《搜索引擎零距离》IRVM 已爬地址处理

roki

浏览: 62706 次
性别:
来自: 上海

最近访客更多访客>>

wind2412

lanyi921

hadasione

whx2010nj

博主相关

博客

微博

相册

留言

关于我

文章分类

社区版块

存档分类

搜索引擎 Hadoop Apache log4j

使用BDB来存储已爬地址，用int status的各个不同的位，来表示不同的访问状态：
访问过；保存过；修改过等等

package com.rayeen.spider.vertical.data;

import java.io.File;
import java.io.IOException;
import java.io.UnsupportedEncodingException;

import org.apache.hadoop.io.MD5Hash;
import org.apache.log4j.Logger;
import com.sleepycat.je.DatabaseException;
import com.sleepycat.je.Environment;
import com.sleepycat.persist.EntityStore;
import com.sleepycat.persist.PrimaryIndex;
import com.sleepycat.persist.SecondaryIndex;
import com.rayeen.spider.vertical.constant.MetResourceTag;
import com.rayeen.spider.vertical.constant.MsgConstant;
import com.rayeen.spider.vertical.util.ResutTree;
import org.apache.commons.lang.*;

public class MetResourceProtocolImpl implements MetResourceProtocol {
	
	
	public MetResourceProtocolImpl(String name){
		this.name=name;
	}
	String name;

	static final Logger LOG = Logger.getLogger(ResutTree.class);

	public static String openMode = "append";

	public static int hitCnt = 0;

	private static MyDbEnv myDbEnv = new MyDbEnv();

	static EntityStore da;

	static PrimaryIndex<String, MetResourceBE> infoMap = null;

	static SecondaryIndex<String, String, MetResourceBE> infoMapByDigest = null;

	static int flushCnt = 0;
 
       //初始化BDB数据库环境
	public synchronized void initilize() throws DatabaseException {

		Environment myDbEnvironment = null;
		
		
		File met = new File("./metResource");
		if(!met.exists()){
			met.mkdir();
		}
		
		
		File file = new File("./metResource/"+name);

		try {
			if (!file.exists()) {

				if (!file.exists()) {
					file.mkdir();
				}

				myDbEnv.close();
				myDbEnv.setupNoTransact(file, false);

			} else {// 
				try {
					myDbEnv.setupAppend(file);
				} catch (Exception e) {
					LOG.error(e.getMessage());
					file.delete();

					file.mkdir();
					myDbEnv.setupNoTransact(file, false);
				}
			}

			myDbEnv.setCacheSize(1024);
		} catch (DatabaseException e1) {
			// TODO Auto-generated catch block
			e1.printStackTrace();
		}

		try {
			da = myDbEnv.getEntityStore();
			infoMap = da.getPrimaryIndex(String.class, MetResourceBE.class);
			infoMapByDigest = da.getSecondaryIndex(infoMap, String.class,
					"digest");

		} catch (DatabaseException e) {
			// TODO Auto-generated catch block
			e.printStackTrace();
		}

	}

	public int error() throws IOException {
		throw new IOException("bobo");
	}

	public int getMetResource(String uri, byte[] content) {
		
		if(null==uri) return 0;
		
		LOG.info("uri:" + uri);
		try {
			MetResourceBE metResource = infoMap.get(uri);

			if (metResource != null) {
				int status = metResource.getStatus();
				String hash = "";
				if (content == null) {
					hash = MD5Hash.digest(uri).toString();
				} else {
					hash = MD5Hash.digest(content).toString();
				}
				int oldStatus = 0;
				String oldHash = metResource.getDigest();
				if (StringUtils.equalsIgnoreCase(oldHash, hash)) {
					status ^= MetResourceBE.MODIFIED;// 去掉“修改过”标志位，未更改过
					status |= MetResourceBE.UNMODIFIED;// 去掉“修改过”标志位，未更改过
				} else {
					status ^= MetResourceBE.UNMODIFIED;//未更改过
					status |= MetResourceBE.MODIFIED;// 更改过
				}
				
