/*
 * FastaParser.java
 */
package org.ngbw.utils;


import java.io.BufferedReader;
import java.io.ByteArrayInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.io.Reader;
import java.io.StringReader;
import java.util.regex.Pattern;


/**
 * 
 * @author Paul Hoover
 *
 */
class FastaParser {

	private static final Pattern m_headerPattern = Pattern.compile("^\\s*>.*$");

	private final BufferedReader m_reader;
	private StringBuilder m_completeSource;
	private String m_line;


	/**
	 * 
	 * @param input
	 */
	public FastaParser(String input)
	{
		this(new StringReader(input));
	}

	/**
	 * 
	 * @param input
	 */
	public FastaParser(byte[] input)
	{
		this(new ByteArrayInputStream(input));
	}

	/**
	 * 
	 * @param input
	 */
	public FastaParser(InputStream input)
	{
		this(new InputStreamReader(input));
	}

	/**
	 * 
	 * @param input
	 */
	public FastaParser(Reader input)
	{
		m_reader = new BufferedReader(input);
	}


	/**
	 * 
	 * @return
	 * @throws IOException
	 */
	public SequenceRecord nextRecord() throws IOException
	{
		if (!findFirstLine())
			return null;

		SequenceRecord record = new SequenceRecord();

		parseHeader(record);

		while (true) {
			if (m_line == null || m_headerPattern.matcher(m_line).matches()) {
				record.completeSource = m_completeSource.toString();

				break;
			}
			else {
				storeLine();
				readLine();
			}
		}

		return record;
	}

	/**
	 * 
	 * @return
	 * @throws IOException
	 */
	private boolean findFirstLine() throws IOException
	{
		m_completeSource = new StringBuilder();

		if (m_line == null)
			readLine();

		while (true) {
			if (m_line == null)
				return false;

			if (m_headerPattern.matcher(m_line).matches()) {
				storeLine();

				break;
			}

			readLine();
		}

		return true;
	}

	/**
	 * 
	 * @throws IOException
	 */
	private void readLine() throws IOException
	{
		m_line = m_reader.readLine();
	}

	/**
	 * 
	 */
	private void storeLine()
	{
		m_completeSource.append(m_line);
		m_completeSource.append('\n');
	}

	/**
	 * 
	 * @param record
	 * @throws IOException
	 */
	private void parseHeader(SequenceRecord record) throws IOException
	{
		String[] subStrings = m_line.substring(1).split("\\|");

		if (subStrings.length == 5) {
			record.primaryId = subStrings[1];
			record.organism = subStrings[3];
			record.name = subStrings[4];
		}

		record.filteredSource = m_line + '\n';

		readLine();
	}
}