def test_perform_block_create(app):
content = 'foo'
with freeze_time('2012-01-01') as frozen_datetime:
eff = app.perform_block_create(EBlockCreate(content))
block_id = perform_sequence([], eff)
assert app.last_modified == Arrow.fromdatetime(frozen_datetime())
eff = app.perform_block_read(EBlockRead(block_id))
block = perform_sequence([], eff)
assert block['content'] == content
def test_perform_block_read(app, app_no_cache):
local_content = 'foo'
eff = app.perform_block_create(EBlockCreate(local_content))
block_id = perform_sequence([], eff)
# Read block in new blocks
eff = app.perform_block_read(EBlockRead(block_id))
block = perform_sequence([], eff)
assert sorted(list(block.keys())) == ['content', 'id']
assert block['id']
assert block['content'] == local_content
remote_content = b'bar'
# Read remote block
assert app.block_cache.currsize == 0
eff = app.perform_block_read(EBlockRead('123'))
sequence = [(EBackendBlockRead('123'), const(Block('123',
remote_content)))]
block = perform_sequence(sequence, eff)
assert sorted(list(block.keys())) == ['content', 'id']
assert block['id']
assert block['content'] == remote_content
assert app.block_cache.currsize == 1
# Read remote block with cache disabled
assert app_no_cache.block_cache.currsize == 0
eff = app_no_cache.perform_block_read(EBlockRead('123'))
sequence = [(EBackendBlockRead('123'), const(Block('123',
remote_content)))]
block = perform_sequence(sequence, eff)
assert sorted(list(block.keys())) == ['content', 'id']
assert block['id']
assert block['content'] == remote_content
assert app_no_cache.block_cache.currsize == 0
# Read block in cache
eff = app.perform_block_read(EBlockRead('123'))
block = perform_sequence([], eff)
assert sorted(list(block.keys())) == ['content', 'id']
assert block['id']
assert block['content'] == remote_content
# Delete block from cache
eff = app.perform_block_delete(EBlockDelete('123'))
perform_sequence([], eff)
# Not found
eff = app.perform_block_read(EBlockRead('123'))
sequence = [(EBackendBlockRead('123'),
conste(BlockNotFound('Block not found.')))]
with pytest.raises(BlockNotFound):
block = perform_sequence(sequence, eff)
eff = app.perform_block_read(EBlockRead('123'))
sequence = [(EBackendBlockRead('123'), conste(BlockError('Block error.'))
) # TODO keep it? usefull with multiple backends...
]
with pytest.raises(BlockNotFound):
block = perform_sequence(sequence, eff)
def read(self, size=None, offset=0):
yield self.flush()
# Get data
matching_blocks = yield self._find_matching_blocks(size, offset)
data = matching_blocks['pre_included_data']
for blocks_and_key in matching_blocks['included_blocks']:
block_key = blocks_and_key['key']
decoded_block_key = from_jsonb64(block_key)
encryptor = load_sym_key(decoded_block_key)
for block_properties in blocks_and_key['blocks']:
block = yield Effect(EBlockRead(block_properties['block']))
# Decrypt
# TODO: clean this hack
if isinstance(block['content'], str):
block_content = from_jsonb64(block['content'])
else:
block_content = from_jsonb64(block['content'].decode())
chunk_data = encryptor.decrypt(block_content)
# Check integrity
assert digest(chunk_data) == block_properties['digest']
assert len(chunk_data) == block_properties['size']
data += chunk_data
data += matching_blocks['post_included_data']
return data
def _find_matching_blocks(self, size=None, offset=0):
if size is None:
size = sys.maxsize
pre_excluded_blocks = []
post_excluded_blocks = []
version = self.get_version()
vlob = yield Effect(EVlobRead(self.id, self.read_trust_seed, version))
blob = vlob['blob']
encrypted_blob = from_jsonb64(blob)
blob = self.encryptor.decrypt(encrypted_blob)
blob = ejson_loads(blob.decode())
pre_excluded_blocks = []
included_blocks = []
post_excluded_blocks = []
cursor = 0
pre_excluded_data = b''
pre_included_data = b''
post_included_data = b''
post_excluded_data = b''
for blocks_and_key in blob:
block_key = blocks_and_key['key']
decoded_block_key = from_jsonb64(block_key)
