def test_download_callbacks_on_dvc_git_fs(tmp_dir, dvc, scm, fs_type):
from dvc.fs.git import GitFileSystem
gen = tmp_dir.scm_gen if fs_type == "git" else tmp_dir.dvc_gen
gen({"dir": {"foo": "foo", "bar": "bar"}, "file": "file"}, commit="gen")
fs = dvc.dvcfs if fs_type == "dvc" else GitFileSystem(scm=scm, rev="HEAD")
callback = fsspec.Callback()
fs.download_file(
"file",
(tmp_dir / "file2").fs_path,
callback=callback,
)
size = os.path.getsize(tmp_dir / "file")
assert (tmp_dir / "file2").read_text() == "file"
assert callback.size == size
assert callback.value == size
callback = fsspec.Callback()
fs.download(
"dir",
(tmp_dir / "dir2").fs_path,
callback=callback,
)
assert (tmp_dir / "dir2").read_text() == {"foo": "foo", "bar": "bar"}
assert callback.size == 2
assert callback.value == 2
def get_fs(self, rev: str, **kwargs):
from dvc.fs.git import GitFileSystem
from .objects import GitTrie
resolved = self.resolve_rev(rev)
tree_obj = self._backend_func("get_tree_obj", rev=resolved)
trie = GitTrie(tree_obj, resolved)
return GitFileSystem(self.root_dir, trie, **kwargs)
def get_fs(self, rev: str):
from dvc.fs.git import GitFileSystem
from .objects import GitTrie
resolved = self.resolve_rev(rev)
tree_obj = self.pygit2.get_tree_obj(rev=resolved)
trie = GitTrie(tree_obj, resolved)
return GitFileSystem(self.root_dir, trie)
def test_ignore_on_branch(tmp_dir, scm, dvc):
from dvc.fs.git import GitFileSystem
tmp_dir.scm_gen({"foo": "foo", "bar": "bar"}, commit="add files")
with tmp_dir.branch("branch", new=True):
tmp_dir.scm_gen(DvcIgnore.DVCIGNORE_FILE, "foo", commit="add ignore")
dvc._reset()
result = walk_files(dvc, dvc.fs, tmp_dir)
assert set(result) == {
(tmp_dir / "foo").fs_path,
(tmp_dir / "bar").fs_path,
(tmp_dir / DvcIgnore.DVCIGNORE_FILE).fs_path,
}
dvc.fs = GitFileSystem(scm=scm, rev="branch")
assert dvc.dvcignore.is_ignored_file(tmp_dir / "foo")
def _get_repo_dirs(
self,
root_dir: str = None,
scm: "Base" = None,
rev: str = None,
uninitialized: bool = False,
):
assert bool(scm) == bool(rev)
from dvc.fs.git import GitFileSystem
from dvc.scm import SCM, Base, Git, SCMError
from dvc.utils.fs import makedirs
dvc_dir = None
tmp_dir = None
try:
fs = (
GitFileSystem(scm=scm, rev=rev)
if isinstance(scm, Git) and rev
else None
)
root_dir = self.find_root(root_dir, fs)
dvc_dir = os.path.join(root_dir, self.DVC_DIR)
tmp_dir = os.path.join(dvc_dir, "tmp")
makedirs(tmp_dir, exist_ok=True)
except NotDvcRepoError:
if not uninitialized:
raise
try:
scm = SCM(root_dir or os.curdir)
except SCMError:
scm = SCM(os.curdir, no_scm=True)
assert isinstance(scm, Base)
root_dir = scm.root_dir
return root_dir, dvc_dir, tmp_dir
def __init__(
self,
root_dir=None,
fs=None,
rev=None,
subrepos=False,
uninitialized=False,
config=None,
url=None,
repo_factory=None,
):
from dvc.config import Config
from dvc.data.db import ODBManager
from dvc.data_cloud import DataCloud
from dvc.fs.git import GitFileSystem
from dvc.fs.local import localfs
from dvc.lock import LockNoop, make_lock
from dvc.repo.live import Live
from dvc.repo.metrics import Metrics
from dvc.repo.params import Params
from dvc.repo.plots import Plots
from dvc.repo.stage import StageLoad
from dvc.scm import SCM
from dvc.stage.cache import StageCache
from dvc.state import State, StateNoop
self.url = url
self._fs_conf = {"repo_factory": repo_factory}
self._fs = fs or localfs
self._scm = None
if rev and not fs:
self._scm = SCM(root_dir or os.curdir)
self._fs = GitFileSystem(scm=self._scm, rev=rev)
self.root_dir, self.dvc_dir, self.tmp_dir = self._get_repo_dirs(
root_dir=root_dir, fs=self.fs, uninitialized=uninitialized)
self.config = Config(self.dvc_dir, fs=self.fs, config=config)
self._uninitialized = uninitialized
# used by RepoFileSystem to determine if it should traverse subrepos
self.subrepos = subrepos
self.cloud = DataCloud(self)
self.stage = StageLoad(self)
if isinstance(self.fs, GitFileSystem) or not self.dvc_dir:
self.lock = LockNoop()
self.state = StateNoop()
self.odb = ODBManager(self)
else:
self.lock = make_lock(
os.path.join(self.tmp_dir, "lock"),
tmp_dir=self.tmp_dir,
hardlink_lock=self.config["core"].get("hardlink_lock", False),
friendly=True,
)
state_db_dir = self._get_database_dir("state")
self.state = State(self.root_dir, state_db_dir, self.dvcignore)
self.odb = ODBManager(self)
self.stage_cache = StageCache(self)
self._ignore()
self.metrics = Metrics(self)
self.plots = Plots(self)
self.params = Params(self)
self.live = Live(self)
self.stage_collection_error_handler: Optional[Callable[
[str, Exception], None]] = None
self._lock_depth = 0
def get_fs(self, rev: str):
from dvc.fs.git import GitFileSystem
return GitFileSystem(scm=self, rev=rev)
def external_repo(url,
rev=None,
for_write=False,
cache_dir=None,
cache_types=None,
**kwargs):
from scmrepo.git import Git
from dvc.config import NoRemoteError
from dvc.fs.git import GitFileSystem
logger.debug("Creating external repo %s@%s", url, rev)
path = _cached_clone(url, rev, for_write=for_write)
# Local HEAD points to the tip of whatever branch we first cloned from
# (which may not be the default branch), use origin/HEAD here to get
# the tip of the default branch
rev = rev or "refs/remotes/origin/HEAD"
cache_config = {
"cache": {
"dir": cache_dir or _get_cache_dir(url),
"type": cache_types
}
}
config = _get_remote_config(url) if os.path.isdir(url) else {}
config.update(cache_config)
if for_write:
root_dir = path
fs = None
else:
root_dir = os.path.realpath(path)
scm = Git(root_dir)
fs = GitFileSystem(scm=scm, rev=rev)
repo_kwargs = dict(
root_dir=root_dir,
url=url,
fs=fs,
config=config,
repo_factory=erepo_factory(url, cache_config),
**kwargs,
)
if "subrepos" not in repo_kwargs:
repo_kwargs["subrepos"] = True
if "uninitialized" not in repo_kwargs:
repo_kwargs["uninitialized"] = True
repo = Repo(**repo_kwargs)
try:
yield repo
except NoRemoteError as exc:
raise NoRemoteInExternalRepoError(url) from exc
except OutputNotFoundError as exc:
if exc.repo is repo:
raise NoOutputInExternalRepoError(exc.output, repo.root_dir,
url) from exc
raise
except FileMissingError as exc:
raise PathMissingError(exc.path, url) from exc
finally:
repo.close()
if for_write:
_remove(path)
