001/* 002 * (C) Copyright 2014 Nuxeo SA (http://nuxeo.com/) and others. 003 * 004 * Licensed under the Apache License, Version 2.0 (the "License"); 005 * you may not use this file except in compliance with the License. 006 * You may obtain a copy of the License at 007 * 008 * http://www.apache.org/licenses/LICENSE-2.0 009 * 010 * Unless required by applicable law or agreed to in writing, software 011 * distributed under the License is distributed on an "AS IS" BASIS, 012 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 013 * See the License for the specific language governing permissions and 014 * limitations under the License. 015 * 016 * Contributors: 017 * bdelbosc 018 */ 019package org.nuxeo.elasticsearch.fetcher; 020 021import java.io.Serializable; 022import java.util.ArrayList; 023import java.util.Collections; 024import java.util.Comparator; 025import java.util.HashMap; 026import java.util.List; 027import java.util.Map; 028 029import org.elasticsearch.action.search.SearchResponse; 030import org.elasticsearch.common.text.Text; 031import org.elasticsearch.search.SearchHit; 032import org.elasticsearch.search.highlight.HighlightField; 033import org.nuxeo.ecm.core.api.CoreInstance; 034import org.nuxeo.ecm.core.api.CoreSession; 035import org.nuxeo.ecm.core.api.DocumentModel; 036import org.nuxeo.ecm.core.api.impl.DocumentModelListImpl; 037import org.nuxeo.ecm.core.query.sql.NXQL; 038import org.nuxeo.ecm.platform.query.api.PageProvider; 039 040/** 041 * @since 6.0 042 */ 043public class VcsFetcher extends Fetcher { 044 045 private static final int CHUNK_SIZE = 100; 046 047 public VcsFetcher(CoreSession session, SearchResponse response, Map<String, String> repoNames) { 048 super(session, response, repoNames); 049 } 050 051 @Override 052 public DocumentModelListImpl fetchDocuments() { 053 Map<String, List<String>> repoHits = getHitsPerRepository(); 054 List<DocumentModel> docs = new ArrayList<>(); 055 String openSessionRepository = getSession().getRepositoryName(); 056 boolean closeSession; 057 CoreSession session; 058 for (String repo : repoHits.keySet()) { 059 if (openSessionRepository.equals(repo)) { 060 session = getSession(); 061 closeSession = false; 062 } else { 063 session = CoreInstance.openCoreSession(repo); 064 closeSession = true; 065 } 066 try { 067 docs.addAll(fetchFromVcs(repoHits.get(repo), session)); 068 } finally { 069 if (closeSession) { 070 session.close(); 071 } 072 } 073 } 074 sortResults(docs); 075 addHighlights(docs); 076 DocumentModelListImpl ret = new DocumentModelListImpl(docs.size()); 077 if (!docs.isEmpty()) { 078 ret.addAll(docs); 079 } 080 return ret; 081 } 082 083 private Map<String, List<String>> getHitsPerRepository() { 084 Map<String, List<String>> ret = new HashMap<>(); 085 for (SearchHit hit : getResponse().getHits()) { 086 String repoName = getRepoForIndex(hit.getIndex()); 087 List<String> docIds = ret.get(repoName); 088 if (docIds == null) { 089 docIds = new ArrayList<>(); 090 ret.put(repoName, docIds); 091 } 092 docIds.add(hit.getId()); 093 } 094 return ret; 095 } 096 097 private List<DocumentModel> fetchFromVcs(List<String> ids, CoreSession session) { 098 List<DocumentModel> ret = null; 099 int size = ids.size(); 100 int start = 0; 101 int end = Math.min(CHUNK_SIZE, size); 102 boolean done = false; 103 while (!done) { 104 List<DocumentModel> docs = fetchFromVcsChunk(ids.subList(start, end), session); 105 if (ret == null) { 106 ret = docs; 107 } else { 108 ret.addAll(docs); 109 } 110 if (end >= ids.size()) { 111 done = true; 112 } else { 113 start = end; 114 end = Math.min(start + CHUNK_SIZE, size); 115 } 116 } 117 return ret; 118 } 119 120 private List<DocumentModel> fetchFromVcsChunk(final List<String> ids, CoreSession session) 121 { 122 StringBuilder sb = new StringBuilder(); 123 sb.append("SELECT * FROM Document, Relation WHERE ecm:uuid IN ("); 124 for (int i = 0; i < ids.size(); i++) { 125 sb.append(NXQL.escapeString(ids.get(i))); 126 if (i < ids.size() - 1) { 127 sb.append(", "); 128 } 129 } 130 sb.append(")"); 131 return session.query(sb.toString()); 132 } 133 134 private void addHighlights(List<DocumentModel> docs) { 135 for (SearchHit hit : getResponse().getHits()) { 136 for (DocumentModel doc : docs) { 137 String docId = doc.getRepositoryName() + doc.getId(); 138 String hitId = getRepoForIndex(hit.getIndex()) + hit.getId(); 139 if (docId.equals(hitId)) { 140 // Add highlight if it exists 141 Map<String, HighlightField> esHighlights = hit.highlightFields(); 142 if (!esHighlights.isEmpty()) { 143 Map<String, List<String>> fields = new HashMap<>(); 144 for (Map.Entry<String, HighlightField> entry : esHighlights.entrySet()) { 145 String field = entry.getKey(); 146 List<String> list = new ArrayList<>(); 147 for (Text fragment : entry.getValue().getFragments()) { 148 list.add(fragment.toString()); 149 } 150 fields.put(field, list); 151 } 152 doc.putContextData(PageProvider.HIGHLIGHT_CTX_DATA, (Serializable) fields); 153 } 154 break; 155 } 156 } 157 } 158 } 159 160 private void sortResults(List<DocumentModel> docs) { 161 final List<String> ids = new ArrayList<>(); 162 for (SearchHit hit : getResponse().getHits()) { 163 ids.add(getRepoForIndex(hit.getIndex()) + hit.getId()); 164 } 165 166 Collections.sort(docs, new Comparator<DocumentModel>() { 167 @Override 168 public int compare(DocumentModel a, DocumentModel b) { 169 return ids.indexOf(a.getRepositoryName() + a.getId()) - ids.indexOf(b.getRepositoryName() + b.getId()); 170 } 171 }); 172 173 } 174 175}