List of usage examples for org.apache.solr.client.solrj SolrQuery setHighlightFragsize
public SolrQuery setHighlightFragsize(int num)
From source file:com.doculibre.constellio.services.FacetServicesImpl.java
License:Open Source License
public static SolrQuery toSolrQuery(SimpleSearch simpleSearch, int start, int row, boolean includeSingleValueFacets, boolean notIncludedOnly, List<String> customFieldFacets, List<String> customQueryFacets, ConstellioUser user) { String solrServerName = simpleSearch.getCollectionName(); RecordCollectionServices collectionServices = ConstellioSpringUtils.getRecordCollectionServices(); RecordCollection collection = collectionServices.get(solrServerName); SolrServices solrServices = ConstellioSpringUtils.getSolrServices(); Boolean usesDisMax = solrServices.usesDisMax(collection); SolrQuery query; if (!collection.isOpenSearch()) { query = SearchServicesImpl.toSolrQuery(simpleSearch, usesDisMax, true, includeSingleValueFacets, notIncludedOnly);//from ww w . j a va 2 s. co m } else { query = SearchServicesImpl.toSolrQuery(simpleSearch, usesDisMax, false, true, false); } query.setParam(ConstellioSolrQueryParams.COLLECTION_NAME, simpleSearch.getCollectionName()); query.setParam(ConstellioSolrQueryParams.LUCENE_QUERY, simpleSearch.getLuceneQuery(includeSingleValueFacets, true)); query.setParam(ConstellioSolrQueryParams.SIMPLE_SEARCH, simpleSearch.toSimpleParams().toString()); if (user != null) { query.setParam(ConstellioSolrQueryParams.USER_ID, "" + user.getId()); } if (StringUtils.isEmpty(query.getQuery())) { query.setQuery(SimpleSearch.SEARCH_ALL); query.setRequestHandler("/elevate"); } query.set("shards.qt", "/elevate"); query.setRequestHandler("/elevate"); query.setRows(row); query.setStart(start); query.setHighlight(true); query.setHighlightFragsize(100); query.setHighlightSnippets(2); query.setFacet(true); query.setFacetLimit(400); query.setFacetMinCount(1); if (collection.isOpenSearch()) { query.setParam("openSearchURL", collection.getOpenSearchURL()); Locale locale = simpleSearch.getSingleSearchLocale(); if (locale != null) { query.setParam("lang", locale.getLanguage()); } } else { for (CollectionFacet collectionFacet : collection.getCollectionFacets()) { if (customFieldFacets == null && collectionFacet.isFieldFacet()) { IndexField indexField = collectionFacet.getFacetField(); String indexFieldName = indexField.getName(); if (!notIncludedOnly) { query.addFacetField(indexFieldName); } else { SearchedFacet searchedFacet = simpleSearch.getSearchedFacet(indexFieldName); if (searchedFacet != null) { if (!searchedFacet.getIncludedValues().isEmpty()) { StringBuffer sbEx = new StringBuffer(); sbEx.append("{!ex=dt}"); // sbEx.append("{!ex="); // boolean first = true; // for (String includedValue : searchedFacet.getIncludedValues()) { // if (first) { // first = false; // } else { // sbEx.append(","); // } // sbEx.append(includedValue); // } // sbEx.append("}"); // query.setParam("facet.field", sbEx.toString() + indexFieldName); query.addFacetField(sbEx.toString() + indexFieldName); } else { query.addFacetField(indexFieldName); } } } } else if (customQueryFacets == null && collectionFacet.isQueryFacet()) { // Modification Rida, remplacement de collectionFacet.getLabels() par // collectionFacet.getLabelledValues() // for (I18NLabel valueLabels : collectionFacet.getLabels()) { for (I18NLabel valueLabels : collectionFacet.getLabelledValues()) { String facetQuery = valueLabels.getKey(); query.addFacetQuery(facetQuery); } } } if (customFieldFacets != null) { for (String facetField : customFieldFacets) { if (!notIncludedOnly) { query.addFacetField(facetField); } else { StringBuffer sbEx = new StringBuffer(); sbEx.append("{!ex=dt}"); // sbEx.append("{!ex="); // boolean first = true; // for (String includedValue : searchedFacet.getIncludedValues()) { // if (first) { // first = false; // } else { // sbEx.append(","); // } // sbEx.append(includedValue); // } // sbEx.append("}"); query.setParam("facet.field", sbEx.toString() + facetField); } } } if (customQueryFacets != null) { for (String facetQuery : customQueryFacets) { query.addFacetQuery(facetQuery); } } } return query; }
From source file:com.doculibre.constellio.services.SearchServicesImpl.java
