81 | {
|
---|
82 | super();
|
---|
83 | start_results = 0;
|
---|
84 | }
|
---|
85 |
|
---|
86 | public void setMaxDocs(int max_docs)
|
---|
87 | {
|
---|
88 | this.max_docs = max_docs;
|
---|
89 | }
|
---|
90 |
|
---|
91 | public void setSolrCore(SolrServer solr_core)
|
---|
92 | {
|
---|
93 | this.solr_core = solr_core;
|
---|
94 | }
|
---|
95 | // make sure its not null.
|
---|
96 | public void setSortField(String sort_field) {
|
---|
97 | if (sort_field != null) {
|
---|
98 | this.sort_field = sort_field;
|
---|
99 | }
|
---|
100 | }
|
---|
101 |
|
---|
102 | public void setSortOrder(String order)
|
---|
103 | {
|
---|
104 | this.sort_order = order;
|
---|
105 | }
|
---|
106 | public void addFacet(String facet)
|
---|
107 | {
|
---|
108 | if (!_facets.contains(facet))
|
---|
109 | {
|
---|
110 | _facets.add(facet);
|
---|
111 | }
|
---|
112 | }
|
---|
113 |
|
---|
114 | public void clearFacets()
|
---|
115 | {
|
---|
116 | _facets.clear();
|
---|
117 | }
|
---|
118 |
|
---|
119 | public void addFacetQuery(String facetQuery)
|
---|
120 | {
|
---|
121 | if (!_facetQueries.contains(facetQuery))
|
---|
122 | {
|
---|
123 | _facetQueries.add(facetQuery);
|
---|
124 | }
|
---|
125 | }
|
---|
126 |
|
---|
127 | public void clearFacetQueries()
|
---|
128 | {
|
---|
129 | _facetQueries.clear();
|
---|
130 | }
|
---|
131 |
|
---|
132 | public boolean initialise()
|
---|
133 | {
|
---|
134 | if (solr_core == null)
|
---|
135 | {
|
---|
136 | utf8out.println("Solr Core not loaded in ");
|
---|
137 | utf8out.flush();
|
---|
138 | return false;
|
---|
139 | }
|
---|
140 | return true;
|
---|
141 | }
|
---|
142 |
|
---|
143 |
|
---|
144 | /** Extracts the query terms from the query string. The query string can be a boolean
|
---|
145 | * combination of the various search fields with their search terms or phrases
|
---|
146 | */
|
---|
147 | public Term[] getTerms(SolrQuery solrQuery, String query_string)
|
---|
148 | {
|
---|
149 | Term terms[] = null;
|
---|
150 |
|
---|
151 | if(solr_core instanceof EmbeddedSolrServer) {
|
---|
152 | EmbeddedSolrServer solrServer = (EmbeddedSolrServer)solr_core;
|
---|
153 |
|
---|
154 | CoreContainer coreContainer = solrServer.getCoreContainer();
|
---|
155 |
|
---|
156 | Collection<SolrCore> solrCores = coreContainer.getCores();
|
---|
157 | if(!solrCores.isEmpty()) {
|
---|
158 | Iterator<SolrCore> coreIterator = solrCores.iterator();
|
---|
159 |
|
---|
160 | // Just use the first core, since the term frequency of any term is the same regardless of core
|
---|
161 | if(coreIterator.hasNext()) {
|
---|
162 | SolrCore solrCore = coreIterator.next();
|
---|
163 |
|
---|
164 |
|
---|
165 | LocalSolrQueryRequest solrQueryRequest = new LocalSolrQueryRequest(solrCore, solrQuery);
|
---|
166 | Query parsedQuery = null;
|
---|
167 |
|
---|
168 | try {
|
---|
169 |
|
---|
170 | // get the qparser, default is LuceneQParserPlugin, which is called "lucene" see http://wiki.apache.org/solr/QueryParser
|
---|
171 | QParser qParser = QParser.getParser(query_string, "lucene", solrQueryRequest);
|
---|
172 | parsedQuery = qParser.getQuery();
|
---|
173 |
|
---|
174 | // For PrefixQuery or WildCardQuery (a subclass of AutomatonQuery, incl RegexpQ),
|
---|
175 | // like ZZ:econom* and ZZ:*date/regex queries, Query.extractTerms() throws an Exception
|
---|
176 | // because it has not done the Query.rewrite() step yet. So do that manually for them.
