ktlili 2005/07/11 17:32:26 CEST
Modified files:
war/src/conf en_US.properties
war/src/java/com/jahia/clipping/struts SelectPartAction.java
SelectPartForm.java
war/src/java/com/jahia/clipping/util DomUtilities.java
war/src/java/com/jahia/clipping/web/html/Impl/ExtractorFilter
AbsctractExtractoreFilter.java
war/src/java/com/jahia/clipping/web/http/impl
HTMLUnitProcessor.java
war/src/webapp/jsp/actions selectPartAction.jsp
Added files:
war/src/java/com/jahia/clipping/util Constants.java
war/src/java/com/jahia/clipping/web/html/Impl/ExtractorFilter
XPathExtractorFilter.java
war/src/java/com/jahia/clipping/web/http/impl
SimpleHTMLUnitProcessor.java
war/src/webapp/jsp/templates webBrowser.jsp
war/src/webapp/jsp testHeader.jsp
Log:
- add XPath based filter
Revision Changes Path
1.19 +2 -0 webclip_builder/war/src/conf/en_US.properties
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/conf/en_US.properties.diff?r1=1.18&r2=1.19&f=h
1.14 +24 -15
webclip_builder/war/src/java/com/jahia/clipping/struts/SelectPartAction.java
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/struts/SelectPartAction.java.diff?r1=1.13&r2=1.14&f=h
1.4 +21 -1
webclip_builder/war/src/java/com/jahia/clipping/struts/SelectPartForm.java
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/struts/SelectPartForm.java.diff?r1=1.3&r2=1.4&f=h
1.1 +101 -0
webclip_builder/war/src/java/com/jahia/clipping/util/Constants.java (new)
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/util/Constants.java?rev=1.1&content-type=text/plain
1.7 +4 -3
webclip_builder/war/src/java/com/jahia/clipping/util/DomUtilities.java
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/util/DomUtilities.java.diff?r1=1.6&r2=1.7&f=h
1.3 +21 -0
webclip_builder/war/src/java/com/jahia/clipping/web/html/Impl/ExtractorFilter/AbsctractExtractoreFilter.java
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/web/html/Impl/ExtractorFilter/AbsctractExtractoreFilter.java.diff?r1=1.2&r2=1.3&f=h
1.1 +113 -0
webclip_builder/war/src/java/com/jahia/clipping/web/html/Impl/ExtractorFilter/XPathExtractorFilter.java
(new)
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/web/html/Impl/ExtractorFilter/XPathExtractorFilter.java?rev=1.1&content-type=text/plain
1.16 +25 -23
webclip_builder/war/src/java/com/jahia/clipping/web/http/impl/HTMLUnitProcessor.java
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/web/http/impl/HTMLUnitProcessor.java.diff?r1=1.15&r2=1.16&f=h
1.1 +548 -0
webclip_builder/war/src/java/com/jahia/clipping/web/http/impl/SimpleHTMLUnitProcessor.java
(new)
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/java/com/jahia/clipping/web/http/impl/SimpleHTMLUnitProcessor.java?rev=1.1&content-type=text/plain
1.6 +8 -1
webclip_builder/war/src/webapp/jsp/actions/selectPartAction.jsp
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/webapp/jsp/actions/selectPartAction.jsp.diff?r1=1.5&r2=1.6&f=h
1.1 +8 -0
webclip_builder/war/src/webapp/jsp/templates/webBrowser.jsp (new)
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/webapp/jsp/templates/webBrowser.jsp?rev=1.1&content-type=text/plain
1.1 +42 -0 webclip_builder/war/src/webapp/jsp/testHeader.jsp (new)
http://jahia.mine.nu:8080/cgi-bin/cvsweb.cgi/webclip_builder/war/src/webapp/jsp/testHeader.jsp?rev=1.1&content-type=text/plain
Index: en_US.properties
===================================================================
RCS file: /home/cvs/repository/webclip_builder/war/src/conf/en_US.properties,v
retrieving revision 1.18
retrieving revision 1.19
diff -u -r1.18 -r1.19
--- en_US.properties 11 Jul 2005 12:31:09 -0000 1.18
+++ en_US.properties 11 Jul 2005 15:32:24 -0000 1.19
@@ -151,6 +151,7 @@
select.clippingMethod = Clipping method
select.typeContent = Type of content
select.tagName = Tag name
+select.xpath.enter = Enter a valid xpath
select.typeContent.dynamic = Dynamic
select.typeContent.static = Static
select.clippingMethod.manual = Manual
@@ -158,6 +159,7 @@
select.clippingMethod.rules = Rules
select.clippingMethod.form = Form extractor
select.clippingMethod.link = Link ectarctor
