File LinkCheckerTransformation.java
Coverage histogram
0% of files have more coverage
Classes
| Class | Line # | Actions | |||||
|---|---|---|---|---|---|---|---|
| LinkCheckerTransformation | 58 | 26 | 0% | 11 | 2 |
Contributing tests
This file is covered by 7 tests.
.
Source view
| 1 | /* | |
| 2 | * See the NOTICE file distributed with this work for additional | |
| 3 | * information regarding copyright ownership. | |
| 4 | * | |
| 5 | * This is free software; you can redistribute it and/or modify it | |
| 6 | * under the terms of the GNU Lesser General Public License as | |
| 7 | * published by the Free Software Foundation; either version 2.1 of | |
| 8 | * the License, or (at your option) any later version. | |
| 9 | * | |
| 10 | * This software is distributed in the hope that it will be useful, | |
| 11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
| 13 | * Lesser General Public License for more details. | |
| 14 | * | |
| 15 | * You should have received a copy of the GNU Lesser General Public | |
| 16 | * License along with this software; if not, write to the Free | |
| 17 | * Software Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA | |
| 18 | * 02110-1301 USA, or see the FSF site: http://www.fsf.org. | |
| 19 | */ | |
| 20 | package org.xwiki.rendering.internal.transformation.linkchecker; | |
| 21 | ||
| 22 | import java.util.LinkedHashMap; | |
| 23 | import java.util.List; | |
| 24 | import java.util.Map; | |
| 25 | import java.util.Queue; | |
| 26 | import java.util.concurrent.ConcurrentLinkedQueue; | |
| 27 | ||
| 28 | import javax.inject.Inject; | |
| 29 | import javax.inject.Named; | |
| 30 | import javax.inject.Provider; | |
| 31 | import javax.inject.Singleton; | |
| 32 | ||
| 33 | import org.xwiki.component.annotation.Component; | |
| 34 | import org.xwiki.component.phase.Initializable; | |
| 35 | import org.xwiki.component.phase.InitializationException; | |
| 36 | import org.xwiki.rendering.block.Block; | |
| 37 | import org.xwiki.rendering.block.LinkBlock; | |
| 38 | import org.xwiki.rendering.block.MetaDataBlock; | |
| 39 | import org.xwiki.rendering.block.match.ClassBlockMatcher; | |
| 40 | import org.xwiki.rendering.block.match.MetadataBlockMatcher; | |
| 41 | import org.xwiki.rendering.listener.MetaData; | |
| 42 | import org.xwiki.rendering.listener.reference.ResourceType; | |
| 43 | import org.xwiki.rendering.transformation.AbstractTransformation; | |
| 44 | import org.xwiki.rendering.transformation.TransformationContext; | |
| 45 | import org.xwiki.rendering.transformation.TransformationException; | |
| 46 | import org.xwiki.rendering.transformation.linkchecker.LinkContextDataProvider; | |
| 47 | ||
| 48 | /** | |
| 49 | * Looks for external URLs in links and verify their status (ok, broken, etc). In order to get good performances this is | |
| 50 | * done asynchronously in a Thread and same links are checked only every N minutes. | |
| 51 | * | |
| 52 | * @version $Id: 92c6ad25b02705d4b7ff4d9877067e3f4ac809c4 $ | |
| 53 | * @since 3.3M1 | |
| 54 | */ | |
| 55 | @Component | |
| 56 | @Named("linkchecker") | |
| 57 | @Singleton | |
| 58 | public class LinkCheckerTransformation extends AbstractTransformation implements Initializable | |
| 59 | { | |
| 60 | /** | |
| 61 | * Anti-flood mechanism. We only allow adding links to check in the queue if it currently has less than | |
| 62 | * MAX_LINKS_IN_QUEUE already. This is a safeguard for the following: | |
| 63 | * - if the checker thread has some issues and thus the queue isn't getting unpiled then this ensures that it won't | |
| 64 | * grow more which would slowly use up all the memory... | |
| 65 | * - we don't swamp the system with links to check (for ex if lots of users go to pages with links the queue can | |
| 66 | * quickly grow large) | |
| 67 | * In any case, the links will be checked again later when the pages are rendered again. | |
| 68 | */ | |
| 69 | static final int MAX_LINKS_IN_QUEUE = 10; | |
| 70 | ||
| 71 | @Inject | |
| 72 | private LinkCheckerThread checkerThread; | |
| 73 | ||
| 74 | @Inject | |
| 75 | private Provider<List<LinkContextDataProvider>> linkContextDataProvidersProvider; | |
| 76 | ||
| 77 | /** | |
| 78 | * The link queue that the checker thread will use to check links. We use a separate checker thread and a queue | |
| 79 | * in order to have good performance so that this transformation doesn't slow down the rendering of content. | |
| 80 | */ | |
| 81 | private Queue<LinkQueueItem> linkQueue = new ConcurrentLinkedQueue<>(); | |
| 82 | ||
| 83 | /** | |
| 84 | * Start a Thread in charge of reading links to check from the Checking queue and checking them. | |
| 85 | * | |
| 86 | * @throws InitializationException not used | |
| 87 | */ | |
| 88 | 8 |  @Override |
| 89 | public void initialize() throws InitializationException | |
| 90 | { | |
