From c5cf05640a6ebcda8cd88d5b0415415bbcacc056 Mon Sep 17 00:00:00 2001 From: "yihua.huang" Date: Thu, 1 Aug 2013 22:53:44 +0800 Subject: [PATCH] processor --- .../main/java/us/codecraft/webmagic/selector/CssSelector.java | 2 +- .../src/test/java/us/codecraft/webmagic/annotation/Blog.java | 4 ++-- .../java/us/codecraft/webmagic/annotation/TestFetcher.java | 4 +++- 3 files changed, 6 insertions(+), 4 deletions(-) diff --git a/webmagic-core/src/main/java/us/codecraft/webmagic/selector/CssSelector.java b/webmagic-core/src/main/java/us/codecraft/webmagic/selector/CssSelector.java index 10dfb62..90a9d1d 100644 --- a/webmagic-core/src/main/java/us/codecraft/webmagic/selector/CssSelector.java +++ b/webmagic-core/src/main/java/us/codecraft/webmagic/selector/CssSelector.java @@ -27,7 +27,7 @@ public class CssSelector implements Selector { public String select(String text) { Document doc = Jsoup.parse(text); Elements elements = doc.select(selectorText); - if (CollectionUtils.isNotEmpty(elements)) { + if (CollectionUtils.isEmpty(elements)) { return null; } return elements.get(0).outerHtml(); diff --git a/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/Blog.java b/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/Blog.java index 6c6e88c..391ce10 100644 --- a/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/Blog.java +++ b/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/Blog.java @@ -5,13 +5,13 @@ package us.codecraft.webmagic.annotation; * @date: 13-8-1
* Time: 下午10:18
*/ -@TargetUrl("http://djjchobits.iteye.com/blog/\\d+") +@TargetUrl("http://my.oschina.net/flashsword/blog/*") public class Blog { @Fetcher("//title") private String title; - @Fetcher(value = "div#main",type = Fetcher.Type.Css) + @Fetcher(value = "div.BlogContent",type = Fetcher.Type.Css) private String content; @Override diff --git a/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/TestFetcher.java b/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/TestFetcher.java index 5318703..e97b5cf 100644 --- a/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/TestFetcher.java +++ b/webmagic-core/src/test/java/us/codecraft/webmagic/annotation/TestFetcher.java @@ -1,5 +1,6 @@ package us.codecraft.webmagic.annotation; +import org.junit.Ignore; import org.junit.Test; import us.codecraft.webmagic.Site; import us.codecraft.webmagic.Spider; @@ -11,9 +12,10 @@ import us.codecraft.webmagic.Spider; */ public class TestFetcher { + @Ignore("takes long") @Test public void test() { - Spider.create(ObjectPageProcessor.create(Site.me().addStartUrl("http://djjchobits.iteye.com/blog/569000"), Blog.class)).run(); + Spider.create(ObjectPageProcessor.create(Site.me().addStartUrl("http://my.oschina.net/flashsword/blog/145796"), Blog.class)).run(); }