zoukankan      html  css  js  c++  java
  • 初版storm项目全流程自动化测试代码实现

      由于项目需要,写了版针对业务的自动化测试代码,主要应用场景在于由于业务日趋复杂,一些公共代码的改动,担心会影响已有业务。还没进行重写,但知识点还是不少的与大家分享实践下。首先,介绍下整个流处理的业务流程。

      首先 从网管实时接入数据到kafka,然后消息接入 进行预处理(这个过程是通过jetty框架,直接用servlet启动的项目,因为考虑到tomcat的并发不够,所以这样用。)随后预处理完 传入kafka,然后storm的不同的topo根据不同的传入类型,进行接入消息的规则匹配,规则是存在于前台的项目中,定时刷入redis(1分钟1刷) 随后加载用户卡数据、用户信息等(这些数据是每晚通过跑mapreduce任务生成的大宽表,直接导入redis),通过redis加载效率非常高,满足实时性(如果redis中不存在数据的情况下,会连接hbase,再进行查询) 随后进行业务处理(有时有些会调各个网管的接口,获取相应业务数据),随后将封装好的数据发总致下游通知拓扑,通知拓扑通过webservice或者restTemple发送值各个其他平台,比如微信,支付宝,短信等,最终将整个运行日志写入hbase。

      首先准备下一些需要的公共类,kafkaclient:

    private Properties properties;
    
    	private String defaultTopic;
    
    	private KafkaProducer<K, V> producer;
    
    	public void setProperties(Properties properties) {
    		this.properties = properties;
    	}
    
    	public void setDefaultTopic(String defaultTopic) {
    		this.defaultTopic = defaultTopic;
    	}
    
    	public void setProducer(KafkaProducer<K, V> producer) {
    		this.producer = producer;
    	}
    
    	public void init() {
    		if (properties == null) {
    			throw new NullPointerException("kafka properties is null.");
    		}
    		this.producer = new KafkaProducer<K, V>(properties);
    	}
    
    	public void syncSend(V value) {
    		ProducerRecord<K, V> producerRecord = new ProducerRecord<K, V>(defaultTopic, value);
    		try {
    			producer.send(producerRecord).get();
    		} catch (Exception e) {
    			throw new RuntimeException(e);
    		}
    	}
    
    	public void asyncSend(V value) {
    		ProducerRecord<K, V> producerRecord = new ProducerRecord<K, V>(defaultTopic, value);
    		producer.send(producerRecord);
    	}
    

      HbaseUtil:

    private static final Logger logger = LoggerFactory.getLogger(HbaseResult.class);
    
    	private Gson gson = new Gson();
    	
    	private HConnection connection = null;
    
    	private Configuration conf = null;
    
    	private String logFile = "D:/error.txt";
    
    	public void init() throws IOException {
    		logger.info("start init HBasehelper...");
    		conf = HBaseConfiguration.create();
    		connection = HConnectionManager.createConnection(conf);
    		logger.info("init HBasehelper successed!");
    	}
    
    	public synchronized HConnection getConnection() throws IOException {
    		if (connection == null) {
    			connection = HConnectionManager.createConnection(conf);
    		}
    		return connection;
    	}
    
    	private synchronized void closeConnection() {
    		if (connection != null) {
    			try {
    				connection.close();
    			} catch (IOException e) {
    			}
    		}
    		connection = null;
    	}
    

      kafkaClient主要负责将读取报文的信息发送至kafka,随之又topo自行运算,最终使用通过调用hbaseUtil,对相应字段的比对查询。

    那么下面对整个自动化测试的流程进行说明:

      一、导入前台活动  由于是自动化测试,我们不可能每次都手工上下线,或在页面配置启用某个活动,所以通过直接调用前台系统 导入功能 的方法,将活动配置写入mysql数据库,并进行状态的切换。s