				LOG.info("status:" + status);
				
				return status;
			}
		} catch (DatabaseException e) {
			// TODO Auto-generated catch block
			e.printStackTrace();
		}
		return 0;
	}

	/**
	 * 
	 * @param uri
	 *            地址
	 * @param content
	 *            内容
	 * @param status
	 *            当前状态(“经过但是不保存”还是“保存”)
	 * @return
	 * @throws DatabaseException
	 */
	public String putMetResource(String uri, byte[] content, int status,
			MetResourceTag type) throws DatabaseException {

		String hash = "";
		//uri = uri.toLowerCase();
		if (content == null) {
			hash = MD5Hash.digest(uri).toString();
		} else {
			hash = MD5Hash.digest(content).toString();
		}

		MetResourceBE metResource = null;
		
		//覆盖
		if (type == MetResourceTag.COVER) {
			
			metResource = infoMap.get(uri);
			metResource = new MetResourceBE();
			metResource.setUri(uri);
			metResource.setDigest(hash);
			metResource.setStatus(status);
			infoMap.putNoReturn(metResource);
			myDbEnv.sync();
			return MsgConstant.SUCCESS;
		}

		int oldStatus = 0;
		if (infoMap.contains(uri)) {
			metResource = infoMap.get(uri);

			oldStatus = metResource.getStatus();
			String oldHash = metResource.getDigest();
			if (StringUtils.equalsIgnoreCase(oldHash, hash)) {
				oldStatus ^= MetResourceBE.MODIFIED;// 去掉“修改过”标志位，未更改过
				oldStatus |= MetResourceBE.UNMODIFIED;// 去掉“修改过”标志位，未更改过
			} else {
				oldStatus ^= MetResourceBE.UNMODIFIED;// 去掉“未修改过”标志位，未更改过
				oldStatus |= MetResourceBE.MODIFIED;// 更改过
			}
			oldStatus |= status;
			metResource.setStatus(oldStatus);

		} else {// 遇到过相同的页面

			if (infoMapByDigest.contains(hash)) {
				oldStatus = MetResourceBE.SAME_CONTENT;// 不同uri相同页面
			} else {// 没有遇到过相同内容的页面
				// 根据status来设置是met还是saved
				oldStatus = status;
			}

			metResource = new MetResourceBE();

			metResource.setUri(uri);
			metResource.setDigest(hash);
			metResource.setStatus(oldStatus);

		}

		if (null != metResource) {
			infoMap.putNoReturn(metResource);
			
			if(flushCnt++ % 5==0){
			    myDbEnv.sync();
			}
		}

		return MsgConstant.SUCCESS;
	}

	public int getMetResource(String uri) {
		// TODO Auto-generated method stub
		return getMetResource(uri, null);
	}

	/**
	 * type: 覆盖还是融合
	 */
	public String putMetResource(String uri, String content, int status,
			MetResourceTag type) throws DatabaseException {
		try {
			putMetResource(uri, content.getBytes("UTF-8"), status, type);
		} catch (UnsupportedEncodingException e) {
			e.printStackTrace();
		} catch (DatabaseException e) {
			e.printStackTrace();
		}
		
		if(flushCnt++ % 5==0){
		    myDbEnv.sync();
		}
		
		return MsgConstant.SUCCESS;
	}

}

其中MetResourceBE的实现：

package com.rayeen.spider.vertical.data;


import java.io.Serializable;
import java.util.logging.Logger;


import com.sleepycat.persist.*;

import com.sleepycat.persist.model.*;

import com.sleepycat.persist.model.Entity;
import com.sleepycat.persist.model.PrimaryKey;
import com.sleepycat.persist.model.SecondaryKey;
import  com.sleepycat.persist.model.Relationship;


@Entity
public class MetResourceBE implements Serializable{

	
	@PrimaryKey(sequence="ID")
	private String  uri;
	 
	@SecondaryKey(relate=Relationship.MANY_TO_ONE)
	private String digest;
	
