import java.io.*;
import java.net.*;
import java.nio.*;
import java.nio.channels.*;
import java.nio.file.*;
import java.util.*;

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


class AMCArtistScrape
{

    private static void printf(String msg, Object... args)
    {
	System.out.println(String.format(msg, args));
    }

    public static void downloadURL(String url_str) {

	try {
	    URL url = new URL(url_str);

	    String local_file_str = url_str.substring(url_str.lastIndexOf('/')+1, url_str.length() );
	    File local_file = new File("audio",local_file_str);

	    if (!local_file.exists()) {
		System.out.printf("Downloading audio to '%s'...\n",local_file_str);
		FileOutputStream fos = new FileOutputStream(local_file);

		ReadableByteChannel rbc = Channels.newChannel(url.openStream());
		fos.getChannel().transferFrom(rbc, 0, Long.MAX_VALUE);

		System.out.println("... done");
	    }
	    else {
		System.out.printf("Skipping '%s' as is already exists\n",local_file_str);
	    }
	    
	}
	catch (Exception e) {
	    e.printStackTrace();
	}
    }
    
    public static ArrayList<MusicRecord> extractArtistRecordsFromPage(Document doc)
    {

	ArrayList<MusicRecord> records = new ArrayList<MusicRecord>();
	
	Elements table_rows = doc.select("table.search_results tr");
	
	for (Element tr : table_rows) {
	    
	    Elements td_cells = tr.select("td");
	    
	    if (td_cells.size() == 6) {
		Element title_anchor = td_cells.get(0).select("a").first();
		String track_url       = title_anchor.attr("href");
		String title           = title_anchor.ownText();
		String instrumentation = td_cells.get(1).ownText();
		String composer        = td_cells.get(2).ownText();
		String year            = td_cells.get(3).ownText();
		
		Element audio_elem = td_cells.get(4).select("audio").first();
		if (audio_elem != null) {
		    String audio_url = audio_elem.attr("src");
		    
		    MusicRecord mr = new MusicRecord(track_url,title,composer,year,audio_url);
		    System.out.println(mr.toJSONString());
		    
		    records.add(mr);
		}
		
	    }					   
	    
	}
	
	return records;
    }

    private static void followArtistPages(String base_domain,String base_url, String artist_suffix)
    {
	String artist_url_str = base_domain+base_url+artist_suffix;

	try {
	    boolean has_next_page = true;
	    int page_number = 1;

	    while (has_next_page) {
		System.out.println("Procesing Artist suffix '" + artist_suffix + "': Page " + page_number);

		Document doc = Jsoup.connect(artist_url_str).get();

		ArrayList<MusicRecord> music_records = extractArtistRecordsFromPage(doc);
		
		for (MusicRecord mr: music_records) {
		    downloadURL(mr.getAudioURL());
		}

		// <ul class="pagination">
		//  ...
		//  <li href="URL-TO-FOLLOW">Next</li>
		// </ul>

		Elements pagination_block = doc.select("ul.pagination > li > a");
		Element last_li_a = pagination_block.last();
		String last_li_a_text = last_li_a.text();
		if (last_li_a_text.equals("Next >")) {
		    artist_url_str = base_domain+last_li_a.attr("href");
		}
		else {
		    has_next_page = false;
		}

		page_number++;

		//System.out.println("*** last li a text = " + last_li_a_text);
		//System.out.println("*** artist_url_str = " + artist_url_str);
	    }
	}
	catch (Exception e) {
	    e.printStackTrace();
	}
    }

    

    public static void main(String[] args)
    {

	// Example artist page at AMC
	//   http://www.australianmusiccentre.com.au/search?type=work&sort=alphaTitleSort&wfc[]=Roger+Dean

	String base_domain = "http://www.australianmusiccentre.com.au";
	String base_url = "/search?type=work&sort=alphaTitleSort";
	base_url += "&wfc[]=";

	followArtistPages(base_domain,base_url,"Roger+Dean");
    }
}