网络爬虫,也叫网络蜘蛛,有的项目也把它称作“walker”。所给的定义是“一种系统地扫描互联网,以获取索引为目的的网络程序”。网络上有很多关于网络爬虫的开源项目,其中比较有名的是和。
有时需要在网上搜集信息,如果需要搜集的是获取方法单一而人工搜集费时费力的信息,比如统计一个网站每个月发了多少篇文章、用了哪些标签,为自然语言处理项目搜集语料,或者为模式识别项目搜集图片等等,就需要爬虫程序来完成这样的任务。而且搜索引擎必不可少的组件之一也是网络爬虫。
很多网络爬虫都是用Python,Java或C#实现的。我这里给出的是Java版本的爬虫程序。为了节省时间和空间,我把程序限制在只扫描本博客地址下的网页(也就是http://johnhan.net/但不包括http://johnhany.net/wp-content/下的内容),并从网址中统计出所用的所有标签。只要稍作修改,去掉代码里的限制条件就能作为扫描整个网络的程序使用。或者对输出格式稍作修改,可以作为生成博客sitemap的工具。
代码也可以在这里下载:。
环境需求
我的开发环境是Windows7 + 。
需要提供通过url访问MySQL数据库的端口。
还要用到三个开源的Java类库:
提供HTTP接口,用来向目标网址提交HTTP请求,以获取网页的内容;
用来解析网页,从DOM节点中提取网址链接;
连接Java程序和MySQL,然后就可以用Java代码操作数据库。
代码
代码位于三个文件中,分别是:crawler.java,httpGet.java和parsePage.java。包名为net.johnhany.wpcrawler。
crawler.java
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 | package net.johnhany.wpcrawler; import java.sql.Connection; import java.sql.DriverManager; import java.sql.ResultSet; import java.sql.SQLException; import java.sql.Statement; public class crawler { public static void main(String args[]) throws Exception { String frontpage = "http://johnhany.net/" ; Connection conn = null ; //connect the MySQL database try { Class.forName( "com.mysql.jdbc.Driver" ); String dburl = "jdbc:mysql://localhost:3306?useUnicode=true&characterEncoding=utf8" ; conn = DriverManager.getConnection(dburl, "root" , "" ); System.out.println( "connection built" ); } catch (SQLException e) { e.printStackTrace(); } catch (ClassNotFoundException e) { e.printStackTrace(); } String sql = null ; String url = frontpage; Statement stmt = null ; ResultSet rs = null ; int count = 0 ; if (conn != null ) { //create database and table that will be needed try { sql = "CREATE DATABASE IF NOT EXISTS crawler" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); sql = "USE crawler" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); sql = "create table if not exists record (recordID int(5) not null auto_increment, URL text not null, crawled tinyint(1) not null, primary key (recordID)) engine=InnoDB DEFAULT CHARSET=utf8" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); sql = "create table if not exists tags (tagnum int(4) not null auto_increment, tagname text not null, primary key (tagnum)) engine=InnoDB DEFAULT CHARSET=utf8" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); } catch (SQLException e) { e.printStackTrace(); } //crawl every link in the database while ( true ) { //get page content of link "url" httpGet.getByString(url,conn); count++; //set boolean value "crawled" to true after crawling this page sql = "UPDATE record SET crawled = 1 WHERE URL = '" + url + "'" ; stmt = conn.createStatement(); if (stmt.executeUpdate(sql) > 0 ) { //get the next page that has not been crawled yet sql = "SELECT * FROM record WHERE crawled = 0" ; stmt = conn.createStatement(); rs = stmt.executeQuery(sql); if (rs.next()) { url = rs.getString( 2 ); } else { //stop crawling if reach the bottom of the list break ; } //set a limit of crawling count if (count > 1000 || url == null ) { break ; } } } conn.close(); conn = null ; System.out.println( "Done." ); System.out.println(count); } } } |
httpGet.java
