flume-ng怎么自定義攔截器

本篇內容主要講解“flume-ng怎么自定義攔截器”，感興趣的朋友不妨來看看。本文介紹的方法操作簡單快捷，實用性強。下面就讓小編來帶大家學習“flume-ng怎么自定義攔截器”吧!

代碼如下:

package com.wy.flume.interceptor;

import java.util.List;
import java.util.Map;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

import org.apache.commons.lang.StringUtils;
import org.apache.flume.Context;
import org.apache.flume.Event;
import org.apache.flume.interceptor.Interceptor;
import org.apache.flume.interceptor.RegexExtractorInterceptorPassThroughSerializer;
import org.apache.flume.interceptor.RegexExtractorInterceptorSerializer;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import com.google.common.base.Charsets;
import com.google.common.base.Preconditions;
import com.google.common.base.Throwables;
import com.google.common.collect.Lists;

public class RegexExtractorHeaderInterceptor implements Interceptor {

    static final String REGEX = "regex";
    static final String SERIALIZERS = "serializers";
    
    
    static final String EXTRACTOR_HEADER = "extractorHeader";  
    static final boolean DEFAULT_EXTRACTOR_HEADER = false;  
    static final String EXTRACTOR_HEADER_KEY = "extractorHeaderKey"; 

    private static final Logger logger = LoggerFactory
        .getLogger(RegexExtractorHeaderInterceptor.class);

    private final Pattern regex;
    private final List<NameAndSerializer> serializers;

    private final boolean extractorHeader;  
    private final String extractorHeaderKey;  
    
    private RegexExtractorHeaderInterceptor(Pattern regex,
        List<NameAndSerializer> serializers,boolean extractorHeader, String extractorHeaderKey) {
      this.regex = regex;
      this.serializers = serializers;
      
      this.extractorHeader = extractorHeader;
      this.extractorHeaderKey = extractorHeaderKey;
      
    }

    @Override
    public void initialize() {
      // NO-OP...
    }

    @Override
    public void close() {
      // NO-OP...
    }

    @Override
    public Event intercept(Event event) {
      String extractorHeaderVal;
      if (extractorHeader){
          
          extractorHeaderVal = event.getHeaders().get(extractorHeaderKey);
          
      }else{
          
          extractorHeaderVal = new String(event.getBody(),Charsets.UTF_8);
          
      }
      
      Matcher matcher = regex.matcher(extractorHeaderVal);
      Map<String, String> headers = event.getHeaders();
      if (matcher.find()) {
        for (int group = 0, count = matcher.groupCount(); group < count; group++) {
          int groupIndex = group + 1;
          if (groupIndex > serializers.size()) {
            if (logger.isDebugEnabled()) {
              logger.debug("Skipping group {} to {} due to missing serializer",
                  group, count);
            }
            break;
          }
          NameAndSerializer serializer = serializers.get(group);
          if (logger.isDebugEnabled()) {
            logger.debug("Serializing {} using {}", serializer.headerName,
                serializer.serializer);
          }
          headers.put(serializer.headerName,
              serializer.serializer.serialize(matcher.group(groupIndex)));
        }
      }
      return event;
    }

    @Override
    public List<Event> intercept(List<Event> events) {
      List<Event> intercepted = Lists.newArrayListWithCapacity(events.size());
      for (Event event : events) {
        Event interceptedEvent = intercept(event);
        if (interceptedEvent != null) {
          intercepted.add(interceptedEvent);
        }
      }
      return intercepted;
    }

    public static class Builder implements Interceptor.Builder {

      private Pattern regex;
      private List<NameAndSerializer> serializerList;
      
      private boolean extractorHeader;
      private String extractorHeaderKey;
      
      private final RegexExtractorInterceptorPassThroughSerializer defaultSerializer = new RegexExtractorInterceptorPassThroughSerializer();
      

      @Override
      public void configure(Context context) {
        String regexString = context.getString(REGEX);
        Preconditions.checkArgument(!StringUtils.isEmpty(regexString),
            "Must supply a valid regex string");
        regex = Pattern.compile(regexString);
        regex.pattern();
        regex.matcher("").groupCount();
        configureSerializers(context);
        
        extractorHeader = context.getBoolean(EXTRACTOR_HEADER,DEFAULT_EXTRACTOR_HEADER);
        
        if (extractorHeader){
            
            extractorHeaderKey = context.getString(EXTRACTOR_HEADER_KEY);
            Preconditions.checkArgument(!StringUtils.isEmpty(extractorHeaderKey),"header key must");
            
        }
        
      }

      private void configureSerializers(Context context) {
        String serializerListStr = context.getString(SERIALIZERS);
        Preconditions.checkArgument(!StringUtils.isEmpty(serializerListStr),
            "Must supply at least one name and serializer");

        String[] serializerNames = serializerListStr.split("\\s+");

        Context serializerContexts =
            new Context(context.getSubProperties(SERIALIZERS + "."));

        serializerList = Lists.newArrayListWithCapacity(serializerNames.length);
        for(String serializerName : serializerNames) {
          Context serializerContext = new Context(
              serializerContexts.getSubProperties(serializerName + "."));
          String type = serializerContext.getString("type", "DEFAULT");
          String name = serializerContext.getString("name");
          Preconditions.checkArgument(!StringUtils.isEmpty(name),
              "Supplied name cannot be empty.");

          if("DEFAULT".equals(type)) {
            serializerList.add(new NameAndSerializer(name, defaultSerializer));
          } else {
            serializerList.add(new NameAndSerializer(name, getCustomSerializer(
                type, serializerContext)));
          }
        }
      }

      private RegexExtractorInterceptorSerializer getCustomSerializer(
          String clazzName, Context context) {
        try {
          RegexExtractorInterceptorSerializer serializer = (RegexExtractorInterceptorSerializer) Class
              .forName(clazzName).newInstance();
          serializer.configure(context);
          return serializer;
        } catch (Exception e) {
          logger.error("Could not instantiate event serializer.", e);
          Throwables.propagate(e);
        }
        return defaultSerializer;
      }

      @Override
      public Interceptor build() {
        Preconditions.checkArgument(regex != null,
            "Regex pattern was misconfigured");
        Preconditions.checkArgument(serializerList.size() > 0,
            "Must supply a valid group match id list");
        return new RegexExtractorHeaderInterceptor(regex, serializerList, extractorHeader, extractorHeaderKey);
      }
    }

    static class NameAndSerializer {
      private final String headerName;
      private final RegexExtractorInterceptorSerializer serializer;

      public NameAndSerializer(String headerName,
          RegexExtractorInterceptorSerializer serializer) {
        this.headerName = headerName;
        this.serializer = serializer;
      }
    }
  }

應用配置：

hdp2.sources.s1.interceptors = i2
hdp2.sources.s1.interceptors.i2.type = com.wy.flume.interceptor.RegexExtractorHeaderInterceptor$Builder
hdp2.sources.s1.interceptors.i2.regex = ([^_]+)_(\\d{8}).*
hdp2.sources.s1.interceptors.i2.extractorHeader = true
hdp2.sources.s1.interceptors.i2.extractorHeaderKey = basename
hdp2.sources.s1.interceptors.i2.serializers = s1 s2
hdp2.sources.s1.interceptors.i2.serializers.s1.name = log_type
hdp2.sources.s1.interceptors.i2.serializers.s2.name = log_day

到此，相信大家對“flume-ng怎么自定義攔截器”有了更深的了解，不妨來實際操作一番吧！這里是億速云網站，更多相關內容可以進入相關頻道進行查詢，關注我們，繼續學習！