|
---|
177 | // This still doesn't provide us with the terms that econom* or *date break down into.
|
---|
178 |
|
---|
179 | //if(parsedQuery instanceof PrefixQuery || parsedQuery instanceof AutomatonQuery) {
|
---|
180 | // Should we just check superclass MultiTermQuery?
|
---|
181 | // Can be a BooleanQuery containing PrefixQuery/WildCardQuery among its clauses, so
|
---|
182 | // just test for * in the query_string to determine if we need to do a rewrite() or not
|
---|
183 | if(query_string.contains("*")) {
|
---|
184 | SolrIndexSearcher searcher = solrQueryRequest.getSearcher();
|
---|
185 | IndexReader indexReader = searcher.getIndexReader(); // returns a DirectoryReader
|
---|
186 | parsedQuery = parsedQuery.rewrite(indexReader); // gets rewritten to ConstantScoreQuery
|
---|
187 | }
|
---|
188 |
|
---|
189 | //System.err.println("#### Query type was: " + parsedQuery.getClass());
|
---|
190 | //logger.error("#### Query type was: " + parsedQuery.getClass());
|
---|
191 |
|
---|
192 | // extract the terms
|
---|
193 | Set<Term> extractedQueryTerms = new HashSet<Term>();
|
---|
194 | parsedQuery.extractTerms(extractedQueryTerms);
|
---|
195 |
|
---|
196 | terms = new Term[extractedQueryTerms.size()];
|
---|
197 |
|
---|
198 | Iterator<Term> termsIterator = extractedQueryTerms.iterator();
|
---|
199 | for(int i = 0; termsIterator.hasNext(); i++) {
|
---|
200 | Term term = termsIterator.next();
|
---|
201 | ///System.err.println("#### Found query term: " + term);
|
---|
202 | ///logger.error("#### Found query term: " + term);
|
---|
203 |
|
---|
204 | terms[i] = term; //(term.field(), term.text());
|
---|
205 | }
|
---|
206 |
|
---|
207 | } catch(Exception queryParseException) {
|
---|
208 | queryParseException.printStackTrace();
|
---|
209 | System.err.println("Exception when parsing query: " + queryParseException.getMessage());
|
---|
210 | System.err.println("#### Query type was: " + parsedQuery.getClass());
|
---|
211 | logger.error("#### Query type was: " + parsedQuery.getClass());
|
---|
212 | }
|
---|
213 | }
|
---|
214 |
|
---|
215 | } else {
|
---|
216 | System.err.println("#### CoreContainer is empty");
|
---|
217 | logger.error("#### CoreContainer is empty");
|
---|
218 | }
|
---|
219 | } else {
|
---|
220 | System.err.println("#### Not an EmbeddedSolrServer. This shouldn't happen." + solr_core.getClass());
|
---|
221 | logger.error("#### Not an EmbeddedSolrServer. This shouldn't happen." + solr_core.getClass());
|
---|
222 | }
|
---|
223 |
|
---|
224 |
|
---|
225 | return terms;
|
---|
226 | }
|
---|
227 |
|
---|
228 | public SharedSoleneQueryResult runQuery(String query_string)
|
---|
229 | {
|
---|
230 | if (query_string == null || query_string.equals(""))
|
---|
231 | {
|
---|
232 | utf8out.println("The query word is not indicated ");
|
---|
233 | utf8out.flush();
|
---|
234 | return null;
|
---|
235 | }
|
---|
236 |
|
---|
237 | SolrQueryResult solr_query_result = new SolrQueryResult();
|
---|
238 | solr_query_result.clear();
|
---|
239 |
|
---|
240 | if (_facetQueries.size() > 0)
|
---|
241 | {
|
---|
242 | HashMap<String, ArrayList<String>> grouping = new HashMap<String, ArrayList<String>>();
|
---|
243 | for (String currentQuery : _facetQueries)
|
---|
244 | {
|
---|
245 | //Facet queries are stored in JSON, so we have to decode it
|
---|
246 | Gson gson = new Gson();
