invite jsoup and cssselector

webmagic-core/pom.xml

@@ -52,6 +52,12 @@
             <version>2.4</version>
         </dependency>
 
+        <dependency>
+            <groupId>org.jsoup</groupId>
+            <artifactId>jsoup</artifactId>
+            <version>1.7.2</version>
+        </dependency>
+
         <dependency>
             <groupId>org.apache.commons</groupId>
             <artifactId>commons-io</artifactId>

webmagic-core/src/main/java/us/codecraft/webmagic/Spider.java

@@ -154,9 +154,11 @@ public class Spider implements Runnable, Task {
                 request = scheduler.poll(this);
             }
         } else {
+            //multi thread
             final AtomicInteger threadAlive = new AtomicInteger(0);
             while (true) {
                 if (request == null) {
+                    //when no request found but some thread is alive, sleep a while.
                     try {
                         Thread.sleep(100);
                     } catch (InterruptedException e) {

webmagic-core/src/main/java/us/codecraft/webmagic/selector/CssSelector.java

@@ -0,0 +1,47 @@
+package us.codecraft.webmagic.selector;
+
+import org.apache.commons.collections.CollectionUtils;
+import org.jsoup.Jsoup;
+import org.jsoup.nodes.Document;
+import org.jsoup.nodes.Element;
+import org.jsoup.select.Elements;
+
+import java.util.ArrayList;
+import java.util.List;
+
+/**
+ * @author code4crafter@gmail.com <br>
+ *         Date: 13-4-21
+ *         Time: 上午9:39
+ */
+public class CssSelector implements Selector {
+
+    private String selectorText;
+
+    public CssSelector(String selectorText) {
+        this.selectorText = selectorText;
+    }
+
+    @Override
+    public String select(String text) {
+        Document doc = Jsoup.parse(text);
+        Elements elements = doc.select(selectorText);
+        if (CollectionUtils.isNotEmpty(elements)) {
+            return null;
+        }
+        return elements.get(0).outerHtml();
+    }
+
+    @Override
+    public List<String> selectList(String text) {
+        List<String> strings = new ArrayList<String>();
+        Document doc = Jsoup.parse(text);
+        Elements elements = doc.select(selectorText);
+        if (CollectionUtils.isNotEmpty(elements)) {
+            for (Element element : elements) {
+                strings.add(element.outerHtml());
+            }
+        }
+        return strings;
+    }
+}

webmagic-core/src/main/java/us/codecraft/webmagic/selector/Html.java

@@ -62,4 +62,10 @@ public class Html extends PlainText {
         return selectList(xpathSelector, strings);
     }
 
+    @Override
+    public Selectable $(String selector) {
+        CssSelector cssSelector = new CssSelector(selector);
+        return selectList(cssSelector,strings);
+    }
+
 }

webmagic-core/src/main/java/us/codecraft/webmagic/selector/PlainText.java

@@ -33,6 +33,11 @@ public class PlainText implements Selectable {
         throw new UnsupportedOperationException();
     }
 
+    @Override
+    public Selectable $(String selector) {
+        throw new UnsupportedOperationException();
+    }
+
     @Override
     public Selectable smartContent() {
         throw new UnsupportedOperationException();

webmagic-core/src/main/java/us/codecraft/webmagic/selector/Selectable.java

@@ -17,6 +17,14 @@ public interface Selectable {
      */
     public Selectable xpath(String xpath);
 
+    /**
+     * select list with jquery selector
+     *
+     * @param
+     * @return
+     */
+    public Selectable $(String selector);
+
     /**
      * select smart content with ReadAbility algorithm
      *

webmagic-core/src/test/java/us/codecraft/webmagic/HtmlTest.java

@@ -17,4 +17,5 @@ public class HtmlTest {
         Assert.assertEquals("abbabbab", (selectable.regex("(.*)").replace("aa(a)", "$1bb").toString()));
 
     }
+
 }

webmagic-core/src/test/java/us/codecraft/webmagic/selector/XpathSelectorTest.java

@@ -1351,7 +1351,7 @@ public class XpathSelectorTest {
     public void testOschina() {
         Html html1 = new Html(html);
         Assert.assertEquals("再次吐槽easyui", html1.xpath(".//*[@class='QTitle']/h1/a").toString());
-        System.out.println(html1.regex("(<body>.*?</body>)").links().toStrings());
+        Assert.assertNotNull(html1.$("a[href]").xpath("//@href").toStrings());
     }
 
 }