fetch: introduce --tree-only

dvc/commands/data_sync.py

@@ -91,6 +91,7 @@ def run(self):
                 with_deps=self.args.with_deps,
                 recursive=self.args.recursive,
                 run_cache=self.args.run_cache,
+                tree_only=self.args.tree_only,
             )
             self.log_summary({"fetched": processed_files_count})
         except DvcException:
@@ -322,6 +323,12 @@ def add_parser(subparsers, _parent_parser):
         default=False,
         help="Fetch run history for all stages.",
     )
+    fetch_parser.add_argument(
+        "--tree-only",
+        action="store_true",
+        default=False,
+        help="Only fetch .dir objects.",
+    )
     fetch_parser.set_defaults(func=CmdDataFetch)
 
     # Status

dvc/dependency/repo.py

@@ -103,7 +103,7 @@ def get_used_objs(self, **kwargs) -> Dict[Optional["ObjectDB"], Set["HashInfo"]]
         return used
 
     def _get_used_and_obj(
-        self, obj_only: bool = False, **kwargs
+        self, obj_only: bool = False, tree_only: bool = False, **kwargs
     ) -> Tuple[Dict[Optional["ObjectDB"], Set["HashInfo"]], "Meta", "HashFile"]:
         from dvc.config import NoRemoteError
         from dvc.exceptions import NoOutputOrStageError, PathMissingError
@@ -126,6 +126,7 @@ def _get_used_and_obj(
                         force=True,
                         jobs=kwargs.get("jobs"),
                         recursive=True,
+                        tree_only=tree_only,
                     ).items():
                         if odb is None:
                             odb = repo.cloud.get_remote_odb()
@@ -152,6 +153,11 @@ def _get_used_and_obj(
             self._objs[rev] = obj
             self._meta[rev] = meta
 
+            if tree_only:
+                if isinstance(obj, Tree):
+                    used_obj_ids[object_store].add(obj.hash_info)
+                return used_obj_ids, meta, obj
+
             used_obj_ids[object_store].add(obj.hash_info)
             if isinstance(obj, Tree):
                 used_obj_ids[object_store].update(oid for _, _, oid in obj)

dvc/output.py

@@ -1065,7 +1065,7 @@ def _collect_used_dir_cache(
         return obj
 
     def get_used_objs(  # noqa: C901, PLR0911
-        self, **kwargs
+        self, tree_only: bool = False, **kwargs
     ) -> Dict[Optional["ObjectDB"], Set["HashInfo"]]:
         """Return filtered set of used object IDs for this out."""
 
@@ -1076,7 +1076,7 @@ def get_used_objs(  # noqa: C901, PLR0911
         if self.stage.is_repo_import:
             if push:
                 return {}
-            return self.get_used_external(**kwargs)
+            return self.get_used_external(tree_only=tree_only, **kwargs)
 
         if push and not self.can_push:
             return {}
@@ -1100,6 +1100,16 @@ def get_used_objs(  # noqa: C901, PLR0911
             logger.warning(msg)
             return {}
 
+        if self.remote:
+            remote = self.repo.cloud.get_remote_odb(name=self.remote)
+        else:
+            remote = None
+
+        if tree_only:
+            if self.is_dir_checksum:
+                return {remote: [self.hash_info]}
+            return {}
+
         obj: Optional["HashFile"]
         if self.is_dir_checksum:
             obj = self._collect_used_dir_cache(**kwargs)
@@ -1111,11 +1121,6 @@ def get_used_objs(  # noqa: C901, PLR0911
         if not obj:
             return {}
 
-        if self.remote:
-            remote = self.repo.cloud.get_remote_odb(name=self.remote)
-        else:
-            remote = None
-
         return {remote: self._named_obj_ids(obj)}
 
     def _named_obj_ids(self, obj):

dvc/repo/__init__.py

@@ -480,6 +480,7 @@ def used_objs(  # noqa: PLR0913
         revs=None,
         num=1,
         push: bool = False,
+        tree_only: bool = False,
     ):
         """Get the stages related to the given target and collect
         the `info` of its outputs.
@@ -515,6 +516,7 @@ def used_objs(  # noqa: PLR0913
                 recursive=recursive,
                 with_deps=with_deps,
                 push=push,
+                tree_only=tree_only,
             ).items():
                 used[odb].update(objs)
 

dvc/repo/fetch.py

@@ -32,6 +32,7 @@ def fetch(  # noqa: C901, PLR0913
     run_cache=False,
     revs=None,
     odb: Optional["HashFileDB"] = None,
+    tree_only: bool = False,
 ) -> int:
     """Download data items from a cloud and imported repositories
 
@@ -96,6 +97,7 @@ def fetch(  # noqa: C901, PLR0913
                 recursive=recursive,
                 revs=revs,
                 odb=odb,
+                tree_only=tree_only,
             )
             result.transferred.update(d)
             result.failed.update(f)

dvc/repo/index.py

@@ -440,6 +440,7 @@ def used_objs(
         recursive: bool = False,
         jobs: Optional[int] = None,
         push: bool = False,
+        tree_only: bool = False,
     ) -> "ObjectContainer":
         from collections import defaultdict
 
@@ -452,6 +453,7 @@ def used_objs(
                 jobs=jobs,
                 filter_info=filter_info,
                 push=push,
+                tree_only=tree_only,
             ).items():
                 used[odb].update(objs)
         return used