public void initialize(InputSplit genericSplit, TaskAttemptContext context) throws IOException {
FileSplit split = (FileSplit) genericSplit;
Configuration job = context.getConfiguration();
m_Sb.setLength(0);
m_Start = split.getStart();
m_End = m_Start + split.getLength();
final Path file = split.getPath();
compressionCodecs = new CompressionCodecFactory(job);
final CompressionCodec codec = compressionCodecs.getCodec(file);
// open the file and seek to the m_Start of the split
FileSystem fs = file.getFileSystem(job);
// getFileStatus fileStatus = fs.getFileStatus(split.getPath());
//noinspection deprecation
@SuppressWarnings(value = "deprecated")
long length = fs.getLength(file);
FSDataInputStream fileIn = fs.open(split.getPath());
if (m_Start > 0) fileIn.seek(m_Start);
if (codec != null) {
CompressionInputStream inputStream = codec.createInputStream(fileIn);
m_Input = new BufferedReader(new InputStreamReader(inputStream));
m_End = length;
} else {
m_Input = new BufferedReader(new InputStreamReader(fileIn));
}
m_Current = m_Start;
m_Key = split.getPath().getName();
}
/**
* A little test program.
*
* @param args
*/
public static void main(String[] args) throws Exception {
Configuration conf = new Configuration();
CompressionCodecFactory factory = new CompressionCodecFactory(conf);
boolean encode = false;
for (int i = 0; i < args.length; ++i) {
if ("-in".equals(args[i])) {
encode = true;
} else if ("-out".equals(args[i])) {
encode = false;
} else {
CompressionCodec codec = factory.getCodec(new Path(args[i]));
if (codec == null) {
System.out.println("Codec for " + args[i] + " not found.");
} else {
if (encode) {
CompressionOutputStream out =
codec.createOutputStream(new java.io.FileOutputStream(args[i]));
byte[] buffer = new byte[100];
String inFilename = removeSuffix(args[i], codec.getDefaultExtension());
java.io.InputStream in = new java.io.FileInputStream(inFilename);
int len = in.read(buffer);
while (len > 0) {
out.write(buffer, 0, len);
len = in.read(buffer);
}
in.close();
out.close();
} else {
CompressionInputStream in =
codec.createInputStream(new java.io.FileInputStream(args[i]));
byte[] buffer = new byte[100];
int len = in.read(buffer);
while (len > 0) {
System.out.write(buffer, 0, len);
len = in.read(buffer);
}
in.close();
}
}
}
}
}
public static void testFinding() {
CompressionCodecFactory factory = new CompressionCodecFactory(new Configuration());
CompressionCodec codec = factory.getCodec(new Path("/tmp/foo.bar"));
assertEquals("default factory foo codec", null, codec);
codec = factory.getCodecByClassName(BarCodec.class.getCanonicalName());
assertEquals("default factory foo codec", null, codec);
codec = factory.getCodec(new Path("/tmp/foo.gz"));
checkCodec("default factory for .gz", GzipCodec.class, codec);
codec = factory.getCodecByClassName(GzipCodec.class.getCanonicalName());
checkCodec("default factory for gzip codec", GzipCodec.class, codec);
codec = factory.getCodecByName("gzip");
checkCodec("default factory for gzip codec", GzipCodec.class, codec);
codec = factory.getCodecByName("GZIP");
checkCodec("default factory for gzip codec", GzipCodec.class, codec);
codec = factory.getCodecByName("GZIPCodec");
checkCodec("default factory for gzip codec", GzipCodec.class, codec);
codec = factory.getCodecByName("gzipcodec");
checkCodec("default factory for gzip codec", GzipCodec.class, codec);
Class klass = factory.getCodecClassByName("gzipcodec");
assertEquals(GzipCodec.class, klass);
codec = factory.getCodec(new Path("/tmp/foo.bz2"));
checkCodec("default factory for .bz2", BZip2Codec.class, codec);
codec = factory.getCodecByClassName(BZip2Codec.class.getCanonicalName());
checkCodec("default factory for bzip2 codec", BZip2Codec.class, codec);
codec = factory.getCodecByName("bzip2");
checkCodec("default factory for bzip2 codec", BZip2Codec.class, codec);
codec = factory.getCodecByName("bzip2codec");
checkCodec("default factory for bzip2 codec", BZip2Codec.class, codec);
codec = factory.getCodecByName("BZIP2");
checkCodec("default factory for bzip2 codec", BZip2Codec.class, codec);
codec = factory.getCodecByName("BZIP2CODEC");
checkCodec("default factory for bzip2 codec", BZip2Codec.class, codec);
codec = factory.getCodecByClassName(DeflateCodec.class.getCanonicalName());
checkCodec("default factory for deflate codec", DeflateCodec.class, codec);
codec = factory.getCodecByName("deflate");
checkCodec("default factory for deflate codec", DeflateCodec.class, codec);
