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.Comparator; 024import java.util.HashMap; 025import java.util.List; 026import java.util.Map; 027 028import org.elasticsearch.action.search.SearchResponse; 029import org.elasticsearch.common.text.Text; 030import org.elasticsearch.search.SearchHit; 031import org.elasticsearch.search.fetch.subphase.highlight.HighlightField; 032import org.nuxeo.ecm.core.api.CoreInstance; 033import org.nuxeo.ecm.core.api.CoreSession; 034import org.nuxeo.ecm.core.api.DocumentModel; 035import org.nuxeo.ecm.core.api.impl.DocumentModelListImpl; 036import org.nuxeo.ecm.core.query.sql.NXQL; 037import org.nuxeo.ecm.platform.query.api.PageProvider; 038 039/** 040 * @since 6.0 041 */ 042public class VcsFetcher extends Fetcher { 043 044 private static final int CHUNK_SIZE = 100; 045 046 public VcsFetcher(CoreSession session, SearchResponse response, Map<String, String> repoNames) { 047 super(session, response, repoNames); 048 } 049 050 @Override 051 public DocumentModelListImpl fetchDocuments() { 052 Map<String, List<String>> repoHits = getHitsPerRepository(); 053 List<DocumentModel> docs = fetchFromVcs(repoHits); 054 sortResults(docs); 055 addHighlights(docs); 056 DocumentModelListImpl ret = new DocumentModelListImpl(docs.size()); 057 if (!docs.isEmpty()) { 058 ret.addAll(docs); 059 } 060 return ret; 061 } 062 063 @SuppressWarnings("resource") // session closed only if we opened it 064 protected List<DocumentModel> fetchFromVcs(Map<String, List<String>> repoHits) { 065 List<DocumentModel> docs = new ArrayList<>(); 066 String openSessionRepository = getSession().getRepositoryName(); 067 for (String repo : repoHits.keySet()) { 068 CoreSession session; 069 if (openSessionRepository.equals(repo)) { 070 session = getSession(); 071 } else { 072 session = CoreInstance.getCoreSession(repo); 073 } 074 docs.addAll(fetchFromVcs(repoHits.get(repo), session)); 075 } 076 return docs; 077 } 078 079 private Map<String, List<String>> getHitsPerRepository() { 080 Map<String, List<String>> ret = new HashMap<>(); 081 for (SearchHit hit : getResponse().getHits()) { 082 String repoName = getRepoForIndex(hit.getIndex()); 083 List<String> docIds = ret.computeIfAbsent(repoName, k -> new ArrayList<>()); 084 docIds.add(hit.getId()); 085 } 086 return ret; 087 } 088 089 private List<DocumentModel> fetchFromVcs(List<String> ids, CoreSession session) { 090 List<DocumentModel> ret = null; 091 int size = ids.size(); 092 int start = 0; 093 int end = Math.min(CHUNK_SIZE, size); 094 boolean done = false; 095 while (!done) { 096 List<DocumentModel> docs = fetchFromVcsChunk(ids.subList(start, end), session); 097 if (ret == null) { 098 ret = docs; 099 } else { 100 ret.addAll(docs); 101 } 102 if (end >= ids.size()) { 103 done = true; 104 } else { 105 start = end; 106 end = Math.min(start + CHUNK_SIZE, size); 107 } 108 } 109 return ret; 110 } 111 112 private List<DocumentModel> fetchFromVcsChunk(final List<String> ids, CoreSession session) { 113 StringBuilder sb = new StringBuilder(); 114 sb.append("SELECT * FROM Document, Relation WHERE ecm:uuid IN ("); 115 for (int i = 0; i < ids.size(); i++) { 116 sb.append(NXQL.escapeString(ids.get(i))); 117 if (i < ids.size() - 1) { 118 sb.append(", "); 119 } 120 } 121 sb.append(")"); 122 return session.query(sb.toString()); 123 } 124 125 private void addHighlights(List<DocumentModel> docs) { 126 for (SearchHit hit : getResponse().getHits()) { 127 for (DocumentModel doc : docs) { 128 String docId = doc.getRepositoryName() + doc.getId(); 129 String hitId = getRepoForIndex(hit.getIndex()) + hit.getId(); 130 if (docId.equals(hitId)) { 131 // Add highlight if it exists 132 Map<String, HighlightField> esHighlights = hit.getHighlightFields(); 133 if (!esHighlights.isEmpty()) { 134 Map<String, List<String>> fields = new HashMap<>(); 135 for (Map.Entry<String, HighlightField> entry : esHighlights.entrySet()) { 136 String field = entry.getKey(); 137 List<String> list = new ArrayList<>(); 138 for (Text fragment : entry.getValue().getFragments()) { 139 list.add(fragment.toString()); 140 } 141 fields.put(field, list); 142 } 143 doc.putContextData(PageProvider.HIGHLIGHT_CTX_DATA, (Serializable) fields); 144 } 145 break; 146 } 147 } 148 } 149 } 150 151 private void sortResults(List<DocumentModel> docs) { 152 final List<String> ids = new ArrayList<>(); 153 for (SearchHit hit : getResponse().getHits()) { 154 ids.add(getRepoForIndex(hit.getIndex()) + hit.getId()); 155 } 156 157 docs.sort(new Comparator<DocumentModel>() { 158 @Override 159 public int compare(DocumentModel a, DocumentModel b) { 160 return ids.indexOf(a.getRepositoryName() + a.getId()) - ids.indexOf(b.getRepositoryName() + b.getId()); 161 } 162 }); 163 164 } 165 166}