License:Open Source License
@Override public QueryResponse search(SimpleSearch simpleSearch, int start, int rows, SearchParams searchParams, ConstellioUser user) {//from w w w.j av a2 s. c o m QueryResponse queryResponse; String collectionName = simpleSearch.getCollectionName(); if (collectionName != null) { RecordCollectionServices collectionServices = ConstellioSpringUtils.getRecordCollectionServices(); RecordServices recordServices = ConstellioSpringUtils.getRecordServices(); RecordCollection collection = collectionServices.get(collectionName); SolrServices solrServices = ConstellioSpringUtils.getSolrServices(); Boolean usesDisMax = solrServices.usesDisMax(collection); SolrQuery query; if (!collection.isOpenSearch()) { query = toSolrQuery(simpleSearch, usesDisMax, true, true); } else { query = toSolrQuery(simpleSearch, usesDisMax, false, true); } // displayQuery(query); String luceneQuery = simpleSearch.getLuceneQuery(); query.setParam(ConstellioSolrQueryParams.LUCENE_QUERY, luceneQuery); query.setParam(ConstellioSolrQueryParams.SIMPLE_SEARCH, simpleSearch.toSimpleParams().toString()); query.setParam(ConstellioSolrQueryParams.COLLECTION_NAME, collectionName); if (user != null) { query.setParam(ConstellioSolrQueryParams.USER_ID, "" + user.getId()); } String queryString = query.getQuery(); if (StringUtils.isEmpty(queryString)) { queryString = SimpleSearch.SEARCH_ALL; } List<Record> pendingExclusions = recordServices.getPendingExclusions(collection); while (!pendingExclusions.isEmpty()) { IndexingManager indexingManager = IndexingManager.get(collection); if (indexingManager.isActive()) { try { Thread.sleep(100); } catch (InterruptedException e) { throw new RuntimeException(e); } pendingExclusions = recordServices.getPendingExclusions(collection); } else { return null; } } // SolrQuery query = new SolrQuery(); query.set("collectionName", simpleSearch.getCollectionName()); // query.setQuery(luceneQuery); query.set("shards.qt", "/elevate"); query.setRequestHandler("/elevate"); // nb rsultats par page query.setRows(rows); // page de dbut query.setStart(start); query.setHighlight(searchParams.isHighlightingEnabled()); if (searchParams.isHighlightingEnabled()) { query.setHighlightFragsize(searchParams.getFragsize()); query.setHighlightSnippets(searchParams.getSnippets()); } if (simpleSearch.getSortField() != null) { ORDER order = SimpleSearch.SORT_DESCENDING.equals(simpleSearch.getSortOrder()) ? ORDER.desc : ORDER.asc; IndexFieldServices indexFieldServices = ConstellioSpringUtils.getIndexFieldServices(); IndexField indexField = indexFieldServices.get(simpleSearch.getSortField(), collection); if (indexField != null) { IndexField sortIndexField = indexFieldServices.getSortFieldOf(indexField); if (sortIndexField != null) { query.setSort(sortIndexField.getName(), order); } } } if (collection.isOpenSearch()) { query.setParam("openSearchURL", collection.getOpenSearchURL()); Locale locale = simpleSearch.getSingleSearchLocale(); if (locale != null) { query.setParam("lang", locale.getLanguage()); } } if (searchParams.getHighlightedFields() == null) { IndexField defaultSearchField = collection.getDefaultSearchIndexField(); query.addHighlightField(defaultSearchField.getName()); for (CopyField copyFieldDest : defaultSearchField.getCopyFieldsDest()) { IndexField copyIndexFieldSource = copyFieldDest.getIndexFieldSource(); if (copyIndexFieldSource != null && !copyIndexFieldSource.isTitleField() && copyIndexFieldSource.isHighlighted()) { query.addHighlightField(copyIndexFieldSource.getName()); } } IndexField titleField = collection.getTitleIndexField(); if (titleField != null && titleField.isHighlighted()) { query.addHighlightField(titleField.getName()); } } else { for (String highlightedField : searchParams.getHighlightedFields()) { IndexField field = collection.getIndexField(highlightedField); if (field != null) { query.addHighlightField(highlightedField); } } } SolrServer server = SolrCoreContext.getSolrServer(collectionName); if (server != null) { try { // displayQuery(query); queryResponse = server.query(query); } catch (SolrServerException e) { queryResponse = null; } } else { queryResponse = null; } // if (queryResponse != null && !collection.isOpenSearch()) { // StatsCompiler statCompiler = StatsCompiler.getInstance(); // try { // statCompiler.saveStats(collectionName, SolrLogContext.getStatsSolrServer(), // SolrLogContext.getStatsCompileSolrServer(), queryResponse, luceneQuery); // } catch (SolrServerException e) { // throw new RuntimeException(e); // } catch (IOException e) { // throw new RuntimeException(e); // } // } } else { queryResponse = null; } // improveQueryResponse(collectionName, queryResponse); // System.out.println("Response size" + queryResponse.getResults().getNumFound()); return queryResponse; }
