def test_compression_equal_regardless_parallel_mode():
a = [b'spam '] * 666
with patch('arctic._compression.ENABLE_PARALLEL', True):
parallel = compress_array(a)
with patch('arctic._compression.ENABLE_PARALLEL', False):
serial = compress_array(a)
assert serial == parallel
def test_compression_equal_regardless_parallel_mode():
a = [b'spam '] * 666
with patch('arctic._compression.ENABLE_PARALLEL', True):
parallel = compress_array(a)
with patch('arctic._compression.ENABLE_PARALLEL', False):
serial = compress_array(a)
assert serial == parallel
def test_compress_array_no_parallel():
a = [b'spam', b'egg', b'spamm', b'spammm']
with patch('arctic._compression.clz4', sentinel.clz4), \
patch('arctic._compression.ENABLE_PARALLEL', False), \
patch('arctic._compression.lz4', wraps=lz4) as patch_lz4:
assert decompress_array(compress_array(a)) == a
assert patch_lz4.compress.call_args_list == [call(x) for x in a]
assert patch_lz4.decompress.call_args_list == [call(compress(x)) for x in a]
def test_compress_array_no_parallel():
a = ["spam", "egg", "spamm", "spammm"]
with patch("arctic._compression.clz4", sentinel.clz4), patch("arctic._compression.ENABLE_PARALLEL", False), patch(
"arctic._compression.lz4", wraps=lz4
) as patch_lz4:
assert decompress_array(compress_array(a)) == a
assert patch_lz4.compress.call_args_list == [call(x) for x in a]
assert patch_lz4.decompress.call_args_list == [call(compress(x)) for x in a]
def test_compress_array_no_parallel():
a = [b'spam', b'egg', b'spamm', b'spammm']
with patch('arctic._compression.clz4', sentinel.clz4), \
patch('arctic._compression.ENABLE_PARALLEL', False), \
patch('arctic._compression.lz4', wraps=lz4) as patch_lz4:
assert decompress_array(compress_array(a)) == a
assert patch_lz4.compress.call_args_list == [call(x) for x in a]
assert patch_lz4.decompress.call_args_list == [
call(compress(x)) for x in a
]
def _do_write(self,
backing_store,
library_name,
version,
symbol,
item,
previous_version,
segment_offset=0):
previous_segment_keys = []
if previous_version:
previous_segment_keys = previous_version['segment_keys']
if segment_offset > 0 and 'segment_index' in previous_version:
existing_index = previous_version['segment_index']
else:
existing_index = None
sze = int(item.dtype.itemsize * np.prod(item.shape[1:]))
length = len(item)
# chunk and store the data by (uncompressed) size
chunk_size = int(backing_store.chunk_size / sze)
# Compress
idxs = xrange(int(np.ceil(float(length) / chunk_size)))
chunks = [(item[i * chunk_size:(i + 1) * chunk_size]).tostring()
for i in idxs]
compressed_segments = compress_array(chunks)
segment_keys = []
raw_segment_index = []
for i, segment_data in zip(idxs, compressed_segments):
segment_idx = min(
(i + 1) * chunk_size - 1, length - 1) + segment_offset
segment_key = backing_store.write_segment(library_name, symbol,
segment_data,
previous_segment_keys)
raw_segment_index.append(segment_idx)
segment_keys.append(segment_key)
segment_index = self._segment_index(item,
existing_index=existing_index,
start=segment_offset,
new_segments=raw_segment_index)
if segment_index:
version['segment_index'] = segment_index
version['raw_segment_index'] = raw_segment_index
version['segment_count'] = len(
segment_keys
) # on appends this value is incorrect but is updated later on
version['append_size'] = 0
version['append_count'] = 0
version['segment_keys'] = segment_keys
def bench_multi(repeats, _strarr, use_HC, pool=None):
measurements = []
for j in range(repeats):
now = dt.now()
if pool:
# Raw LZ4 lib
if use_HC:
res = pool.map(c.lz4_compressHC, _strarr)
else:
res = pool.map(c.lz4_compress, _strarr)
else:
# Arctic's compression layer
if use_HC:
res = c.compressHC_array(_strarr)
else:
res = c.compress_array(_strarr, withHC=False)
sample = (dt.now() - now).total_seconds()
assert len(res) == len(_strarr)
assert all(res)
measurements.append(sample)
return measurements
def bench_multi(repeats, _strarr, use_HC, pool=None):
measurements = []
for j in range(repeats):
now = dt.now()
if pool:
# Raw LZ4 lib
if use_HC:
res = pool.map(c.lz4_compressHC, _strarr)
else:
res = pool.map(c.lz4_compress, _strarr)
else:
# Arctic's compression layer
if use_HC:
res = c.compressHC_array(_strarr)
else:
res = c.compress_array(_strarr, withHC=False)
sample = (dt.now() - now).total_seconds()
assert len(res) == len(_strarr)
assert all(res)
measurements.append(sample)
return measurements
def test_decompress_array():
ll = ['foo%s' % i for i in range(100)]
assert decompress_array(compress_array(ll)) == ll
def test_compress_array_LZ4_sequential():
use_lz4hc(False)
cfn = Mock()
with patch('arctic._compression.clz4.compress', cfn):
compress_array(["foo"] * 49)
assert cfn.call_count == 49
def test_compress_array_usesLZ4():
use_lz4hc(False)
cfn = Mock()
with patch('arctic._compression.clz4.compressarr', cfn):
compress_array(["foo"] * 100)
assert cfn.call_count == 1
def test_compressarr_LZ4HC():
assert len(compress_array(["foobar" * 10])) > 0
assert isinstance(compress_array(["foobar" * 10]), list)
def test_decompress_array():
ll = ['foo%s' % i for i in range(100)]
assert decompress_array(compress_array(ll)) == ll
def test_compress_array_LZ4_sequential():
cfn = Mock()
with patch('arctic._compression.lz4_compress', cfn):
compress_array([b"foo"] * 49)
assert len(cfn.call_args_list) == 49
def test_compress_array_LZ4_sequential():
cfn = Mock()
with patch('arctic._compression.clz4.compress', cfn):
compress_array([b"foo"] * 49)
assert cfn.call_count == 49
def test_decompress_array():
ll = [('foo%s' % i).encode('ascii') for i in range(100)]
assert decompress_array(compress_array(ll)) == ll
def test_compress_array_LZ4_sequential():
cfn = Mock()
with patch('arctic._compression.lz4_compress', cfn):
compress_array([b"foo"] * 49)
assert len(cfn.call_args_list) == 49
def test_compress_array_usesLZ4():
cfn = Mock()
with patch('arctic._compression.lz4_compress', cfn):
compress_array([b"foo"] * 100)
assert len(cfn.call_args_list) == 100 # call_count is not thread safe
def test_roundtrip_arr(n, length):
_strarr = [random_string(length) for _ in range(n)]
cstr = c.compress_array(_strarr)
assert _strarr == c.decompress_array(cstr)
def test_compressarr():
assert len(compress_array([b"foobar"*10])) > 0
assert isinstance(compress_array([b"foobar"*10]), list)
def test_compress_array_usesLZ4():
cfn = Mock()
with patch('arctic._compression.clz4.compressarr', cfn):
compress_array([b"foo"] * 100)
assert cfn.call_count == 1
def test_compress_array_usesLZ4():
cfn = Mock()
with patch('arctic._compression.lz4_compress', cfn):
compress_array([b"foo"] * 100)
assert len(cfn.call_args_list) == 100 # call_count is not thread safe
def test_decompress_array():
ll = [('foo%s' % i).encode('ascii') for i in range(100)]
assert decompress_array(compress_array(ll)) == ll
def test_compress_array_usesLZ4HC():
use_lz4hc(True)
cfn = Mock()
with patch('arctic._compression.clz4.compressarrHC', cfn):
compress_array(["foo"] * 100)
assert cfn.call_count == 1
def test_roundtrip_arr(n, length):
_strarr = [random_string(length) for _ in range(n)]
cstr = c.compress_array(_strarr)
assert _strarr == c.decompress_array(cstr)
def test_compress_array_LZ4HC_sequential():
use_lz4hc(True)
cfn = Mock()
with patch('arctic._compression.clz4.compressHC', cfn):
compress_array(["foo"] * 4)
assert cfn.call_count == 4
