VcfValidator.java

package org.molgenis.annotation.cmd.utils;

import java.io.*;
import java.util.Date;
import java.util.Scanner;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

import static org.apache.commons.io.FilenameUtils.removeExtension;

public class VcfValidator
{
	private String vcfValidatorLocation;

	public VcfValidator(String vcfValidatorLocation)
	{
		this.vcfValidatorLocation = vcfValidatorLocation;
	}

	/**
	 * Validation method that calls the perl executable of the vcf-validator. Logs vcf validation output into a log file
	 *
	 * @param vcfFile The VCF file generated by the annotation process
	 * @return Success or fail message
	 * @throws IOException
	 */
	public String validateVCF(File vcfFile) throws IOException
	{
		BufferedWriter bufferedWriter = null;
		Scanner inputScanner = null;
		Scanner errorScanner = null;

		try
		{
			ProcessBuilder processBuilder = new ProcessBuilder(vcfValidatorLocation, vcfFile.getAbsolutePath(), "-u",
					"-d").directory(new File(vcfValidatorLocation).getParentFile());

			Process proc = processBuilder.start();

			// Checks if vcf-tools is present
			if (vcfValidatorLocation == null || !new File(vcfValidatorLocation).exists())
			{
				return "No vcf-validator present, skipping validation.";
			}

			String line = "";
			Integer errorCount = null;
			Pattern p = Pattern.compile("(\\d+)\\s*errors\\s*total");

			String fileNameWithoutExtension = removeExtension(vcfFile.getName());
			File logFile = new File(fileNameWithoutExtension + "-validation.log");
			if (!logFile.exists())
			{
				logFile.createNewFile();
			}

			Date date = new Date();
			bufferedWriter = new BufferedWriter(
					new OutputStreamWriter(new FileOutputStream(logFile.getAbsolutePath()), "UTF-8"));

			bufferedWriter.write("### Validation report for " + vcfFile.getName() + "\n");
			bufferedWriter.write("### Validation date: " + date + "\n");

			inputScanner = new Scanner(proc.getInputStream(), "UTF-8");
			errorScanner = new Scanner(proc.getErrorStream(), "UTF-8");

			while (proc.isAlive() || inputScanner.hasNext() || errorScanner.hasNext())
			{
				while (errorScanner.hasNext())
				{
					line = errorScanner.nextLine();
					bufferedWriter.append("ERR> " + line + "\n");
				}
				while (inputScanner.hasNext())
				{
					line = inputScanner.nextLine();
					bufferedWriter.append(line + "\n");
					Matcher m = p.matcher(line);
					if (m.find())
					{
						errorCount = Integer.parseInt(m.group(1));
					}
				}
			}

			bufferedWriter.write("\n##################################################\n");

			if (errorCount != null && errorCount == 0)
			{
				return "VCF file [" + vcfFile.getName() + "] passed validation.";
			}
			else
			{
				return "VCF file [" + vcfFile.getName() + "] did not pass validation, see the log for more details.";
			}
		}
		catch (IOException e)
		{
			throw new RuntimeException("Something went wrong: " + e);
		}
		finally
		{
			tryClose(bufferedWriter);
			tryClose(inputScanner);
			tryClose(errorScanner);
		}
	}

	private void tryClose(Closeable closeable)
	{
		if (closeable != null)
		{
			try
			{
				closeable.close();
			}
			catch (Exception ignore)
			{

			}
		}
	}
}