*) adding function to query the pause/resume state of the crawling queues

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@2958 6c8d7289-2bf4-0310-a012-ef5d649a1542
2024-09-19 00:01:41 +02:00 · 2006-11-13 06:25:15 +00:00 · 2006-11-13 06:25:15 +00:00 · 203f2bde9a
commit 203f2bde9a
parent 6d3a130878
3 changed files with 175 additions and 39 deletions
--- a/source/de/anomic/soap/services/CrawlService.java
+++ b/source/de/anomic/soap/services/CrawlService.java
@ -45,8 +45,11 @@

 package de.anomic.soap.services;

+import javax.xml.parsers.ParserConfigurationException;
+
 import org.apache.axis.AxisFault;
 import org.w3c.dom.Document;
+import org.w3c.dom.Element;

 import de.anomic.plasma.plasmaSwitchboard;
 import de.anomic.server.serverObjects;
@ -54,7 +57,13 @@ import de.anomic.soap.AbstractService;

 public class CrawlService extends AbstractService {
    
-    /**
+    private static final String GLOBALCRAWLTRIGGER = "globalcrawltrigger";
+	private static final String REMOTETRIGGEREDCRAWL = "remotetriggeredcrawl";
+	private static final String LOCAL_CRAWL = "localCrawl";
+	private static final String CRAWL_STATE = "crawlState";
+	
+	
+	/**
     * Constant: template for crawling
     */    
    private static final String TEMPLATE_CRAWLING = "QuickCrawlLink_p.xml";    
@ -168,5 +177,50 @@ public class CrawlService extends AbstractService {
            }
        }        
    }
+    
+    /**
+     * Function to query the current state of the following crawling queues:
+     * <ul>
+     * 	<li>local crawl jobs</li>
+     * 	<li>remote crawl jobs</li>
+     * 	<li>of remote crawl job triggers</li>
+     * </ul>
+     * @return returns a XML document in the following format
+     * <pre>
+     * &lt;?xml version="1.0" encoding="UTF-8"?&gt;
+     * &lt;crawlState&gt;
+     * 	&lt;localCrawl&gt;true&lt;/localCrawl&gt;
+     * 	&lt;remotetriggeredcrawl&gt;false&lt;/remotetriggeredcrawl&gt;
+     * 	&lt;globalcrawltrigger&gt;false&lt;/globalcrawltrigger&gt;
+     * &lt;/crawlState&gt;
+     * </pre>
+     * @throws AxisFault if authentication failed
+     * @throws ParserConfigurationException if xml generation failed
+     */
+    public Document getCrawlPauseResumeState() throws AxisFault, ParserConfigurationException {
+    	
+        // extracting the message context
+        extractMessageContext(AUTHENTICATION_NEEDED);             
+    	plasmaSwitchboard sb = (plasmaSwitchboard)this.switchboard;
+        
+        // creating XML document
+        Element xmlElement = null;
+    	Document xmlDoc = createNewXMLDocument(CRAWL_STATE);
+    	Element xmlRoot = xmlDoc.getDocumentElement();        
+    	
+    	xmlElement = xmlDoc.createElement(LOCAL_CRAWL);
+    	xmlElement.appendChild(xmlDoc.createTextNode(Boolean.toString(sb.crawlJobIsPaused(plasmaSwitchboard.CRAWLJOB_LOCAL_CRAWL))));
+    	xmlRoot.appendChild(xmlElement);    	
+    	
+    	xmlElement = xmlDoc.createElement(REMOTETRIGGEREDCRAWL);
+    	xmlElement.appendChild(xmlDoc.createTextNode(Boolean.toString(sb.crawlJobIsPaused(plasmaSwitchboard.CRAWLJOB_REMOTE_TRIGGERED_CRAWL))));
+    	xmlRoot.appendChild(xmlElement);       	
+    	
+    	xmlElement = xmlDoc.createElement(GLOBALCRAWLTRIGGER);
+    	xmlElement.appendChild(xmlDoc.createTextNode(Boolean.toString(sb.crawlJobIsPaused(plasmaSwitchboard.CRAWLJOB_GLOBAL_CRAWL_TRIGGER))));
+    	xmlRoot.appendChild(xmlElement);         	
+    	
+    	return xmlDoc;
+    }    

