def collect_used_dir_cache(
self, remote=None, force=False, jobs=None, filter_info=None
):
"""Get a list of `info`s related to the given directory.
- Pull the directory entry from the remote cache if it was changed.
Example:
Given the following commands:
$ echo "foo" > directory/foo
$ echo "bar" > directory/bar
$ dvc add directory
It will return a NamedCache like:
nc = NamedCache()
nc.add(self.scheme, 'c157a79031e1', 'directory/foo')
nc.add(self.scheme, 'd3b07384d113', 'directory/bar')
"""
cache = NamedCache()
try:
self.get_dir_cache(jobs=jobs, remote=remote)
except DvcException:
logger.debug(f"failed to pull cache for '{self}'")
try:
objects.check(self.odb, self.odb.get(self.hash_info))
except (FileNotFoundError, ObjectFormatError):
msg = (
"Missing cache for directory '{}'. "
"Cache for files inside will be lost. "
"Would you like to continue? Use '-f' to force."
)
if not force and not prompt.confirm(msg.format(self.path_info)):
raise CollectCacheError(
"unable to fully collect used cache"
" without cache for directory '{}'".format(self)
)
return cache
path = str(self.path_info)
filter_path = str(filter_info) if filter_info else None
for entry_key, entry_obj in self.obj:
entry_path = os.path.join(path, *entry_key)
if (
not filter_path
or entry_path == filter_path
or entry_path.startswith(filter_path + os.sep)
):
cache.add(self.scheme, entry_obj.hash_info.value, entry_path)
return cache
def test_status_download_optimization(mocker, dvc):
"""When comparing the status to pull a remote cache,
And the desired files to fetch are already on the local cache,
Don't check the existence of the desired files on the remote cache
"""
odb = LocalObjectDB(LocalFileSystem())
infos = NamedCache()
infos.add("local", "acbd18db4cc2f85cedef654fccc4a4d8", "foo")
infos.add("local", "37b51d194a7513e45b56f6524f2d51f2", "bar")
local_exists = list(infos["local"])
mocker.patch.object(odb, "hashes_exist", return_value=local_exists)
other_remote = mocker.Mock()
other_remote.url = "other_remote"
other_remote.hashes_exist.return_value = []
other_remote.index = RemoteIndexNoop()
other_remote.status(odb, infos, download=True)
assert other_remote.hashes_exist.call_count == 0
