/// <exception cref="System.IO.IOException"/>
public virtual void TestGzipCodecWrite(bool useNative)
{
// Create a gzipped file using a compressor from the CodecPool,
// and try to read it back via the regular GZIPInputStream.
// Use native libs per the parameter
Configuration conf = new Configuration();
conf.SetBoolean(CommonConfigurationKeys.IoNativeLibAvailableKey, useNative);
if (useNative)
{
if (!ZlibFactory.IsNativeZlibLoaded(conf))
{
Log.Warn("testGzipCodecWrite skipped: native libs not loaded");
return;
}
}
else
{
NUnit.Framework.Assert.IsFalse("ZlibFactory is using native libs against request"
, ZlibFactory.IsNativeZlibLoaded(conf));
}
// Ensure that the CodecPool has a BuiltInZlibDeflater in it.
Compressor zlibCompressor = ZlibFactory.GetZlibCompressor(conf);
NUnit.Framework.Assert.IsNotNull("zlibCompressor is null!", zlibCompressor);
Assert.True("ZlibFactory returned unexpected deflator", useNative
? zlibCompressor is ZlibCompressor : zlibCompressor is BuiltInZlibDeflater);
CodecPool.ReturnCompressor(zlibCompressor);
// Create a GZIP text file via the Compressor interface.
CompressionCodecFactory ccf = new CompressionCodecFactory(conf);
CompressionCodec codec = ccf.GetCodec(new Path("foo.gz"));
Assert.True("Codec for .gz file is not GzipCodec", codec is GzipCodec
);
string msg = "This is the message we are going to compress.";
string tmpDir = Runtime.GetProperty("test.build.data", "/tmp/");
string fileName = new Path(new Path(tmpDir), "testGzipCodecWrite.txt.gz").ToString
();
BufferedWriter w = null;
Compressor gzipCompressor = CodecPool.GetCompressor(codec);
if (null != gzipCompressor)
{
// If it gives us back a Compressor, we should be able to use this
// to write files we can then read back with Java's gzip tools.
OutputStream os = new CompressorStream(new FileOutputStream(fileName), gzipCompressor
);
w = new BufferedWriter(new OutputStreamWriter(os));
w.Write(msg);
w.Close();
CodecPool.ReturnCompressor(gzipCompressor);
VerifyGzipFile(fileName, msg);
}
// Create a gzip text file via codec.getOutputStream().
w = new BufferedWriter(new OutputStreamWriter(codec.CreateOutputStream(new FileOutputStream
(fileName))));
w.Write(msg);
w.Close();
VerifyGzipFile(fileName, msg);
}
/// <exception cref="System.IO.IOException"/>
private static void CodecTestWithNOCompression(Configuration conf, string codecClass
)
{
// Create a compressor with NO_COMPRESSION and make sure that
// output is not compressed by comparing the size with the
// original input
CompressionCodec codec = null;
ZlibFactory.SetCompressionLevel(conf, ZlibCompressor.CompressionLevel.NoCompression
);
try
{
codec = (CompressionCodec)ReflectionUtils.NewInstance(conf.GetClassByName(codecClass
), conf);
}
catch (TypeLoadException)
{
throw new IOException("Illegal codec!");
}
Compressor c = codec.CreateCompressor();
// ensure same compressor placed earlier
ByteArrayOutputStream bos = new ByteArrayOutputStream();
CompressionOutputStream cos = null;
// write trivially compressable data
byte[] b = new byte[1 << 15];
Arrays.Fill(b, unchecked ((byte)43));
try
{
cos = codec.CreateOutputStream(bos, c);
cos.Write(b);
}
finally
{
if (cos != null)
{
cos.Close();
}
}
byte[] outbytes = bos.ToByteArray();
// verify data were not compressed
Assert.True("Compressed bytes contrary to configuration(NO_COMPRESSION)"
, outbytes.Length >= b.Length);
}
/// <exception cref="System.IO.IOException"/>
private static void GzipReinitTest(Configuration conf, CompressionCodec codec)
{
// Add codec to cache
ZlibFactory.SetCompressionLevel(conf, ZlibCompressor.CompressionLevel.BestCompression
);
ZlibFactory.SetCompressionStrategy(conf, ZlibCompressor.CompressionStrategy.DefaultStrategy
);
Compressor c1 = CodecPool.GetCompressor(codec);
CodecPool.ReturnCompressor(c1);
// reset compressor's compression level to perform no compression
ZlibFactory.SetCompressionLevel(conf, ZlibCompressor.CompressionLevel.NoCompression
);
Compressor c2 = CodecPool.GetCompressor(codec, conf);
// ensure same compressor placed earlier
Assert.True("Got mismatched ZlibCompressor", c1 == c2);
ByteArrayOutputStream bos = new ByteArrayOutputStream();
CompressionOutputStream cos = null;
// write trivially compressable data
byte[] b = new byte[1 << 15];
Arrays.Fill(b, unchecked ((byte)43));
try
{
cos = codec.CreateOutputStream(bos, c2);
cos.Write(b);
}
finally
{
if (cos != null)
{
cos.Close();
}
CodecPool.ReturnCompressor(c2);
}
byte[] outbytes = bos.ToByteArray();
// verify data were not compressed
Assert.True("Compressed bytes contrary to configuration", outbytes
.Length >= b.Length);
}
/// <summary>Write infLen bytes (deflated) to file in test dir using codec.</summary>
/// <remarks>
/// Write infLen bytes (deflated) to file in test dir using codec.