+select.clippingMethod.xPath = XPath
select.clippingMethod.table = Table extrator
select.clippingMethod.filter = Filter
select.changeClippingMethod = changeClippingMethod
Index: SelectPartAction.java
===================================================================
RCS file:
/home/cvs/repository/webclip_builder/war/src/java/com/jahia/clipping/struts/SelectPartAction.java,v
retrieving revision 1.13
retrieving revision 1.14
diff -u -r1.13 -r1.14
--- SelectPartAction.java 11 Jul 2005 12:31:10 -0000 1.13
+++ SelectPartAction.java 11 Jul 2005 15:32:24 -0000 1.14
@@ -14,7 +14,6 @@
import com.jahia.clipping.web.html.Impl.HTMLParser.*;
import com.jahia.clipping.web.html.Impl.ExtractorFilter.*;
-
/**
* Description of the Class
*
@@ -53,13 +52,23 @@
return map;
}
- public ActionForward view(ActionMapping actionMapping, ActionForm
actionForm, HttpServletRequest httpServletRequest, HttpServletResponse
httpServletResponse) {
- logger.debug("[ View ]");
- AbstractWizardForm form = (AbstractWizardForm) actionForm;
-
getWebBrowserForm(httpServletRequest).setShow(Constants.WEB_BROWSER_SHOW_LAST_DOCUMENT);
- return actionMapping.getInputForward();
- }
+ /**
+ * Description of the Method
+ *
+ [EMAIL PROTECTED] actionMapping Description of Parameter
+ [EMAIL PROTECTED] actionForm Description of Parameter
+ [EMAIL PROTECTED] httpServletRequest Description of Parameter
+ [EMAIL PROTECTED] httpServletResponse Description of Parameter
+ [EMAIL PROTECTED] Description of the Returned
Value
+ */
+ public ActionForward view(ActionMapping actionMapping, ActionForm
actionForm, HttpServletRequest httpServletRequest, HttpServletResponse
httpServletResponse) {
+ logger.debug("[ View ]");
+ AbstractWizardForm form = (AbstractWizardForm) actionForm;
+
getWebBrowserForm(httpServletRequest).setShow(Constants.WEB_BROWSER_SHOW_LAST_DOCUMENT);
+
+ return actionMapping.getInputForward();
+ }
@@ -112,7 +121,9 @@
}
//table
else if (typeFilter.equalsIgnoreCase("4")) {
- filter = new SimpleExtractorFilter(selectedContent);
+ filter = new XPathExtractorFilter();
+ String xPath = selectPartForm.getWebClippingXPath();
+ ((XPathExtractorFilter) filter).setXPahKey(xPath);
}
//link
else {
@@ -120,8 +131,8 @@
}
logger.debug("[ Selected content is " + selectedContent + " ]");
- // int mode = Integer.parseInt(typeContent);
- // filter.setMode(mode);
+ // int mode = Integer.parseInt(typeContent);
+ // filter.setMode(mode);
bean.setFilter(filter);
return actionMapping.findForward("preview");
@@ -150,7 +161,6 @@
ClipperBean bean = getClipperBean(httpServletRequest);
-
//set the selected part
String selectedContent = "view filter not available";
selectPartForm.setSelectedContent(selectedContent);
@@ -162,9 +172,9 @@
ChewExtractorFilter filter = new ChewExtractorFilter();
//set the key of the filter
String hashKey = hashSelectedPart;
- String tagName = selectPartForm.getWebClippingTagName();
+ String tagName = selectPartForm.getWebClippingTagName();
filter.setHashKey(hashKey);
- filter.setTagNameKey(tagName);
+ filter.setTagNameKey(tagName);
// set the filter
bean.setFilter(filter);
@@ -223,7 +233,7 @@
// get the document
//HTMLDocument htmlDocument =
getHTMLDocument(httpServletRequest, httpServletResponse);
//selectPartForm.setBodyContent(htmlDocument.getTransformedDocumentAsString());
-
getWebBrowserForm(httpServletRequest).setShow(Constants.WEB_BROWSER_SHOW_LAST_DOCUMENT);
+
getWebBrowserForm(httpServletRequest).setShow(Constants.WEB_BROWSER_SHOW_LAST_DOCUMENT);
return actionMapping.getInputForward();
}
@@ -289,7 +299,6 @@
-
/**
* Description of the Method
*
Index: SelectPartForm.java
===================================================================
RCS file:
/home/cvs/repository/webclip_builder/war/src/java/com/jahia/clipping/struts/SelectPartForm.java,v
retrieving revision 1.3
retrieving revision 1.4
diff -u -r1.3 -r1.4
--- SelectPartForm.java 11 Jul 2005 12:31:10 -0000 1.3
+++ SelectPartForm.java 11 Jul 2005 15:32:24 -0000 1.4
@@ -20,6 +20,7 @@
private String webClippingTypeContent;
private String webClippingPartHashCode;
private String webClippingTagName;
