HtmlParser使用指南

需要做一个垂直搜索引擎，比较了nekohtml和htmlparser 的功能，尽管nekohtml在容错性、性能等方面的口碑好像比htmlparser好（htmlunit也用的是nekohtml），但感觉nekohtml的测试用例和文档都比htmlparser都少，而且htmlparser基本上能够满足垂直搜索引擎页面处理分析的需求，因此先研究一下htmlparser的使用，有空再研究nekohtml和mozilla html parser的使用。
    html的功能还是官方说得最为清楚，

    HTML Parser is a Java library used to parse HTML in either a linear or nested fashion. Primarily used for transformation or extraction, it features filters, visitors, custom tags and easy to use JavaBeans. It is a fast, robust and well tested package.

    The two fundamental use-cases that are handled by the parser are extraction and transformation (the syntheses use-case, where HTML pages are created from scratch, is better handled by other tools closer to the source of data). While prior versions concentrated on data extraction from web pages, Version 1.4 of the HTMLParser has substantial improvements in the area of transforming web pages, with simplified tag creation and editing, and verbatim toHtml() method output.

    研究的重点还是extraction的使用，有空再研究transformation的使用。

1、htmlparser对html页面处理的数据结构


如图所示，HtmlParser采用了经典的Composite模式，通过RemarkNode、TextNode、TagNode、AbstractNode和Tag来描述HTML页面各元素。

•org.htmlparser.Node：
    Node接口定义了进行树形结构节点操作的各种典型操作方法，包括：

    节点到html文本、text文本的方法：toPlainTextString、toHtml

   典型树形结构遍历的方法：getParent、getChildren、getFirstChild、getLastChild、getPreviousSibling、getNextSibling、getText

    获取节点对应的树形结构结构的顶级节点Page对象方法：getPage

    获取节点起始位置的方法：getStartPosition、getEndPosition

   Visitor方法遍历节点时候方法：accept (NodeVisitor visitor)

    Filter方法：collectInto (NodeList list, NodeFilter filter)

    Object方法：toString、clone

•org.htmlparser.nodes.AbstractNode：
    AbstractNode是形成HTML树形结构抽象基类，实现了Node接口。

    在htmlparser中，Node分成三类：

    RemarkNode：代表Html中的注释

    TagNode：标签节点。

    TextNode：文本节点

    这三类节点都继承AbstractNode。

•org.htmlparser.nodes.TagNode：
    TagNode包含了对HTML处理的核心的各个类，是所有TAG的基类，其中有分为包含其他TAG的复合节点ComositeTag和不包含其他TAG的叶子节点Tag。

    复合节点CompositeTag：   

        AppletTag,BodyTag,Bullet,BulletList,DefinitionList,DefinitionListBullet,Div,FormTag,FrameSetTag,HeadingTag,

        HeadTag,Html,LabelTag,LinkTag,ObjectTag,ParagraphTag,ScriptTag,SelectTag,Span,StyleTag,TableColumn,

       TableHeader,TableRow,TableTag,TextareaTag,TitleTag

    叶子节点TAG：

        BaseHrefTag,DoctypeTag,FrameTag,ImageTag,InputTag,JspTag,MetaTag,ProcessingInstructionTag,

2、htmlparser对html页面处理的算法
主要是如下几种方式

•采用Visitor方式访问Html
try {
    Parser parser = new Parser();
    parser.setURL(”http://www.google.com”);
    parser.setEncoding(parser.getEncoding());
    NodeVisitor visitor = new NodeVisitor() {
        public void visitTag(Tag tag) {
            logger.fatal(”testVisitorAll()  Tag name is :”
                    + tag.getTagName() + ” \n Class is :”
                    + tag.getClass());
        }

    };

    parser.visitAllNodesWith(visitor);
} catch (ParserException e) {
    e.printStackTrace();
}

