def test_repotree_cache_save(tmp_dir, dvc, scm, erepo_dir, setup_remote):
with erepo_dir.chdir():
erepo_dir.gen({"dir": {"subdir": {"foo": "foo"}, "bar": "bar"}})
erepo_dir.dvc_add("dir/subdir", commit="subdir")
erepo_dir.scm_add("dir", commit="dir")
setup_remote(erepo_dir.dvc)
erepo_dir.dvc.push()
# test only cares that either fetch or stream are set so that DVC dirs are
# walked.
#
# for this test, all file objects are being opened() and copied from tree
# into dvc.cache, not fetched or streamed from a remote
tree = RepoTree(erepo_dir.dvc, stream=True)
expected = [
tree.get_file_checksum(erepo_dir / path)
for path in ("dir/bar", "dir/subdir/foo")
]
with erepo_dir.dvc.state:
cache = dvc.cache.local
with cache.state:
cache.save(PathInfo(erepo_dir / "dir"), None, tree=tree)
for checksum in expected:
assert os.path.exists(cache.checksum_to_path_info(checksum))
def _get_checksum(self, locked=True):
from dvc.repo.tree import RepoTree
with self._make_repo(locked=locked) as repo:
try:
return repo.find_out_by_relpath(self.def_path).info["md5"]
except OutputNotFoundError:
path = PathInfo(os.path.join(repo.root_dir, self.def_path))
# we want stream but not fetch, so DVC out directories are
# walked, but dir contents is not fetched
tree = RepoTree(repo, stream=True)
# We are polluting our repo cache with some dir listing here
if tree.isdir(path):
return self.repo.cache.local.get_checksum(path, tree)
return tree.get_file_checksum(path)