 }
--- a/source/de/anomic/soap/services/crawl.wsdl
+++ b/source/de/anomic/soap/services/crawl.wsdl
@ -1,48 +1,65 @@
 <?xml version="1.0" encoding="UTF-8"?>
-<wsdl:definitions targetNamespace="http://yacy:8080/soap/crawl" xmlns:apachesoap="http://xml.apache.org/xml-soap" xmlns:impl="http://yacy:8080/soap/crawl" xmlns:intf="http://yacy:8080/soap/crawl" xmlns:soapenc="http://schemas.xmlsoap.org/soap/encoding/" xmlns:wsdl="http://schemas.xmlsoap.org/wsdl/" xmlns:wsdlsoap="http://schemas.xmlsoap.org/wsdl/soap/" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><!--WSDL created by Apache Axis version: 1.2RC2
-Built on Nov 16, 2004 (12:19:44 EST)-->
-  <wsdl:message name="pauseCrawlingResponse">
-  </wsdl:message>
+<wsdl:definitions targetNamespace="http://yacy:8080/soap/crawl" xmlns:apachesoap="http://xml.apache.org/xml-soap" xmlns:impl="http://yacy:8080/soap/crawl" xmlns:intf="http://yacy:8080/soap/crawl" xmlns:soapenc="http://schemas.xmlsoap.org/soap/encoding/" xmlns:wsdl="http://schemas.xmlsoap.org/wsdl/" xmlns:wsdlsoap="http://schemas.xmlsoap.org/wsdl/soap/" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><!--WSDL created by Apache Axis version: 1.4
+Built on Apr 22, 2006 (06:55:48 PDT)-->
  <wsdl:message name="pauseResumeCrawlingResponse">
  </wsdl:message>
+  <wsdl:message name="crawlSingleUrlRequest">
+    <wsdl:part name="crawlingURL" type="xsd:string"/>
+  </wsdl:message>
+  <wsdl:message name="createNewXMLDocumentResponse">
+    <wsdl:part name="createNewXMLDocumentReturn" type="apachesoap:Document"/>
+  </wsdl:message>
+
+  <wsdl:message name="resumeCrawlingResponse">
+  </wsdl:message>
+  <wsdl:message name="pauseResumeCrawlingRequest">
+    <wsdl:part name="localCrawl" type="xsd:boolean"/>
+    <wsdl:part name="remoteTriggeredCrawl" type="xsd:boolean"/>
+    <wsdl:part name="globalCrawlTrigger" type="xsd:boolean"/>
+  </wsdl:message>
+  <wsdl:message name="crawlingResponse">
+    <wsdl:part name="crawlingReturn" type="apachesoap:Document"/>
+
+  </wsdl:message>
+  <wsdl:message name="getCrawlPauseResumeStateRequest">
+  </wsdl:message>
+  <wsdl:message name="getCrawlPauseResumeStateResponse">
+    <wsdl:part name="getCrawlPauseResumeStateReturn" type="apachesoap:Document"/>
+  </wsdl:message>
+  <wsdl:message name="pauseCrawlingResponse">
+  </wsdl:message>
  <wsdl:message name="resumeCrawlingRequest">
+
  </wsdl:message>
  <wsdl:message name="crawlSingleUrlResponse">
    <wsdl:part name="crawlSingleUrlReturn" type="apachesoap:Document"/>
  </wsdl:message>
-  <wsdl:message name="crawlSingleUrlRequest">
-    <wsdl:part name="crawlingURL" type="soapenc:string"/>
+  <wsdl:message name="createNewXMLDocumentRequest">
+    <wsdl:part name="rootElementName" type="xsd:string"/>
  </wsdl:message>
  <wsdl:message name="crawlingRequest">
-    <wsdl:part name="crawlingURL" type="soapenc:string"/>
-    <wsdl:part name="crawljobTitel" type="soapenc:string"/>
-    <wsdl:part name="crawlingDepth" type="soapenc:int"/>
-    <wsdl:part name="crawlingFilter" type="soapenc:string"/>
-    <wsdl:part name="localIndexing" type="soapenc:boolean"/>
-    <wsdl:part name="crawlingQ" type="soapenc:boolean"/>
-    <wsdl:part name="storeHTCache" type="soapenc:boolean"/>
-    <wsdl:part name="crawlOrder" type="soapenc:boolean"/>
-    <wsdl:part name="crawlOrderIntention" type="soapenc:string"/>
-    <wsdl:part name="xsstopw" type="soapenc:boolean"/>
-  </wsdl:message>
-  <wsdl:message name="resumeCrawlingResponse">
-  </wsdl:message>
-  <wsdl:message name="pauseResumeCrawlingRequest">
-    <wsdl:part name="localCrawl" type="soapenc:boolean"/>
-    <wsdl:part name="remoteTriggeredCrawl" type="soapenc:boolean"/>
-    <wsdl:part name="globalCrawlTrigger" type="soapenc:boolean"/>
+    <wsdl:part name="crawlingURL" type="xsd:string"/>
+
+    <wsdl:part name="crawljobTitel" type="xsd:string"/>
+    <wsdl:part name="crawlingDepth" type="xsd:int"/>
+    <wsdl:part name="crawlingFilter" type="xsd:string"/>