•采用Filter方式访问html
try {

    NodeFilter filter = new NodeClassFilter(LinkTag.class);
    Parser parser = new Parser();
    parser.setURL(”http://www.google.com”);
    parser.setEncoding(parser.getEncoding());
    NodeList list = parser.extractAllNodesThatMatch(filter);
    for (int i = 0; i < list.size(); i++) {
        LinkTag node = (LinkTag) list.elementAt(i);
        logger.fatal(”testLinkTag() Link is :” + node.extractLink());
    }
} catch (Exception e) {
    e.printStackTrace();
}

•采用org.htmlparser.beans方式
另外htmlparser 还在org.htmlparser.beans中对一些常用的方法进行了封装，以简化操作，例如：

Parser parser = new Parser();

LinkBean linkBean = new LinkBean();
linkBean.setURL(”http://www.google.com”);
URL[] urls = linkBean.getLinks();

for (int i = 0; i < urls.length; i++) {
    URL url = urls[i];
    logger.fatal(”testLinkBean() -url  is :” + url);
}



3、htmlparser关键包结构说明
    htmlparser其实核心代码并不多，好好研究一下其代码，弥补文档不足的问题。同时htmlparser的代码注释和单元测试用例还是很齐全的，也有助于了解htmlparser的用法。




3.1、org.htmlparser


    定义了htmlparser的一些基础类。其中最为重要的是Parser类。

    Parser是htmlparser的最核心的类，其构造函数提供了如下：Parser.createParser (String html, String charset)、 Parser ()、Parser (Lexer lexer, ParserFeedback fb)、Parser (URLConnection connection, ParserFeedback fb)、Parser (String resource, ParserFeedback feedback)、 Parser (String resource)

  各构造函数的具体用法及含义可以查看其代码，很容易理解。

  Parser常用的几个方法：

•  elements获取元素
    Parser parser = new Parser (”http://www.google.com”);
    for (NodeIterator i = parser.elements (); i.hasMoreElements (); )
      processMyNodes (i.nextNode ());

•parse (NodeFilter filter)：通过NodeFilter方式获取
•visitAllNodesWith (NodeVisitor visitor)：通过Nodevisitor方式
•extractAllNodesThatMatch (NodeFilter filter)：通过NodeFilter方式
3.2、org.htmlparser.beans
    对Visitor和Filter的方法进行了封装，定义了针对一些常用html元素操作的bean，简化对常用元素的提取操作。

    包括：FilterBean、HTMLLinkBean、HTMLTextBean、LinkBean、StringBean、BeanyBaby等。

3.3、org.htmlparser.nodes
    定义了基础的node，包括：AbstractNode、RemarkNode、TagNode、TextNode等。

3.4、org.htmlparser.tags
    定义了htmlparser的各种tag。

3.5、org.htmlparser.filters
    定义了htmlparser所提供的各种filter，主要通过extractAllNodesThatMatch (NodeFilter filter)来对html页面指定类型的元素进行过滤，包括：AndFilter、CssSelectorNodeFilter、HasAttributeFilter、HasChildFilter、HasParentFilter、HasSiblingFilter、IsEqualFilter、LinkRegexFilter、LinkStringFilter、NodeClassFilter、NotFilter、OrFilter、RegexFilter、StringFilter、TagNameFilter、XorFilter

3.6、org.htmlparser.visitors
   定义了htmlparser所提供的各种visitor，主要通过visitAllNodesWith (NodeVisitor visitor)来对html页面元素进行遍历，包括：HtmlPage、LinkFindingVisitor、NodeVisitor、ObjectFindingVisitor、StringFindingVisitor、TagFindingVisitor、TextExtractingVisitor、UrlModifyingVisitor



3.7、org.htmlparser.parserapplications
   定义了一些实用的工具，包括LinkExtractor、SiteCapturer、StringExtractor、WikiCapturer，这几个类也可以作为htmlparser使用样例。