/// Records are of the form
/// <i><b64 rand><i+i><b64 rand>
/// </remarks>
/// <exception cref="System.IO.IOException"/>
private static Path WriteSplitTestFile(FileSystem fs, Random rand, CompressionCodec
codec, long infLen)
{
int RecSize = 1024;
Path wd = new Path(new Path(Runtime.GetProperty("test.build.data", "/tmp")).MakeQualified
(fs), codec.GetType().Name);
Path file = new Path(wd, "test" + codec.GetDefaultExtension());
byte[] b = new byte[RecSize];
Base64 b64 = new Base64(0, null);
DataOutputStream fout = null;
Compressor cmp = CodecPool.GetCompressor(codec);
try
{
fout = new DataOutputStream(codec.CreateOutputStream(fs.Create(file, true), cmp));
DataOutputBuffer dob = new DataOutputBuffer(RecSize * 4 / 3 + 4);
int seq = 0;
while (infLen > 0)
{
rand.NextBytes(b);
byte[] b64enc = b64.Encode(b);
// ensures rand printable, no LF
dob.Reset();
dob.WriteInt(seq);
System.Array.Copy(dob.GetData(), 0, b64enc, 0, dob.GetLength());
fout.Write(b64enc);
fout.Write('\n');
++seq;
infLen -= b64enc.Length;
}
Log.Info("Wrote " + seq + " records to " + file);
}
finally
{
IOUtils.Cleanup(Log, fout);
CodecPool.ReturnCompressor(cmp);
}
return(file);
}
/// <summary>Create an output stream with a codec taken from the global CodecPool.</summary>
/// <param name="codec">The codec to use to create the output stream.</param>
/// <param name="conf">The configuration to use if we need to create a new codec.</param>
/// <param name="out">The output stream to wrap.</param>
/// <returns>The new output stream</returns>
/// <exception cref="System.IO.IOException"/>
internal static CompressionOutputStream CreateOutputStreamWithCodecPool(CompressionCodec
codec, Configuration conf, OutputStream @out)
{
Compressor compressor = CodecPool.GetCompressor(codec, conf);
CompressionOutputStream stream = null;
try
{
stream = codec.CreateOutputStream(@out, compressor);
}
finally
{
if (stream == null)
{
CodecPool.ReturnCompressor(compressor);
}
else
{
stream.SetTrackedCompressor(compressor);
}
}
return(stream);
}
/// <exception cref="System.IO.IOException"/>
private static void CodecTest(Configuration conf, int seed, int count, string codecClass
)
{
// Create the codec
CompressionCodec codec = null;
try
{
codec = (CompressionCodec)ReflectionUtils.NewInstance(conf.GetClassByName(codecClass
), conf);
}
catch (TypeLoadException)
{
throw new IOException("Illegal codec!");
}
Log.Info("Created a Codec object of type: " + codecClass);
// Generate data
DataOutputBuffer data = new DataOutputBuffer();
RandomDatum.Generator generator = new RandomDatum.Generator(seed);
for (int i = 0; i < count; ++i)
{
generator.Next();
RandomDatum key = generator.GetKey();
RandomDatum value = generator.GetValue();
key.Write(data);
value.Write(data);
}
Log.Info("Generated " + count + " records");
// Compress data
DataOutputBuffer compressedDataBuffer = new DataOutputBuffer();
CompressionOutputStream deflateFilter = codec.CreateOutputStream(compressedDataBuffer
);
DataOutputStream deflateOut = new DataOutputStream(new BufferedOutputStream(deflateFilter
));
deflateOut.Write(data.GetData(), 0, data.GetLength());
deflateOut.Flush();
deflateFilter.Finish();
Log.Info("Finished compressing data");