+    <wsdl:part name="localIndexing" type="xsd:boolean"/>
+    <wsdl:part name="crawlingQ" type="xsd:boolean"/>
+    <wsdl:part name="storeHTCache" type="xsd:boolean"/>
+    <wsdl:part name="crawlOrder" type="xsd:boolean"/>
+    <wsdl:part name="crawlOrderIntention" type="xsd:string"/>
+    <wsdl:part name="xsstopw" type="xsd:boolean"/>
+
  </wsdl:message>
  <wsdl:message name="pauseCrawlingRequest">
  </wsdl:message>
-  <wsdl:message name="crawlingResponse">
-    <wsdl:part name="crawlingReturn" type="apachesoap:Document"/>
-  </wsdl:message>
  <wsdl:portType name="CrawlService">
    <wsdl:operation name="crawlSingleUrl" parameterOrder="crawlingURL">
      <wsdl:input message="impl:crawlSingleUrlRequest" name="crawlSingleUrlRequest"/>
      <wsdl:output message="impl:crawlSingleUrlResponse" name="crawlSingleUrlResponse"/>
    </wsdl:operation>
    <wsdl:operation name="crawling" parameterOrder="crawlingURL crawljobTitel crawlingDepth crawlingFilter localIndexing crawlingQ storeHTCache crawlOrder crawlOrderIntention xsstopw">
+
      <wsdl:input message="impl:crawlingRequest" name="crawlingRequest"/>
      <wsdl:output message="impl:crawlingResponse" name="crawlingResponse"/>
    </wsdl:operation>
@ -50,18 +67,29 @@ Built on Nov 16, 2004 (12:19:44 EST)-->
      <wsdl:input message="impl:pauseCrawlingRequest" name="pauseCrawlingRequest"/>
      <wsdl:output message="impl:pauseCrawlingResponse" name="pauseCrawlingResponse"/>
    </wsdl:operation>
+    <wsdl:operation name="pauseResumeCrawling" parameterOrder="localCrawl remoteTriggeredCrawl globalCrawlTrigger">
+      <wsdl:input message="impl:pauseResumeCrawlingRequest" name="pauseResumeCrawlingRequest"/>
+
+      <wsdl:output message="impl:pauseResumeCrawlingResponse" name="pauseResumeCrawlingResponse"/>
+    </wsdl:operation>
    <wsdl:operation name="resumeCrawling">
      <wsdl:input message="impl:resumeCrawlingRequest" name="resumeCrawlingRequest"/>
      <wsdl:output message="impl:resumeCrawlingResponse" name="resumeCrawlingResponse"/>
    </wsdl:operation>
-    <wsdl:operation name="pauseResumeCrawling" parameterOrder="localCrawl remoteTriggeredCrawl globalCrawlTrigger">
-      <wsdl:input message="impl:pauseResumeCrawlingRequest" name="pauseResumeCrawlingRequest"/>
-      <wsdl:output message="impl:pauseResumeCrawlingResponse" name="pauseResumeCrawlingResponse"/>
+    <wsdl:operation name="getCrawlPauseResumeState">
+      <wsdl:input message="impl:getCrawlPauseResumeStateRequest" name="getCrawlPauseResumeStateRequest"/>
+      <wsdl:output message="impl:getCrawlPauseResumeStateResponse" name="getCrawlPauseResumeStateResponse"/>
+
+    </wsdl:operation>
+    <wsdl:operation name="createNewXMLDocument" parameterOrder="rootElementName">
+      <wsdl:input message="impl:createNewXMLDocumentRequest" name="createNewXMLDocumentRequest"/>
+      <wsdl:output message="impl:createNewXMLDocumentResponse" name="createNewXMLDocumentResponse"/>
    </wsdl:operation>
  </wsdl:portType>
  <wsdl:binding name="crawlSoapBinding" type="impl:CrawlService">
    <wsdlsoap:binding style="rpc" transport="http://schemas.xmlsoap.org/soap/http"/>
    <wsdl:operation name="crawlSingleUrl">
+
      <wsdlsoap:operation soapAction=""/>
      <wsdl:input name="crawlSingleUrlRequest">
        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://services.soap.anomic.de" use="encoded"/>
@ -71,6 +99,7 @@ Built on Nov 16, 2004 (12:19:44 EST)-->
      </wsdl:output>
    </wsdl:operation>
    <wsdl:operation name="crawling">
+
      <wsdlsoap:operation soapAction=""/>
      <wsdl:input name="crawlingRequest">
        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://services.soap.anomic.de" use="encoded"/>
@ -80,6 +109,7 @@ Built on Nov 16, 2004 (12:19:44 EST)-->
      </wsdl:output>
    </wsdl:operation>
    <wsdl:operation name="pauseCrawling">
+
      <wsdlsoap:operation soapAction=""/>
      <wsdl:input name="pauseCrawlingRequest">
