def _check_can_merge(self, out):
if self.scheme != out.scheme:
raise MergeError("unable to auto-merge outputs of different types")
my = self.dumpd()
other = out.dumpd()
ignored = [
self.fs.PARAM_CHECKSUM,
HashInfo.PARAM_SIZE,
HashInfo.PARAM_NFILES,
]
for opt in ignored:
my.pop(opt, None)
other.pop(opt, None)
if my != other:
raise MergeError(
"unable to auto-merge outputs with different options"
)
if not out.is_dir_checksum:
raise MergeError(
"unable to auto-merge outputs that are not directories"
)
def _check_can_merge(stage, ancestor_out=None):
if isinstance(stage, PipelineStage):
raise MergeError("unable to auto-merge pipeline stages")
if not stage.is_data_source or stage.deps or len(stage.outs) > 1:
raise MergeError("unable to auto-merge DVC files that weren't "
"created by `dvc add`")
if ancestor_out and not stage.outs:
raise MergeError(
"unable to auto-merge DVC files with deleted outputs")
def merge(self, ancestor, other):
from dvc_data.objects.tree import MergeError as TreeMergeError
from dvc_data.objects.tree import du, merge
assert other
if ancestor:
self._check_can_merge(ancestor)
ancestor_info = ancestor.hash_info
else:
ancestor_info = None
self._check_can_merge(self)
self._check_can_merge(other)
try:
merged = merge(
self.odb, ancestor_info, self.hash_info, other.hash_info
)
except TreeMergeError as exc:
raise MergeError(str(exc)) from exc
self.odb.add(merged.path, merged.fs, merged.oid)
self.hash_info = merged.hash_info
self.meta = Meta(
size=du(self.odb, merged),
nfiles=len(merged),
)
def _check_can_merge(self, out):
if self.scheme != out.scheme:
raise MergeError("unable to auto-merge outputs of different types")
my = self.dumpd()
other = out.dumpd()
my.pop(self.tree.PARAM_CHECKSUM)
other.pop(self.tree.PARAM_CHECKSUM)
if my != other:
raise MergeError(
"unable to auto-merge outputs with different options")
if not out.is_dir_checksum:
raise MergeError(
"unable to auto-merge outputs that are not directories")
def _diff(ancestor, other, allow_removed=False):
from dictdiffer import diff
allowed = ["add"]
if allow_removed:
allowed.append("remove")
result = list(diff(ancestor, other))
for typ, _, _ in result:
if typ not in allowed:
raise MergeError(
"unable to auto-merge directories with diff that contains "
f"'{typ}'ed files")
return result