package kratav.BGC.scrapper;

import java.io.File;
import java.io.FileOutputStream;
import java.io.OutputStreamWriter;
import java.io.Writer;
import java.util.ArrayList;

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

public class MusicCords {
	static int x = 1;
	static String title = null;
	private static String path = "/home/aayush/Development/BollyGuitarCords/Files";
	private static String base_file_name = "/home/aayush/Development/BollyGuitarCords/Files/G";
	private static ArrayList<String> illegal;
	private static String []ignore = {
		"http://tabandchord.com/2014/08/veerey-di-wedding-chords-entertainment/",
		"http://tabandchord.com/2014/08/independence-day-songs-chords-desh-bhakti-song-chords/",
		"http://tabandchord.com/2014/08/independence-day-songs-tab-desh-bhakti-song-tab/",
		"http://tabandchord.com/2014/02/samjho-na-kuchh-toh-samjho-na-tab-aapka-surror/",
		"http://tabandchord.com/2014/02/samjho-na-kuchh-toh-samjho-na-chords-aapka-surror/",
	};
	public static void main(String... args) throws Exception {

		for (int i = 1; i <= 23; i++) {
			Document doc = Jsoup
					.connect(
							"http://tabandchord.com/category/chord/chord-hindi/page/"
									+ i + "/").userAgent("Mozilla")
					.timeout(15 * 1000).get();
			Elements e = doc.select(".entry-content");
			Elements links = e.select("a[href]");


		outer:	for (Element link : links) {
				if (link.toString().contains("201")) {
					System.out.println(link.attr("href"));

					if(!link.text().contains("–")){
						System.out.println(link.text());
						continue;
					}
					for(String s :ignore){
						if(link.attr("href").equals(s)){
							continue outer;
						}
					}
					if(!writeToFile(x, link.attr("href"),link.text())){
						continue;
					}
					x++;
				}
			}
		}

		System.out.println("==== end of road ====");
		for(String u : illegal){
			System.out.println(u);
		}
		// Parse Data from each page and write that to a file

	}

	public static boolean writeToFile(int x, String url, String title) throws Exception {

		Document doc = Jsoup.connect(url).userAgent("Mozilla")
				.timeout(15 * 1000).get();
		String e = doc.select(".entry-content p").toString();

		System.out.println(e);
		if(e.length()==0){
			return false;
		}
		String g;
		try {
			if (e.toString().contains("</strong>") == true) {
				Elements s = doc.select(".entry-content").select("strong");

				Elements a = s.select("a");

				if(a.size()>0){
//					illegal.add(url);
					illegal = new ArrayList<String>();
					for (Element el : a) {
						Element culprit = el.parent().parent();
						illegal.add(culprit.html());
					}

				}

				System.out.println("\n\n" + s + "\n\n\n");

				String s1[] = s.toString().split("</strong>");
				String K0 = s1[0];
				g = e.toString().replace("<p>" + K0 + "</strong></p>", "");
				try {
					String K1 = s1[1];
					g = g.replace("<p>" + K1.substring(1) + "</strong></p>", "");
				} catch (Exception m) {
				}
			}
		} catch (Exception ex) {
			ex.printStackTrace();
			ex.getMessage();
			System.exit(0);
		}
		e= e.replace("<p>.</p>", "");
		if(illegal != null){
			for (String  hate: illegal) {
				e = e.replace("<p>"+hate+"</p>", "");
			}

		}
		if (title.contains("- Tab And Chord"))
			title = title.replace(" - Tab And Chord", "");

		title = title.split("\\(")[0];
		title.trim();

		new File(path).mkdir();
		FileOutputStream fos = new FileOutputStream(base_file_name
				 + (x) + " - "
						+ title + ".html");
		Writer out = new OutputStreamWriter(fos, "UTF-8");
		System.out.println(e.toString());

		StringBuilder sb = new StringBuilder();
		sb.append(
				"<meta name=\"viewport\" content=\"width=device-width, initial-scale=1\">")
				.append(e.toString());

		out.write(sb.toString());
		out.close();
		return true;

	}
}