君子博学而日参省乎己 则知明而行无过矣

博客园 首页 新随笔 联系 订阅 管理
solr dataimport 数据导入源码分析（十四）

在solr的数据导入源码里面，DocBuilder类用于创建solr的Document，具体过程不在本文分析，最后调用SolrWriter类的对象执行SolrInputDocument doc对象的提交、修改、删除等操作

本人先来分析一下SolrWriter对象的创建，在DataImportHandler类（数据导入请求处理）的void handleRequestBody(SolrQueryRequest req, SolrQueryResponse rsp)方法里面

UpdateRequestProcessorChain processorChain =
                req.getCore().getUpdateProcessingChain(SolrPluginUtils.resolveUpdateChainParam(params, LOG));
        UpdateRequestProcessor processor = processorChain.createProcessor(req, rsp);
        SolrResourceLoader loader = req.getCore().getResourceLoader();
        SolrWriter sw = getSolrWriter(processor, loader, requestParams);

创建SolrWriter对象的方法如下

private SolrWriter getSolrWriter(final UpdateRequestProcessor processor,
                                   final SolrResourceLoader loader, final DataImporter.RequestParams requestParams) {

    return new SolrWriter(processor) {

      public boolean upload(SolrInputDocument document) {
        try {
          if (requestParams.debug) {
            debugDocuments.add(document);
          }
          return super.upload(document);
        } catch (RuntimeException e) {
          LOG.error( "Exception while adding: " + document, e);
          return false;
        }
      }
    };
  }

 我们浏览一下相关类的UML模型

DIHWriter接口定义了各种操作SolrInputDocument doc对象的方法，SolrWriter类提供了具体实现

/**
 * <p> Writes documents to SOLR. </p>
 * <p/>
 * <b>This API is experimental and may change in the future.</b>
 *
 * @version $Id: SolrWriter.java 1303792 2012-03-22 14:11:16Z jdyer $
 * @since solr 1.3
 */
public class SolrWriter extends DIHWriterBase implements DIHWriter {
  private static final Logger log = LoggerFactory.getLogger(SolrWriter.class);

  static final String LAST_INDEX_KEY = "last_index_time";

  private final UpdateRequestProcessor processor;

  public SolrWriter(UpdateRequestProcessor processor) {
    this.processor = processor;
  }

  public void close() {
      try {
          processor.finish();
      } catch (IOException e) {
          throw new DataImportHandlerException(DataImportHandlerException.SEVERE,
                  "Unable to call finish() on UpdateRequestProcessor", e);
      }
  }
  public boolean upload(SolrInputDocument d) {
    try {
      AddUpdateCommand command = new AddUpdateCommand();
      command.solrDoc = d;
      command.allowDups = false;
      command.overwritePending = true;
      command.overwriteCommitted = true;
      processor.processAdd(command);
    } catch (Exception e) {
      log.warn("Error creating document : " + d, e);
      return false;
    }

    return true;
  }
  
  public void deleteDoc(Object id) {
    try {
      log.info("Deleting document: " + id);
      DeleteUpdateCommand delCmd = new DeleteUpdateCommand();
      delCmd.id = id.toString();
      delCmd.fromPending = true;
      delCmd.fromCommitted = true;
      processor.processDelete(delCmd);
    } catch (IOException e) {
      log.error("Exception while deleteing: " + id, e);
    }
  }
      
    public void deleteByQuery(String query) {
    try {
      log.info("Deleting documents from Solr with query: " + query);
      DeleteUpdateCommand delCmd = new DeleteUpdateCommand();
      delCmd.query = query;
      delCmd.fromCommitted = true;
      delCmd.fromPending = true;
      processor.processDelete(delCmd);
    } catch (IOException e) {
      log.error("Exception while deleting by query: " + query, e);
    }
  }

    public void commit(boolean optimize) {
    try {
      CommitUpdateCommand commit = new CommitUpdateCommand(optimize);
      processor.processCommit(commit);
    } catch (Throwable t) {
      log.error("Exception while solr commit.", t);
    }
  }

    public void rollback() {
    try {
      RollbackUpdateCommand rollback = new RollbackUpdateCommand();
      processor.processRollback(rollback);
    } catch (Throwable t) {
      log.error("Exception while solr rollback.", t);
    }
  }

    public void doDeleteAll() {
    try {
      DeleteUpdateCommand deleteCommand = new DeleteUpdateCommand();
      deleteCommand.query = "*:*";
      deleteCommand.fromCommitted = true;
      deleteCommand.fromPending = true;
      processor.processDelete(deleteCommand);
    } catch (IOException e) {
      throw new DataImportHandlerException(DataImportHandlerException.SEVERE,
              "Exception in full dump while deleting all documents.", e);
    }
  }

  static String getResourceAsString(InputStream in) throws IOException {
    ByteArrayOutputStream baos = new ByteArrayOutputStream(1024);
    byte[] buf = new byte[1024];
    int sz = 0;
    try {
      while ((sz = in.read(buf)) != -1) {
        baos.write(buf, 0, sz);
      }
    } finally {
      try {
        in.close();
      } catch (Exception e) {

      }
    }
    return new String(baos.toByteArray(), "UTF-8");
  }

  static String getDocCount() {
    if (DocBuilder.INSTANCE.get() != null) {
      return ""
              + (DocBuilder.INSTANCE.get().importStatistics.docCount.get() + 1);
    } else {
      return null;
    }
  }
    public void init(Context context) {
        /* NO-OP */        
    }    
}

构造方法初始化成员变量UpdateRequestProcessor processor，针对各种不同的操作构建不同的UpdateCommand对象的具体类，最后调用UpdateRequestProcessor processor对象的不同方法操作具体的UpdateCommand对象(这些相关类已经是solr的成员类了)

UpdateRequestProcessor类为针对SolrInputDocument doc对象的请求处理器，该类为抽象类，具体继承类实现相应操作

/**
 * This is a good place for subclassed update handlers to process the document before it is 
 * indexed.  You may wish to add/remove fields or check if the requested user is allowed to 
 * update the given document...
 * 
 * Perhaps you continue adding an error message (without indexing the document)...
 * perhaps you throw an error and halt indexing (remove anything already indexed??)
 * 
 * By default, this just passes the request to the next processor in the chain.
 * 
 * @since solr 1.3
 */
public abstract class UpdateRequestProcessor {
  protected final Logger log = LoggerFactory.getLogger(getClass());

  protected final UpdateRequestProcessor next;

  public UpdateRequestProcessor( UpdateRequestProcessor next) {
    this.next = next;
  }

  public void processAdd(AddUpdateCommand cmd) throws IOException {
    if (next != null) next.processAdd(cmd);
  }

  public void processDelete(DeleteUpdateCommand cmd) throws IOException {
    if (next != null) next.processDelete(cmd);
  }

  public void processMergeIndexes(MergeIndexesCommand cmd) throws IOException {
    if (next != null) next.processMergeIndexes(cmd);
  }

  public void processCommit(CommitUpdateCommand cmd) throws IOException
  {
    if (next != null) next.processCommit(cmd);
  }

  /**
   * @since Solr 1.4
   */
  public void processRollback(RollbackUpdateCommand cmd) throws IOException
  {
    if (next != null) next.processRollback(cmd);
  }

  public void finish() throws IOException {
    if (next != null) next.finish();    
  }
}

执行SolrInputDocument doc对象的增删改操作的具体继承类为RunUpdateProcessor

class RunUpdateProcessor extends UpdateRequestProcessor 
{
  private final SolrQueryRequest req;
  private final UpdateHandler updateHandler;

  public RunUpdateProcessor(SolrQueryRequest req, UpdateRequestProcessor next) {
    super( next );
    this.req = req;
    this.updateHandler = req.getCore().getUpdateHandler();
  }

  @Override
  public void processAdd(AddUpdateCommand cmd) throws IOException {
    cmd.doc = DocumentBuilder.toDocument(cmd.getSolrInputDocument(), req.getSchema());
    updateHandler.addDoc(cmd);
    super.processAdd(cmd);
  }

  @Override
  public void processDelete(DeleteUpdateCommand cmd) throws IOException {
    if( cmd.id != null ) {
      updateHandler.delete(cmd);
    }
    else {
      updateHandler.deleteByQuery(cmd);
    }
    super.processDelete(cmd);
  }