        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://services.soap.anomic.de" use="encoded"/>
@ -88,16 +118,8 @@ Built on Nov 16, 2004 (12:19:44 EST)-->
        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://yacy:8080/soap/crawl" use="encoded"/>
      </wsdl:output>
    </wsdl:operation>
-    <wsdl:operation name="resumeCrawling">
-      <wsdlsoap:operation soapAction=""/>
-      <wsdl:input name="resumeCrawlingRequest">
-        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://services.soap.anomic.de" use="encoded"/>
-      </wsdl:input>
-      <wsdl:output name="resumeCrawlingResponse">
-        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://yacy:8080/soap/crawl" use="encoded"/>
-      </wsdl:output>
-    </wsdl:operation>
    <wsdl:operation name="pauseResumeCrawling">
+
      <wsdlsoap:operation soapAction=""/>
      <wsdl:input name="pauseResumeCrawlingRequest">
        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://services.soap.anomic.de" use="encoded"/>
@ -106,7 +128,38 @@ Built on Nov 16, 2004 (12:19:44 EST)-->
        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://yacy:8080/soap/crawl" use="encoded"/>
      </wsdl:output>
    </wsdl:operation>
+    <wsdl:operation name="resumeCrawling">
+
+      <wsdlsoap:operation soapAction=""/>
+      <wsdl:input name="resumeCrawlingRequest">
+        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://services.soap.anomic.de" use="encoded"/>
+      </wsdl:input>
+      <wsdl:output name="resumeCrawlingResponse">
+        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://yacy:8080/soap/crawl" use="encoded"/>
+      </wsdl:output>
+    </wsdl:operation>
+    <wsdl:operation name="getCrawlPauseResumeState">
+
+      <wsdlsoap:operation soapAction=""/>
+      <wsdl:input name="getCrawlPauseResumeStateRequest">
+        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://services.soap.anomic.de" use="encoded"/>
+      </wsdl:input>
+      <wsdl:output name="getCrawlPauseResumeStateResponse">
+        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://yacy:8080/soap/crawl" use="encoded"/>
+      </wsdl:output>
+    </wsdl:operation>
+    <wsdl:operation name="createNewXMLDocument">
+
+      <wsdlsoap:operation soapAction=""/>
+      <wsdl:input name="createNewXMLDocumentRequest">
+        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://soap.anomic.de" use="encoded"/>
+      </wsdl:input>
+      <wsdl:output name="createNewXMLDocumentResponse">
+        <wsdlsoap:body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/" namespace="http://yacy:8080/soap/crawl" use="encoded"/>
+      </wsdl:output>
+    </wsdl:operation>
  </wsdl:binding>
+
  <wsdl:service name="CrawlServiceService">
    <wsdl:port binding="impl:crawlSoapBinding" name="crawl">
      <wsdlsoap:address location="http://yacy:8080/soap/crawl"/>
--- a/test/de/anomic/soap/services/CrawlServiceTest.java
+++ b/test/de/anomic/soap/services/CrawlServiceTest.java
@ -0,0 +1,29 @@
+package de.anomic.soap.services;
+
+import java.rmi.RemoteException;
+
+import javax.xml.parsers.ParserConfigurationException;
+import javax.xml.rpc.ServiceException;
+
+import org.apache.axis.AxisFault;
+import org.apache.axis.utils.XMLUtils;
+import org.w3c.dom.Document;
+
+import yacy.soap.crawl.CrawlService;
+import yacy.soap.crawl.CrawlServiceServiceLocator;
+
+public class CrawlServiceTest extends AbstractServiceTest {
+
+	protected void createServiceClass() throws ServiceException {
+		// construct Soap object
+		CrawlServiceServiceLocator locator = new CrawlServiceServiceLocator();
+		locator.setcrawlEndpointAddress(getBaseServiceURL() + "crawl");
+		
+		service = locator.getcrawl();	
+	}
+
+	public void testGetCrawlPauseResumeState() throws RemoteException  {
+		Document xml = ((CrawlService)service).getCrawlPauseResumeState();
+		System.out.println(XMLUtils.DocumentToString(xml));
+	}
+}