3.8、org.htmlparser.tests
   对各种功能的单元测试用例，也可以作为htmlparser使用的样例。



4、htmlparser的使用样例


import java.net.URL;

import junit.framework.TestCase;

import org.apache.log4j.Logger;
import org.htmlparser.Node;
import org.htmlparser.NodeFilter;
import org.htmlparser.Parser;
import org.htmlparser.Tag;
import org.htmlparser.beans.LinkBean;
import org.htmlparser.filters.NodeClassFilter;
import org.htmlparser.filters.OrFilter;
import org.htmlparser.filters.TagNameFilter;
import org.htmlparser.tags.HeadTag;
import org.htmlparser.tags.ImageTag;
import org.htmlparser.tags.InputTag;
import org.htmlparser.tags.LinkTag;
import org.htmlparser.tags.OptionTag;
import org.htmlparser.tags.SelectTag;
import org.htmlparser.tags.TableColumn;
import org.htmlparser.tags.TableRow;
import org.htmlparser.tags.TableTag;
import org.htmlparser.tags.TitleTag;
import org.htmlparser.util.NodeIterator;
import org.htmlparser.util.NodeList;
import org.htmlparser.util.ParserException;
import org.htmlparser.visitors.HtmlPage;
import org.htmlparser.visitors.NodeVisitor;
import org.htmlparser.visitors.ObjectFindingVisitor;

public class ParserTestCase extends TestCase {

    private static final Logger logger = Logger.getLogger(ParserTestCase.class);

    public ParserTestCase(String name) {
        super(name);
    }
    /*
     * 测试ObjectFindVisitor的用法
     */
    public void testImageVisitor() {
        try {
            ImageTag imgLink;
            ObjectFindingVisitor visitor = new ObjectFindingVisitor(
                    ImageTag.class);
            Parser parser = new Parser();
            parser.setURL(”http://www.google.com”);
            parser.setEncoding(parser.getEncoding());
            parser.visitAllNodesWith(visitor);
            Node[] nodes = visitor.getTags();
            for (int i = 0; i < nodes.length; i++) {
                imgLink = (ImageTag) nodes[i];
                logger.fatal(”testImageVisitor() ImageURL = “
                        + imgLink.getImageURL());
                logger.fatal(”testImageVisitor() ImageLocation = “
                        + imgLink.extractImageLocn());
                logger.fatal(”testImageVisitor() SRC = “
                        + imgLink.getAttribute(”SRC”));
            }
        }
        catch (Exception e) {
            e.printStackTrace();
        }
    }
    /*
     * 测试TagNameFilter用法
     */
    public void testNodeFilter() {
        try {
            NodeFilter filter = new TagNameFilter(”IMG”);
            Parser parser = new Parser();
            parser.setURL(”http://www.google.com”);
            parser.setEncoding(parser.getEncoding());
            NodeList list = parser.extractAllNodesThatMatch(filter);
            for (int i = 0; i < list.size(); i++) {
                logger.fatal(”testNodeFilter() ” + list.elementAt(i).toHtml());
            }
        } catch (Exception e) {
            e.printStackTrace();
        }

    }
    /*
     * 测试NodeClassFilter用法
     */
    public void testLinkTag() {
        try {

            NodeFilter filter = new NodeClassFilter(LinkTag.class);
            Parser parser = new Parser();
            parser.setURL(”http://www.google.com”);
            parser.setEncoding(parser.getEncoding());
            NodeList list = parser.extractAllNodesThatMatch(filter);
            for (int i = 0; i < list.size(); i++) {
                LinkTag node = (LinkTag) list.elementAt(i);
                logger.fatal(”testLinkTag() Link is :” + node.extractLink());
            }
        } catch (Exception e) {
            e.printStackTrace();
        }

    }
    /*
     * 测试<link href=” text=’text/css’ rel=’stylesheet’ />用法
     */
    public void testLinkCSS() {
        try {

            Parser parser = new Parser();
            parser
                    .setInputHTML(”<head><title>Link Test</title>”
                            + “<link href=’/test01/css.css’ text=’text/css’ rel=’stylesheet’ />”
                            + “<link href=’/test02/css.css’ text=’text/css’ rel=’stylesheet’ />”
                            + “</head>” + “<body>”);
            parser.setEncoding(parser.getEncoding());
            NodeList nodeList = null;

            for (NodeIterator e = parser.elements(); e.hasMoreNodes();) {
                Node node = e.nextNode();
                logger
                        .fatal(”testLinkCSS()” + node.getText()
                                + node.getClass());

