使用storm实现实时GPS数据客流特征分析系统：源码分析

最新推荐文章于 2018-01-12 18:45:37 发布

雪峰庙

最新推荐文章于 2018-01-12 18:45:37 发布

阅读量2k

点赞数

分类专栏： STORM

STORM 专栏收录该内容

5 篇文章 0 订阅

订阅专栏

问题导读
1.项目背景是什么？
2.nextTuple() 的作用是什么？
3.declareOutputFields(OutputFieldsDeclarer declarer)的作用是什么？
4.excute() 和declareOutputFields(OutputFieldsDeclarer declarer) 的作用是什么？

项目背景：实时GPS数据客流特征分析系统，项目来源于深圳交委，数据来源是深圳大约5万两出租车和公交车的车载GPS仪，其目的是要研究出行者的出行特征、实时路况、客流特征等。

开发环境：请参考： storm的开发环境部署指导

函数详解：
一、GPSReceiverSpout

最重要的两个函数是 nextTuple() 和 declareOutputFields(OutputFieldsDeclarer declarer)。

nextTuple()告诉storm下一个tuple是什么内容，其处理过程是先用一个socket函数接受来自网络的实时GPS数据，用lineSplit()将GPS以逗号分隔成字符串数组，发送给下一个处理单元 DistrictMatchingBolt。
例如一条原始GPS记录：粤BXXXXX,114.121765,22.569218,2013-02-08 17:29:58,1065382,28,101,0,蓝色
_collector.emit(new Values(GPSRecord[0],GPSRecord[3],GPSRecord[7],GPSRecord[5],GPSRecord[6] , GPSRecord[2],GPSRecord[1])) 这一条语句则提取了GPS记录中的第0、3、7、5、6、2、1列字符串发送给下一个处理单元。

declareOutputFields（）告诉下一个处理单元DistrictMatchingBolt： spout的输出数据即DistrictMatchingBolt的输入数据格式的列数和内容，即："vehicle_number","date_time","occupied","speed","bearing","lantitude","longitude" 共7列。

public class GPSReceiverSpout implements IRichSpout {
private static final long serialVersionUID = 1L;
private SpoutOutputCollector _collector;
private BufferedReader fileReader;
//private TopologyContext context;
//private String file="/home/ghchen/2013-01-05.1/2013-01-05--11_05_48.txt";
private TupleInfo tupleInfo=new TupleInfo();
static Socket sock=null;
@Override
public void close() {
}
@Override
public void open(Map conf, TopologyContext context, SpoutOutputCollector collector)
{
_collector = collector;
System.out.println("This is open function in FieldSpout !");
}
@SuppressWarnings("unused")
@Override
public void nextTuple() {
int count=0;
int ch=0;
int err=0;
try {
if(sock==null){
sock=new Socket("172.20.14.XXX",portXXXX);}
while(true){
byte[] b3= new byte[3];
if(sock!=null ){
try{
sock.getInputStream().read(b3,0,3);
ch=b3[0];
}catch ( Exception e){
System.out.println("connection reset, reconnecting ...");
sock.close();
Thread.sleep(100);
sock=new Socket("172.20.14.XXX",portXXXX);;
}
}else{
sock=new Socket("172.20.14.XXX",portXXXX);;
break ;
}
int len=SocketJava.bytesToShort(b3, 1);
if(len<0) break;
byte[] bytelen= new byte[len];
sock.getInputStream().read(bytelen);
if(bytelen==null){
System.out.println("read the second part from byte from socket failed ! ");
break;
}
sock.getInputStream().markSupported();
sock.getInputStream().mark(3);
String gpsString=SocketJava.DissectOneMessage(ch,bytelen);
String[] GPSRecord=null;
if(gpsString!=null){
GPSRecord =gpsString.split(TupleInfo.getDelimiter());
_collector.emit(new Values(GPSRecord[0],GPSRecord[3],GPSRecord[7],GPSRecord[5],
GPSRecord[6] , GPSRecord[2],GPSRecord[1]));
//}
}else{
break;
}
}
} catch (IOException e) {
e.printStackTrace();
} catch (Exception e) {
e.printStackTrace();
}
}
@Override
public void ack(Object id) {
System.out.println("OK:"+id);
}
@Override
public void fail(Object id) {
System.out.println("Fail:"+id);
}
@Override
public void declareOutputFields(OutputFieldsDeclarer declarer) {
TupleInfo tuple = new TupleInfo();
Fields fieldsArr;
try {
fieldsArr= tuple.getFieldList();
declarer.declare(fieldsArr);
} catch (Exception e) {
throw new RuntimeException("error:fail to new Tuple object in declareOutputFields, tuple is null",e);
}
}
@Override
public void activate() {
}
@Override
public void deactivate() {
}
@Override
public Map<String, Object> getComponentConfiguration() {
return null;
}
static int count=0;
public static void writeToFile(String fileName, Object obj){
try {
FileWriter fwriter;
fwriter= new FileWriter(fileName,true);
BufferedWriter writer= new BufferedWriter(fwriter);
writer.write(obj.toString());
writer.close();
} catch (IOException e1) {
e1.printStackTrace();
}
}
}