encryptor = load_sym_key(decoded_block_key)
for block_properties in blocks_and_key['blocks']:
cursor += block_properties['size']
if cursor <= offset:
if len(pre_excluded_blocks) and pre_excluded_blocks[-1]['key'] == block_key:
pre_excluded_blocks[-1]['blocks'].append(block_properties)
else:
pre_excluded_blocks.append({'blocks': [block_properties], 'key': block_key})
elif cursor > offset and cursor - block_properties['size'] < offset:
delta = cursor - offset
block = yield Effect(EBlockRead(block_properties['block']))
content = from_jsonb64(block['content'])
block_data = encryptor.decrypt(content)
pre_excluded_data = block_data[:-delta]
pre_included_data = block_data[-delta:][:size]
if size < len(block_data[-delta:]):
post_excluded_data = block_data[-delta:][size:]
elif cursor > offset and cursor <= offset + size:
if len(included_blocks) and included_blocks[-1]['key'] == block_key:
included_blocks[-1]['blocks'].append(block_properties)
else:
included_blocks.append({'blocks': [block_properties], 'key': block_key})
elif cursor > offset + size and cursor - block_properties['size'] < offset + size:
delta = offset + size - (cursor - block_properties['size'])
block = yield Effect(EBlockRead(block_properties['block']))
content = from_jsonb64(block['content'])
block_data = encryptor.decrypt(content)
post_included_data = block_data[:delta]
post_excluded_data = block_data[delta:]
else:
if len(post_excluded_blocks) and post_excluded_blocks[-1]['key'] == block_key:
post_excluded_blocks[-1]['blocks'].append(block_properties)
else:
post_excluded_blocks.append({'blocks': [block_properties],
'key': block_key})
return {
'pre_excluded_blocks': pre_excluded_blocks,
'pre_excluded_data': pre_excluded_data,
'pre_included_data': pre_included_data,
'included_blocks': included_blocks,
'post_included_data': post_included_data,
'post_excluded_data': post_excluded_data,
'post_excluded_blocks': post_excluded_blocks
}
def test_find_matching_blocks(self, file):
vlob_id = '1234'
block_size = 4096
# Contents
contents = {}
total_length = 0
for index, length in enumerate([
block_size + 1, block_size - 1, block_size, 2 * block_size + 2,
2 * block_size - 2, 2 * block_size
]):
content = b''.join(
[str(random.randint(1, 9)).encode() for i in range(0, length)])
contents[index] = content
total_length += length
# Blocks
def generator():
i = 2000
while True:
yield str(i)
i += 1
gen = generator()
blocks = {}
block_contents = {}
block_id = 2000
for index, content in contents.items():
chunks = [
content[i:i + block_size]
for i in range(0, len(content), block_size)
]
if not chunks:
chunks = [b'']
sequence = []
for chunk in chunks:
encoded_chunk = to_jsonb64(chunk)
sequence.append((EBlockCreate(encoded_chunk),
lambda id=id: next(gen))) # TODO dirty
block_contents[str(block_id)] = encoded_chunk
block_id += 1
blocks[index] = perform_sequence(sequence,
file._build_file_blocks(content))
# Create file
blob = ejson_dumps([blocks[i] for i in range(0, len(blocks))]).encode()
blob = to_jsonb64(blob)
# All matching blocks
sequence = [(EVlobRead(vlob_id, '42', 1),
const({
'id': vlob_id,
'blob': blob,
'version': 1
}))]
matching_blocks = perform_sequence(sequence,
file._find_matching_blocks())
assert matching_blocks == {
'pre_excluded_blocks': [],
'pre_excluded_data': b'',
'pre_included_data': b'',
'included_blocks': [blocks[i] for i in range(0, len(blocks))],
'post_included_data': b'',
'post_excluded_data': b'',
'post_excluded_blocks': []
}
# With offset
delta = 10
offset = (blocks[0]['blocks'][0]['size'] +
blocks[0]['blocks'][1]['size'] +
blocks[1]['blocks'][0]['size'] +
blocks[2]['blocks'][0]['size'] - delta)
sequence = [(EVlobRead(vlob_id, '42', 1),
const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
(EBlockRead('2003'),
const({
'content': block_contents['2003'],
'creation_date': '2012-01-01T00:00:00'
}))]
matching_blocks = perform_sequence(
sequence, file._find_matching_blocks(None, offset))