            }
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
    /*
     * 测试OrFilter的用法
     */
    public void testOrFilter() {
        NodeFilter inputFilter = new NodeClassFilter(InputTag.class);
        NodeFilter selectFilter = new NodeClassFilter(SelectTag.class);
        Parser myParser;
        NodeList nodeList = null;

        try {
            Parser parser = new Parser();
            parser
                    .setInputHTML(”<head><title>OrFilter Test</title>”
                            + “<link href=’/test01/css.css’ text=’text/css’ rel=’stylesheet’ />”
                            + “<link href=’/test02/css.css’ text=’text/css’ rel=’stylesheet’ />”
                            + “</head>”
                            + “<body>”
                            + “<input type=’text’ value=’text1′ name=’text1′/>”
                            + “<input type=’text’ value=’text2′ name=’text2′/>”
                            + “<select><option id=’1′>1</option><option id=’2′>2</option><option id=’3′></option></select>”
                            + “<a href=’http://www.yeeach.com’>yeeach.com</a>”
                            + “</body>”);

            parser.setEncoding(parser.getEncoding());
            OrFilter lastFilter = new OrFilter();
            lastFilter.setPredicates(new NodeFilter[] { selectFilter,
                    inputFilter });
            nodeList = parser.parse(lastFilter);
            for (int i = 0; i <= nodeList.size(); i++) {
                if (nodeList.elementAt(i) instanceof InputTag) {
                    InputTag tag = (InputTag) nodeList.elementAt(i);
                    logger.fatal(”OrFilter tag name is :” + tag.getTagName()
                            + ” ,tag value is:” + tag.getAttribute(”value”));
                }
                if (nodeList.elementAt(i) instanceof SelectTag) {
                    SelectTag tag = (SelectTag) nodeList.elementAt(i);
                    NodeList list = tag.getChildren();

                    for (int j = 0; j < list.size(); j++) {
                        OptionTag option = (OptionTag) list.elementAt(j);
                        logger
                                .fatal(”OrFilter Option”
                                        + option.getOptionText());
                    }

                }
            }

        } catch (ParserException e) {
            e.printStackTrace();
        }
    }
    /*
     * 测试对<table><tr><td></td></tr></table>的解析
     */
    public void testTable() {
        Parser myParser;
        NodeList nodeList = null;
        myParser = Parser.createParser(”<body> ” + “<table id=’table1′ >”
                + “<tr><td>1-11</td><td>1-12</td><td>1-13</td>”
                + “<tr><td>1-21</td><td>1-22</td><td>1-23</td>”
                + “<tr><td>1-31</td><td>1-32</td><td>1-33</td></table>”
                + “<table id=’table2′ >”
                + “<tr><td>2-11</td><td>2-12</td><td>2-13</td>”
                + “<tr><td>2-21</td><td>2-22</td><td>2-23</td>”
                + “<tr><td>2-31</td><td>2-32</td><td>2-33</td></table>”
                + “</body>”, “GBK”);
        NodeFilter tableFilter = new NodeClassFilter(TableTag.class);
        OrFilter lastFilter = new OrFilter();
        lastFilter.setPredicates(new NodeFilter[] { tableFilter });
        try {
            nodeList = myParser.parse(lastFilter);
            for (int i = 0; i <= nodeList.size(); i++) {
                if (nodeList.elementAt(i) instanceof TableTag) {
                    TableTag tag = (TableTag) nodeList.elementAt(i);
                    TableRow[] rows = tag.getRows();

                    for (int j = 0; j < rows.length; j++) {
                        TableRow tr = (TableRow) rows[j];
                        TableColumn[] td = tr.getColumns();
                        for (int k = 0; k < td.length; k++) {
                            logger.fatal(”<td>” + td[k].toPlainTextString());
                        }

                    }

                }
            }

        } catch (ParserException e) {
            e.printStackTrace();
        }
    }
    /*
     * 测试NodeVisitor的用法，遍历所有节点
     */
    public void testVisitorAll() {
        try {
            Parser parser = new Parser();
            parser.setURL(”http://www.google.com”);
            parser.setEncoding(parser.getEncoding());
            NodeVisitor visitor = new NodeVisitor() {
                public void visitTag(Tag tag) {
                    logger.fatal(”testVisitorAll()  Tag name is :”
                            + tag.getTagName() + ” \n Class is :”
                            + tag.getClass());
                }