From source file:com.liferay.portal.search.solr.internal.SolrIndexSearcher.java
License:Open Source License
protected void addHighlights(SolrQuery solrQuery, QueryConfig queryConfig) { if (!queryConfig.isHighlightEnabled()) { return;// w ww .j a v a 2s. com } solrQuery.setHighlight(true); solrQuery.setHighlightFragsize(queryConfig.getHighlightFragmentSize()); solrQuery.setHighlightSimplePost(HighlightUtil.HIGHLIGHT_TAG_CLOSE); solrQuery.setHighlightSimplePre(HighlightUtil.HIGHLIGHT_TAG_OPEN); solrQuery.setHighlightSnippets(queryConfig.getHighlightSnippetSize()); for (String highlightFieldName : queryConfig.getHighlightFieldNames()) { addHighlightedField(solrQuery, queryConfig, highlightFieldName); } solrQuery.setHighlightRequireFieldMatch(queryConfig.isHighlightRequireFieldMatch()); }
From source file:com.liferay.portal.search.solr.SolrIndexSearcher.java
License:Open Source License
protected void addHighlights(SolrQuery solrQuery, QueryConfig queryConfig) { if (!queryConfig.isHighlightEnabled()) { return;// w ww . ja va 2s. com } solrQuery.setHighlight(true); solrQuery.setHighlightFragsize(queryConfig.getHighlightFragmentSize()); solrQuery.setHighlightSimplePost(SearchUtil.HIGHLIGHT_TAG_CLOSE); solrQuery.setHighlightSimplePre(SearchUtil.HIGHLIGHT_TAG_OPEN); solrQuery.setHighlightSnippets(queryConfig.getHighlightSnippetSize()); for (String highlightFieldName : queryConfig.getHighlightFieldNames()) { addHighlightedField(solrQuery, queryConfig, highlightFieldName); } solrQuery.setHighlightRequireFieldMatch(queryConfig.isHighlightRequireFieldMatch()); }
From source file:com.liferay.portal.search.solr.SolrIndexSearcherImpl.java
License:Open Source License
protected SolrQuery translateQuery(long companyId, Query query, Sort[] sorts, int start, int end) throws Exception { QueryTranslatorUtil.translateForSolr(query); String queryString = query.toString(); QueryConfig queryConfig = query.getQueryConfig(); SolrQuery solrQuery = new SolrQuery(); solrQuery.setHighlight(queryConfig.isHighlightEnabled()); solrQuery.setHighlightFragsize(queryConfig.getHighlightFragmentSize()); solrQuery.setHighlightSnippets(queryConfig.getHighlightSnippetSize()); solrQuery.setIncludeScore(queryConfig.isScoreEnabled()); solrQuery.setQuery(queryString);// ww w .j a va 2 s. com if ((start == QueryUtil.ALL_POS) && (end == QueryUtil.ALL_POS)) { solrQuery.setRows(0); } else { solrQuery.setRows(end - start); solrQuery.setStart(start); } if ((sorts != null) && (sorts.length > 0)) { for (Sort sort : sorts) { if (sort == null) { continue; } String sortFieldName = sort.getFieldName(); if (ArrayUtil.contains(PortletPropsValues.SOLR_SORTABLE_TEXT_FIELDS, sortFieldName)) { sortFieldName = GetterUtil.getString( PortletProps.get(PortletPropsKeys.SOLR_COPY_FIELDS, new Filter(sortFieldName))); } ORDER order = ORDER.asc; if (sortFieldName == null) { sortFieldName = "score"; order = ORDER.desc; } else if (sort.isReverse()) { order = ORDER.desc; } solrQuery.addSortField(sortFieldName, order); } } return solrQuery; }
From source file:eu.prestoprime.search.util.SolrQueryUtils.java
License:Open Source License
/** * Enable Solr's Highlighting Component on this query object. Search terms * in the map (as values) are only highlighted in the belonging search * fields.//from ww w . j av a 2s . c o m * * @param solrQuery * @param fieldMap */ public static void enableHighlight(SolrQuery solrQuery, final Map<Schema.searchField, String> fieldMap) { // enable highlighting on Query solrQuery.setHighlight(true); // unlimited fragment size and hightlights solrQuery.setHighlightFragsize(0); StringBuilder logMsg = new StringBuilder("Enable highlighting on fields: "); // use these to set a start/end tag for highlights other than <em> // (defined in Solr Config) // solrQuery.setHighlightSimplePre(Constants.getString("HIGHLIGHT_PRE")); // solrQuery.setHighlightSimplePost(Constants.getString("HIGHLIGHT_POST")); if (fieldMap == null || fieldMap.isEmpty()) { // dismax query -> enable // highlighting on all // fields solrQuery.set("hl.highlightMultiTerm", true); for (searchField s : Schema.catchAll) { solrQuery.addHighlightField(s.getFieldName()); logMsg.append(s.getFieldName() + ", "); } LOGGER.debug(logMsg.toString().substring(0, logMsg.length() - 2)); } else { // highlight matching words in queried fields only solrQuery.set("hl.usePhraseHighlighter", true); solrQuery.setHighlightRequireFieldMatch(true); for (Entry<Schema.searchField, String> e : fieldMap.entrySet()) { // highlighting on range fields does no work. Just use text // fields if (e.getKey().getType().equals(Schema.FieldType.P4_TEXT) || e.getKey().getType().equals(Schema.FieldType.STRING)) { logMsg.append(e.getKey().getFieldName() + ", "); solrQuery.addHighlightField(e.getKey().getFieldName()); } } LOGGER.debug(logMsg.toString().substring(0, logMsg.length() - 2)); } }
From source file:fr.paris.lutece.plugins.search.solr.business.SolrSearchEngine.java
License:Open Source License
/** * Return the result with facets. Does NOT support authentification yet. * @param strQuery the query//from ww w . java 2s . co m * @param facetQueries The selected facets * @param sortName The facet name to sort by * @param sortOrder "asc" or "desc" * @param nLimit Maximal number of results. * @return the result with facets */ public SolrFacetedResult getFacetedSearchResults(String strQuery, String[] facetQueries, String sortName, String sortOrder, int nLimit, int nCurrentPageIndex, int nItemsPerPage, Boolean bSpellCheck) { SolrFacetedResult facetedResult = new SolrFacetedResult(); SolrClient solrServer = SolrServerService.getInstance().getSolrServer(); List<SolrSearchResult> results = new ArrayList<SolrSearchResult>(); Hashtable<Field, List<String>> myValuesList = new Hashtable<Field, List<String>>(); if (solrServer != null) { SolrQuery query = new SolrQuery(strQuery); query.setHighlight(true); query.setHighlightSimplePre(SOLR_HIGHLIGHT_PRE); query.setHighlightSimplePost(SOLR_HIGHLIGHT_POST); query.setHighlightSnippets(SOLR_HIGHLIGHT_SNIPPETS); query.setHighlightFragsize(SOLR_HIGHLIGHT_FRAGSIZE); query.setFacet(true); query.setFacetLimit(SOLR_FACET_LIMIT); // query.setFacetMinCount( 1 ); for (Field field : SolrFieldManager.getFacetList().values()) { //Add facet Field if (field.getEnableFacet()) { if (field.getName().equalsIgnoreCase("date") || field.getName().toLowerCase().endsWith("_date")) { query.setParam("facet.date", field.getName()); query.setParam("facet.date.start", SOLR_FACET_DATE_START); query.setParam("facet.date.gap", SOLR_FACET_DATE_GAP); query.setParam("facet.date.end", SOLR_FACET_DATE_END); query.setParam("facet.date.mincount", "0"); } else { query.addFacetField(field.getSolrName()); query.setParam("f." + field.getSolrName() + ".facet.mincount", String.valueOf(field.getFacetMincount())); } myValuesList.put(field, new ArrayList<String>()); } } //Facet intersection List<String> treeParam = new ArrayList<String>(); for (FacetIntersection intersect : SolrFieldManager.getIntersectionlist()) { treeParam.add(intersect.getField1().getSolrName() + "," + intersect.getField2().getSolrName()); } //(String []) al.toArray (new String [0]); query.setParam("facet.tree", (String[]) treeParam.toArray(new String[0])); query.setParam("spellcheck", bSpellCheck); //sort order if ((sortName != null) && !"".equals(sortName)) { if (sortOrder.equals("asc")) { query.setSort(sortName, ORDER.asc); } else { query.setSort(sortName, ORDER.desc); } } else { for (Field field : SolrFieldManager.getSortList()) { if (field.getDefaultSort()) { query.setSort(field.getName(), ORDER.desc); } } } //Treat HttpRequest //FacetQuery if (facetQueries != null) { for (String strFacetQuery : facetQueries) { // if ( strFacetQuery.contains( DATE_COLON ) ) // { // query.addFilterQuery( strFacetQuery ); // } // else // { String myValues[] = strFacetQuery.split(":", 2); if (myValues != null && myValues.length == 2) { myValuesList = getFieldArrange(myValues, myValuesList); } //strFacetQueryWithColon = strFacetQuery.replaceFirst( SolrConstants.CONSTANT_COLON, COLON_QUOTE ); //strFacetQueryWithColon += SolrConstants.CONSTANT_QUOTE; // query.addFilterQuery( strFacetQuery ); // } } for (Field tmpFieldValue : myValuesList.keySet()) { List<String> strValues = myValuesList.get(tmpFieldValue); String strFacetString = ""; if (strValues.size() > 0) { strFacetString = extractQuery(strValues, tmpFieldValue.getOperator()); if (tmpFieldValue.getName().equalsIgnoreCase("date") || tmpFieldValue.getName().toLowerCase().endsWith("_date")) { strFacetString = strFacetString.replaceAll("\"", ""); } query.addFilterQuery(tmpFieldValue.getName() + ":" + strFacetString); } } } try { // count query query.setRows(0); QueryResponse response = solrServer.query(query); int nResults = (int) response.getResults().getNumFound(); facetedResult.setCount(nResults > nLimit ? nLimit : nResults); query.setStart((nCurrentPageIndex - 1) * nItemsPerPage); query.setRows(nItemsPerPage > nLimit ? nLimit : nItemsPerPage); query.setParam("defType", DEF_TYPE); String strWeightValue = generateQueryWeightValue(); query.setParam("qf", strWeightValue); response = solrServer.query(query); //HighLight Map<String, Map<String, List<String>>> highlightsMap = response.getHighlighting(); SolrHighlights highlights = null; if (highlightsMap != null) { highlights = new SolrHighlights(highlightsMap); } //resultList List<SolrItem> itemList = response.getBeans(SolrItem.class); results = SolrUtil.transformSolrItemsToSolrSearchResults(itemList, highlights); //set the spellcheckresult facetedResult.setSolrSpellCheckResponse(response.getSpellCheckResponse()); //Date facet if ((response.getFacetDates() != null) && !response.getFacetDates().isEmpty()) { facetedResult.setFacetDateList(response.getFacetDates()); } //FacetField facetedResult.setFacetFields(response.getFacetFields()); //Facet intersection (facet tree) NamedList<Object> resp = (NamedList<Object>) response.getResponse().get("facet_counts"); if (resp != null) { NamedList<NamedList<NamedList<Integer>>> trees = (NamedList<NamedList<NamedList<Integer>>>) resp .get("trees"); Map<String, ArrayList<FacetField>> treesResult = new HashMap<String, ArrayList<FacetField>>(); if (trees != null) { for (Entry<String, NamedList<NamedList<Integer>>> selectedFacet : trees) { //Selected Facet (ex : type,categorie ) //System.out.println(selectedFacet.getKey()); ArrayList<FacetField> facetFields = new ArrayList<FacetField>( selectedFacet.getValue().size()); for (Entry<String, NamedList<Integer>> facetField : selectedFacet.getValue()) { FacetField ff = new FacetField(facetField.getKey()); //System.out.println("\t" + facetField.getKey()); for (Entry<String, Integer> value : facetField.getValue()) { // Second Level ff.add(value.getKey(), value.getValue()); //System.out.println("\t\t" + value.getKey() + " : " + value.getValue()); } facetFields.add(ff); } treesResult.put(selectedFacet.getKey(), facetFields); } } facetedResult.setFacetIntersection(treesResult); } } catch (SolrServerException e) { AppLogService.error(e.getMessage(), e); } catch (IOException e) { AppLogService.error(e.getMessage(), e); } } else { facetedResult.setFacetFields(new ArrayList<FacetField>()); } facetedResult.setSolrSearchResults(results); return facetedResult; }
From source file:fr.paris.lutece.plugins.search.solr.business.SolrSearchEngine.java
License:Open Source License
public String getDocumentHighLighting(String strDocumentId, String terms) { String strDocumentIdPrefixed = SolrIndexerService.getWebAppName() + SolrConstants.CONSTANT_UNDERSCORE + strDocumentId;/* w w w .j a va 2 s .co m*/ String xmlContent = null; SolrClient solrServer = SolrServerService.getInstance().getSolrServer(); SolrQuery query = new SolrQuery(terms); query.setHighlight(true); query.setHighlightSimplePre(SOLR_HIGHLIGHT_PRE); query.setHighlightSimplePost(SOLR_HIGHLIGHT_POST); query.setHighlightFragsize(0); //return all the content, not fragments query.setParam("hl.fl", SolrItem.FIELD_XML_CONTENT); //return only the field xml_content HighLighting query.setFields(SearchItem.FIELD_UID); //return only the field uid query.setRows(1); query.addFilterQuery(SearchItem.FIELD_UID + ":" + strDocumentIdPrefixed); try { QueryResponse response = solrServer.query(query); if (response.getResults().size() == 1) { SolrHighlights highlights = new SolrHighlights(response.getHighlighting()); if (highlights.getHighlights(strDocumentIdPrefixed).getMap().size() > 0) { xmlContent = highlights.getHighlights(strDocumentIdPrefixed).getMap() .get(SolrItem.FIELD_XML_CONTENT).get(0); } } } catch (SolrServerException e) { AppLogService.error(e.getMessage(), e); } catch (IOException e) { AppLogService.error(e.getMessage(), e); } return xmlContent; }
