parsedhtml.java

一个比较不错的java分页标签,有源代码,开发者 可以学习学习

/*
 * fgiust
 * PATCHED to add support for table rows
 */
package com.meterware.httpunit;

import java.io.IOException;
import java.net.URL;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;

import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.Node;
import org.w3c.dom.NodeList;

import com.meterware.httpunit.scripting.ScriptableDelegate;


/**
 * @author <a href="mailto:[e-mail]">Russell Gold</a>
 * @author <a href="mailto:[e-mail]">Benoit Xhenseval</a>
 */
class ParsedHTML
{

    final static private HTMLElement[] NO_ELEMENTS = new HTMLElement[0];

    final static private String[] TEXT_ELEMENTS = {"p", "h1", "h2", "h3", "h4", "h5", "h6"};

    private Node _rootNode;

    protected URL _baseURL;

    protected FrameSelector _frame;

    protected String _baseTarget;

    protected String _characterSet;

    private WebResponse _response;

    private boolean _updateElements = true;

    private boolean _enableNoScriptNodes;

    /** map of element IDs to elements. * */
    private HashMap _elementsByID = new HashMap();

    /** map of element names to lists of elements. * */
    private HashMap _elementsByName = new HashMap();

    /** map of DOM elements to HTML elements * */
    private HashMap _elements = new HashMap();

    private ArrayList _formsList = new ArrayList();

    private WebForm[] _forms;

    private WebForm _activeForm;

    private ArrayList _imagesList = new ArrayList();

    private WebImage[] _images;

    private ArrayList _linkList = new ArrayList();

    private WebLink[] _links;

    private ArrayList _blocksList = new ArrayList();

    private TextBlock[] _blocks;

    private ArrayList _appletList = new ArrayList();

    private WebApplet[] _applets;

    private ArrayList _tableList = new ArrayList();

    private WebTable[] _tables;

    private ArrayList _frameList = new ArrayList();

    private WebFrame[] _frames;

    ParsedHTML(
        WebResponse response,
        FrameSelector frame,
        URL baseURL,
        String baseTarget,
        Node rootNode,
        String characterSet)
    {
        _response = response;
        _frame = frame;
        _baseURL = baseURL;
        _baseTarget = baseTarget;
        _rootNode = rootNode;
        _characterSet = characterSet;
    }

    /**
     * Returns the forms found in the page in the order in which they appear.
     */
    public WebForm[] getForms()
    {
        if (_forms == null)
        {
            loadElements();
            _forms = (WebForm[]) _formsList.toArray(new WebForm[_formsList.size()]);
        }
        return _forms;
    }

    /**
     * Returns the links found in the page in the order in which they appear.
     */
    public WebLink[] getLinks()
    {
        if (_links == null)
        {
            loadElements();
            _links = (WebLink[]) _linkList.toArray(new WebLink[_linkList.size()]);
        }
        return _links;
    }

    /**
     * Returns a proxy for each applet found embedded in this page.
     */
    public WebApplet[] getApplets()
    {
        if (_applets == null)
        {
            loadElements();
            _applets = (WebApplet[]) _appletList.toArray(new WebApplet[_appletList.size()]);
        }
        return _applets;
    }

    /**
     * Returns the images found in the page in the order in which they appear.
     */
    public WebImage[] getImages()
    {
        if (_images == null)
        {
            loadElements();
            _images = (WebImage[]) _imagesList.toArray(new WebImage[_imagesList.size()]);
        }
        return _images;
    }

    /**
     * Returns the top-level block elements found in the page in the order in which they appear.
     */
    public TextBlock[] getTextBlocks()
    {
        if (_blocks == null)
        {
            loadElements();
            _blocks = (TextBlock[]) _blocksList.toArray(new TextBlock[_blocksList.size()]);
        }
        return _blocks;
    }