       List<String> codeList = new ArrayList<String>();
         	 List<String> activityIdList = new ArrayList<String>();
            try {
              FileBody bin = new FileBody(new File("D:/activityTest/activity.ac"));
              InputStream in = bin.getInputStream();
              BufferedReader br = new BufferedReader(new InputStreamReader(in));
              String tr = null;
              while((tr = br.readLine())!=null){
                  HttpPost httppost = new HttpPost("http://*********:8088/***/***/***/import");
                  CloseableHttpClient httpclient = HttpClients.createDefault();
        	    ObjectMapper mapper = new ObjectMapper();
        	      ActivityConfig cloneActivity = null;
        	      cloneActivity = mapper.readValue(tr.toString(),ActivityConfig.class);
        	      List<ActivityConfig> cloneActivitys = new ArrayList<ActivityConfig>();//存放所有的活动
        	      cloneActivitys.add(cloneActivity);
        			
        	      for (ActivityConfig cloneActivity1 : cloneActivitys) {
                    String code = cloneActivity1.getCode();
        	          codeList.add(code);
        	      }
                    HttpEntity reqEntity = MultipartEntityBuilder.create()
                            .addPart("file", bin)
                            .build();
                    httppost.setEntity(reqEntity);
                    System.out.println("executing request " + httppost.getRequestLine());
                    CloseableHttpResponse response = httpclient.execute(httppost);
                    System.out.println(response.getStatusLine());
                    HttpEntity resEntity = response.getEntity();
                    if (resEntity != null) {
                        System.out.println("Response content length: " + resEntity.getContentLength());
                    }
                    EntityUtils.consume(resEntity);
                    response.close();
                    httpclient.close();
              }
              for(String code : codeList){
                String code1 = "'" + code + "'";
                  if(StringUtils.isNotEmpty(activityCode)){
                    activityCode.append(",");
                  }
              activityCode.append(code1);
               }
          }
          return activityIdList;
        ]

    二、读取准备好的报文数据(xml形式需通过解析,数据分隔符格式读取后直接发送至kafka)

    public String readTxt() throws IOException{
            StringBuffer sendMessage = new StringBuffer();
            BufferedReader br = null;
            try {
                br = new BufferedReader(
                        new InputStreamReader(new FileInputStream(MessageText), "UTF-8"));
                String line = "";
                while((line = br.readLine()) != null){
                    if (line.contains("<?xml")) {
                        int beginIndex = line.indexOf("<?xml");
                        line = line.substring(beginIndex);
                    }
                    sendMessage.append(line);
                }
            } catch (UnsupportedEncodingException e) {
                e.printStackTrace();
            } catch (FileNotFoundException e) {
                e.printStackTrace();
            }finally{
                br.close();
            }
            return sendMessage.toString();
        }

    三、下来,我们需要将解析后的报文数据写入hbase的相应用户宽表、卡宽表中,以便storm拓扑中进行用户数据的加载,这里的rowkey为预分区过的。

    HbaseResult baseHelper = new HbaseResult();
    baseHelper.init();
    tableName = "CARD****";
    rowkey = HTableManager.generatRowkey(cardNo);
    data.put("*****", "10019");
    data.put("*****", cardNo);
    data.put("*****", certNo);
    data.put("*****", "A"); 
    data.put("*****", "1019"); 
    data.put("*****", supplementCardNo);
    data.put("*****", "10020");
    data.put("*****", certNo);
    data.put("*****", cardType);
    data.put("*****", cardType);
    data.put("*****", cardNo.substring(12,16));
    data.put("*****", "F");			
    data.put("*****", "ysy");
    Put put = new Put(Bytes.toBytes(rowkey));
    for (Entry<String, String> rs : data.entrySet()) {
        put.add(HTableManager.DEFAULT_FAMILY_NAME, Bytes.toBytes(rs.getKey()), Bytes.toBytes(rs.getValue()));
    }	
    baseHelper.put(tableName, put);
    System.out.println("rowkey:"+rowkey);
    data.clear();
    

    四、随后就可进行消息的发送(发送至集群的kafka)

    KafkaInit();
    FSTConfiguration fstConf = FSTConfiguration.getDefaultConfiguration();
    kafkaClient.syncSend(fstConf.asByteArray(kafkaData));
    

    五、最终进行发送数据的字段对比(通过报文中的,预设的数据字段 与 最终输出的字段或结果进行对比,随后追加写入输出文件)