	
	
	private java.util.Date insertTime;
	private int status;
	
	final static public int MET=0x1;//url遇到过
	final static public int SAVED=0x2;//url保存过
	final static public int MODIFIED=0x4;//内容更新过
	
	final static public int UNMET=0x8;//url未遇到过
	final static public int UNSAVED=0x10;//url未保存过
	final static public int UNMODIFIED=0x20;//内容未更新
	
	final static public int BLANK=0x0;//未遇到过，未保存过，未更新过
	final static public int FULL=MET|SAVED|MODIFIED|UNMET|UNSAVED|UNMODIFIED;// 遇到过， 保存过， 更新过
//	
	

	

	//final static public int MET_MODIFIED=0x5;//遇到url相同，内容不同的页面
	
	
//	final static public int UNMET=0x0;//url遇到过，未更新过，未保存过
//	final static public int UNSAVED_UNMODIFY=0x4;//遇到过，更新或未更新过，未保存过
//	final static public int UNSAVED_MODIFY=0x6;//遇到过，更新或未更新过，未保存过
	
	
	//final static public int UNSAVED=0x8;//url保存过

	
	
	final static public int SAME_CONTENT=0x10;//遇到url不同，内容相同的页面
	final static public int SAME_URL=0x20;//遇到url相同，内容不同的页面
//	final static public int MODIFIED=0x40;//遇到url相同，内容不同的页面
//	final static public int UNMODIFIED=0x80;//遇到url相同，内容不同的页面

	



	
	public java.util.Date getInsertTime() {
		return insertTime;
	}

	public void setInsertTime(java.util.Date insertTime) {
		this.insertTime = insertTime;
	}

	public String getDigest() {
		return digest;
	}

	public void setDigest(String digest) {
		this.digest = digest;
	}

	public String getUri() {
		return uri;
	}

	public void setUri(String uri) {
		this.uri = uri;
	}

	public int getStatus() {
		return status;
	}

	public void setStatus(int status) {
		this.status = status;
	}

}

判断是否爬过某个页面的代码片段：

	// 在fetchedList非空的情况下才处理以下逻辑
			if (null != fetchedList) {
				int curStatus = fetchedList.getMetResource(realUrl, content
						.getBytes("UTF-8"));

				// 如果不应该处理这个页面，那么直接返回
				if (ParseUtils.EntranceCantProcess(processStandard, curStatus)) {
					log(MetResourceUtil.explainMetResourceReason(
							url.getToUrl(), processStandard, curStatus),
							LogType.ENTRANCE_CONDITIONAL);
					return curStatus;
				}
				// 否则，记录这个页面的状态(met和unmet状态转换)
				fetchedList.putMetResource(url.getToUrl(), content
						.getBytes("UTF-8"), MetResourceBE.UNMET
						| MetResourceBE.MET, MetResourceTag.MERGE);

			}

保存数据之后，修改URL状态的代码片段：

public void save(String saveStat) throws SemanticException {

		// curHierarchyResultMap中保存着到上级页面为止的入口处的信息
		// 之前的若干步骤都是为了填充curHierarchyResultMap里的数据
		// 这里的curHierarchyResultMap是上层的crawl函数设进参数的HierarchyResultMap
		// 本层的所有save函数共用这个curHierarchyResultMap

		// curUri会被enter之后的处理过程还原出来,强行设置为
		VMUtils.save(curUri, curPage, curHierarchyResultMap, curCrsc, saveStat);

		// 保存“已保存过标记”
		if (null != fetchedList) {
			try {
				fetchedList.putMetResource(curUrl.getToUrl(), curContent,
						MetResourceBE.SAVED | MetResourceBE.UNSAVED,
						MetResourceTag.MERGE);
			} catch (DatabaseException e) {
				e.printStackTrace();
			}
		}

