def test_hashed_stream_reader(tmp_dir):
tmp_dir.gen({"foo": "foo"})
foo = tmp_dir / "foo"
with open(foo, "rb") as fobj:
stream_reader = HashedStreamReader(fobj)
assert stream_reader.read(3) == b"foo"
hex_digest = file_md5(foo, LocalFileSystem(None, {}))
assert stream_reader.is_text_file
assert hex_digest == stream_reader.hash_info.value
def test_hashed_stream_reader_as_chunks(tmp_dir):
tmp_dir.gen({"foo": b"foo \x00" * 16})
foo = tmp_dir / "foo"
with open(foo, "rb") as fobj:
stream_reader = HashedStreamReader(fobj)
while True:
chunk = stream_reader.read(16)
if not chunk:
break
hex_digest = file_md5(foo, LocalFileSystem(None, {}))
assert not stream_reader.is_text_file
assert hex_digest == stream_reader.hash_info.value
def test_hashed_stream_reader_compatibility(tmp_dir, contents):
# Always read more than the DEFAULT_CHUNK_SIZE (512 bytes).
# This imitates the read actions performed by upload_fobj.
chunk_size = DEFAULT_CHUNK_SIZE * 2
tmp_dir.gen("data", contents)
data = tmp_dir / "data"
with open(data, "rb") as fobj:
stream_reader = HashedStreamReader(fobj)
stream_reader.read(chunk_size)
local_fs = LocalFileSystem()
hex_digest = file_md5(data, local_fs)
assert stream_reader.is_text_file is istextfile(data, local_fs)
assert stream_reader.hash_info.value == hex_digest
def _upload_file(path_info, fs, odb, upload_odb):
from dvc.utils import tmp_fname
from dvc.utils.stream import HashedStreamReader
tmp_info = upload_odb.path_info / tmp_fname()
with fs.open(path_info, mode="rb", chunk_size=fs.CHUNK_SIZE) as stream:
stream = HashedStreamReader(stream)
size = fs.getsize(path_info)
upload_odb.fs.upload(stream, tmp_info, desc=path_info.name, total=size)
odb.add(tmp_info, upload_odb.fs, stream.hash_info)
return path_info, odb.get(stream.hash_info)
def test_hashed_stream_reader_as_chunks(tmp_dir):
tmp_dir.gen({"foo": b"foo \x00" * 16})
foo = tmp_dir / "foo"
actual_size = len(foo.read_bytes())
with open(foo, "rb") as fobj:
stream_reader = HashedStreamReader(fobj)
total_read = 0
while True:
chunk = stream_reader.read(16)
total_read += len(chunk)
assert stream_reader.tell() == total_read
if not chunk:
break
assert stream_reader.tell() == actual_size == total_read
hex_digest = file_md5(foo, LocalFileSystem(None, {}))
assert not stream_reader.is_text_file
assert hex_digest == stream_reader.hash_info.value
def _upload_file(path_info, fs, odb):
from dvc.utils import tmp_fname
from dvc.utils.stream import HashedStreamReader
tmp_info = odb.path_info / tmp_fname()
with fs.open(path_info, mode="rb", chunk_size=fs.CHUNK_SIZE) as stream:
stream = HashedStreamReader(stream)
odb.fs.upload_fobj(stream,
tmp_info,
desc=path_info.name,
size=fs.getsize(path_info))
obj = HashFile(tmp_info, odb.fs, stream.hash_info)
return path_info, obj
def _upload_file(from_fs_path, fs, odb, upload_odb):
from dvc.utils import tmp_fname
from dvc.utils.stream import HashedStreamReader
fs_path = upload_odb.fs.path
tmp_info = fs_path.join(upload_odb.fs_path, tmp_fname())
with fs.open(from_fs_path, mode="rb", chunk_size=fs.CHUNK_SIZE) as stream:
stream = HashedStreamReader(stream)
size = fs.getsize(from_fs_path)
upload_odb.fs.upload(
stream, tmp_info, desc=fs_path.name(from_fs_path), total=size
)
odb.add(tmp_info, upload_odb.fs, stream.hash_info)
meta = Meta(size=size)
return from_fs_path, meta, odb.get(stream.hash_info)
def test_hashed_stream_reader(tmp_dir):
tmp_dir.gen({"foo": "foo"})
foo = tmp_dir / "foo"
with open(foo, "rb") as fobj:
stream_reader = HashedStreamReader(fobj)
assert stream_reader.readable()
assert not stream_reader.seekable()
assert stream_reader.read(2) == b"fo"
assert stream_reader.tell() == 2
assert stream_reader.read(1) == b"o"
assert stream_reader.tell() == 3
hex_digest = file_md5(foo, LocalFileSystem())
assert stream_reader.is_text_file
assert hex_digest == stream_reader.hash_info.value
def _transfer_file(odb, from_fs, from_info):
from dvc.utils import tmp_fname
from dvc.utils.stream import HashedStreamReader
tmp_info = odb.fs.path_info / tmp_fname()
with from_fs.open(from_info, mode="rb",
chunk_size=from_fs.CHUNK_SIZE) as stream:
stream_reader = HashedStreamReader(stream)
# Since we don't know the hash beforehand, we'll
# upload it to a temporary location and then move
# it.
odb.fs.upload_fobj(
stream_reader,
tmp_info,
total=from_fs.getsize(from_info),
desc=from_info.name,
)
hash_info = stream_reader.hash_info
return tmp_info, hash_info
def _transfer_file(self, from_tree, from_info):
from dvc.utils import tmp_fname
from dvc.utils.stream import HashedStreamReader
tmp_info = self.tree.path_info / tmp_fname()
with from_tree.open(from_info,
mode="rb",
chunk_size=from_tree.CHUNK_SIZE) as stream:
stream_reader = HashedStreamReader(stream)
# Since we don't know the hash beforehand, we'll
# upload it to a temporary location and then move
# it.
self.tree.upload_fobj(
stream_reader,
tmp_info,
total=from_tree.getsize(from_info),
desc=from_info.name,
)
hash_info = stream_reader.hash_info
self.move(tmp_info, self.tree.hash_to_path_info(hash_info.value))
return hash_info