            };

            parser.visitAllNodesWith(visitor);
        } catch (ParserException e) {
            e.printStackTrace();
        }
    }
    /*
     * 测试对指定Tag的NodeVisitor的用法
     */
    public void testTagVisitor() {
        try {

            Parser parser = new Parser(
                    “<head><title>dddd</title>”
                            + “<link href=’/test01/css.css’ text=’text/css’ rel=’stylesheet’ />”
                            + “<link href=’/test02/css.css’ text=’text/css’ rel=’stylesheet’ />”
                            + “</head>” + “<body>”
                            + “<a href=’http://www.yeeach.com’>yeeach.com</a>”
                            + “</body>”);
            NodeVisitor visitor = new NodeVisitor() {
                public void visitTag(Tag tag) {
                    if (tag instanceof HeadTag) {
                        logger.fatal(”visitTag() HeadTag : Tag name is :”
                                + tag.getTagName() + ” \n Class is :”
                                + tag.getClass() + “\n Text is :”
                                + tag.getText());
                    } else if (tag instanceof TitleTag) {
                        logger.fatal(”visitTag() TitleTag : Tag name is :”
                                + tag.getTagName() + ” \n Class is :”
                                + tag.getClass() + “\n Text is :”
                                + tag.getText());

                    } else if (tag instanceof LinkTag) {
                        logger.fatal(”visitTag() LinkTag : Tag name is :”
                                + tag.getTagName() + ” \n Class is :”
                                + tag.getClass() + “\n Text is :”
                                + tag.getText() + ” \n getAttribute is :”
                                + tag.getAttribute(”href”));
                    } else {
                        logger.fatal(”visitTag() : Tag name is :”
                                + tag.getTagName() + ” \n Class is :”
                                + tag.getClass() + “\n Text is :”
                                + tag.getText());
                    }

                }

            };

            parser.visitAllNodesWith(visitor);
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
    /*
     * 测试HtmlPage的用法
     */
    public void testHtmlPage() {
        String inputHTML = “<html>” + “<head>”
                + “<title>Welcome to the HTMLParser website</title>”
                + “</head>” + “<body>” + “Welcome to HTMLParser”
                + “<table id=’table1′ >”
                + “<tr><td>1-11</td><td>1-12</td><td>1-13</td>”
                + “<tr><td>1-21</td><td>1-22</td><td>1-23</td>”
                + “<tr><td>1-31</td><td>1-32</td><td>1-33</td></table>”
                + “<table id=’table2′ >”
                + “<tr><td>2-11</td><td>2-12</td><td>2-13</td>”
                + “<tr><td>2-21</td><td>2-22</td><td>2-23</td>”
                + “<tr><td>2-31</td><td>2-32</td><td>2-33</td></table>”
                + “</body>” + “</html>”;
        Parser parser = new Parser();
        try {
            parser.setInputHTML(inputHTML);
            parser.setEncoding(parser.getURL());
            HtmlPage page = new HtmlPage(parser);
            parser.visitAllNodesWith(page);
            logger.fatal(”testHtmlPage -title is :” + page.getTitle());
            NodeList list = page.getBody();

            for (NodeIterator iterator = list.elements(); iterator
                    .hasMoreNodes();) {
                Node node = iterator.nextNode();
                logger.fatal(”testHtmlPage -node  is :” + node.toHtml());
            }

        } catch (ParserException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
    }
    /*
     * 测试LinkBean的用法
     */
    public void testLinkBean() {
        Parser parser = new Parser();

        LinkBean linkBean = new LinkBean();
        linkBean.setURL(”http://www.google.com”);
        URL[] urls = linkBean.getLinks();

        for (int i = 0; i < urls.length; i++) {
            URL url = urls[i];
            logger.fatal(”testLinkBean() -url  is :” + url);
        }

    }

}