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 | package net.johnhany.wpcrawler; import java.io.IOException; import java.sql.Connection; import org.apache.http.HttpEntity; import org.apache.http.HttpResponse; import org.apache.http.client.ClientProtocolException; import org.apache.http.client.ResponseHandler; import org.apache.http.client.methods.HttpGet; import org.apache.http.impl.client.CloseableHttpClient; import org.apache.http.impl.client.HttpClients; import org.apache.http.util.EntityUtils; public class httpGet { public final static void getByString(String url, Connection conn) throws Exception { CloseableHttpClient httpclient = HttpClients.createDefault(); try { HttpGet httpget = new HttpGet(url); System.out.println( "executing request " + httpget.getURI()); ResponseHandler<String> responseHandler = new ResponseHandler<String>() { public String handleResponse( final HttpResponse response) throws ClientProtocolException, IOException { int status = response.getStatusLine().getStatusCode(); if (status >= 200 && status < 300 ) { HttpEntity entity = response.getEntity(); return entity != null ? EntityUtils.toString(entity) : null ; } else { throw new ClientProtocolException( "Unexpected response status: " + status); } } }; String responseBody = httpclient.execute(httpget, responseHandler); /* //print the content of the page System.out.println("----------------------------------------"); System.out.println(responseBody); System.out.println("----------------------------------------"); */ parsePage.parseFromString(responseBody,conn); } finally { httpclient.close(); } } } |
parsePage.java
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 | package net.johnhany.wpcrawler; import java.sql.Connection; import java.sql.PreparedStatement; import java.sql.ResultSet; import java.sql.SQLException; import java.sql.Statement; import org.htmlparser.Node; import org.htmlparser.Parser; import org.htmlparser.filters.HasAttributeFilter; import org.htmlparser.tags.LinkTag; import org.htmlparser.util.NodeList; import org.htmlparser.util.ParserException; import java.net.URLDecoder; public class parsePage { public static void parseFromString(String content, Connection conn) throws Exception { Parser parser = new Parser(content); HasAttributeFilter filter = new HasAttributeFilter( "href" ); try { NodeList list = parser.parse(filter); int count = list.size(); //process every link on this page for ( int i= 0 ; i<count; i++) { Node node = list.elementAt(i); if (node instanceof LinkTag) { LinkTag link = (LinkTag) node; String nextlink = link.extractLink(); String mainurl = "http://johnhany.net/" ; String wpurl = mainurl + "wp-content/" ; //only save page from "http://johnhany.net" if (nextlink.startsWith(mainurl)) { String sql = null ; ResultSet rs = null ; PreparedStatement pstmt = null ; Statement stmt = null ; String tag = null ; //do not save any page from "wp-content" if (nextlink.startsWith(wpurl)) { continue ; } try { //check if the link already exists in the database sql = "SELECT * FROM record WHERE URL = '" + nextlink + "'" ; stmt = conn.createStatement(ResultSet.TYPE_FORWARD_ONLY,ResultSet.CONCUR_UPDATABLE); rs = stmt.executeQuery(sql); if (rs.next()) { } else { //if the link does not exist in the database, insert it sql = "INSERT INTO record (URL, crawled) VALUES ('" + nextlink + "',0)" ; pstmt = conn.prepareStatement(sql, Statement.RETURN_GENERATED_KEYS); pstmt.execute(); System.out.println(nextlink); //use substring for better comparison performance nextlink = nextlink.substring(mainurl.length()); //System.out.println(nextlink); if (nextlink.startsWith( "tag/" )) { tag = nextlink.substring( 4 , nextlink.length()- 1 ); //decode in UTF-8 for Chinese characters tag = URLDecoder.decode(tag, "UTF-8" ); sql = "INSERT INTO tags (tagname) VALUES ('" + tag + "')" ; pstmt = conn.prepareStatement(sql, Statement.RETURN_GENERATED_KEYS); //if the links are different from each other, the tags must be different //so there is no need to check if the tag already exists pstmt.execute(); } } } catch (SQLException e) { //handle the exceptions System.out.println( "SQLException: " + e.getMessage()); System.out.println( "SQLState: " + e.getSQLState()); System.out.println( "VendorError: " + e.getErrorCode()); } finally { //close and release the resources of PreparedStatement, ResultSet and Statement if (pstmt != null ) { try { pstmt.close(); } catch (SQLException e2) {} } pstmt = null ; if (rs != null ) { try { rs.close(); } catch (SQLException e1) {} } rs = null ; if (stmt != null ) { try { stmt.close(); } catch (SQLException e3) {} } stmt = null ; } } } } } catch (ParserException e) { e.printStackTrace(); } } } |
程序原理
所谓“互联网”,是网状结构,任意两个节点间都有可能存在路径。爬虫程序对互联网的扫描,在图论角度来讲,就是对有向图的遍历(链接是从一个网页指向另一个网页,所以是有向的)。常见的遍历方法有深度优先和广度优先两种。相关理论知识可以参考树的遍历:和。我的程序采用的是广度优先方式。
程序从crawler.java的main()开始运行。
1 2 3 4 | Class.forName( "com.mysql.jdbc.Driver" ); String dburl = "jdbc:mysql://localhost:3306?useUnicode=true&characterEncoding=utf8" ; conn = DriverManager.getConnection(dburl, "root" , "" ); System.out.println( "connection built" ); |
首先,调用DriverManager连接MySQL服务。这里使用的是XAMPP的默认MySQL端口3306,端口值可以在XAMPP主界面看到:
Apache和MySQL都启动之后,在浏览器地址栏输入“http://localhost/phpmyadmin/”就可以看到数据库了。等程序运行完之后可以在这里检查一下运行是否正确。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | sql = "CREATE DATABASE IF NOT EXISTS crawler" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); sql = "USE crawler" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); sql = "create table if not exists record (recordID int(5) not null auto_increment, URL text not null, crawled tinyint(1) not null, primary key (recordID)) engine=InnoDB DEFAULT CHARSET=utf8" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); sql = "create table if not exists tags (tagnum int(4) not null auto_increment, tagname text not null, primary key (tagnum)) engine=InnoDB DEFAULT CHARSET=utf8" ; stmt = conn.createStatement(); stmt.executeUpdate(sql); |
连接好数据库后,建立一个名为“crawler”的数据库,在库里建两个表,一个叫“record”,包含字段“recordID”,“URL”和“crawled”,分别记录地址编号、链接地址和地址是否被扫描过;另一个叫“tags”,包含字段“tagnum”和“tagname”,分别记录标签编号和标签名。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 | while ( true ) { httpGet.getByString(url,conn); count++; sql = "UPDATE record SET crawled = 1 WHERE URL = '" + url + "'" ; stmt = conn.createStatement(); if (stmt.executeUpdate(sql) > 0 ) { sql = "SELECT * FROM record WHERE crawled = 0" ; stmt = conn.createStatement(); rs = stmt.executeQuery(sql); if (rs.next()) { url = rs.getString( 2 ); } else { break ; } } } |