pre_excluded_data = contents[2][:blocks[2]['blocks'][0]['size'] -
delta]
pre_included_data = contents[2][-delta:]
assert matching_blocks == {
'pre_excluded_blocks': [blocks[0], blocks[1]],
'pre_excluded_data': pre_excluded_data,
'pre_included_data': pre_included_data,
'included_blocks': [blocks[i] for i in range(3, 6)],
'post_included_data': b'',
'post_excluded_data': b'',
'post_excluded_blocks': []
}
# With small size
delta = 10
size = 5
offset = (blocks[0]['blocks'][0]['size'] +
blocks[0]['blocks'][1]['size'] +
blocks[1]['blocks'][0]['size'] +
blocks[2]['blocks'][0]['size'] - delta)
sequence = [(EVlobRead(vlob_id, '42', 1),
const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
(EBlockRead(id='2003'),
const({
'content': block_contents['2003'],
'creation_date': '2012-01-01T00:00:00'
}))]
matching_blocks = perform_sequence(
sequence, file._find_matching_blocks(size, offset))
pre_excluded_data = contents[2][:blocks[2]['blocks'][0]['size'] -
delta]
pre_included_data = contents[2][-delta:][:size]
post_excluded_data = contents[2][-delta:][size:]
assert matching_blocks == {
'pre_excluded_blocks': [blocks[0], blocks[1]],
'pre_excluded_data': pre_excluded_data,
'pre_included_data': pre_included_data,
'included_blocks': [],
'post_included_data': b'',
'post_excluded_data': post_excluded_data,
'post_excluded_blocks': [blocks[i] for i in range(3, 6)]
}
# With big size
delta = 10
size = delta
size += blocks[3]['blocks'][0]['size']
size += blocks[3]['blocks'][1]['size']
size += blocks[3]['blocks'][2]['size']
size += 2 * delta
offset = (blocks[0]['blocks'][0]['size'] +
blocks[0]['blocks'][1]['size'] +
blocks[1]['blocks'][0]['size'] +
blocks[2]['blocks'][0]['size'] - delta)
sequence = [(EVlobRead(vlob_id, '42', 1),
const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
(EBlockRead('2003'),
const({
'content': block_contents['2003'],
'creation_date': '2012-01-01T00:00:00'
})),
(EBlockRead('2007'),
const({
'content': block_contents['2007'],
'creation_date': '2012-01-01T00:00:00'
}))]
matching_blocks = perform_sequence(
sequence, file._find_matching_blocks(size, offset))
pre_excluded_data = contents[2][:-delta]
pre_included_data = contents[2][-delta:]
post_included_data = contents[4][:2 * delta]
post_excluded_data = contents[4][:block_size][2 * delta:]
partial_block_4 = deepcopy(blocks[4])
del partial_block_4['blocks'][0]
assert matching_blocks == {
'pre_excluded_blocks': [blocks[0], blocks[1]],
'pre_excluded_data': pre_excluded_data,
'pre_included_data': pre_included_data,
'included_blocks': [blocks[3]],
'post_included_data': post_included_data,
'post_excluded_data': post_excluded_data,
'post_excluded_blocks': [partial_block_4, blocks[5]]
}
# With big size and no delta
size = blocks[3]['blocks'][0]['size']
size += blocks[3]['blocks'][1]['size']
size += blocks[3]['blocks'][2]['size']
offset = (blocks[0]['blocks'][0]['size'] +
blocks[0]['blocks'][1]['size'] +
blocks[1]['blocks'][0]['size'] +
blocks[2]['blocks'][0]['size'])
sequence = [
(EVlobRead(vlob_id, '42',
1), const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
]
matching_blocks = perform_sequence(
sequence, file._find_matching_blocks(size, offset))
assert matching_blocks == {
'pre_excluded_blocks': [blocks[0], blocks[1], blocks[2]],
'pre_excluded_data': b'',
'pre_included_data': b'',
'included_blocks': [blocks[3]],
'post_included_data': b'',
'post_excluded_data': b'',
'post_excluded_blocks': [blocks[4], blocks[5]]
}
# With total size
sequence = [
(EVlobRead(vlob_id, '42',
1), const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
]
matching_blocks = perform_sequence(
sequence, file._find_matching_blocks(total_length, 0))
assert matching_blocks == {
'pre_excluded_blocks': [],
'pre_excluded_data': b'',
'pre_included_data': b'',
'included_blocks': [blocks[i] for i in range(0, 6)],
'post_included_data': b'',
'post_excluded_data': b'',
'post_excluded_blocks': []
}
def test_commit(self, file):
vlob_id = '1234'
content = b'This is a test content.'
