def test_verify_checksums(tmp_dir, scm, dvc, mocker, tmp_path_factory):
tmp_dir.dvc_gen({"file": "file1 content"}, commit="add file")
tmp_dir.dvc_gen({"dir": {"subfile": "file2 content"}}, commit="add dir")
RemoteConfig(dvc.config).add(
"local_remote",
fspath(tmp_path_factory.mktemp("local_remote")),
default=True,
)
dvc.push()
# remove artifacts and cache to trigger fetching
os.remove("file")
shutil.rmtree("dir")
shutil.rmtree(dvc.cache.local.cache_dir)
checksum_spy = mocker.spy(dvc.cache.local, "get_file_checksum")
dvc.pull()
assert checksum_spy.call_count == 0
# Removing cache will invalidate existing state entries
shutil.rmtree(dvc.cache.local.cache_dir)
dvc.config.set(
Config.SECTION_REMOTE_FMT.format("local_remote"),
Config.SECTION_REMOTE_VERIFY,
"True",
)
dvc.pull()
assert checksum_spy.call_count == 3
def test_partial_push_n_pull(tmp_dir, dvc, tmp_path_factory):
remote_config = RemoteConfig(dvc.config)
remote_config.add("upstream",
fspath(tmp_path_factory.mktemp("upstream")),
default=True)
foo = tmp_dir.dvc_gen({"foo": "foo content"})[0].outs[0]
bar = tmp_dir.dvc_gen({"bar": "bar content"})[0].outs[0]
# Faulty upload version, failing on foo
original = RemoteLOCAL._upload
def unreliable_upload(self, from_file, to_info, name=None, **kwargs):
if name == "foo":
raise Exception("stop foo")
return original(self, from_file, to_info, name, **kwargs)
with patch.object(RemoteLOCAL, "_upload", unreliable_upload):
with pytest.raises(UploadError) as upload_error_info:
dvc.push()
assert upload_error_info.value.amount == 1
remote = dvc.cloud.get_remote("upstream")
assert not remote.exists(remote.checksum_to_path_info(foo.checksum))
assert remote.exists(remote.checksum_to_path_info(bar.checksum))
# Push everything and delete local cache
dvc.push()
shutil.rmtree(dvc.cache.local.cache_dir)
with patch.object(RemoteLOCAL, "_download", side_effect=Exception):
with pytest.raises(DownloadError) as download_error_info:
dvc.pull()
assert download_error_info.value.amount == 2
def _external_repo(url=None, rev=None, cache_dir=None):
from dvc.config import Config
from dvc.cache import CacheConfig
from dvc.repo import Repo
key = (url, rev, cache_dir)
if key in REPO_CACHE:
return REPO_CACHE[key]
new_path = tempfile.mkdtemp("dvc-erepo")
# Copy and adjust existing clone
if (url, None, None) in REPO_CACHE:
old_path = REPO_CACHE[url, None, None]
# This one unlike shutil.copytree() works with an existing dir
copy_tree(old_path, new_path)
else:
# Create a new clone
_clone_repo(url, new_path)
# Save clean clone dir so that we will have access to a default branch
clean_clone_path = tempfile.mkdtemp("dvc-erepo")
copy_tree(new_path, clean_clone_path)
REPO_CACHE[url, None, None] = clean_clone_path
# Adjust new clone/copy to fit rev and cache_dir
# Checkout needs to be done first because current branch might not be
# DVC repository
if rev is not None:
_git_checkout(new_path, rev)
repo = Repo(new_path)
try:
# check if the URL is local and no default remote is present
# add default remote pointing to the original repo's cache location
if os.path.isdir(url):
rconfig = RemoteConfig(repo.config)
if not _default_remote_set(rconfig):
original_repo = Repo(url)
try:
rconfig.add(
"auto-generated-upstream",
original_repo.cache.local.cache_dir,
default=True,
level=Config.LEVEL_LOCAL,
)
finally:
original_repo.close()
if cache_dir is not None:
cache_config = CacheConfig(repo.config)
cache_config.set_dir(cache_dir, level=Config.LEVEL_LOCAL)
finally:
# Need to close/reopen repo to force config reread
repo.close()
REPO_CACHE[key] = new_path
return new_path
def test_raise_on_too_many_open_files(tmp_dir, dvc, tmp_path_factory, mocker):
storage = tmp_path_factory.mktemp("test_remote_base")
remote_config = RemoteConfig(dvc.config)
remote_config.add("local_remote", fspath(storage), default=True)
tmp_dir.dvc_gen({"file": "file content"})
mocker.patch.object(
RemoteLOCAL,
"_upload",
side_effect=OSError(errno.EMFILE, "Too many open files"),
)
with pytest.raises(OSError) as e:
dvc.push()
assert e.errno == errno.EMFILE
def _set_upstream(self):
# check if the URL is local and no default remote is present
# add default remote pointing to the original repo's cache location
if os.path.isdir(self.url):
rconfig = RemoteConfig(self.config)
if not rconfig.has_default():
src_repo = Repo(self.url)
try:
rconfig.add(
"auto-generated-upstream",
src_repo.cache.local.cache_dir,
default=True,
level=Config.LEVEL_LOCAL,
)
finally:
src_repo.close()
def _external_repo(url=None, rev=None, cache_dir=None):
from dvc.config import Config
from dvc.cache import CacheConfig
from dvc.repo import Repo
key = (url, rev, cache_dir)
if key in REPO_CACHE:
return REPO_CACHE[key]
new_path = cached_clone(url, rev=rev)
repo = Repo(new_path)
try:
# check if the URL is local and no default remote is present
# add default remote pointing to the original repo's cache location
if os.path.isdir(url):
rconfig = RemoteConfig(repo.config)
if not _default_remote_set(rconfig):
original_repo = Repo(url)
try:
rconfig.add(
"auto-generated-upstream",
original_repo.cache.local.cache_dir,
default=True,
level=Config.LEVEL_LOCAL,
)
finally:
original_repo.close()
if cache_dir is not None:
cache_config = CacheConfig(repo.config)
cache_config.set_dir(cache_dir, level=Config.LEVEL_LOCAL)
finally:
# Need to close/reopen repo to force config reread
repo.close()
REPO_CACHE[key] = new_path
return new_path
def test_pre_push_hook(self, tmp_dir, scm, dvc, tmp_path_factory):
scm.install()
temp = tmp_path_factory.mktemp("external")
git_remote = temp / "project.git"
storage_path = temp / "dvc_storage"
RemoteConfig(dvc.config).add("store",
fspath(storage_path),
default=True)
tmp_dir.dvc_gen("file", "file_content", "commit message")
file_checksum = file_md5("file")[0]
expected_storage_path = (storage_path / file_checksum[:2] /
file_checksum[2:])
scm.repo.clone(fspath(git_remote))
scm.repo.create_remote("origin", fspath(git_remote))
assert not expected_storage_path.is_file()
scm.repo.git.push("origin", "master")
assert expected_storage_path.is_file()
assert expected_storage_path.read_text() == "file_content"
def test_modify_missing_remote(dvc):
remote_config = RemoteConfig(dvc.config)
with pytest.raises(ConfigError, match=r"Unable to find remote section"):
remote_config.modify("myremote", "gdrive_client_id", "xxx")