  @Override
  public void processMergeIndexes(MergeIndexesCommand cmd) throws IOException {
    updateHandler.mergeIndexes(cmd);
    super.processMergeIndexes(cmd);
  }

  @Override
  public void processCommit(CommitUpdateCommand cmd) throws IOException
  {
    updateHandler.commit(cmd);
    super.processCommit(cmd);
  }

  /**
   * @since Solr 1.4
   */
  @Override
  public void processRollback(RollbackUpdateCommand cmd) throws IOException
  {
    updateHandler.rollback(cmd);
    super.processRollback(cmd);
  }
}

调用成员UpdateHandler updateHandler对象的相应方法，UpdateHandler类为抽象类，由继承类DirectUpdateHandler2实现具体操作

最后调用成员IndexWriter writer对象操作lucene的Document 

// iwCommit protects internal data and open/close of the IndexWriter and
  // is a mutex. Any use of the index writer should be protected by iwAccess, 
  // which admits multiple simultaneous acquisitions.  iwAccess is 
  // mutually-exclusive with the iwCommit lock.
  protected final Lock iwAccess, iwCommit;

  protected IndexWriter writer;

  public DirectUpdateHandler2(SolrCore core) throws IOException {
    super(core);

    // Pass fairness=true so commit request is not starved
    // when add/updates are running hot (SOLR-2342):
    ReadWriteLock rwl = new ReentrantReadWriteLock(true);
    iwAccess = rwl.readLock();
    iwCommit = rwl.writeLock();

    commitTracker = new CommitTracker("commitTracker", core,
        core.getSolrConfig().getUpdateHandlerInfo().autoCommmitMaxDocs,
        core.getSolrConfig().getUpdateHandlerInfo().autoCommmitMaxTime, true, false);
  }

  // must only be called when iwCommit lock held
  private void deleteAll() throws IOException {
    SolrCore.log.info(core.getLogId()+"REMOVING ALL DOCUMENTS FROM INDEX");
    closeWriter();
    writer = createMainIndexWriter("DirectUpdateHandler2", true);
  }

  // must only be called when iwCommit lock held
  protected void openWriter() throws IOException {
    if (writer==null) {
      writer = createMainIndexWriter("DirectUpdateHandler2", false);
    }
  }

  // must only be called when iwCommit lock held
  protected void closeWriter() throws IOException {
    try {
      numDocsPending.set(0);
      if (writer!=null) writer.close();
    } finally {
      // if an exception causes the writelock to not be
      // released, we could try and delete it here
      writer=null;
    }
  }

  // must only be called when iwCommit lock held
  protected void rollbackWriter() throws IOException {
    try {
      numDocsPending.set(0);
      if (writer!=null) writer.rollback();
    } finally {
      writer = null;
    }
  }

  @Override
  public int addDoc(AddUpdateCommand cmd) throws IOException {
    addCommands.incrementAndGet();
    addCommandsCumulative.incrementAndGet();
    int rc=-1;

    // if there is no ID field, use allowDups
    if( idField == null ) {
      cmd.allowDups = true;
      cmd.overwriteCommitted = false;
      cmd.overwritePending = false;
    }

    iwAccess.lock();
    try {

      // We can't use iwCommit to protect internal data here, since it would
      // block other addDoc calls.  Hence, we synchronize to protect internal
      // state.  This is safe as all other state-changing operations are
      // protected with iwCommit (which iwAccess excludes from this block).
      synchronized (this) {
        // adding document -- prep writer
        openWriter();
        commitTracker.addedDocument( cmd.commitWithin );
      } // end synchronized block