block_ids = ['4567', '5678', '6789']
new_vlob = {
'id': '2345',
'read_trust_seed': 'ABC',
'write_trust_seed': 'DEF'
}
# Original content
chunk_1 = content[:5]
chunk_2 = content[5:14]
chunk_3 = content[14:]
blob = [{
'blocks': [{
'block': block_ids[0],
'digest': digest(chunk_1),
'size': len(chunk_1)
}, {
'block': block_ids[1],
'digest': digest(chunk_2),
'size': len(chunk_2)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000003>')
}, {
'blocks': [{
'block': block_ids[2],
'digest': digest(chunk_3),
'size': len(chunk_3)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000004>')
}]
blob = ejson_dumps(blob).encode()
blob = to_jsonb64(blob)
# New content after truncate
new_chuck_2 = b'is a'
new_block_id = '7654'
new_blob = [{
'blocks': [{
'block': block_ids[0],
'digest': digest(chunk_1),
'size': len(chunk_1)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000003>')
}, {
'blocks': [{
'block': new_block_id,
'digest': digest(new_chuck_2),
'size': len(new_chuck_2)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000003>')
}]
new_blob = ejson_dumps(new_blob).encode()
new_blob = to_jsonb64(new_blob)
file.truncate(9)
sequence = [
(EVlobRead('1234', '42',
1), const({
'id': '1234',
'blob': blob,
'version': 1
})),
(EVlobRead('1234', '42',
1), const({
'id': '1234',
'blob': blob,
'version': 1
})),
(EBlockRead(block_ids[1]),
const({
'content': to_jsonb64(chunk_2),
'creation_date': '2012-01-01T00:00:00'
})), (EBlockCreate(to_jsonb64(new_chuck_2)), const(new_block_id)),
(EVlobUpdate(vlob_id, '43', 1, new_blob), noop),
(EVlobRead('1234', '42',
1), const({
'id': '1234',
'blob': new_blob,
'version': 1
})),
(EBlockDelete('5678'), conste(BlockNotFound('Block not found.'))),
(EBlockDelete('6789'), noop),
(EVlobRead('1234', '42',
1), const({
'id': '1234',
'blob': new_blob,
'version': 1
})), (EBlockSynchronize('4567'), const(True)),
(EBlockSynchronize('7654'), const(False)),
(EVlobSynchronize('1234'), const(new_vlob))
]
ret = perform_sequence(sequence, file.commit())
new_vlob['key'] = to_jsonb64(b'<dummy-key-00000000000000000002>')
assert ret == new_vlob
assert file.dirty is False
assert file.version == 1
def test_flush(self, file):
file.truncate(9)
file.write(b'IS', 5)
file.write(b'IS a nice test content.', 5)
file.dirty = False
file.version = 2
vlob_id = '1234'
content = b'This is a test content.'
block_ids = ['4567', '5678', '6789']
# Original content
chunk_1 = content[:5]
chunk_2 = content[5:14]
chunk_3 = content[14:]
blob = [{
'blocks': [{
'block': block_ids[0],
'digest': digest(chunk_1),
'size': len(chunk_1)
}, {
'block': block_ids[1],
'digest': digest(chunk_2),
'size': len(chunk_2)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000001>')
}, {
'blocks': [{
'block': block_ids[2],
'digest': digest(chunk_3),
'size': len(chunk_3)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000002>')
}]
blob = ejson_dumps(blob).encode()
blob = to_jsonb64(blob)
# New content after truncate
new_chuck_2 = b'is a'
new_block_id = '7654'
new_blob = [{
'blocks': [{
'block': block_ids[0],
'digest': digest(chunk_1),
'size': len(chunk_1)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000001>')
}, {
'blocks': [{
'block': new_block_id,
'digest': digest(new_chuck_2),
'size': len(new_chuck_2)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000003>')
}]
new_blob = ejson_dumps(new_blob).encode()
new_blob = to_jsonb64(new_blob)
# New content after write
new_block_2_id = '6543'
new_chunk_4 = b'IS a nice test content.'