|
---|
247 | Type type = new TypeToken<List<String>>()
|
---|
248 | {
|
---|
249 | }.getType();
|
---|
250 | List<String> queryElems = gson.fromJson(currentQuery, type);
|
---|
251 |
|
---|
252 | //Group each query segment by the index it uses
|
---|
253 | for (String currentQueryElement : queryElems)
|
---|
254 | {
|
---|
255 | String decodedQueryElement = null;
|
---|
256 | try
|
---|
257 | {
|
---|
258 | decodedQueryElement = URLDecoder.decode(currentQueryElement, "UTF-8");
|
---|
259 | }
|
---|
260 | catch (Exception ex)
|
---|
261 | {
|
---|
262 | continue;
|
---|
263 | }
|
---|
264 |
|
---|
265 | int colonIndex = currentQueryElement.indexOf(":");
|
---|
266 | String indexShortName = currentQueryElement.substring(0, colonIndex);
|
---|
267 |
|
---|
268 | if (grouping.get(indexShortName) == null)
|
---|
269 | {
|
---|
270 | grouping.put(indexShortName, new ArrayList<String>());
|
---|
271 | }
|
---|
272 | grouping.get(indexShortName).add(decodedQueryElement);
|
---|
273 | }
|
---|
274 | }
|
---|
275 |
|
---|
276 | //Construct the facet query string to add to the regular query string
|
---|
277 | StringBuilder facetQueryString = new StringBuilder();
|
---|
278 | int keysetCounter = 0;
|
---|
279 | for (String key : grouping.keySet())
|
---|
280 | {
|
---|
281 | StringBuilder currentFacetString = new StringBuilder("(");
|
---|
282 | int groupCounter = 0;
|
---|
283 | for (String queryElem : grouping.get(key))
|
---|
284 | {
|
---|
285 | currentFacetString.append(queryElem);
|
---|
286 |
|
---|
287 | groupCounter++;
|
---|
288 | if (groupCounter < grouping.get(key).size())
|
---|
289 | {
|
---|
290 | currentFacetString.append(" OR ");
|
---|
291 | }
|
---|
292 | }
|
---|
293 | currentFacetString.append(")");
|
---|
294 |
|
---|
295 | facetQueryString.append(currentFacetString);
|
---|
296 |
|
---|
297 | keysetCounter++;
|
---|
298 | if (keysetCounter < grouping.keySet().size())
|
---|
299 | {
|
---|
300 | facetQueryString.append(" AND ");
|
---|
301 | }
|
---|
302 | }
|
---|
303 |
|
---|
304 | if (facetQueryString.length() > 0)
|
---|
305 | {
|
---|
306 | query_string += " AND " + facetQueryString;
|
---|
307 | }
|
---|
308 | }
|
---|
309 |
|
---|
310 |
|
---|
311 | SolrQuery solrQuery = new SolrQuery(query_string);
|
---|
312 | solrQuery.addSort(this.sort_field, SolrQuery.ORDER.valueOf(this.sort_order)); // sort param, like "score desc" or "byORG asc"
|
---|
313 | solrQuery.setStart(start_results); // which result to start from
|
---|
314 | solrQuery.setRows((end_results - start_results) + 1); // how many results per "page"
|
---|
315 |
|
---|
316 | // http://lucene.472066.n3.nabble.com/get-term-frequency-just-only-keywords-search-td4084510.html
|
---|
317 | // WORKS (search didx core):
|
---|
318 | //TI:farming
|
---|
319 | //docOID,score,termfreq(TI,'farming'),totaltermfreq(TI,'farming')
|
---|
320 |
|
---|
321 |
|
---|
322 | // which fields to return for each document, we'll add the request for totaltermfreq later
|
---|
323 | // fl=docOID score termfreq(TI,'farming') totaltermfreq(TI,'farming')
|
---|
324 | solrQuery.setFields("docOID", "score"); //solrParams.set("fl", "docOID score totaltermfreq(field,'queryterm')");
|
---|
325 |
|
---|
326 | //solrQuery.setTerms(true); // turn on the termsComponent
|
---|
327 | //solrQuery.set("terms.fl", "ZZ"); // which field to get the terms from. ModifiableSolrParams method
