def _decompress(self, key_offset):
# Copy of `_read_key_value`, but uses memoryview
pos = key_offset
key_size = struct.unpack_from(">i", self._buffer, pos)[0]
pos += self.KEY_LENGTH
if key_size != -1:
pos += key_size
value_size = struct.unpack_from(">i", self._buffer, pos)[0]
pos += self.VALUE_LENGTH
if value_size == -1:
raise CorruptRecordException("Value of compressed message is None")
else:
data = self._buffer[pos:pos + value_size]
compression_type = self.compression_type
if compression_type == self.CODEC_GZIP:
uncompressed = gzip_decode(data)
elif compression_type == self.CODEC_SNAPPY:
uncompressed = snappy_decode(data.tobytes())
elif compression_type == self.CODEC_LZ4:
if self._magic == 0:
uncompressed = lz4_decode_old_kafka(data.tobytes())
else:
uncompressed = lz4_decode(data.tobytes())
return uncompressed
def _decompress(self, key_offset):
# Copy of `_read_key_value`, but uses memoryview
pos = key_offset
key_size = struct.unpack_from(">i", self._buffer, pos)[0]
pos += self.KEY_LENGTH
if key_size != -1:
pos += key_size
value_size = struct.unpack_from(">i", self._buffer, pos)[0]
pos += self.VALUE_LENGTH
if value_size == -1:
raise CorruptRecordException("Value of compressed message is None")
else:
data = self._buffer[pos:pos + value_size]
compression_type = self.compression_type
if compression_type == self.CODEC_GZIP:
uncompressed = gzip_decode(data)
elif compression_type == self.CODEC_SNAPPY:
uncompressed = snappy_decode(data.tobytes())
elif compression_type == self.CODEC_LZ4:
if self._magic == 0:
uncompressed = lz4_decode_old_kafka(data.tobytes())
else:
uncompressed = lz4_decode(data.tobytes())
return uncompressed
def decompress(self):
codec = self.attributes & self.CODEC_MASK
assert codec in (
self.CODEC_GZIP,
self.CODEC_SNAPPY,
self.CODEC_LZ4,
self.CODEC_ZSTD,
)
if codec == self.CODEC_GZIP:
assert has_gzip(), "Gzip decompression unsupported"
raw_bytes = gzip_decode(self.value)
elif codec == self.CODEC_SNAPPY:
assert has_snappy(), "Snappy decompression unsupported"
raw_bytes = snappy_decode(self.value)
elif codec == self.CODEC_LZ4:
assert has_lz4(), "LZ4 decompression unsupported"
if self.magic == 0:
raw_bytes = lz4_decode_old_kafka(self.value)
else:
raw_bytes = lz4_decode(self.value)
elif codec == self.CODEC_ZSTD:
assert has_zstd(), "ZSTD decompression unsupported"
raw_bytes = zstd_decode(self.value)
else:
raise Exception("This should be impossible")
return MessageSet.decode(raw_bytes, bytes_to_read=len(raw_bytes))
def test_lz4_incremental():
for i in range(1000):
# lz4 max single block size is 4MB
# make sure we test with multiple-blocks
b1 = random_string(100).encode('utf-8') * 50000
b2 = lz4_decode(lz4_encode(b1))
assert len(b1) == len(b2)
assert b1 == b2
def test_lz4_incremental():
for i in xrange(1000):
# lz4 max single block size is 4MB
# make sure we test with multiple-blocks
b1 = random_string(100).encode('utf-8') * 50000
b2 = lz4_decode(lz4_encode(b1))
assert len(b1) == len(b2)
assert b1 == b2
def _maybe_uncompress(self):
if not self._decompressed:
compression_type = self.compression_type
if compression_type != self.CODEC_NONE:
data = memoryview(self._buffer)[self._pos:]
if compression_type == self.CODEC_GZIP:
uncompressed = gzip_decode(data)
if compression_type == self.CODEC_SNAPPY:
uncompressed = snappy_decode(data.tobytes())
if compression_type == self.CODEC_LZ4:
uncompressed = lz4_decode(data.tobytes())
self._buffer = bytearray(uncompressed)
self._pos = 0
self._decompressed = True
def decompress(self):
codec = self.attributes & self.CODEC_MASK
assert codec in (self.CODEC_GZIP, self.CODEC_SNAPPY, self.CODEC_LZ4)
if codec == self.CODEC_GZIP:
assert has_gzip(), 'Gzip decompression unsupported'
raw_bytes = gzip_decode(self.value)
elif codec == self.CODEC_SNAPPY:
assert has_snappy(), 'Snappy decompression unsupported'
raw_bytes = snappy_decode(self.value)
elif codec == self.CODEC_LZ4:
assert has_lz4(), 'LZ4 decompression unsupported'
if self.magic == 0:
raw_bytes = lz4_decode_old_kafka(self.value)
else:
raw_bytes = lz4_decode(self.value)
else:
raise Exception('This should be impossible')
return MessageSet.decode(raw_bytes, bytes_to_read=len(raw_bytes))
def test_lz4():
for i in range(1000):
b1 = random_string(100).encode('utf-8')
b2 = lz4_decode(lz4_encode(b1))
assert len(b1) == len(b2)
assert b1 == b2
def test_lz4():
for i in xrange(1000):
b1 = random_string(100).encode('utf-8')
b2 = lz4_decode(lz4_encode(b1))
assert len(b1) == len(b2)
assert b1 == b2