接着在一个while循环内依次处理表record内的每个地址。每次处理时,把地址url传递给httpGet.getByString(),然后在表record中把crawled改为true,表明这个地址已经处理过。然后寻找下一个crawled为false的地址,继续处理,直到处理到表尾。
这里需要注意的细节是,执行executeQuery()后,得到了一个ResultSet结构rs,rs包含SQL查询返回的所有行和一个指针,指针指向结果中第一行之前的位置,需要执行一次rs.next()才能让rs的指针指向第一个结果,同时返回true,之后每次执行rs.next()都会把指针移到下一个结果上并返回true,直至再也没有结果时,rs.next()的返回值变成了false。
还有一个细节,在执行建库建表、INSERT、UPDATE时,需要用executeUpdate();在执行SELECT时,需要使用executeQuery()。executeQuery()总是返回一个ResultSet,executeUpdate()返回符合查询的行数。
httpGet.java的getByString()类负责向所给的网址发送请求,然后下载网页内容。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | HttpGet httpget = new HttpGet(url); System.out.println( "executing request " + httpget.getURI()); ResponseHandler<String> responseHandler = new ResponseHandler<String>() { public String handleResponse( final HttpResponse response) throws ClientProtocolException, IOException { int status = response.getStatusLine().getStatusCode(); if (status >= 200 && status < 300 ) { HttpEntity entity = response.getEntity(); return entity != null ? EntityUtils.toString(entity) : null ; } else { throw new ClientProtocolException( "Unexpected response status: " + status); } } }; String responseBody = httpclient.execute(httpget, responseHandler); |
这段代码是HTTPComponents的HTTP Client组件中给出的样例,在很多情况下可以直接使用。这部分代码获得了一个字符串responseBody,里面保存着网页中的全部字符。
接着,就需要把responseBody传递给parsePage.java的parseFromString类提取链接。
1 2 3 4 5 6 7 8 9 10 11 | Parser parser = new Parser(content); HasAttributeFilter filter = new HasAttributeFilter( "href" ); try { NodeList list = parser.parse(filter); int count = list.size(); //process every link on this page for ( int i= 0 ; i<count; i++) { Node node = list.elementAt(i); if (node instanceof LinkTag) { |
在HTML文件中,链接一般都在a标签的href属性中,所以需要创建一个属性过滤器。NodeList保存着这个HTML文件中的所有DOM节点,通过在for循环中依次处理每个节点寻找符合要求的标签,可以把网页中的所有链接提取出来。
然后通过nextlink.startsWith()进一步筛选,只处理以“http://johnhany.net/”开头的链接并跳过以“http://johnhany.net/wp-content/”开头的链接。
1 2 3 4 5 6 7 8 9 10 11 | sql = "SELECT * FROM record WHERE URL = '" + nextlink + "'" ; stmt = conn.createStatement(ResultSet.TYPE_FORWARD_ONLY,ResultSet.CONCUR_UPDATABLE); rs = stmt.executeQuery(sql); if (rs.next()) { } else { //if the link does not exist in the database, insert it sql = "INSERT INTO record (URL, crawled) VALUES ('" + nextlink + "',0)" ; pstmt = conn.prepareStatement(sql, Statement.RETURN_GENERATED_KEYS); pstmt.execute(); |
在表record中查找是否已经存在这个链接,如果存在(rs.next()==true),不做任何处理;如果不存在(rs.next()==false),在表中插入这个地址并把crawled置为false。因为之前recordID设为AUTO_INCREMENT,所以要用 Statement.RETURN_GENERATED_KEYS获取适当的编号。
1 2 3 4 5 6 7 8 | nextlink = nextlink.substring(mainurl.length()); if (nextlink.startsWith( "tag/" )) { tag = nextlink.substring( 4 , nextlink.length()- 1 ); tag = URLDecoder.decode(tag, "UTF-8" ); sql = "INSERT INTO tags (tagname) VALUES ('" + tag + "')" ; pstmt = conn.prepareStatement(sql, Statement.RETURN_GENERATED_KEYS); pstmt.execute(); |
去掉链接开头的“http://johnhany.net/”几个字符,提高字符比较的速度。如果含有“tag/”说明其后的字符是一个标签的名字,把这给名字提取出来,用UTF-8编码,保证汉字的正常显示,然后存入表tags。类似地还可以加入判断“article/”,“author/”,或“2013/11/”等对其他链接进行归类。
结果
这是两张数据库的截图,显示了程序的部分结果:
在可以获得全部输出结果。可以与本博客的比较一下,看看如果想在其基础上实现sitemap生成工具,还要做哪些修改。