|
---|
328 |
|
---|
329 | // http://wiki.apache.org/solr/TermVectorComponent and https://cwiki.apache.org/confluence/display/solr/The+Term+Vector+Component
|
---|
330 | // http://lucene.472066.n3.nabble.com/get-term-frequency-just-only-keywords-search-td4084510.html
|
---|
331 | // http://stackoverflow.com/questions/13031534/word-frequency-in-solr
|
---|
332 | // http://wiki.apache.org/solr/FunctionQuery#tf and #termfreq and #totaltermfreq
|
---|
333 | // https://wiki.apache.org/solr/TermsComponent
|
---|
334 |
|
---|
335 | //solrParams.set("tv.tf", true);// turn on the terms vector Component
|
---|
336 | //solrParams.set("tv.fl", "ZZ");// which field to get the terms from /// ZZ
|
---|
337 |
|
---|
338 |
|
---|
339 | if (_facets.size() > 0)
|
---|
340 | {
|
---|
341 | // enable facet counts in the query response
|
---|
342 | solrQuery.setFacet(true); //solrParams.set("facet", "true");
|
---|
343 | for (int i = 0; i < _facets.size(); i++)
|
---|
344 | {
|
---|
345 | // add this field as a facet
|
---|
346 | solrQuery.addFacetField(_facets.get(i)); // solrParams.add("facet.field", _facets.get(i));
|
---|
347 | }
|
---|
348 | }
|
---|
349 |
|
---|
350 | // get the individual terms that make up the query, then request solr to return the totaltermfreq for each term
|
---|
351 | Term[] terms = getTerms(solrQuery, query_string);
|
---|
352 | if(terms != null) {
|
---|
353 | for(int i = 0; i < terms.length; i++) {
|
---|
354 | Term term = terms[i];
|
---|
355 | String field = term.field();
|
---|
356 | String queryTerm = term.text();
|
---|
357 | // totaltermfreq(TI, 'farming') termfreq(TI, 'farming')
|
---|
358 |
|
---|
359 | solrQuery.addField("totaltermfreq(" + field + ",'" + queryTerm + "')");
|
---|
360 | solrQuery.addField("termfreq(" + field + ",'" + queryTerm + "')");
|
---|
361 | }
|
---|
362 | }
|
---|
363 |
|
---|
364 | // do the query
|
---|
365 | try
|
---|
366 | {
|
---|
367 | QueryResponse solrResponse = solr_core.query(solrQuery); //solr_core.query(solrParams);
|
---|
368 | SolrDocumentList hits = solrResponse.getResults();
|
---|
369 | //TermsResponse termResponse = solrResponse.getTermsResponse(); // null unless termvectors=true in schema.xml
|
---|
370 |
|
---|
371 | if (hits != null)
|
---|
372 | {
|
---|
373 | logger.info("*** hits size = " + hits.size());
|
---|
374 | logger.info("*** num docs found = " + hits.getNumFound());
|
---|
375 |
|
---|
376 | logger.info("*** start results = " + start_results);
|
---|
377 | logger.info("*** end results = " + end_results);
|
---|
378 | logger.info("*** max docs = " + max_docs);
|
---|
379 |
|
---|
380 | // numDocsFound is the total number of matching docs in the collection
|
---|
381 | // as opposed to the number of documents returned in the hits list
|
---|
382 |
|
---|
383 | solr_query_result.setTotalDocs((int) hits.getNumFound());
|
---|
384 |
|
---|
385 | solr_query_result.setStartResults(start_results);
|
---|
386 | solr_query_result.setEndResults(start_results + hits.size());
|
---|
387 |
|
---|
388 |
|
---|
389 | // get the first field we're searching in, this will be the fallback field
|
---|
390 | int sepIndex = query_string.indexOf(":");
|
---|
391 | String defaultField = query_string.substring(0, sepIndex);
|
---|
392 | //String query = query_string.substring(sepIndex + 2, query_string.length() - 1); // Replaced by call to getTerms()
