
com.jaeksoft.searchlib.parser.htmlParser.JSoupParser Maven / Gradle / Ivy
/**
* License Agreement for OpenSearchServer
*
* Copyright (C) 2012-2013 Emmanuel Keller / Jaeksoft
*
* http://www.open-search-server.com
*
* This file is part of OpenSearchServer.
*
* OpenSearchServer is free software: you can redistribute it and/or
* modify it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* OpenSearchServer is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with OpenSearchServer.
* If not, see .
**/
package com.jaeksoft.searchlib.parser.htmlParser;
import java.io.IOException;
import java.io.InputStream;
import javax.xml.parsers.ParserConfigurationException;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Node;
import org.xml.sax.SAXException;
import com.jaeksoft.searchlib.util.IOUtils;
public class JSoupParser extends HtmlDocumentProvider {
public JSoupParser() {
super(HtmlParserEnum.JSoupParser);
}
@Override
protected HtmlNodeAbstract> getDocument(String charset,
InputStream inputStream) throws SAXException, IOException,
ParserConfigurationException {
Node node = Jsoup.parse(IOUtils.toString(inputStream, charset));
return new JSoupHtmlNode(node);
}
@Override
protected HtmlNodeAbstract> getDocument(String htmlSource)
throws IOException, ParserConfigurationException {
Node node = Jsoup.parse(htmlSource);
return new JSoupHtmlNode(node);
}
@Override
public boolean isXPathSupported() {
return false;
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy