def test_different_compression():
all_ascii = (string.ascii_letters + string.digits).encode()
v1 = all_ascii * 1000
v2 = all_ascii * 900 + string.ascii_letters.encode() * 100 + all_ascii * 100
delta_a = xdelta3.encode(v1, v2, xdelta3.Flags.COMPLEVEL_9)
v2_a = xdelta3.decode(v1, delta_a)
assert v2 == v2_a
delta_b = xdelta3.encode(v1, v2, xdelta3.Flags.COMPLEVEL_1)
v2_b = xdelta3.decode(v1, delta_b)
assert v2 == v2_b
assert len(delta_a) < len(delta_b)
def write_content_diff(out, in1, in2):
content_id_map = {
iterdiff.PROLOGUE: CONTENT_PROLOGUE,
iterdiff.EPILOGUE: CONTENT_EPILOGUE,
CONTENT_PROLOGUE: None, # fail-fast on struct.pack
CONTENT_EPILOGUE: None, # fail-fast on struct.pack
}
# Default compresslevel of 9 is okay as gzip costs nothing compared to XML parsing of source.
# Note, xdelta from b'' may be still smaller than the original blob.
with gzip.GzipFile(fileobj=out, mode='wb') as gz:
for k, v1, v2 in iterdiff.iter_content_diff(in1, in2):
k = content_id_map.get(k, k)
try:
delta = xdelta3.encode(v1, v2)
# FIXME: workaround for bug https://github.com/samuelcolvin/xdelta3-python/issues/2
try:
good = xdelta3.decode(v1,
delta) == v2 # Epic Fail if happens.
except xdelta3.XDeltaError:
good = False
magic = MAGIC_XDELTA3 if good else MAGIC_UGLY
except xdelta3.NoDeltaFound:
magic, delta = MAGIC_RAW, None
blob = delta if magic == MAGIC_XDELTA3 else v2
gz.write(CDIFF_HEAD.pack(magic, k, len(v1), len(blob)))
gz.write(v1)
gz.write(blob)
def test_readme():
value_one = b'wonderful string to demonstrate xdelta3, much of these two strings is the same.'
value_two = b'different string to demonstrate xdelta3, much of these two strings is the same.'
delta = xdelta3.encode(value_one, value_two)
value_two_rebuilt = xdelta3.decode(value_one, delta)
assert value_two_rebuilt == value_two
def xdelta3gen(source_bytes, target_bytes):
#if source bytes finished, everything remaining is a target bytes
if not source_bytes:
sources_bytes = b''
#if target bytes finished, we should stop
if not target_bytes:
return None
#these are memoryviews and python xdelta3 lib doesn't like that...
return xdelta3.encode(bytes(source_bytes), bytes(target_bytes))
def test_large_decode():
this_dir = Path(__file__).parent
try:
b1 = (this_dir / 'b1.bin').read_bytes()
b2 = (this_dir / 'b2.bin').read_bytes()
except FileNotFoundError as e:
raise RuntimeError(
'file required for test not found, run `make download-test-files`'
) from e
d = xdelta3.encode(b1, b2)
b3 = xdelta3.decode(b1, d)
assert b2 == b3
curl https://ocw.mit.edu/ans7870/6/6.006/s08/lecturenotes/files/t8.shakespeare.txt > shakespeare.txt
cp shakespeare.txt shakespeare_changed.txt
vim shakespeare_changed.txt (and make some changes to shakespeare_changed.txt)
python performance.py
"""
from pathlib import Path
from statistics import mean, stdev
from time import time
import xdelta3
v1 = Path('shakespeare.txt').read_bytes()
v2 = Path('shakespeare_changed.txt').read_bytes()
times = []
for i in range(50):
start = time()
delta = xdelta3.encode(v1, v2, xdelta3.Flags.COMPLEVEL_1)
v22 = xdelta3.decode(v1, delta)
time_taken = (time() - start) * 1000
times.append(time_taken)
print(f'{i + 1:3} result_match={v2 == v22} time={time_taken:0.1f}ms')
print(f'\noriginal length: {len(v1)}')
print(f'changed length: {len(v2)}')
print(f'delta length: {len(delta)}')
print(
f'mean time taken to encode and decode: {mean(times):0.3f}ms, stdev {stdev(times):0.3f}ms'
)
def test_no_delta():
with pytest.raises(xdelta3.NoDeltaFound) as exc_info:
xdelta3.encode(b'hello', b'goodbye')
assert exc_info.value.args[
0] == 'No delta found shorter than the input value'
def test_long_random():
v1 = base64.b32encode(os.urandom(1000))
v2 = b'x' + v1 + b'x'
delta = xdelta3.encode(v1, v2)
v22 = xdelta3.decode(v1, delta)
assert v2 == v22
def test_encode_decode():
delta = xdelta3.encode(value_one, value_two)
assert delta == expected_delta
value_two2 = xdelta3.decode(value_one, delta)
assert value_two == value_two2