	}

一些位操作的辅助函数：



	static TObjectIntHashMap bitmap = new TObjectIntHashMap();

	static {
		bitmap.put("遇到过", MetResourceBE.MET);
		bitmap.put("更新过", MetResourceBE.MODIFIED);
		bitmap.put("保存过", MetResourceBE.SAVED);

		bitmap.put("未遇到过", MetResourceBE.UNMET);
		bitmap.put("未更新过", MetResourceBE.UNMODIFIED);
		bitmap.put("未保存过", MetResourceBE.UNSAVED);
	}
	/**
	 * 处理前三位数据 111(未遇到过||更新过||未保存过)& 001(未保存过) 返回前3位数据（位置上的后三位，逻辑上的前3位）
	 * 
	 * @param standard
	 * @return
	 * @throws SemanticException
	 */
	public static int parseProcessStardard(String standard)
			throws SemanticException {
		if (null == standard)
			return -1;

		int idx = standard.indexOf(ConfConstant.PROCESS_STANDARD);
		int status = 0;

		if (idx != -1) {
			standard = standard.substring(idx
					+ ConfConstant.PROCESS_STANDARD.length() + 1);
		}

		String[] stdsOR = Pattern.compile("||", Pattern.LITERAL).split(
				standard, 0);
		int or = MetResourceBE.BLANK;
		for (String strOR : stdsOR) {
			strOR = strOR.trim();
			int and = MetResourceBE.BLANK;
			String[] stdsAND = Pattern.compile("&&", Pattern.LITERAL).split(
					strOR, 0);
			for (String strAnd : stdsAND) {
				strAnd = strAnd.trim();
				if (!bitmap.containsKey(strAnd)) {
					ParalleIRVirtualMachine.error(
							"error enterance strandard grammer:"
									+ strAnd.substring(1), ErrorType.GRAMMER);
				}
				and |= bitmap.get(strAnd);
			}

			or = or | and;
		}
		status = or;

		return status;
	}

	static public boolean canProceess(int standard, int status) {

		return ((standard | status) & MetResourceBE.FULL) > 0;
	}

	static public boolean PageCanProceess(String standardStr, int status)
			throws SemanticException {

		int standard = parseProcessStardard(standardStr);
		return ((standard | status) & MetResourceBE.FULL) > 0;
	}

	// 如果：所有遇到过的uri都不进入（不探测是否更新），则在met的情况下，cantEnter返回true
	static public boolean EntranceCantEnter(String standardStr, int status)
			throws SemanticException {
		int standard = parseProcessStardard(standardStr);
		return EntranceCantEnter(standard, status);
	}

	static public boolean EntranceCantEnter(int standard, int status) {

		if (standard == -1)
			return false;

		// url遇到过，未更新过，未保存过
		if (standard == MetResourceBE.UNMET && (status & MetResourceBE.MET) > 0) {
			return true;
		}
		// 未保存过+更新过+遇到过+<->当前状态:未保存过+未更新过+遇到过+
		//
		return false;
	}

	// 如果标准是:更新过，而status是未更新过
	// 或者标准是:未保存过，而status是保存过，那么，本页不需要处理

	// 如果标准是:更新过,未保存过,status是未更新过，未保存过，那么，没有不能完全符合，也不能处理
	// 不考虑是否遇到过这个页面

	// 只处理标准中提到的信息,没提到的无所谓
	// 是否需要多加一倍的字段数? 由3位01变成6位01?
	static public boolean EntranceCantProcess(int standard, int status) {

		if (standard == -1)
			return false;

		// "未遇到过"也就是,status中的"未遇到过"位必须是1
		if ((standard & MetResourceBE.UNMET) > 0
				&& (status & MetResourceBE.UNMET) > 0) {
			return true;
		}

		// 是否更新过
		if (((standard & MetResourceBE.MODIFIED) ^ (status & MetResourceBE.MODIFIED)) != 0) {
			return true;
		}

		// 如果标准是:更新过||未保存过||未遇到过 , 而status是未更新过&&未保存过&&遇到过，
		// 则不处理这个页面
		if (standard == (MetResourceBE.MODIFIED | MetResourceBE.SAVED | MetResourceBE.MET)
				&& status == (MetResourceBE.MODIFIED | MetResourceBE.SAVED)) {
			return true;
		}
		return false;
	}

分享到：