Commit ecb61d13 authored by yihua.huang's avatar yihua.huang

update pipeline

parent 755b9aa8
......@@ -10,6 +10,9 @@ import us.codecraft.webmagic.processor.PageProcessor;
import us.codecraft.webmagic.schedular.QueueSchedular;
import us.codecraft.webmagic.schedular.Schedular;
import java.util.ArrayList;
import java.util.List;
/**
* User: cairne
* Date: 13-4-21
......@@ -19,7 +22,7 @@ public class Spider implements Runnable {
private Downloader downloader = new HttpClientDownloader();
private Pipeline pipeline = new ConsolePipeline();
private List<Pipeline> pipelines = new ArrayList<Pipeline>();
private PageProcessor pageProcessor;
......@@ -47,7 +50,7 @@ public class Spider implements Runnable {
}
public Spider pipeline(Pipeline pipeline) {
this.pipeline = pipeline;
this.pipelines.add(pipeline);
return this;
}
......@@ -56,6 +59,9 @@ public class Spider implements Runnable {
public void run() {
Site site = pageProcessor.getSite();
Request request = schedular.poll(site);
if (pipelines.isEmpty()){
pipelines.add(new ConsolePipeline());
}
while (request != null) {
Page page = downloader.download(request,site);
if (page == null) {
......@@ -64,7 +70,9 @@ public class Spider implements Runnable {
}
pageProcessor.process(page);
addRequest(page);
pipeline.process(page,site);
for (Pipeline pipeline : pipelines) {
pipeline.process(page,site);
}
sleep(site.getSleepTime());
request = schedular.poll(site);
}
......
package us.codecraft.webmagic.pipeline;
import org.apache.commons.codec.digest.DigestUtils;
import org.apache.commons.lang3.StringUtils;
import us.codecraft.webmagic.Page;
import us.codecraft.webmagic.Site;
import us.codecraft.webmagic.selector.Selectable;
import us.codecraft.webmagic.utils.UrlUtils;
import java.io.File;
import java.io.FileWriter;
......@@ -21,7 +21,7 @@ public class FilePipeline implements Pipeline {
private String path = "/data/temp/webmagic/";
public FilePipeline(){
public FilePipeline() {
}
......@@ -30,11 +30,9 @@ public class FilePipeline implements Pipeline {
}
@Override
public void process(Page page,Site site) {
public void process(Page page, Site site) {
String domain = site.getDomain();
domain = StringUtils.removeStart(domain, "http://");
domain = StringUtils.removeStart(domain, "https://");
domain = StringUtils.replace(domain, "/", "");
domain = UrlUtils.getDomain(domain);
String path = this.path + "" + domain + "/";
File file = new File(path);
if (!file.exists()) {
......
......@@ -23,6 +23,9 @@ public class RegexSelector implements Selector {
if (StringUtils.isBlank(regexStr)){
throw new IllegalArgumentException("regex must not be empty");
}
if (!StringUtils.contains(regexStr,"(")&&!StringUtils.contains(regexStr,")")){
regexStr="("+regexStr+")";
}
if (!StringUtils.contains(regexStr,"(")||!StringUtils.contains(regexStr,")")){
throw new IllegalArgumentException("regex must have capture group 1");
}
......
......@@ -20,6 +20,11 @@
<version>4.7</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.freemarker</groupId>
<artifactId>freemarker</artifactId>
<version>2.3.19</version>
</dependency>
</dependencies>
<build>
......
package us.codecraft.webmagic.pipeline;
import freemarker.template.Configuration;
import freemarker.template.Template;
import freemarker.template.TemplateException;
import org.apache.commons.codec.digest.DigestUtils;
import us.codecraft.webmagic.Page;
import us.codecraft.webmagic.Site;
import us.codecraft.webmagic.utils.UrlUtils;
import java.io.*;
/**
* User: cairne
* Date: 13-6-8
* Time: 下午9:00
*/
public class FreemarkerPipeline implements Pipeline {
private Configuration configuration;
private Template template;
private String path = "/data/temp/webmagic/ftl/";
public FreemarkerPipeline(String template, String path) throws IOException {
configuration = new Configuration();
configuration.setDirectoryForTemplateLoading(new File(this.getClass().getClassLoader().getResource("ftl/").getFile()));
this.template = configuration.getTemplate(template);
this.path = path;
File file = new File(path);
if (!file.exists()) {
file.mkdir();
}
}
public FreemarkerPipeline(String template) throws IOException {
this(template, "/data/temp/webmagic/ftl/");
}
@Override
public void process(Page page, Site site) {
String domain = site.getDomain();
domain = UrlUtils.getDomain(domain);
String path = this.path + "" + domain + "/";
try {
PrintWriter printWriter = new PrintWriter(new FileWriter(path + DigestUtils.md5Hex(page.getUrl().toString()) + ".html"));
template.process(page.getFields(), printWriter);
printWriter.close();
} catch (TemplateException e) {
e.printStackTrace(); //To change body of catch statement use File | Settings | File Templates.