    Result result = baseHelper.getResult("EVENT_LOG_DH", messageKey);
    //对比字段
    baseHelper.compareData(dataMap, result,activityCode);
    
    public Result getResult(String tableName, String rowKey) throws IOException {
        Get get = new Get(Bytes.toBytes(rowKey));
        Result result = null;
        HTableInterface tableInterface = null;
        try {
            tableInterface = getConnection().getTable(tableName);
            result = tableInterface.get(get);
            return result;
             } catch (Exception e) {
                 closeConnection();
            logger.error("", e);
             } finally {
                 if (tableInterface != null) {
                tableInterface.close();
             }
    }
    
    
    
    public void compareData(Map<String,Object> messageData, Result res,List<String> activityCode) throws IOException{
            List<String> Messages = new ArrayList<String>();
            for (Cell cell : res.rawCells()) {
                String qualifier = Bytes.toString(CellUtil.cloneQualifier(cell));
                if(Bytes.toString(CellUtil.cloneQualifier(cell)).equalsIgnoreCase("VARIABLESETS")){
                    System.out.println(qualifier + "[" + new Gson().fromJson(Bytes.toString(CellUtil.cloneValue(cell)), Map.class) + "]");
                    @SuppressWarnings("unchecked")
                    Map<String,Object> data = gson.fromJson(Bytes.toString(CellUtil.cloneValue(cell)), Map.class);
                    String message = "";
                    for(String datakey : data.keySet()){
                        if(messageData.containsKey(datakey)){
                            
                            String dataValue = getString(data,datakey);
                            String messageValue = getString(messageData,datakey);
                            
                            if(datakey.equals("dh22")){
                                dataValue = dataValue.substring(0,dataValue.indexOf("."));
                                messageValue = messageValue.substring(0,messageValue.indexOf("."));
                            }
                            if(dataValue.equals(messageValue)){
                                message = datakey + " = " + dataValue + " 与报文中的 " + dataValue + "对比相同";
                                Messages.add(message);
                            }else{
                                message = datakey + " = " + dataValue + " 与报文中的 " + dataValue + "不一致!!!";
                                Messages.add(message);
                            }
                        }
                    }
                    
                }
                if(Bytes.toString(CellUtil.cloneQualifier(cell)).equalsIgnoreCase("NOTIFY__")){
                    
                }
            }
    
            if(Messages.size() > 0){
                StringBuffer sb = new StringBuffer();
                for(String error : Messages){
                    sb.append(error).append("
    ");
                }
                FileWriter fw = new FileWriter(logFile,true);
                fw.write("
    ----------------------");
                fw.write(sb.toString());
                fw.flush();
                fw.close();
            }else{
                String sb = "没有对不上的字段呀";
                FileWriter fw = new FileWriter(logFile);
                fw.write(sb);
                fw.flush();
                fw.close();
            }
        }

    六、清除导入的数据等信息,整个流程结束~

    public void delHbaseData(String cardNo,String certNo) throws IOException{
        String rowkeyCard = HTableManager.generatRowkey(cardNo) ;
        String rowKeyUse =  HTableManager.generatRowkey(certNo) ;
        Delete delData = null;
        HTableInterface tableInterface = null;
        String tableName = "";
        try {
            tableInterface = getConnection().getTable(tableName);
            tableInterface.delete(delData);
            }
                
            return;
        } catch (Exception e) {
            closeConnection();
            logger.error("", e);
            } finally {
        if (tableInterface != null) {
            tableInterface.close();
        }
        }
    }

      

  • 相关阅读:
    一些数据集
    经典的图像匹配算法----SIFT
    LDA处理文档主题分布代码
    Dirichlet Process
    主题模型-LDA浅析
    matplotlib —— 添加文本信息(text)
    xgboost原理及应用
    XGBoost参数调优
    MySQL中实现Oracle里面 rank()over ( PARTITION BY ORDER BY) 分类分组功能
    4.12 省选模拟赛 LCA on tree 树链剖分 树状数组 分析答案变化量
  • 原文地址:https://www.cnblogs.com/yangsy0915/p/5767599.html
Copyright © 2011-2022 走看看