      // this is the only unsynchronized code in the iwAccess block, which
      // should account for most of the time
            Term updateTerm = null;

      if (cmd.overwriteCommitted || cmd.overwritePending) {
        if (cmd.indexedId == null) {
          cmd.indexedId = getIndexedId(cmd.doc);
        }
        Term idTerm = this.idTerm.createTerm(cmd.indexedId);
        boolean del = false;
        if (cmd.updateTerm == null) {
          updateTerm = idTerm;
        } else {
          del = true;
            updateTerm = cmd.updateTerm;
        }

        writer.updateDocument(updateTerm, cmd.getLuceneDocument(schema));
        if(del) { // ensure id remains unique
          BooleanQuery bq = new BooleanQuery();
          bq.add(new BooleanClause(new TermQuery(updateTerm), Occur.MUST_NOT));
          bq.add(new BooleanClause(new TermQuery(idTerm), Occur.MUST));
          writer.deleteDocuments(bq);
        }
      } else {
        // allow duplicates
        writer.addDocument(cmd.getLuceneDocument(schema));
      }

      rc = 1;
    } finally {
      iwAccess.unlock();
      if (rc!=1) {
        numErrors.incrementAndGet();
        numErrorsCumulative.incrementAndGet();
      } else {
        numDocsPending.incrementAndGet();
      }
    }

    return rc;
  }


  // could return the number of docs deleted, but is that always possible to know???
  @Override
  public void delete(DeleteUpdateCommand cmd) throws IOException {
    deleteByIdCommands.incrementAndGet();
    deleteByIdCommandsCumulative.incrementAndGet();

    if (!cmd.fromPending && !cmd.fromCommitted) {
      numErrors.incrementAndGet();
      numErrorsCumulative.incrementAndGet();
      throw new SolrException( SolrException.ErrorCode.BAD_REQUEST,"meaningless command: " + cmd);
    }
    if (!cmd.fromPending || !cmd.fromCommitted) {
      numErrors.incrementAndGet();
      numErrorsCumulative.incrementAndGet();
      throw new SolrException( SolrException.ErrorCode.BAD_REQUEST,"operation not supported" + cmd);
    }

    iwCommit.lock();
    try {
      openWriter();
      commitTracker.deletedDocument( cmd.commitWithin );
      writer.deleteDocuments(idTerm.createTerm(idFieldType.toInternal(cmd.id)));
    } finally {
      iwCommit.unlock();
    }

    commitTracker.scheduleCommitWithin(commitTracker.getTimeUpperBound());
  }

  // why not return number of docs deleted?
  // Depending on implementation, we may not be able to immediately determine the num...
   @Override
  public void deleteByQuery(DeleteUpdateCommand cmd) throws IOException {
     deleteByQueryCommands.incrementAndGet();
     deleteByQueryCommandsCumulative.incrementAndGet();

     if (!cmd.fromPending && !cmd.fromCommitted) {
       numErrors.incrementAndGet();
       numErrorsCumulative.incrementAndGet();
       throw new SolrException( SolrException.ErrorCode.BAD_REQUEST,"meaningless command: " + cmd);
     }
     if (!cmd.fromPending || !cmd.fromCommitted) {
       numErrors.incrementAndGet();
       numErrorsCumulative.incrementAndGet();
       throw new SolrException( SolrException.ErrorCode.BAD_REQUEST,"operation not supported" + cmd);
     }

    boolean madeIt=false;
    boolean delAll=false;
    try {
     Query q = QueryParsing.parseQuery(cmd.query, schema);
     delAll = MatchAllDocsQuery.class == q.getClass();

     iwCommit.lock();
     try {
       commitTracker.deletedDocument(cmd.commitWithin);
       if (delAll) {
         deleteAll();
       } else {
        openWriter();
        writer.deleteDocuments(q);
       }
     } finally {
       iwCommit.unlock();
     }

     madeIt=true;

     commitTracker.scheduleCommitWithin(commitTracker.getTimeUpperBound());

    } finally {
      if (!madeIt) {
        numErrors.incrementAndGet();
        numErrorsCumulative.incrementAndGet();
      }
    }
  }

  @Override
  public int mergeIndexes(MergeIndexesCommand cmd) throws IOException {
    mergeIndexesCommands.incrementAndGet();
    int rc = -1;

    iwCommit.lock();
    try {
      log.info("start " + cmd);