From source file:net.yacy.peers.Protocol.java
License:Open Source License
protected static int solrQuery(final SearchEvent event, final SolrQuery solrQuery, final int offset, final int count, final Seed target, final int partitions, final Blacklist blacklist) { if (event.query.getQueryGoal().getQueryString(false) == null || event.query.getQueryGoal().getQueryString(false).length() == 0) { return -1; // we cannot query solr only with word hashes, there is no clear text string }// www . ja v a 2s.c o m event.addExpectedRemoteReferences(count); if (partitions > 0) solrQuery.set("partitions", partitions); solrQuery.setStart(offset); solrQuery.setRows(count); // set highlighting query attributes if (event.query.contentdom == Classification.ContentDomain.TEXT || event.query.contentdom == Classification.ContentDomain.ALL) { solrQuery.setHighlight(true); solrQuery.setHighlightFragsize(SearchEvent.SNIPPET_MAX_LENGTH); //solrQuery.setHighlightRequireFieldMatch(); solrQuery.setHighlightSimplePost("</b>"); solrQuery.setHighlightSimplePre("<b>"); solrQuery.setHighlightSnippets(5); for (CollectionSchema field : snippetFields) solrQuery.addHighlightField(field.getSolrFieldName()); } else { solrQuery.setHighlight(false); } boolean localsearch = target == null || target.equals(event.peers.mySeed()); Map<String, ReversibleScoreMap<String>> facets = new HashMap<String, ReversibleScoreMap<String>>( event.query.facetfields.size()); Map<String, LinkedHashSet<String>> snippets = new HashMap<String, LinkedHashSet<String>>(); // this will be a list of urlhash-snippet entries final QueryResponse[] rsp = new QueryResponse[] { null }; final SolrDocumentList[] docList = new SolrDocumentList[] { null }; String ip = target.getIP(); {// encapsulate expensive solr QueryResponse object if (localsearch && !Switchboard.getSwitchboard() .getConfigBool(SwitchboardConstants.DEBUG_SEARCH_REMOTE_SOLR_TESTLOCAL, false)) { // search the local index try { SolrConnector sc = event.getQuery().getSegment().fulltext().getDefaultConnector(); if (!sc.isClosed()) { rsp[0] = sc.getResponseByParams(solrQuery); docList[0] = rsp[0].getResults(); } } catch (final Throwable e) { Network.log.info("SEARCH failed (solr), localpeer (" + e.getMessage() + ")", e); return -1; } } else { try { final boolean myseed = target == event.peers.mySeed(); if (!myseed && !target.getFlagSolrAvailable()) { // skip if peer.dna has flag that last try resulted in error Network.log.info("SEARCH skip (solr), remote Solr interface not accessible, peer=" + target.getName()); return -1; } final String address = myseed ? "localhost:" + target.getPort() : target.getPublicAddress(ip); final int solrtimeout = Switchboard.getSwitchboard() .getConfigInt(SwitchboardConstants.FEDERATED_SERVICE_SOLR_INDEXING_TIMEOUT, 6000); Thread remoteRequest = new Thread() { @Override public void run() { this.setName("Protocol.solrQuery(" + solrQuery.getQuery() + " to " + target.hash + ")"); try { RemoteInstance instance = new RemoteInstance("http://" + address, null, "solr", solrtimeout); // this is a 'patch configuration' which considers 'solr' as default collection try { SolrConnector solrConnector = new RemoteSolrConnector(instance, myseed ? true : target.getVersion() >= 1.63, "solr"); if (!solrConnector.isClosed()) try { rsp[0] = solrConnector.getResponseByParams(solrQuery); docList[0] = rsp[0].getResults(); } catch (Throwable e) { } finally { solrConnector.close(); } } catch (Throwable ee) { } finally { instance.close(); } } catch (Throwable eee) { } } }; remoteRequest.start(); remoteRequest.join(solrtimeout); // just wait until timeout appears if (remoteRequest.isAlive()) { try { remoteRequest.interrupt(); } catch (Throwable e) { } Network.log.info("SEARCH failed (solr), remote Peer: " + target.getName() + "/" + target.getPublicAddress(ip) + " does not answer (time-out)"); target.setFlagSolrAvailable(false || myseed); return -1; // give up, leave remoteRequest abandoned. } // no need to close this here because that sends a commit to remote solr which is not wanted here } catch (final Throwable e) { Network.log.info("SEARCH failed (solr), remote Peer: " + target.getName() + "/" + target.getPublicAddress(ip) + " (" + e.getMessage() + ")"); target.setFlagSolrAvailable(false || localsearch); return -1; } } if (rsp[0] == null || docList[0] == null) { Network.log.info("SEARCH failed (solr), remote Peer: " + target.getName() + "/" + target.getPublicAddress(ip) + " returned null"); target.setFlagSolrAvailable(false || localsearch); return -1; } // evaluate facets for (String field : event.query.facetfields) { FacetField facet = rsp[0].getFacetField(field); ReversibleScoreMap<String> result = new ClusteredScoreMap<String>(UTF8.insensitiveUTF8Comparator); List<Count> values = facet == null ? null : facet.getValues(); if (values == null) continue; for (Count ff : values) { int c = (int) ff.getCount(); if (c == 0) continue; if (ff.getName().length() == 0) continue; // facet entry without text is not useful result.set(ff.getName(), c); } if (result.size() > 0) facets.put(field, result); } // evaluate snippets Map<String, Map<String, List<String>>> rawsnippets = rsp[0].getHighlighting(); // a map from the urlhash to a map with key=field and value = list of snippets if (rawsnippets != null) { nextsnippet: for (Map.Entry<String, Map<String, List<String>>> re : rawsnippets.entrySet()) { Map<String, List<String>> rs = re.getValue(); for (CollectionSchema field : snippetFields) { if (rs.containsKey(field.getSolrFieldName())) { List<String> s = rs.get(field.getSolrFieldName()); if (s.size() > 0) { LinkedHashSet<String> ls = new LinkedHashSet<String>(); ls.addAll(s); snippets.put(re.getKey(), ls); continue nextsnippet; } } } // no snippet found :( --we don't assign a value here by default; that can be done as an evaluation outside this method } } rsp[0] = null; } // evaluate result if (docList == null || docList[0].size() == 0) { Network.log.info("SEARCH (solr), returned 0 out of 0 documents from " + (target == null ? "shard" : ("peer " + target.hash + ":" + target.getName())) + " query = " + solrQuery.toString()); return 0; } List<URIMetadataNode> container = new ArrayList<URIMetadataNode>(); Network.log.info("SEARCH (solr), returned " + docList[0].size() + " out of " + docList[0].getNumFound() + " documents and " + facets.size() + " facets " + facets.keySet().toString() + " from " + (target == null ? "shard" : ("peer " + target.hash + ":" + target.getName()))); int term = count; Collection<SolrInputDocument> docs; if (event.addResultsToLocalIndex) { // only needed to store remote results docs = new ArrayList<SolrInputDocument>(docList[0].size()); } else docs = null; for (final SolrDocument doc : docList[0]) { if (term-- <= 0) { break; // do not process more that requested (in case that evil peers fill us up with rubbish) } // get one single search result if (doc == null) { continue; } URIMetadataNode urlEntry = new URIMetadataNode(doc); if (blacklist.isListed(BlacklistType.SEARCH, urlEntry.url())) { if (Network.log.isInfo()) { if (localsearch) { Network.log.info("local search (solr): filtered blacklisted url " + urlEntry.url()); } else { Network.log.info("remote search (solr): filtered blacklisted url " + urlEntry.url() + " from " + (target == null ? "shard" : ("peer " + target.hash + ":" + target.getName()))); } } continue; // block with blacklist } final String urlRejectReason = Switchboard.getSwitchboard().crawlStacker .urlInAcceptedDomain(urlEntry.url()); if (urlRejectReason != null) { if (Network.log.isInfo()) { if (localsearch) { Network.log.info("local search (solr): rejected url '" + urlEntry.url() + "' (" + urlRejectReason + ")"); } else { Network.log.info("remote search (solr): rejected url '" + urlEntry.url() + "' (" + urlRejectReason + ") from peer " + target.getName()); } } continue; // reject url outside of our domain } // passed all checks, store url if (!localsearch) { // put the remote documents to the local index. We must convert the solr document to a