|
---|
393 |
|
---|
394 | //solr_query_result.addTerm(query, field, (int) hits.getNumFound(), -1);
|
---|
395 |
|
---|
396 | // Output the matching documents
|
---|
397 | for (int i = 0; i < hits.size(); i++)
|
---|
398 | {
|
---|
399 | SolrDocument doc = hits.get(i);
|
---|
400 |
|
---|
401 | // Need to think about how to support document term frequency. Make zero for now
|
---|
402 | int doc_term_freq = 0;
|
---|
403 | String docOID = (String) doc.get("docOID");
|
---|
404 | Float score = (Float) doc.get("score");
|
---|
405 |
|
---|
406 | logger.info("**** docOID = " + docOID);
|
---|
407 | logger.info("**** score = " + score);
|
---|
408 |
|
---|
409 |
|
---|
410 | // solr returns each term's totaltermfreq, ttf, at the document level, even though
|
---|
411 | // the ttf is the same for each document. So extract this information just for the first document
|
---|
412 | if(i == 0) { // first document
|
---|
413 |
|
---|
414 | if(terms != null) {
|
---|
415 | for(int j = 0; j < terms.length; j++) {
|
---|
416 | Term term = terms[j];
|
---|
417 | String field = term.field();
|
---|
418 | String queryTerm = term.text();
|
---|
419 |
|
---|
420 | // totaltermfreq(TI, 'farming') termfreq(TI, 'farming')
|
---|
421 | Long totaltermfreq = (Long)doc.get("totaltermfreq("+field+",'"+queryTerm+"')");
|
---|
422 | Integer termfreq = (Integer)doc.get("termfreq("+field+",'"+queryTerm+"')");
|
---|
423 |
|
---|
424 | //System.err.println("**** ttf = " + totaltermfreq);
|
---|
425 | //System.err.println("**** tf = " + termfreq);
|
---|
426 | //logger.info("**** ttf = " + totaltermfreq);
|
---|
427 | //logger.info("**** tf = " + termfreq);
|
---|
428 |
|
---|
429 | solr_query_result.addTerm(queryTerm, field, (int) hits.getNumFound(), totaltermfreq.intValue()); // long totaltermfreq to int
|
---|
430 | }
|
---|
431 | } else { // no terms extracted from query_string
|
---|
432 | solr_query_result.addTerm(query_string, defaultField, (int) hits.getNumFound(), -1); // no terms
|
---|
433 | }
|
---|
434 | }
|
---|
435 |
|
---|
436 | solr_query_result.addDoc(docOID, score.floatValue(), doc_term_freq); // doc_termfreq for which term????
|
---|
437 | }
|
---|
438 | }
|
---|
439 | else
|
---|
440 | {
|
---|
441 | solr_query_result.setTotalDocs(0);
|
---|
442 |
|
---|
443 | solr_query_result.setStartResults(0);
|
---|
444 | solr_query_result.setEndResults(0);
|
---|
445 | }
|
---|
446 |
|
---|
447 | solr_query_result.setFacetResults(solrResponse.getFacetFields());
|
---|
448 | }
|
---|
449 | catch (SolrServerException server_exception)
|
---|
450 | {
|
---|
451 | server_exception.printStackTrace();
|
---|
452 | solr_query_result.setError(SolrQueryResult.SERVER_ERROR);
|
---|
453 | }
|
---|
454 |
|
---|
455 | return solr_query_result;
|
---|
456 | }
|
---|
457 |
|
---|
458 | //Greenstone universe operates with a base of 1 for "start_results"
|
---|
459 | //But Solr operates from 0
|
---|
460 | public void setStartResults(int start_results)
|
---|
461 | {
|
---|
462 | if (start_results < 0)
|
---|
463 | {
|
---|
464 | start_results = 0;
|
---|
465 | }
|
---|
466 | this.start_results = start_results - 1;
|
---|
467 | }
|
---|
468 |
|
---|
469 | public void cleanUp()
|
---|
470 | {
|
---|
471 | super.cleanUp();
|
---|
472 | }
|
---|
473 |
|
---|
474 | }
|
---|