复制代码

二、DistrictMatchingBolt

最重要的两个函数是 excute() 和declareOutputFields(OutputFieldsDeclarer declarer)。

excute() 将GPSReceiverSpout接受到的数据提取经度和纬度，调用Sects类 Sect.fetchSect(GPSrecord)方法，查询本地的地理信息数据库，返回该条GPS记录所在的区域标号 districtID,并将这个字段添加到GPS后面，发射给下一个bolt ： countBolt。

declareOutputFields（）告诉下一个countBolt ，这个DistrcitMatchingBolt的输出数据格式是："viechleID", "dateTime", "occupied", "speed","bearing", "latitude", "longitude", "districtID"。

需要说明的是Sects类调用了开源的地理信息系统工具geotools，感兴趣的朋友可以去http://www.geotools.org/ 下载安装包，并将相关的jar包全部添加到Eclipse 的building path里面，就可以调用geotools查询本地的地理信息数据库了。

package main.java.realODMatrix.bolt;
import java.io.IOException;
import java.util.List;
import java.util.Map;
import backtype.storm.task.OutputCollector;
import backtype.storm.task.TopologyContext;
import backtype.storm.topology.IRichBolt;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Values;
import backtype.storm.tuple.Tuple;
import main.java.realODMatrix.spout.FieldListenerSpout;
import main.java.realODMatrix.struct.*;
public class DistrictMatchingBolt implements IRichBolt {
private static final long serialVersionUID = -433427751113113358L;
private OutputCollector _collector;
Integer districtID ;
GPSRcrd record;
Map<GPSRcrd, Integer> gpsMatch; //map
Integer taskID;
String taskname;
List<Object> inputLine;
Fields matchBoltDeclare=null;
static String path = "/home/ghchen/sects/sects.shp";
static Sects sects=null ;
int count=0;
@Override
public void prepare(Map stormConf, TopologyContext context,
OutputCollector collector) {
// TODO Auto-generated method stub
this._collector=collector;
this.taskID=context.getThisTaskId();
this.taskname=context.getThisComponentId();
}
@Override
public void execute(Tuple input) {
try {
if(sects==null){
sects= new Sects(path);
}
List<Object> inputLine = input.getValues();//getFields();
Fields inputLineFields = input.getFields();
record=new GPSRcrd(Double.parseDouble((String) inputLine.get(6)),
Double.parseDouble((String) inputLine.get(5)), Integer.parseInt((String) inputLine.get(3)),
Integer.parseInt((String) inputLine.get(4)));
if( Double.parseDouble((String) inputLine.get(6)) > 114.5692938 ||
Double.parseDouble((String) inputLine.get(6)) < 113.740000 ||
Double.parseDouble((String) inputLine.get(5)) > 22.839945 ||
Double.parseDouble((String) inputLine.get(5)) < 22.44
) return;
districtID = sects.fetchSect(record);
if(districtID!=-1)
{
System.out.println(count++ +": GPS Point falls into Sect No. :" + districtID);
inputLine.add(Integer.toString(districtID));
//input.getFields().toList().add("districtID");
List<String> fieldList= input.getFields().toList();
fieldList.add("districtID");
matchBoltDeclare=new Fields(fieldList);
//FieldListenerSpout.writeToFile("/home/ghchen/output","matchBoltDeclare="+matchBoltDeclare);
String[] obToStrings=new String[inputLine.size()];
obToStrings=inputLine.toArray(obToStrings);
_collector.emit(new Values(obToStrings));
//_collector.emit(new Values(inputLine));
}
} catch (Exception e) {
e.printStackTrace();
}
_collector.ack(input);
}
@Override
public void cleanup() {
// TODO Auto-generated method stub
System.out.println("-- District Mathchier ["+taskname+"-"+districtID+"] --");
for(Map.Entry<GPSRcrd, Integer> entry : gpsMatch.entrySet()){
System.out.println(entry.getKey()+": "+entry.getValue());
}
}
@Override
public void declareOutputFields(OutputFieldsDeclarer declarer) {
declarer.declare(new Fields ("viechleID", "dateTime", "occupied", "speed",
"bearing", "latitude", "longitude", "districtID"));
}
@Override
public Map<String, Object> getComponentConfiguration() {
// TODO Auto-generated method stub
return null;
}
}

雪峰庙

关注

0
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
使用storm实现实时GPS数据客流特征分析系统：源码分析

问题导读1.项目背景是什么？2.nextTuple() 的作用是什么？3.declareOutputFields(OutputFieldsDeclarer declarer)的作用是什么？4.excute() 和declareOutputFields(OutputFieldsDeclarer declarer) 的作用是什么？项目背景：实时GPS数据客流特征
复制链接

扫一扫