001: /*
002: * File : $Source: /usr/local/cvs/opencms/src/org/opencms/util/I_CmsHtmlNodeVisitor.java,v $
003: * Date : $Date: 2008-02-27 12:05:36 $
004: * Version: $Revision: 1.6 $
005: *
006: * This library is part of OpenCms -
007: * the Open Source Content Management System
008: *
009: * Copyright (c) 2002 - 2008 Alkacon Software GmbH (http://www.alkacon.com)
010: *
011: * This library is free software; you can redistribute it and/or
012: * modify it under the terms of the GNU Lesser General Public
013: * License as published by the Free Software Foundation; either
014: * version 2.1 of the License, or (at your option) any later version.
015: *
016: * This library is distributed in the hope that it will be useful,
017: * but WITHOUT ANY WARRANTY; without even the implied warranty of
018: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
019: * Lesser General Public License for more details.
020: *
021: * For further information about Alkacon Software GmbH, please see the
022: * company website: http://www.alkacon.com
023: *
024: * For further information about OpenCms, please see the
025: * project website: http://www.opencms.org
026: *
027: * You should have received a copy of the GNU Lesser General Public
028: * License along with this library; if not, write to the Free Software
029: * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
030: */
031:
032: package org.opencms.util;
033:
034: import java.util.List;
035:
036: import org.htmlparser.Remark;
037: import org.htmlparser.Tag;
038: import org.htmlparser.Text;
039: import org.htmlparser.util.ParserException;
040:
041: /**
042: *
043: * Interface for a combination of a visitor of HTML documents along with the hook to start the
044: * parser / lexer that triggers the visit.
045: * <p>
046: *
047: *
048: *
049: * @author Alexander Kandzior
050: *
051: * @version $Revision: 1.6 $
052: *
053: * @since 6.1.3
054: *
055: */
056: public interface I_CmsHtmlNodeVisitor {
057:
058: /**
059: * Returns the configuartion String of this visitor or the empty String if was not provided
060: * before.
061: * <p>
062: *
063: * @return the configuartion String of this visitor - by this contract never null but an empty
064: * String if not provided.
065: *
066: * @see #setConfiguration(String)
067: */
068: String getConfiguration();
069:
070: /**
071: * Returns the text extraction result.
072: * <p>
073: *
074: * @return the text extraction result
075: */
076: String getResult();
077:
078: /**
079: * Extracts the text from the given html content, assuming the given html encoding.
080: * <p>
081: *
082: * @param html the content to extract the plain text from
083: * @param encoding the encoding to use
084: *
085: * @return the text extracted from the given html content
086: *
087: * @throws ParserException if something goes wrong
088: */
089: String process(String html, String encoding) throws ParserException;
090:
091: /**
092: * Set a configuartion String for this visitor.
093: * <p>
094: *
095: * This will most likely be done with data from an xsd, custom jsp tag, ...
096: * <p>
097: *
098: * @param configuration the configuration of this visitor to set.
099: */
100: void setConfiguration(String configuration);
101:
102: /**
103: * Sets a list of upper case tag names for which parsing / visitng should not correct missing closing tags.<p>
104: *
105: * This has to be used before <code>{@link #process(String, String)}</code> is invoked to take an effect.<p>
106: *
107: * @param noAutoCloseTags a list of upper case tag names for which parsing / visiting
108: * should not correct missing closing tags to set.
109: */
110: void setNoAutoCloseTags(List noAutoCloseTags);
111:
112: /**
113: * Visitor method (callback) invoked when a closing Tag is encountered.
114: * <p>
115: *
116: * @param tag the tag that is ended.
117: *
118: * @see org.htmlparser.visitors.NodeVisitor#visitEndTag(org.htmlparser.Tag)
119: */
120: void visitEndTag(Tag tag);
121:
122: /**
123: * Visitor method (callback) invoked when a remark Tag (HTML comment) is encountered.
124: * <p>
125: *
126: * @param remark the remark Tag to visit.
127: *
128: * @see org.htmlparser.visitors.NodeVisitor#visitRemarkNode(org.htmlparser.Remark)
129: */
130: void visitRemarkNode(Remark remark);
131:
132: /**
133: *
134: * Visitor method (callback) invoked when a remark Tag (HTML comment) is encountered.
135: * <p>
136: *
137: * @param text the text that is visited.
138: *
139: * @see org.htmlparser.visitors.NodeVisitor#visitStringNode(org.htmlparser.Text)
140: */
141: void visitStringNode(Text text);
142:
143: /**
144: * Visitor method (callback) invoked when a starting Tag (HTML comment) is encountered.
145: * <p>
146: *
147: * @param tag the tag that is visited.
148: *
149: * @see org.htmlparser.visitors.NodeVisitor#visitTag(org.htmlparser.Tag)
150: */
151: void visitTag(Tag tag);
152:
153: }
|