solr input document: if (event.addResultsToLocalIndex) { final SolrInputDocument sid = event.query.getSegment().fulltext().getDefaultConfiguration() .toSolrInputDocument(doc); // the input document stays untouched because it contains top-level cloned objects docs.add(sid); // will be stored to index, and is a full solr document, can be added to firstseen event.query.getSegment().setFirstSeenTime(urlEntry.hash(), Math.min(urlEntry.moddate().getTime(), System.currentTimeMillis())); } // after this conversion we can remove the largest and not used field text_t and synonyms_sxt from the document // because that goes into a search cache and would take a lot of memory in the search cache //doc.removeFields(CollectionSchema.text_t.getSolrFieldName()); doc.removeFields(CollectionSchema.synonyms_sxt.getSolrFieldName()); ResultURLs.stack(ASCII.String(urlEntry.url().hash()), urlEntry.url().getHost(), event.peers.mySeed().hash.getBytes(), UTF8.getBytes(target.hash), EventOrigin.QUERIES); } // add the url entry to the word indexes container.add(urlEntry); } final int dls = docList[0].size(); final int numFound = (int) docList[0].getNumFound(); docList[0].clear(); docList[0] = null; if (localsearch) { event.addNodes(container, facets, snippets, true, "localpeer", numFound); event.addFinalize(); event.addExpectedRemoteReferences(-count); Network.log.info( "local search (solr): localpeer sent " + container.size() + "/" + numFound + " references"); } else { if (event.addResultsToLocalIndex) { for (SolrInputDocument doc : docs) { event.query.getSegment().putDocument(doc); } docs.clear(); docs = null; } event.addNodes(container, facets, snippets, false, target.getName() + "/" + target.hash, numFound); event.addFinalize(); event.addExpectedRemoteReferences(-count); Network.log.info("remote search (solr): peer " + target.getName() + " sent " + (container.size() == 0 ? 0 : container.size()) + "/" + numFound + " references"); } return dls; }
From source file:org.ala.dao.FulltextSearchDaoImplSolr.java
License:Open Source License
/** * Helper method to create SolrQuery object and add facet settings * * @return solrQuery the SolrQuery/*w ww .j av a 2 s. c o m*/ */ protected SolrQuery initSolrQuery(String[] facets) { SolrQuery solrQuery = new SolrQuery(); solrQuery.setQueryType("standard"); if (facets == null) { //use the default set solrQuery.setFacet(true); solrQuery.addFacetField("idxtype"); solrQuery.addFacetField("australian_s"); solrQuery.addFacetField("speciesGroup"); solrQuery.addFacetField("speciesSubgroup"); //solrQuery.addFacetField("kingdom"); solrQuery.addFacetField("rank"); //solrQuery.addFacetField("rankId"); //solrQuery.addFacetField("pestStatus"); // solrQuery.addFacetField("conservationStatus"); solrQuery.addFacetField("conservationStatusAUS"); solrQuery.addFacetField("conservationStatusACT"); solrQuery.addFacetField("conservationStatusNSW"); solrQuery.addFacetField("conservationStatusNT"); solrQuery.addFacetField("conservationStatusQLD"); solrQuery.addFacetField("conservationStatusSA"); solrQuery.addFacetField("conservationStatusTAS"); solrQuery.addFacetField("conservationStatusVIC"); solrQuery.addFacetField("conservationStatusWA"); solrQuery.addFacetField("category_m_s"); solrQuery.addFacetField("category_NSW_m_s"); solrQuery.addFacetField("category_ACT_m_s"); solrQuery.addFacetField("category_QLD_m_s"); solrQuery.addFacetField("category_SA_m_s"); solrQuery.addFacetField("category_NT_m_s"); solrQuery.addFacetField("category_TAS_m_s"); solrQuery.addFacetField("category_WA_m_s"); solrQuery.addFacetField("category_VIC_m_s"); } else { solrQuery.addFacetField(facets); } solrQuery.setFacetMinCount(1); solrQuery.setRows(10); solrQuery.setStart(0); //add highlights solrQuery.setHighlight(true); solrQuery.setHighlightFragsize(80); solrQuery.setHighlightSnippets(2); solrQuery.setHighlightSimplePre("<strong>"); solrQuery.setHighlightSimplePost("</strong>"); solrQuery.add("hl.usePhraseHighlighter", "true"); solrQuery.addHighlightField("commonName"); solrQuery.addHighlightField("scientificName"); solrQuery.addHighlightField("pestStatus"); solrQuery.addHighlightField("conservationStatus"); solrQuery.addHighlightField("simpleText"); solrQuery.addHighlightField("content"); return solrQuery; }