codec = factory.getCodecByName("deflatecodec");
checkCodec("default factory for deflate codec", DeflateCodec.class, codec);
codec = factory.getCodecByName("DEFLATE");
checkCodec("default factory for deflate codec", DeflateCodec.class, codec);
codec = factory.getCodecByName("DEFLATECODEC");
checkCodec("default factory for deflate codec", DeflateCodec.class, codec);
factory = setClasses(new Class[0]);
// gz, bz2, snappy, lz4 are picked up by service loader, but bar isn't
codec = factory.getCodec(new Path("/tmp/foo.bar"));
assertEquals("empty factory bar codec", null, codec);
codec = factory.getCodecByClassName(BarCodec.class.getCanonicalName());
assertEquals("empty factory bar codec", null, codec);
codec = factory.getCodec(new Path("/tmp/foo.gz"));
checkCodec("empty factory gz codec", GzipCodec.class, codec);
codec = factory.getCodecByClassName(GzipCodec.class.getCanonicalName());
checkCodec("empty factory gz codec", GzipCodec.class, codec);
codec = factory.getCodec(new Path("/tmp/foo.bz2"));
checkCodec("empty factory for .bz2", BZip2Codec.class, codec);
codec = factory.getCodecByClassName(BZip2Codec.class.getCanonicalName());
checkCodec("empty factory for bzip2 codec", BZip2Codec.class, codec);
codec = factory.getCodec(new Path("/tmp/foo.snappy"));
checkCodec("empty factory snappy codec", SnappyCodec.class, codec);
codec = factory.getCodecByClassName(SnappyCodec.class.getCanonicalName());
checkCodec("empty factory snappy codec", SnappyCodec.class, codec);
codec = factory.getCodec(new Path("/tmp/foo.lz4"));
checkCodec("empty factory lz4 codec", Lz4Codec.class, codec);
codec = factory.getCodecByClassName(Lz4Codec.class.getCanonicalName());
checkCodec("empty factory lz4 codec", Lz4Codec.class, codec);
factory = setClasses(new Class[] {BarCodec.class, FooCodec.class, FooBarCodec.class});
codec = factory.getCodec(new Path("/tmp/.foo.bar.gz"));
checkCodec("full factory gz codec", GzipCodec.class, codec);
codec = factory.getCodecByClassName(GzipCodec.class.getCanonicalName());
checkCodec("full codec gz codec", GzipCodec.class, codec);
codec = factory.getCodec(new Path("/tmp/foo.bz2"));
checkCodec("full factory for .bz2", BZip2Codec.class, codec);
codec = factory.getCodecByClassName(BZip2Codec.class.getCanonicalName());
checkCodec("full codec bzip2 codec", BZip2Codec.class, codec);
codec = factory.getCodec(new Path("/tmp/foo.bar"));
checkCodec("full factory bar codec", BarCodec.class, codec);
codec = factory.getCodecByClassName(BarCodec.class.getCanonicalName());
checkCodec("full factory bar codec", BarCodec.class, codec);
codec = factory.getCodecByName("bar");
checkCodec("full factory bar codec", BarCodec.class, codec);
codec = factory.getCodecByName("BAR");
checkCodec("full factory bar codec", BarCodec.class, codec);
codec = factory.getCodec(new Path("/tmp/foo/baz.foo.bar"));
checkCodec("full factory foo bar codec", FooBarCodec.class, codec);
codec = factory.getCodecByClassName(FooBarCodec.class.getCanonicalName());
checkCodec("full factory foo bar codec", FooBarCodec.class, codec);
codec = factory.getCodecByName("foobar");
checkCodec("full factory foo bar codec", FooBarCodec.class, codec);
codec = factory.getCodecByName("FOOBAR");
checkCodec("full factory foo bar codec", FooBarCodec.class, codec);
codec = factory.getCodec(new Path("/tmp/foo.foo"));
checkCodec("full factory foo codec", FooCodec.class, codec);
codec = factory.getCodecByClassName(FooCodec.class.getCanonicalName());
checkCodec("full factory foo codec", FooCodec.class, codec);
codec = factory.getCodecByName("foo");
checkCodec("full factory foo codec", FooCodec.class, codec);
codec = factory.getCodecByName("FOO");
checkCodec("full factory foo codec", FooCodec.class, codec);
factory = setClasses(new Class[] {NewGzipCodec.class});
codec = factory.getCodec(new Path("/tmp/foo.gz"));
checkCodec("overridden factory for .gz", NewGzipCodec.class, codec);
codec = factory.getCodecByClassName(NewGzipCodec.class.getCanonicalName());
checkCodec("overridden factory for gzip codec", NewGzipCodec.class, codec);
}
/**
* Returns a factory for a given set of codecs
*
* @param classes the codec classes to include
* @return a new factory
*/
private static CompressionCodecFactory setClasses(Class[] classes) {
Configuration conf = new Configuration();
CompressionCodecFactory.setCodecClasses(conf, Arrays.asList(classes));
return new CompressionCodecFactory(conf);
}