Java HTML Parse Jsoup parseInfoHeader(Element element)

Description

parse Info Header

License

Apache License

Declaration

public static synchronized void parseInfoHeader(Element element)
            throws JSONException 

Method Source Code

//package com.java2s;
//License from project: Apache License 

import org.json.JSONException;
import org.json.JSONObject;

import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;

public class Main {
    public static synchronized void parseInfoHeader(Element element)
            throws JSONException {
        Element infotemplatebox = element.getElementsByClass(
                "infotemplatebox").first();
        if (infotemplatebox != null) {
            parseTemplate1_1(element);/*from ww w . ja  va2s .  com*/
            return;
        }
        infotemplatebox = element.select("table").first();
        if (infotemplatebox != null) {
            parseTemplate1_2(element);
            return;
        }
    }

    public static synchronized void parseTemplate1_1(Element element)
            throws JSONException {
        JSONObject jsonObject = new JSONObject();
        JSONObject jsonItem = new JSONObject();
        Element infotemplatebox = element.getElementsByClass(
                "infotemplatebox").first();
        Elements elements = infotemplatebox.select("tr");
        for (int i = 0, total = elements.size(); i < total; i++) {
            Element item = elements.get(i);
            //            System.err.println(item);
            if (i == 0) {
                jsonObject.put("cover", item.select("img").attr("src"));
            } else if (i > total - 3) {
                if (i == total - 1) {
                    jsonItem.put(elements.get(i - 1).text(), item.text());
                }
            } else {
                Elements th = item.select("th");
                Elements td = item.select("td");
                if (th.size() == 0 || td.size() == 0) {
                    jsonObject.put(td.size() == 0 ? th.text() : td.text(),
                            jsonItem);
                } else {
                    jsonItem.put(th.text(), td.text());
                }
            }
        }

        System.err.println(jsonObject.toString());

        element.getElementsByClass("infoBox").remove();
        element.getElementsByClass("infotemplatebox").remove();
        elements = element.select("p");

        //        System.err.println(elements.text());
    }

    public static synchronized void parseTemplate1_2(Element element) {
        Element infotemplatebox = element.select("table").first();
        Elements elements = infotemplatebox.select("tr");
        for (int i = 0, total = elements.size(); i < total; i++) {
            Element item = elements.get(i);
            if (i == 1) {
                System.err.println(item.select("img").attr("src"));
            } else {
                System.err.println(item.text());
            }
        }
        element.getElementsByClass("infoBox").remove();
        element.getElementsByClass("infotemplatebox").remove();
        elements = element.select("p");

        System.err.println(elements.text());
    }
}

Related

1. parse(URL url, int timeout)
2. parseByteData(ByteBuffer byteData, String charsetName, String baseUri, Parser parser)
3. parseEmail(String content)
4. parseFile(String filePath)
5. parseInfoBody(Element element)
6. parsePropertyTable(Element table)
7. parseTable2ArrayList(Document doc, String selectorRow, String selectorCol)
8. parseTemplate1_1(Element element)
9. parseTemplate1_2(Element element)