+ private String webClippingXPath;
/**
@@ -94,6 +95,16 @@
/**
+ * Sets the WebClippingXPath attribute of the SelectPartForm object
+ *
+ [EMAIL PROTECTED] webClippingXPath The new WebClippingXPath value
+ */
+ public void setWebClippingXPath(String webClippingXPath) {
+ this.webClippingXPath = webClippingXPath;
+ }
+
+
+ /**
* Gets the Id attribute of the SelectPartForm object
*
[EMAIL PROTECTED] The Id value
@@ -185,6 +196,16 @@
/**
+ * Gets the WebClippingXPath attribute of the SelectPartForm object
+ *
+ [EMAIL PROTECTED] The WebClippingXPath value
+ */
+ public String getWebClippingXPath() {
+ return webClippingXPath;
+ }
+
+
+ /**
* Description of the Method
*
[EMAIL PROTECTED] actionMapping Description of Parameter
@@ -218,5 +239,4 @@
}
-
}
Index: Constants.java
====================================================================
package com.jahia.clipping.util;
public class Constants {
public static final String CLIPPER_BEAN = "clipperBean";
public static final String URL_MAP = "urlMap";
public static final String FIRST_URL_HASH = "userHash";
public static final String FIRST_URL_FROM = "userFrom";
//encoding
public static String ENCODING_ISO = "ISO-8859-1";
public static String ENCODING_UTF = "UTF-8";
// Attribut identifier
public static String TARGET_URL = "targetUrl";
public static String RECORDING = "recording";
public static String HTTPSTATE = "webClipping_httpState";
public static String SIMPLE_WEBCLIENT = "webClipping_webClient";
public static String ADVANCED_WEBCLIENT = "webClipping_webClient";
public static String LAST_URL = "webClipping_lastUrl";
public static String USER = "user";
public static Boolean TRUE = Boolean.TRUE;
public static Boolean FALSE = Boolean.FALSE;
public static String JAVASCRIPT_LOCATION = "location";
public static String[] JAVASCRIPT_EVENT_NAMES_ARRAY = {
"onAbort",
// Executes JavaScript code when the user aborts the loading of
an image.
"onBlur",
// Executes JavaScript code when a form element loses focus or
when a window or frame loses focus.
"onChange",
// Executes JavaScript code when a Select, Text, or Textarea
field loses focus and its value has been modified
"onClick",
// Executes JavaScript code when an object on a form is clicked.
"onDblClick",
// Executes JavaScript code when the user double-clicks a form
element or a link.
"onDragDrop",
// Executes JavaScript code when the user drops an object onto
the browser window, such as dropping a file.
"onError",
// Executes JavaScript code when the loading of a document or
image causes an error.
"onFocus",
// Executes JavaScript code when a window, frame, or frameset
receives focus or when a form element receives input focus.
"onKeyDown",
// Executes JavaScript code when the user depresses a key.
"onKeyPress",
// Executes JavaScript code when the user presses or holds down a
key.
"onKeyUp",
// Executes JavaScript code when the user releases a key.
"onLoad",
// Executes JavaScript code when the browser finishes loading a
window or all frames within a FRAMESET tag.
"onMouseDown",
// Executes JavaScript code when the user depresses a mouse
button.
"onMouseMove",
// Executes JavaScript code when the user moves the cursor.
"onMouseOut",
// Executes JavaScript code each time the mouse pointer leaves an
area (client-side image map) or link from inside that area or link.
"onMouseOver",
// Executes JavaScript code once each time the mouse pointer
moves over an object or area from outside that object or area.
"onMouseUp",
// Executes JavaScript code when the user releases a mouse button.
"onMove",
// Executes JavaScript code when the user or script moves a
window or frame.
"onReset",
// Executes JavaScript code when a user resets a form (clicks a
Reset button).
"onResize",
// Executes JavaScript code when a user or script resizes a
window or frame.
"onSelect",
// Executes JavaScript code when a user selects some of the text
within a text or textarea field.
"onSubmit",
// Executes JavaScript code when a user submits a form.
"onUnload",
// Executes JavaScript code when the user exits a document.
};
public static String NOT_URL[] = {"javascript:", "mailto:";, "about:"};
public static int URLENCODER_TYPE_LINK = 0;
public static int URLENCODER_TYPE_FORM = 1;
public static int URLENCODER_TYPE_JAVASCRIPT = 2;
public static int URLENCODER_TYPE_METALINK = 3;
//Value for from
public static String FROM_LINK = "link";
public static String FROM_FORM = "submittedForm";
public static String FROM_META = "metaLink";
//possible value of type atrtibute of input element
public static String NO_VALUE_FOUNDED = "no_value_Found";
public static String TYPE_TEXT = "text";
public static String TYPE_BUTTON = "text";
public static String TYPE_PASSWORD = "password";
public static String TYPE_CHECKBOX = "checkbox";
public static String TYPE_RADIO = "radio";
public static String TYPE_SUBMIT = "submit";
public static String TYPE_RESET = "reset";
public static String TYPE_FILE = "file";
public static String TYPE_HIDDEN = "hidden";
public static String TYPE_IMAGE = "image";
public static String TYPE_SELECT = "select";
}
Index: DomUtilities.java
===================================================================
RCS file:
/home/cvs/repository/webclip_builder/war/src/java/com/jahia/clipping/util/DomUtilities.java,v
retrieving revision 1.6
retrieving revision 1.7
diff -u -r1.6 -r1.7
--- DomUtilities.java 28 Jun 2005 15:42:30 -0000 1.6
+++ DomUtilities.java 11 Jul 2005 15:32:24 -0000 1.7
@@ -1,14 +1,15 @@
package com.jahia.clipping.util;
+import java.io.*;
+
+import javax.xml.parsers.*;
import javax.xml.transform.*;
import javax.xml.transform.dom.*;
import javax.xml.transform.stream.*;
import javax.xml.xpath.*;
-import org.w3c.dom.*;
-import javax.xml.parsers.*;
+
import org.w3c.dom.*;
import org.xml.sax.*;
-import java.io.*;
/**
* Utilities for Dom
Index: XPathExtractorFilter.java
====================================================================
package com.jahia.clipping.web.html.Impl.ExtractorFilter;
import java.io.*;
import com.jahia.clipping.web.html.*;
import com.jahia.clipping.web.html.Impl.*;
import org.xml.sax.*;
import com.jahia.clipping.util.*;
import javax.xml.parsers.*;
import javax.xml.transform.*;
import javax.xml.transform.dom.*;
import javax.xml.transform.stream.*;
import javax.xml.xpath.*;
import org.apache.xerces.dom.*;
import org.w3c.dom.*;
import org.xml.sax.*;
/**
* Description of the Class
*
[EMAIL PROTECTED] Tlili Khaled
*/
public class XPathExtractorFilter extends AbsctractExtractoreFilter {
private int mode;
private static final String XPATH_KEY = "xPathKey";
private static org.apache.log4j.Logger logger =
org.apache.log4j.Logger.getLogger(XPathExtractorFilter.class);
/**
* Sets the Mode attribute of the XPathExtractorFilter object
*
[EMAIL PROTECTED] mode The new Mode value
*/
public void setMode(int mode) {
this.mode = mode;
}
/**
* Sets the XPahKey attribute of the XPathExtractorFilter object
*
[EMAIL PROTECTED] xPathKey The new XPahKey value
*/
public void setXPahKey(String xPathKey) {
super.addKeyPart(XPATH_KEY, xPathKey);
}
/**
* Gets the Mode attribute of the XPathExtractorFilter object
*
[EMAIL PROTECTED] The Mode value
*/
public int getMode() {
return mode;
}
/**
* Gets the XPahKey attribute of the XPathExtractorFilter object
*
[EMAIL PROTECTED] The XPahKey value
*/
public String getXPahKey() {
return super.getKeyPart(XPATH_KEY);
}
/**
* Gets the SelectedPart attribute of the XPathExtractorFilter object
*
[EMAIL PROTECTED] doc Description of Parameter
[EMAIL PROTECTED] The SelectedPart value
*/
public String getSelectedPart(HTMLDocument doc) {
String html = doc.getUserDocumentAsString();
try {
// Get default parser
HTMLParser parser = new DefaultHTMLParser();
// build the transformed document
org.w3c.dom.Document w3cDoc = parser.parse(html);
if (w3cDoc == null) {
logger.error("[Parsing failed !!!!]");
}
else {
logger.debug("[Parsing finish whith succes
!!!!]");
}
NodeList list = DomUtilities.getNodeListByXPath(w3cDoc,
this.getXPahKey());
Document selectedPartDoc = new
org.apache.xerces.dom.DocumentImpl();
Element root = selectedPartDoc.createElement("div");
for (int i = 0; i < list.getLength(); i++) {
Node node = list.item(i).cloneNode(true);
Node clone = selectedPartDoc.importNode(node,
true);
root.appendChild(clone);
}
return
DomUtilities.getDocumentAsString(selectedPartDoc);
}
catch (SAXException ex) {
ex.printStackTrace();
return "Error has occured: " + this.getXPahKey();
}
catch (IOException ex) {
ex.printStackTrace();
return "Error has occured: " + this.getXPahKey();
}
}
}
Index: AbsctractExtractoreFilter.java
===================================================================
RCS file:
/home/cvs/repository/webclip_builder/war/src/java/com/jahia/clipping/web/html/Impl/ExtractorFilter/AbsctractExtractoreFilter.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- AbsctractExtractoreFilter.java 11 Jul 2005 12:31:12 -0000 1.2