      IndexReader[] readers = cmd.readers;
      if (readers != null && readers.length > 0) {
        openWriter();
        writer.addIndexes(readers);
        rc = 1;
      } else {
        rc = 0;
      }
      log.info("end_mergeIndexes");
    } finally {
      iwCommit.unlock();
    }

    if (rc == 1 && commitTracker.getTimeUpperBound() > 0) {
      commitTracker.scheduleCommitWithin(commitTracker.getTimeUpperBound());
    }

    return rc;
  }

   public void forceOpenWriter() throws IOException  {
    iwCommit.lock();
    try {
      openWriter();
    } finally {
      iwCommit.unlock();
    }
  }

  @Override
  public void commit(CommitUpdateCommand cmd) throws IOException {

    if (cmd.optimize) {
      optimizeCommands.incrementAndGet();
    } else {
      commitCommands.incrementAndGet();
      if (cmd.expungeDeletes) expungeDeleteCommands.incrementAndGet();
    }

    Future[] waitSearcher = null;
    if (cmd.waitSearcher) {
      waitSearcher = new Future[1];
    }

    boolean error=true;
    iwCommit.lock();
    try {
      log.info("start "+cmd);

      if (cmd.optimize) {
        openWriter();
        writer.forceMerge(cmd.maxOptimizeSegments);
      } else if (cmd.expungeDeletes) {
        openWriter();
        writer.forceMergeDeletes();
      }
      
      closeWriter();

      callPostCommitCallbacks();
      if (cmd.optimize) {
        callPostOptimizeCallbacks();
      }
      // open a new searcher in the sync block to avoid opening it
      // after a deleteByQuery changed the index, or in between deletes
      // and adds of another commit being done.
      core.getSearcher(true,false,waitSearcher);

      // reset commit tracking
      commitTracker.didCommit();

      log.info("end_commit_flush");

      error=false;
    }
    finally {
      iwCommit.unlock();
      addCommands.set(0);
      deleteByIdCommands.set(0);
      deleteByQueryCommands.set(0);
      numErrors.set(error ? 1 : 0);
    }

    // if we are supposed to wait for the searcher to be registered, then we should do it
    // outside of the synchronized block so that other update operations can proceed.
    if (waitSearcher!=null && waitSearcher[0] != null) {
       try {
        waitSearcher[0].get();
      } catch (InterruptedException e) {
        SolrException.log(log,e);
      } catch (ExecutionException e) {
        SolrException.log(log,e);
      }
    }
  }

  /**
   * @since Solr 1.4
   */
  @Override
  public void rollback(RollbackUpdateCommand cmd) throws IOException {

    rollbackCommands.incrementAndGet();

    boolean error=true;
    iwCommit.lock();
    try {
      log.info("start "+cmd);

      rollbackWriter();

      //callPostRollbackCallbacks();

      // reset commit tracking
      commitTracker.didRollback();

      log.info("end_rollback");

      error=false;
    }
    finally {
      iwCommit.unlock();
      addCommandsCumulative.set(
          addCommandsCumulative.get() - addCommands.getAndSet( 0 ) );
      deleteByIdCommandsCumulative.set(
          deleteByIdCommandsCumulative.get() - deleteByIdCommands.getAndSet( 0 ) );
      deleteByQueryCommandsCumulative.set(
          deleteByQueryCommandsCumulative.get() - deleteByQueryCommands.getAndSet( 0 ) );
      numErrors.set(error ? 1 : 0);
    }
  }


  @Override
  public void close() throws IOException {
    log.info("closing " + this);
    iwCommit.lock();
    try{
      commitTracker.close();
      closeWriter();
    } finally {
      iwCommit.unlock();
    }
    log.info("closed " + this);
  }

IndexWriter writer对象的创建方法如下（SolrIndexWriter类为lucene的IndexWriter类的继承类）

protected SolrIndexWriter createMainIndexWriter(String name, boolean removeAllExisting) throws IOException {
    return new SolrIndexWriter(name,core.getNewIndexDir(), core.getDirectoryFactory(), removeAllExisting, schema, core.getSolrConfig().indexConfig, core.getDeletionPolicy());
  }

---------------------------------------------------------------------------

本系列solr dataimport 数据导入源码分析系本人原创

转载请注明出处 博客园 刺猬的温驯

本文链接 http://www.cnblogs.com/chenying99/archive/2013/05/04/3059443.html

posted on 2013-05-04 21:43  刺猬的温驯  阅读(2476)  评论(0)  编辑  收藏  举报