    /**
     * Returns the first text block found in the page which matches the specified predicate and value.
     */
    public TextBlock getFirstMatchingTextBlock(HTMLElementPredicate predicate, Object criteria)
    {
        TextBlock[] blocks = getTextBlocks();
        for (int i = 0; i < blocks.length; i++)
        {
            if (predicate.matchesCriteria(blocks[i], criteria))
                return blocks[i];
        }
        return null;
    }

    public TextBlock getNextTextBlock(TextBlock block)
    {
        int index = _blocksList.indexOf(block);
        if (index < 0 || index == _blocksList.size() - 1)
            return null;
        return (TextBlock) _blocksList.get(index + 1);
    }

    /**
     * Returns the top-level tables found in the page in the order in which they appear.
     */
    public WebTable[] getTables()
    {
        if (_tables == null)
        {
            loadElements();
            _tables = (WebTable[]) _tableList.toArray(new WebTable[_tableList.size()]);
        }
        return _tables;
    }

    /**
     * Returns the HTMLElement with the specified ID.
     */
    public HTMLElement getElementWithID(String id)
    {
        return (HTMLElement) getElementWithID(id, HTMLElement.class);
    }

    /**
     * Returns the HTML elements with the specified name.
     */
    public HTMLElement[] getElementsWithName(String name)
    {
        loadElements();
        ArrayList elements = (ArrayList) _elementsByName.get(name);
        return elements == null ? NO_ELEMENTS : (HTMLElement[]) elements.toArray(new HTMLElement[elements.size()]);
    }

    /**
     * Returns the HTML elements with an attribute with the specified name and value.
     */
    public HTMLElement[] getElementsWithAttribute(String name, String value)
    {
        loadElements();
        ArrayList elements = new ArrayList();
        for (Iterator i = _elements.values().iterator(); i.hasNext();)
        {
            HTMLElement element = (HTMLElement) i.next();
            if (value.equals(element.getAttribute(name)))
                elements.add(element);
        }
        return (HTMLElement[]) elements.toArray(new HTMLElement[elements.size()]);
    }

    /**
     * Returns a list of HTML element names contained in this HTML section.
     */
    public String[] getElementNames()
    {
        loadElements();
        return (String[]) _elementsByName.keySet().toArray(new String[_elementsByName.size()]);
    }

    HTMLElement[] getElementsByTagName(Node dom, String name)
    {
        loadElements();
        if (dom instanceof Element)
        {
            return getElementsFromList(((Element) dom).getElementsByTagName(name));
        }
        else
        {
            return getElementsFromList(((Document) dom).getElementsByTagName(name));
        }
    }

    private HTMLElement[] getElementsFromList(NodeList nl)
    {
        HTMLElement[] elements = new HTMLElement[nl.getLength()];
        for (int i = 0; i < elements.length; i++)
        {
            Node node = nl.item(i);
            elements[i] = (HTMLElement) _elements.get(node);
            if (elements[i] == null)
            {
                elements[i] = toDefaultElement((Element) node);
                _elements.put(node, elements[i]);
            }
        }
        return elements;
    }

    /**
     * Returns the form found in the page with the specified ID.
     */
    public WebForm getFormWithID(String id)
    {
        return (WebForm) getElementWithID(id, WebForm.class);
    }

    /**
     * Returns the link found in the page with the specified ID.
     */
    public WebLink getLinkWithID(String id)
    {
        return (WebLink) getElementWithID(id, WebLink.class);

    }

    private Object getElementWithID(String id, final Class klass)
    {
        loadElements();
        return whenCast(_elementsByID.get(id), klass);
    }

    private Object whenCast(Object o, Class klass)
    {
        return klass.isInstance(o) ? o : null;
    }

    /**
     * Returns the first link found in the page matching the specified criteria.
     */
    public WebForm getFirstMatchingForm(HTMLElementPredicate predicate, Object criteria)
    {
        WebForm[] forms = getForms();
        for (int i = 0; i < forms.length; i++)
        {
            if (predicate.matchesCriteria(forms[i], criteria))
                return forms[i];
        }
        return null;
    }