// De-compress data
DataInputBuffer deCompressedDataBuffer = new DataInputBuffer();
deCompressedDataBuffer.Reset(compressedDataBuffer.GetData(), 0, compressedDataBuffer
.GetLength());
CompressionInputStream inflateFilter = codec.CreateInputStream(deCompressedDataBuffer
);
DataInputStream inflateIn = new DataInputStream(new BufferedInputStream(inflateFilter
));
// Check
DataInputBuffer originalData = new DataInputBuffer();
originalData.Reset(data.GetData(), 0, data.GetLength());
DataInputStream originalIn = new DataInputStream(new BufferedInputStream(originalData
));
for (int i_1 = 0; i_1 < count; ++i_1)
{
RandomDatum k1 = new RandomDatum();
RandomDatum v1 = new RandomDatum();
k1.ReadFields(originalIn);
v1.ReadFields(originalIn);
RandomDatum k2 = new RandomDatum();
RandomDatum v2 = new RandomDatum();
k2.ReadFields(inflateIn);
v2.ReadFields(inflateIn);
Assert.True("original and compressed-then-decompressed-output not equal"
, k1.Equals(k2) && v1.Equals(v2));
// original and compressed-then-decompressed-output have the same hashCode
IDictionary <RandomDatum, string> m = new Dictionary <RandomDatum, string>();
m[k1] = k1.ToString();
m[v1] = v1.ToString();
string result = m[k2];
Assert.Equal("k1 and k2 hashcode not equal", result, k1.ToString
());
result = m[v2];
Assert.Equal("v1 and v2 hashcode not equal", result, v1.ToString
());
}
// De-compress data byte-at-a-time
originalData.Reset(data.GetData(), 0, data.GetLength());
deCompressedDataBuffer.Reset(compressedDataBuffer.GetData(), 0, compressedDataBuffer
.GetLength());
inflateFilter = codec.CreateInputStream(deCompressedDataBuffer);
// Check
originalIn = new DataInputStream(new BufferedInputStream(originalData));
int expected;
do
{
expected = originalIn.Read();
Assert.Equal("Inflated stream read by byte does not match", expected
, inflateFilter.Read());
}while (expected != -1);
Log.Info("SUCCESS! Completed checking " + count + " records");
}
/// <summary>A little test program.</summary>
/// <param name="args"/>
/// <exception cref="System.Exception"/>
public static void Main(string[] args)
{
Configuration conf = new Configuration();
Org.Apache.Hadoop.IO.Compress.CompressionCodecFactory factory = new Org.Apache.Hadoop.IO.Compress.CompressionCodecFactory
(conf);
bool 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.Console.Out.WriteLine("Codec for " + args[i] + " not found.");
}
else
{
if (encode)
{
CompressionOutputStream @out = null;
InputStream @in = null;
try
{
@out = codec.CreateOutputStream(new FileOutputStream(args[i]));
byte[] buffer = new byte[100];
string inFilename = RemoveSuffix(args[i], codec.GetDefaultExtension());
@in = new FileInputStream(inFilename);
int len = @in.Read(buffer);
while (len > 0)
{
@out.Write(buffer, 0, len);
len = @in.Read(buffer);
}
}
finally
{
if (@out != null)
{
@out.Close();
}
if (@in != null)
{
@in.Close();
}
}
}
else
{
CompressionInputStream @in = null;
try
{
@in = codec.CreateInputStream(new FileInputStream(args[i]));
byte[] buffer = new byte[100];
int len = @in.Read(buffer);
while (len > 0)
{
System.Console.Out.Write(buffer, 0, len);
len = @in.Read(buffer);
}
}
finally
{
if (@in != null)
{
@in.Close();
}
}
}
}
}
}
}
}