} catch (IOException e) {
e.printStackTrace(); //To change body of catch statement use File | Settings | File Templates.
}
}
}
<item>
<title>$it.Title</title>
<link>http://127.0.0.1/wordpress/?p=$it.Id</link>
<pubDate>${date}</pubDate>
<dc:creator>admin</dc:creator>
<guid isPermaLink="false">http://127.0.0.1/wordpress/?p=$it.Id</guid>
<description></description>
<content:encoded><![CDATA[${text}]]></content:encoded>
<excerpt:encoded><![CDATA[]]></excerpt:encoded>
<#--<wp:post_id>$it.Id</wp:post_id>-->
<wp:post_date>${date}</wp:post_date>
<wp:post_date_gmt>${date}</wp:post_date_gmt>
<wp:comment_status>open</wp:comment_status>
<wp:ping_status>open</wp:ping_status>
<wp:post_name>${title}</wp:post_name>
<wp:status>publish</wp:status>
<wp:post_parent>0</wp:post_parent>
<wp:menu_order>0</wp:menu_order>
<wp:post_type>post</wp:post_type>
<wp:post_password></wp:post_password>
<wp:is_sticky>0</wp:is_sticky>
$tags
</item>
\ No newline at end of file
package us.codecraft.webmagic;
import org.junit.Test;
import us.codecraft.webmagic.pipeline.FreemarkerPipeline;
import java.io.IOException;
/**
* User: cairne
* Date: 13-6-9
* Time: 上午7:14
*/
public class FreemarkerPipelineTest {
@Test
public void test() throws IOException {
FreemarkerPipeline freemarkerPipeline = new FreemarkerPipeline("wordpress.ftl");
}
}
<item>
<title>$it.Title</title>
<link>http://127.0.0.1/wordpress/?p=$it.Id</link>
<pubDate>${date}</pubDate>
<dc:creator>admin</dc:creator>
<guid isPermaLink="false">http://127.0.0.1/wordpress/?p=$it.Id</guid>
<description></description>
<content:encoded><![CDATA[${text}]]></content:encoded>
<excerpt:encoded><![CDATA[]]></excerpt:encoded>
<#--<wp:post_id>$it.Id</wp:post_id>-->
<wp:post_date>${date}</wp:post_date>
<wp:post_date_gmt>${date}</wp:post_date_gmt>
<wp:comment_status>open</wp:comment_status>
<wp:ping_status>open</wp:ping_status>
<wp:post_name>${title}</wp:post_name>
<wp:status>publish</wp:status>
<wp:post_parent>0</wp:post_parent>
<wp:menu_order>0</wp:menu_order>
<wp:post_type>post</wp:post_type>
<wp:post_password></wp:post_password>
<wp:is_sticky>0</wp:is_sticky>
$tags
</item>
\ No newline at end of file
......@@ -15,6 +15,11 @@
<artifactId>webmagic-core</artifactId>
<version>0.0.1-SNAPSHOT</version>
</dependency>
<dependency>
<groupId>us.codecraft</groupId>
<artifactId>webmagic-plugin</artifactId>
<version>0.0.1-SNAPSHOT</version>
</dependency>
<dependency>
<groupId>junit</groupId>
<artifactId>junit</artifactId>
......
