def calculate_md5(file_path, offset, chunk):
"""
Calculate the base64 encoded MD5 hash for a part of a file.
:param str file_path: Path to the file
:param int offset: Offset where the part starts in the file
:param int chunk: Number of bytes in the part
:return: Base64 encoded MD5 hash
:rtype: str
"""
# Determine if we can use the hashlib version of md5 or the bundled
# version of md5
if is_fips_mode():
md5 = MD5.md5
else:
md5 = hashlib.md5
m = md5()
with io.open(file_path, mode='rb') as f:
bpr = BufferedPartReader(f, offset, chunk)
while True:
part = bpr.read(READ_BUFFER_SIZE)
if part == b'':
break
m.update(part)
return base64.b64encode(m.digest()).decode("utf-8")
def calculate_md5(file_path, offset, chunk):
"""
Calculate the base64 encoded MD5 hash for a part of a file.
:param str file_path: Path to the file
:param int offset: Offset where the part starts in the file
:param int chunk: Number of bytes in the part
:return: Base64 encoded MD5 hash
:rtype: str
"""
# Determine if we can use the hashlib version of md5 or the bundled
# version of md5
if is_fips_mode():
try:
md5 = MD5.md5()
except InternalError as ex:
logger.warning(
"An exception occur due to {}. Fallback to using hashlib.new('md5', usedforsecurity=false) for md5."
.format(ex))
md5 = hashlib.new('md5', usedforsecurity=False)
else:
md5 = hashlib.md5()
m = md5
with io.open(file_path, mode='rb') as f:
bpr = BufferedPartReader(f, offset, chunk)
while True:
part = bpr.read(READ_BUFFER_SIZE)
if part == b'':
break
m.update(part)
return base64.b64encode(m.digest()).decode("utf-8")
def _upload_stream_part(self, part_num, part_bytes, **kwargs):
try:
if is_fips_mode():
try:
m = MD5.md5()
except InternalError as ex:
logger.warning(
"An exception occur due to {}. Fallback to using hashlib.new('md5', usedforsecurity=false) for md5."
.format(ex))
m = hashlib.new('md5', usedforsecurity=False)
else:
m = hashlib.md5()
m.update(part_bytes)
new_kwargs = {
'content_md5': base64.b64encode(m.digest()).decode("utf-8")
}
if 'opc_client_request_id' in kwargs:
new_kwargs['opc_client_request_id'] = kwargs[
'opc_client_request_id']
# A call to upload_part can be retried because we've already read() the data and retrying the
# request with same data part_bytes
remaining_tries = self.max_retries
while remaining_tries > 0:
try:
response = self.object_storage_client.upload_part(
self.manifest["namespace"],
self.manifest["bucketName"],
self.manifest["objectName"],
self.manifest["uploadId"],
part_num +
1, # Internally this is 0-based but object storage is 1-based
io.BytesIO(part_bytes),
**new_kwargs)
except Exception as e:
if self._is_exception_retryable(e) and remaining_tries > 1:
remaining_tries -= 1
else:
if 'shared_dict' in kwargs:
kwargs['shared_dict']['should_continue'] = False
kwargs['shared_dict']['exceptions'].put(e)
raise
else:
break
if response.status == 200:
self.manifest['parts'].append({
'etag':
response.headers['etag'],
'opc_md5':
str(response.headers['opc-content-md5']),
'part_num':
part_num
})
if 'progress_callback' in kwargs:
kwargs['progress_callback'](len(part_bytes))
except Exception as e:
if 'shared_dict' in kwargs:
kwargs['shared_dict']['should_continue'] = False
kwargs['shared_dict']['exceptions'].put(e)
raise
finally:
if 'semaphore' in kwargs:
kwargs['semaphore'].release()
