getcmiqikan.java

本程序是专门用于从网页上自动收集cmi,cnki上的被引文献的数据

package cn.ac.cintcm.spider.cmi;

import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

import org.htmlparser.Node;
import org.htmlparser.NodeFilter;
import org.htmlparser.beans.FilterBean;
import org.htmlparser.filters.TagNameFilter;
import org.htmlparser.tags.InputTag;
import org.htmlparser.util.NodeList;

public class GetCmiQiKan {
	
	public static void main(String[] args) throws IOException {
		//String resource = (new GetUrlContent("http://cnki.cintcm.ac.cn:1012/cdweb/page/kanmin.cbs?db=mulu")).getContent();
       List<String> qikanList = null;
		//Parser parser = new Parser(resource);
        FilterBean bean = new FilterBean ();
        bean.setFilters(getFilter());
        bean.setURL("http://cnki.cintcm.ac.cn:1012/cdweb/page/kanmin.cbs?db=mulu");
        NodeList list = bean.getNodes();
        qikanList = new ArrayList<String>();
        for (Node node : list.toNodeArray()) {
        	InputTag tag = new InputTag();
        	tag.setText(node.toHtml());
        	String name = tag.getAttribute("value");
        	qikanList.add(name);
        	System.out.println(name);
        }
	}

	public static NodeFilter[] getFilter() {
        TagNameFilter filter0 = new TagNameFilter ();
        filter0.setName ("option");
        NodeFilter[] array0 = new NodeFilter[1];
        array0[0] = filter0;
        return array0;
	}
}