| 91 | 8 | this.checkerThread.setName("Link Checker Thread"); |
| 92 | 8 | this.checkerThread.startProcessing(getLinkQueue()); |
| 93 | } | |
| 94 | ||
| 95 | 14 | @Override |
| 96 | public void transform(Block source, TransformationContext context) throws TransformationException | |
| 97 | { | |
| 98 | // Note that we don't check for pages to excludes here because this transformation is running in the main | |
| 99 | // thread and is executed for each page view and thus needs to be as fast as possible. The exclusion handling | |
| 100 | // is thus done in the Link Checker Thread. | |
| 101 | ||
| 102 | // Anti-flood mechanism, only add items in the queue if there are less than MAX_LINKS_IN_QUEUE. This means that | |
| 103 | // if the queue has MAX_LINKS_IN_QUEUE or more elements already the links from the current page being rendered | |
| 104 | // will not be verified. They'll get their chance the next time the page is visited again... | |
| 105 | 14 | if (getLinkQueue().size() < MAX_LINKS_IN_QUEUE) { |
| 106 | 13 | for (LinkBlock linkBlock : source.<LinkBlock>getBlocks( |
| 107 | new ClassBlockMatcher(LinkBlock.class), Block.Axes.DESCENDANT)) | |
| 108 | { | |
| 109 | 24 | if (linkBlock.getReference().getType().equals(ResourceType.URL)) { |
| 110 | // This is a link pointing to an external URL, add it to the queue for processing (i.e. checking). | |
| 111 | 24 | String linkReference = linkBlock.getReference().getReference(); |
| 112 | 24 | String contentReference = extractSourceContentReference(linkBlock); |
| 113 | // If there's no content reference then use a default name of "default" | |
| 114 | 24 | if (contentReference == null) { |
| 115 | 20 | contentReference = "default"; |
| 116 | } | |
| 117 | // Add Link Context Data | |
| 118 | 24 | Map<String, Object> linkContextData = createLinkContextData(linkReference, contentReference); |
| 119 | 24 | this.linkQueue.add(new LinkQueueItem(linkReference, contentReference, linkContextData)); |
| 120 | } | |
| 121 | } | |
| 122 | } | |
| 123 | } | |
| 124 | ||
| 125 | /** | |
| 126 | * Stops the checking thread. Should be called when the application is stopped for a clean shutdown. | |
| 127 | * | |
| 128 | * @throws InterruptedException if the thread failed to be stopped | |
| 129 | */ | |
| 130 | 7 | public void stopLinkCheckerThread() throws InterruptedException |
| 131 | { | |
| 132 | 7 | this.checkerThread.stopProcessing(); |
| 133 | // Wait till the thread goes away | |
| 134 | 7 | this.checkerThread.join(); |
| 135 | } | |
| 136 | ||
| 137 | /** | |
| 138 | * @return the checker queue containing all pending links to check | |
| 139 | */ | |
| 140 | 24 | public Queue<LinkQueueItem> getLinkQueue() |
| 141 | { | |
| 142 | 24 | return this.linkQueue; |
| 143 | } | |
| 144 | ||
| 145 | /** | |
| 146 | * @param linkReference the reference to the link to check (usually a URL) | |
| 147 | * @param contentReference the reference to the content containing the link to check | |
| 148 | * @return context data to provide more information about the link being checked (for example it could be useful | |
| 149 | * in some situations to store the HTTP request leading to the link being checked since there could be | |
| 150 | * HTTP query string parameters useful to see to understand why such a link was generated in the content) | |
| 151 | */ | |
| 152 | 24 | private Map<String, Object> createLinkContextData(String linkReference, String contentReference) |
| 153 | { | |
| 154 | // For performance reason we don't want to store an empty map in the Link state cache when there are no | |
| 155 | // context data. | |
| 156 | 24 | Map<String, Object> linkContextData = null; |
| 157 | 24 | for (LinkContextDataProvider linkContextDataProvider : this.linkContextDataProvidersProvider.get()) { |
| 158 | 4 | Map<String, Object> contextData = |
| 159 | linkContextDataProvider.getContextData(linkReference, contentReference); | |
| 160 | 4 | if (linkContextData == null) { |
| 161 | 4 | linkContextData = new LinkedHashMap<>(contextData.size()); |
| 162 | } | |
| 163 | 4 | linkContextData.putAll(contextData); |
| 164 | } | |
| 165 | 24 | return linkContextData; |
| 166 | } | |
| 167 | ||
| 168 | /** | |
| 169 | * @param source the blocks from where to try to extract the source content | |
| 170 | * @return the source content reference or null if none is found | |
| 171 | */ | |
| 172 | 24 | private String extractSourceContentReference(Block source) |
| 173 | { | |
| 174 | 24 | String contentSource = null; |
| 175 | 24 | MetaDataBlock metaDataBlock = |
| 176 | source.getFirstBlock(new MetadataBlockMatcher(MetaData.SOURCE), Block.Axes.ANCESTOR); | |
| 177 | 24 | if (metaDataBlock != null) { |
| 178 | 4 | contentSource = (String) metaDataBlock.getMetaData().getMetaData(MetaData.SOURCE); |
| 179 | } | |
| 180 | 24 | return contentSource; |
| 181 | } | |
| 182 | } |