+++ AbsctractExtractoreFilter.java 11 Jul 2005 15:32:25 -0000 1.3
@@ -10,6 +10,17 @@
*/
public abstract class AbsctractExtractoreFilter implements ExtractorFilter {
private Map keyMap = new Hashtable();
+ private String name;
+
+
+ /**
+ * Sets the Name attribute of the AbsctractExtractoreFilter object
+ *
+ [EMAIL PROTECTED] name The new Name value
+ */
+ public void setName(String name) {
+ this.name = name;
+ }
/**
@@ -34,6 +45,16 @@
/**
+ * Gets the Name attribute of the AbsctractExtractoreFilter object
+ *
+ [EMAIL PROTECTED] The Name value
+ */
+ public String getName() {
+ return name;
+ }
+
+
+ /**
* Adds a feature to the Key attribute of the AbsctractExtractoreFilter
* object
*
Index: SimpleHTMLUnitProcessor.java
====================================================================
package com.jahia.clipping.web.http.impl;
import java.io.*;
import java.net.*;
import java.util.*;
import com.gargoylesoftware.htmlunit.*;
import com.gargoylesoftware.htmlunit.html.*;
import com.jahia.clipping.Bean.*;
import com.jahia.clipping.util.*;
import com.jahia.clipping.web.http.*;
import org.apache.commons.httpclient.*;
/**
* Description of the Class
*
[EMAIL PROTECTED] Tlili Khaled
*/
public class SimpleHTMLUnitProcessor implements HTMLClient {
// Url to process
/**
* Description of the Field
*/
protected UrlBean urlBean;
/**
* Description of the Field
*/
protected boolean javascriptEnable;
//Proxy
/**
* Description of the Field
*/
protected String proxyHost;
/**
* Description of the Field
*/
protected int proxyPort = 0;
// Client to use
/**
* Description of the Field
*/
protected WebClient webClient = null;
/**
* Description of the Field
*/
protected RefreshHandler refreshHandler_;
/**
* Description of the Field
*/
protected Header[] headers;
// Resulting page;
/**
* Description of the Field
*/
protected HtmlPage currentPage_;
//List of name of the frame includind in the page
/**
* Description of the Field
*/
protected List allNameFrame = new ArrayList();
//Method to call: Get, post,...
/**
* Description of the Field
*/
protected SubmitMethod method = null;
// resultiing htmlDocument
/**
* Description of the Field
*/
protected String finalUrl;
/**
* Description of the Field
*/
protected String htmlAsString;
/**
* Description of the Field
*/
protected String responseCharSet;
/**
* Description of the Field
*/
protected boolean errorHasOccured = false;
/**
* Description of the Field
*/
protected static org.apache.log4j.Logger logger =
org.apache.log4j.Logger.getLogger(SimpleHTMLUnitProcessor.class);
/**
* Constructor for the HTMLUnitProcessor object
*
[EMAIL PROTECTED] headers Description of Parameter
[EMAIL PROTECTED] httpMethod Description of Parameter
[EMAIL PROTECTED] uBean Description of Parameter
*/
public SimpleHTMLUnitProcessor(Header[] headers, String httpMethod,
UrlBean uBean) {
this.setHeaders(headers);
setMethod(httpMethod);
setUrlBean(uBean);
}
/**
* Sets the UrlBean attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] urlBean The new UrlBean value
*/
public void setUrlBean(UrlBean urlBean) {
this.urlBean = urlBean;
}
/**
* Sets the Method attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] method The new Method value
*/
public void setMethod(String method) {
if (method.equalsIgnoreCase("GET")) {
this.method = SubmitMethod.GET;
}
else if (method.equalsIgnoreCase("POST")) {
this.method = SubmitMethod.POST;
}
else {
logger.error("[ Method " + method + " unknown]");
}
}
/**
* Sets the WebClient attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] webClient The new WebClient value
*/
public void setWebClient(Object webClient) {
this.webClient = (WebClient) webClient;
}
/**
* Sets the ProxyPort attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] proxyPort The new ProxyPort value
*/
public void setProxyPort(int proxyPort) {
this.proxyPort = proxyPort;
}
/**
* Sets the Proxy attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] proxy The new Proxy value
*/
public void setProxyHost(String proxy) {
this.proxyHost = proxy;
}
/**
* Sets the HtmlAsString attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] htmlAsString The new HtmlAsString value
*/
public void setHtmlAsString(String htmlAsString) {
this.htmlAsString = htmlAsString;
}
/**
* Sets the FinalUrl attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] finalUrl The new FinalUrl value
*/
public void setFinalUrl(String finalUrl) {
this.finalUrl = finalUrl;
}
/**
* Sets the Header attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] headers The new Headers value
*/
public void setHeaders(Header[] headers) {
this.headers = headers;
}
/**
* Sets the ResponseCharSet attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] responseCharSet The new ResponseCharSet value
*/
public void setResponseCharSet(String responseCharSet) {
this.responseCharSet = responseCharSet;
}
/**
* Gets the UrlBean attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The UrlBean value
*/
public UrlBean getUrlBean() {
return urlBean;
}
/**
* Gets the Method attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The Method value
*/
public SubmitMethod getSubmitMethod() {
return method;
}
/**
* Gets the Method attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The Method value
*/
public String getMethod() {
return method.getName();
}
/**
* Gets the WebClient attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The WebClient value
*/
public WebClient getWebClient() {
return webClient;
}
/**
* Gets the ProxyPort attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The ProxyPort value
*/
public int getProxyPort() {
return proxyPort;
}
/**
* Gets the Proxy attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The Proxy value
*/
public String getProxyHost() {
return proxyHost;
}
/**
* Gets the AllNameFrame attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The AllNameFrame value
*/
public List getAllNameFrame() {
return allNameFrame;
}
/**
* Gets the HtmlAsString attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The HtmlAsString value
*/
public String getHtmlAsString() {
return htmlAsString;
}
/**
* Gets the FinalUrl attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The FinalUrl value
*/
public String getFinalUrl() {
return finalUrl;
}
/**
* Gets the RequestedUrlFromUrlBean attribute of the HTMLUnitProcessor
* object
*
[EMAIL PROTECTED] The RequestedUrlFromUrlBean value
*/
public String getRequestedUrlFromUrlBean() {
return getUrlBean().getRelativeUrlValue();
}
/**
* Gets the JavascriptEnable attribute of the HTMLClient object
*
[EMAIL PROTECTED] The JavascriptEnable value
*/
public boolean isJavascriptEnable() {
return this.javascriptEnable;
}
/**
* Gets the ResponseCharSet attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The ResponseCharSet value
*/
public String getResponseCharSet() {
return responseCharSet;
}
/**
* Gets the Header attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The Header value
*/
public Header[] getHeader() {
return headers;
}
/**
* Adds a feature to the NameFrame attribute of the HTMLUnitProcessor
object
*
[EMAIL PROTECTED] name The feature to be added to the NameFrame
attribute
*/
public void addNameFrame(String name) {
getAllNameFrame().add(name);
}
/**
* Description of the Method
*
[EMAIL PROTECTED] WebClippingException Description of Exception
*/
public void execute() throws WebClippingException {
String html = "";
try {
// Assert a non null webClient
assertNotNullWebClient();
//Get the requested url
URL url = new URL(getRequestedUrlFromUrlBean());
//set cookies
webClient.getWebConnection().getStateForUrl(url);
// add header
//addHeaders();
// Get the paramter
SubmitMethod httpMethod = getSubmitMethod();
ArrayList paramsList =
getNameValuePairListFromURLBean();
// execute
currentPage_ = (HtmlPage) webClient.getPage(url,
httpMethod, paramsList);
// Get the response
WebResponse response = currentPage_.getWebResponse();
String charSet = response.getContentCharSet();
logger.debug("CharSet: " + charSet);
// update url after all redirections
logger.debug("[ URL before process is: " +
getUrlBean().getAbsoluteURL().toExternalForm() + " ]");
setFinalUrl(response.getUrl().toExternalForm());
getUrlBean().setRedirectUrl(URLUtilities.getURL(getFinalUrl()));
logger.debug("[ URL after process is: " +
getUrlBean().getAbsoluteURL().toExternalForm() + " ]");
// set the reponse as string
html = response.getContentAsString();
logger.debug("[ HTML :\n " + html + " ]");
logger.debug("[ Encoding of the response is: " +
response.getContentCharSet() + " ]");
// set the html
setHtmlAsString(html);
}
catch (FailingHttpStatusCodeException ex) {
logger.error("[ FailingHttpStatusCodeException: " +
ex.getMessage() + " ]");
ex.printStackTrace();
throw new WebClippingException("Failing Http Status
Code", ex);
}
catch (MalformedURLException ex) {
logger.error("[ MalformedURLException: " +
ex.getMessage() + " ]");
ex.printStackTrace();
throw new WebClippingException("Malformed URL", ex);
}
catch (IOException ex) {
logger.error("[ IOException occured: " +
ex.getMessage() + " ]");
ex.printStackTrace();
throw new WebClippingException("IOException", ex);
}
catch (Exception ex) {
logger.debug("[ Exception occured: " + ex.getMessage()
+ " ]");
throw new WebClippingException("Exception", ex);
}
}
/**
* Allow to know if there is frame on the taget file.
*
[EMAIL PROTECTED] Description of the Returned Value
*/
public boolean foundFrame() {
HtmlPage htmlPage = (HtmlPage) currentPage_;
List frameList = htmlPage.getFrames();
return !frameList.isEmpty();
}
/**
* Enable/disable javascript processing
*
[EMAIL PROTECTED] enable Description of Parameter
*/
public void enabledJavascript(boolean enable) {
javascriptEnable = enable;
//update the refesh handler of the web client
WebClient webClient = this.getWebClient();
webClient.setJavaScriptEnabled(isJavascriptEnable());
}
/**
* Remove the last recorded url
*/
public void removeLastUrl() {
// do nothing.
}
/**
* Gets the NameValuePairListFromURLBean attribute of the
HTMLUnitProcessor
* object
*
[EMAIL PROTECTED] The NameValuePairListFromURLBean value
*/
private ArrayList getNameValuePairListFromURLBean() {
logger.debug("[ Convert Map To List NameValuePair ]");
// retrieve the query parameters
if (getUrlBean() == null) {
logger.error("[ URL is not set]");
return null;
}
List queryParamBeanList = getUrlBean().getQueryParamBeanList();
//build list
ArrayList listNamePairAndValue = new ArrayList();
for (int i = 0; i < queryParamBeanList.size(); i++) {
QueryParamBean qBean = (QueryParamBean)
queryParamBeanList.get(i);
String name = qBean.getName();
String value = qBean.getDefaultValue();
logger.debug("[ Found Param: " + name + " with value "
+ value + " ]");
NameValuePair paramAndValue = new NameValuePair(name,
value);
listNamePairAndValue.add(paramAndValue);
}
return listNamePairAndValue;
}
/**
* Gets a default WebClient attribute of the HTMLUnitProcessor object
*
[EMAIL PROTECTED] The NewWebClient value
*/
private WebClient getNewWebClient() {
//Get Proxy parameter
String proxyHost = getProxyHost();
int proxyPort = getProxyPort();
// build the webClient
WebClient webClient = null;
if (proxyHost != null && (proxyPort != 0)) {
webClient = new
WebClient(BrowserVersion.INTERNET_EXPLORER_6_0, proxyHost, proxyPort);
}
else {
webClient = new
WebClient(BrowserVersion.INTERNET_EXPLORER_6_0);
}
// disable throws exception when there is javacscript error
webClient.setThrowExceptionOnScriptError(false);
webClient.setThrowExceptionOnFailingStatusCode(false);
// Disable validation. It will be done by the HTMLParser
webClient.setRedirectEnabled(true);
webClient.setJavaScriptEnabled(isJavascriptEnable());
return webClient;
}
/**
* Adds a feature to the Headers attribute of the HTMLUnitProcessor
object
*/
private void addHeaders() {
logger.debug("[ Add headers ]");
Header[] header = this.getHeader();
if (header != null) {
for (int i = 0; i < header.length; i++) {
Header h = header[i];
if (h != null) {
webClient.addRequestHeader(h.getName(),
h.getValue());
}
}
}
}
/**
* Build a defaultClient if webClient is null
*/
private void assertNotNullWebClient() {
if (webClient == null) {
setWebClient(getNewWebClient());
}
}
}
Index: HTMLUnitProcessor.java
===================================================================
RCS file:
/home/cvs/repository/webclip_builder/war/src/java/com/jahia/clipping/web/http/impl/HTMLUnitProcessor.java,v
retrieving revision 1.15
retrieving revision 1.16
diff -u -r1.15 -r1.16
--- HTMLUnitProcessor.java 11 Jul 2005 09:15:00 -0000 1.15
+++ HTMLUnitProcessor.java 11 Jul 2005 15:32:25 -0000 1.16
@@ -86,7 +86,10 @@
}
}
else {
+ logger.debug("Simlpe HtmlUnit client was
executed");
super.execute();
+ addPage(currentPage_);
+ return;
}
// Get the response
@@ -225,31 +228,30 @@
[EMAIL PROTECTED] The SubmittedForm value
[EMAIL PROTECTED] ElementNotFoundException Description of Exception
*/
- private HtmlForm getSubmittedForm() throws ElementNotFoundException {
- String fromHash = getUrlBean().getHash();
- logger.debug("From Hash: " + fromHash);
- HtmlForm form = null;
- String submittedFormName =
HashUtilities.getFormNameFromHash(fromHash);
- String submittedFormId =
HashUtilities.getFormIdFromHash(fromHash);
- // get the submiited form
- if (submittedFormName == null ||
submittedFormName.equalsIgnoreCase("")) {
- if (submittedFormId == null ||
submittedFormId.equalsIgnoreCase("")) {
- List formList = currentPage_.getForms();
- int pos =
Integer.parseInt(HashUtilities.getFormPositionFromHash(fromHash));
- form = (HtmlForm) formList.get(pos);
-
- }
- else {
- form = (HtmlForm)
currentPage_.getHtmlElementById(submittedFormId);
- }
- }
- else {
- form = (HtmlForm)
currentPage_.getFormByName(submittedFormName);
- }
+ private HtmlForm getSubmittedForm() throws ElementNotFoundException {
+ String fromHash = getUrlBean().getHash();
+ logger.debug("From Hash: " + fromHash);
+ HtmlForm form = null;
+ String submittedFormName =
HashUtilities.getFormNameFromHash(fromHash);
+ String submittedFormId =
HashUtilities.getFormIdFromHash(fromHash);
+ // get the submiited form
+ if (submittedFormName == null ||
submittedFormName.equalsIgnoreCase("")) {
+ if (submittedFormId == null ||
submittedFormId.equalsIgnoreCase("")) {
+ List formList = currentPage_.getForms();
+ int pos =
Integer.parseInt(HashUtilities.getFormPositionFromHash(fromHash));
+ form = (HtmlForm) formList.get(pos);
- return form;
- }
+ }
+ else {
+ form = (HtmlForm)
currentPage_.getHtmlElementById(submittedFormId);
+ }
+ }
+ else {
+ form = (HtmlForm)
currentPage_.getFormByName(submittedFormName);
+ }
+ return form;
+ }
Index: selectPartAction.jsp
===================================================================
RCS file:
/home/cvs/repository/webclip_builder/war/src/webapp/jsp/actions/selectPartAction.jsp,v
retrieving revision 1.5
retrieving revision 1.6
diff -u -r1.5 -r1.6
--- selectPartAction.jsp 11 Jul 2005 12:31:13 -0000 1.5
+++ selectPartAction.jsp 11 Jul 2005 15:32:25 -0000 1.6
@@ -105,7 +105,7 @@
<bean:message key="select.clippingMethod.form"/>
</html:option>
<html:option value="4">
- <bean:message key="select.clippingMethod.link"/>
+ <bean:message key="select.clippingMethod.xPath"/>
</html:option>
<html:option value="5">
<bean:message key="select.clippingMethod.table"/>
@@ -136,6 +136,13 @@
</html:select>
</td>
</logic:equal>
+ <logic:equal name="selectPartForm" value="4"
property="clippingMethod">
+ <!-- case: chew-->
+ <td class="topmenugreen bold waBG">
+ <bean:message key="select.xpath.enter"/>
+ <html:text property="webClippingXPath" size="1" value="0"/>
+ </td>
+ </logic:equal>
</tr>
</table>
</html:form>
Index: webBrowser.jsp
====================================================================
<[EMAIL PROTECTED] uri="/WEB-INF/struts-tiles.tld" prefix="tiles"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-nested.tld" prefix="nested"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-logic.tld" prefix="logic"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-template.tld" prefix="template"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-bean.tld" prefix="bean"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-html.tld" prefix="html"%>
<bean:write name="webBrowserForm" property="content" filter="false"/>
Index: testHeader.jsp
====================================================================
<[EMAIL PROTECTED] uri="/WEB-INF/struts-tiles.tld" prefix="tiles"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-nested.tld" prefix="nested"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-logic.tld" prefix="logic"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-template.tld" prefix="template"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-bean.tld" prefix="bean"%>
<[EMAIL PROTECTED] uri="/WEB-INF/struts-html.tld" prefix="html"%>
<table align="center" width="100%">
<!-- Head wizard-->
<tr valign="middle">
<td class="boxtitlecolor1">
<bean:message key="test.title"/>
</td>
<td class="boxtitleborder1" width="10%">
<html:link action="/manageAction.do?webClippingAction=init">
<bean:message key="test.quit"/>
</html:link>
</td>
</tr>
<!-- Wizar action-->
<tr valign="middle">
<td>
<table>
<!-- Show only the step that has been performed -->
<tr>
<!-- Preview -->
<logic:present name="previewForm">
<td>
<html:form action="saveClipper" method="post">
<html:image styleClass="wizard" pageKey="image.preview"/>
<html:hidden property="webClippingAction" value="wizard"/>
</html:form>
</td>
</logic:present>
</tr>
</table>
</td>
</tr>
</table>
<br/>