new_blob_2 = [{
'blocks': [{
'block': block_ids[0],
'digest': digest(chunk_1),
'size': len(chunk_1)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000001>')
}, {
'blocks': [{
'block': new_block_2_id,
'digest': digest(new_chunk_4),
'size': len(new_chunk_4)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000004>')
}]
new_blob_2 = ejson_dumps(new_blob_2).encode()
new_blob_2 = to_jsonb64(new_blob_2)
sequence = [
(
EVlobRead(vlob_id, '42', 2), # Get blocks
const({
'id': vlob_id,
'blob': blob,
'version': 2
})),
(
EVlobRead(vlob_id, '42', 2), # Matching blocks
const({
'id': vlob_id,
'blob': blob,
'version': 2
})),
(EBlockRead(block_ids[1]),
const({
'content': to_jsonb64(chunk_2),
'creation_date': '2012-01-01T00:00:00'
})),
(EBlockCreate(to_jsonb64(new_chuck_2)), const(new_block_id)),
(EVlobUpdate(vlob_id, '43', 3, new_blob), noop),
(
EVlobRead(vlob_id, '42', 3), # Matching blocks
const({
'id': vlob_id,
'blob': new_blob,
'version': 3
})),
(EBlockCreate(to_jsonb64(new_chunk_4)), const(new_block_2_id)),
(EVlobUpdate(vlob_id, '43', 3, new_blob_2), noop),
(EVlobRead(vlob_id, '42', 3),
const({
'id': vlob_id,
'blob': new_blob_2,
'version': 3
})),
(EBlockDelete('5678'), conste(BlockNotFound('Block not found.'))),
(EBlockDelete('6789'), noop),
]
ret = perform_sequence(sequence, file.flush())
assert ret is None
assert file.dirty is True
assert file.version == 2
def test_read(self, file):
file.dirty = False
file.version = 1
# Empty file
vlob_id = '1234'
chunk_digest = digest(b'')
blob = [{
'blocks': [{
'block': '4567',
'digest': chunk_digest,
'size': 0
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000001>')
}]
blob = ejson_dumps(blob).encode()
blob = to_jsonb64(blob)
sequence = [
(EVlobRead(vlob_id, '42',
1), const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
]
read_content = perform_sequence(sequence, file.read())
assert read_content == b''
# Not empty file
content = b'This is a test content.'
block_ids = ['4567', '5678', '6789']
chunk_1 = content[:5]
chunk_2 = content[5:14]
chunk_3 = content[14:]
blob = [{
'blocks': [{
'block': block_ids[0],
'digest': digest(chunk_1),
'size': len(chunk_1)
}, {
'block': block_ids[1],
'digest': digest(chunk_2),
'size': len(chunk_2)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000001>')
}, {
'blocks': [{
'block': block_ids[2],
'digest': digest(chunk_3),
'size': len(chunk_3)
}],
'key':
to_jsonb64(b'<dummy-key-00000000000000000002>')
}]
blob = ejson_dumps(blob).encode()
blob = to_jsonb64(blob)
sequence = [(EVlobRead(vlob_id, '42', 1),
const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
(EBlockRead(block_ids[0]),
const({
'content': to_jsonb64(chunk_1),
'creation_date': '2012-01-01T00:00:00'
})),
(EBlockRead(block_ids[1]),
const({
'content': to_jsonb64(chunk_2),
'creation_date': '2012-01-01T00:00:00'
})),
(EBlockRead(block_ids[2]),
const({
'content': to_jsonb64(chunk_3),
'creation_date': '2012-01-01T00:00:00'
}))]
read_content = perform_sequence(sequence, file.read())
assert read_content == content
# Offset
offset = 5
sequence = [(EVlobRead(vlob_id, '42', 1),
const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
(EBlockRead(block_ids[1]),
const({
'content': to_jsonb64(chunk_2),
'creation_date': '2012-01-01T00:00:00'
})),
(EBlockRead(block_ids[2]),
const({
'content': to_jsonb64(chunk_3),
'creation_date': '2012-01-01T00:00:00'
}))]
read_content = perform_sequence(sequence, file.read(offset=offset))
assert read_content == content[offset:]
# Size
size = 9
sequence = [(EVlobRead(vlob_id, '42', 1),
const({
'id': vlob_id,
'blob': blob,
'version': 1
})),
(EBlockRead(block_ids[1]),
const({
'content': to_jsonb64(chunk_2),
'creation_date': '2012-01-01T00:00:00'
}))]
read_content = perform_sequence(sequence,
file.read(offset=offset, size=size))
assert read_content == content[offset:][:size]
assert file.dirty is False
assert file.version == 1