    /**
     * Returns all links found in the page matching the specified criteria.
     */
    public WebForm[] getMatchingForms(HTMLElementPredicate predicate, Object criteria)
    {
        ArrayList matches = new ArrayList();
        WebForm[] forms = getForms();
        for (int i = 0; i < forms.length; i++)
        {
            if (predicate.matchesCriteria(forms[i], criteria))
                matches.add(forms[i]);
        }
        return (WebForm[]) matches.toArray(new WebForm[matches.size()]);
    }

    /**
     * Returns the form found in the page with the specified name.
     */
    public WebForm getFormWithName(String name)
    {
        return getFirstMatchingForm(WebForm.MATCH_NAME, name);
    }

    private void interpretScriptElement(Element element)
    {
        String script = getScript(element);
        if (script != null)
        {
            try
            {
                _updateElements = false;
                String language = NodeUtils.getNodeAttribute(element, "language", null);
                if (!getResponse().getScriptableObject().supportsScript(language))
                    _enableNoScriptNodes = true;
                getResponse().getScriptableObject().runScript(language, script);
            }
            finally
            {
                setRootNode(_rootNode);
            }
        }
    }

    private String getScript(Node scriptNode)
    {
        String scriptLocation = NodeUtils.getNodeAttribute(scriptNode, "src", null);
        if (scriptLocation == null)
        {
            return NodeUtils.asText(scriptNode.getChildNodes());
        }
        else
        {
            try
            {
                return getIncludedScript(scriptLocation);
            }
            catch (IOException e)
            {
                throw new RuntimeException("Error loading included script: " + e);
            }
        }
    }

    /**
     * Returns the contents of an included script, given its src attribute.
     * @param srcAttribute
     * @return the contents of the script.
     * @throws java.io.IOException if there is a problem retrieving the script
     */
    String getIncludedScript(String srcAttribute) throws IOException
    {
        WebRequest req = new GetMethodWebRequest(getBaseURL(), srcAttribute);
        WebWindow window = getResponse().getWindow();
        if (window == null)
            throw new IllegalStateException(
                "Unable to retrieve script included by this response, since it was loaded by getResource(). Use getResponse() instead.");
        return window.getResource(req).getText();
    }

    /**
     * If noscript node content is enabled, returns null - otherwise returns a concealing element.
     */
    private HTMLElement toNoscriptElement(Element element)
    {
        return _enableNoScriptNodes ? null : new NoScriptElement(element);
    }

    static class HtmlElementRecorder
    {

        protected void recordHtmlElement(NodeUtils.PreOrderTraversal pot, Node node, HTMLElement htmlElement)
        {
            if (htmlElement != null)
            {
                addToMaps(pot, node, htmlElement);
                addToLists(pot, htmlElement);
            }
        }

        protected void addToLists(NodeUtils.PreOrderTraversal pot, HTMLElement htmlElement)
        {
            for (Iterator i = pot.getContexts(); i.hasNext();)
            {
                Object o = i.next();
                if (o instanceof ParsedHTML)
                    ((ParsedHTML) o).addToList(htmlElement);
            }
        }

        protected void addToMaps(NodeUtils.PreOrderTraversal pot, Node node, HTMLElement htmlElement)
        {
            for (Iterator i = pot.getContexts(); i.hasNext();)
            {
                Object o = i.next();
                if (o instanceof ParsedHTML)
                    ((ParsedHTML) o).addToMaps(node, htmlElement);
            }
        }

    }

    abstract static class HTMLElementFactory extends HtmlElementRecorder
    {

        abstract HTMLElement toHTMLElement(NodeUtils.PreOrderTraversal pot, ParsedHTML parsedHTML, Element element);

        void recordElement(NodeUtils.PreOrderTraversal pot, Element element, ParsedHTML parsedHTML)
        {
            HTMLElement htmlElement = toHTMLElement(pot, parsedHTML, element);
            recordHtmlElement(pot, element, htmlElement);
        }