......@@ -3,6 +3,7 @@ package us.codecraft.webmagic.samples;
import us.codecraft.webmagic.Page;
import us.codecraft.webmagic.Site;
import us.codecraft.webmagic.processor.PageProcessor;
import us.codecraft.webmagic.selector.PlainText;
import java.util.List;
......@@ -14,20 +15,21 @@ import java.util.List;
public class DiaoyuwengProcessor implements PageProcessor {
@Override
public void process(Page page) {
//http://progressdaily.diandian.com/post/2013-01-24/40046867275
List<String> requests = page.getHtml().rs("<a[^<>]*href=[\"']{1}(/shop/.*?)[\"']{1}").toStrings();
List<String> requests = page.getHtml().as().rs("(http://www\\.diaoyuweng\\.com/home\\.php\\?mod=space&uid=88304&do=thread&view=me&type=thread&order=dateline&from=space&page=\\d+)").toStrings();
page.addTargetRequests(requests);
requests = page.getHtml().rs("<a[^<>]*href=[\"']{1}(/search/category/.*?)[\"']{1}").toStrings();
requests = page.getHtml().as().rs("(http://www\\.diaoyuweng\\.com/thread-\\d+-1-1.html)").toStrings();
page.addTargetRequests(requests);
if (page.getUrl().toString().contains("shop")){
page.putField("title", page.getHtml().x("//h1[@class='shop-title']"));
page.putField("content", page.getHtml().sc());
if (page.getUrl().toString().contains("thread")){
page.putField("title", page.getHtml().x("//a[@id='thread_subject']"));
page.putField("content", page.getHtml().x("//div[@class='pcb']//tbody"));
page.putField("date",page.getHtml().r("发表于 (\\d{4}-\\d+-\\d+ \\d+:\\d+:\\d+)"));
page.putField("id",new PlainText("1000"+page.getUrl().r("http://www\\.diaoyuweng\\.com/thread-(\\d+)-1-1.html").toString()));
}
}
@Override
public Site getSite() {
return Site.me().setDomain("www.dianping.com").setStartUrl("http://www.dianping.com/").
setUserAgent("Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_2) AppleWebKit/537.31 (KHTML, like Gecko) Chrome/26.0.1410.65 Safari/537.31");
return Site.me().setDomain("www.diaoyuweng.com").setStartUrl("http://www.diaoyuweng.com/home.php?mod=space&uid=88304&do=thread&view=me&type=thread&from=space").
setUserAgent("Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_2) AppleWebKit/537.31 (KHTML, like Gecko) Chrome/26.0.1410.65 Safari/537.31").setEncoding("GBK").setSleepTime(500);
}
}
package us.codecraft.webmagic.processor;
import org.junit.Test;
import us.codecraft.webmagic.Spider;
import us.codecraft.webmagic.pipeline.FilePipeline;
import us.codecraft.webmagic.pipeline.FreemarkerPipeline;
import us.codecraft.webmagic.samples.DiaoyuwengProcessor;
import us.codecraft.webmagic.schedular.FileCacheQueueSchedular;
import java.io.IOException;
/**
* User: cairne
* Date: 13-6-9
* Time: 上午8:02
*/
public class DiaoyuwengProcessorTest {
@Test
public void test() throws IOException {
DiaoyuwengProcessor diaoyuwengProcessor = new DiaoyuwengProcessor();
FreemarkerPipeline pipeline = new FreemarkerPipeline("wordpress.ftl");
Spider.me().pipeline(new FilePipeline()).pipeline(pipeline).schedular(new FileCacheQueueSchedular(diaoyuwengProcessor.getSite(), "/data/temp/webmagic/cache/")).
processor(diaoyuwengProcessor).run();
}
}
<item>
<title>${title}</title>
<link>http://127.0.0.1/wordpress/?p=${id}</link>
<pubDate>${date}</pubDate>
<dc:creator>admin</dc:creator>
<guid isPermaLink="false">http://127.0.0.1/wordpress/?p=${id}</guid>
<description></description>
<content:encoded><![CDATA[${content}]]></content:encoded>
<excerpt:encoded><![CDATA[]]></excerpt:encoded>
<#--<wp:post_id>$it.Id</wp:post_id>-->
<wp:post_date>${date}</wp:post_date>
<wp:post_date_gmt>${date}</wp:post_date_gmt>
<wp:comment_status>open</wp:comment_status>
<wp:ping_status>open</wp:ping_status>
<wp:post_name>${title}</wp:post_name>
<wp:status>publish</wp:status>
<wp:post_parent>0</wp:post_parent>
<wp:menu_order>0</wp:menu_order>
<wp:post_type>post</wp:post_type>
<wp:post_password></wp:post_password>
<wp:is_sticky>0</wp:is_sticky>
</item>
\ No newline at end of file
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment