dvc: separate data and objects

Initial simple separation to get the process of extracting dvc-objects
and dvc-data packages out of dvc.

dvc-objects is for:
 * defining base Object (aka HashFile) and ObjectDB
 * providing tools to work with those objects (check/transfer/etc)
 * etc

dvc-data is for:
 * defining higher level objects (e.g. Tree)
 * bulding objects from the workspace (aka staging)
 * projecting objects back into the workspace (aka checkout)
 * comparing Tree objects (aka diff)
 * defining misc object databases (e.g LocalObjectDB)
 * higher-level transfer (e.g. with Tree indexes)
 * etc

For example, the objects we use in run-cache are based on dvc-objects
and not on dvc-data, as they have absolutely nothing to do with data
management.

There are still some mixups in the code that will be addressed in the
followups.

dvc/data/__init__.py

@@ -5,9 +5,8 @@ from .tree import Tree
 
 if TYPE_CHECKING:
     from dvc.hash_info import HashInfo
-
-    from .db.base import ObjectDB
-    from .file import HashFile
+    from dvc.objects.db import ObjectDB
+    from dvc.objects.file import HashFile
 
 logger = logging.getLogger(__name__)
 

dvc/data/checkout.py

@@ -5,14 +5,13 @@ from dvc import prompt
 from dvc.exceptions import CacheLinkError, CheckoutError, ConfirmRemoveError
 from dvc.fs.utils import test_links, transfer
 from dvc.ignore import DvcIgnoreFilter
-from dvc.objects.db.slow_link_detection import (  # type: ignore[attr-defined]
-    slow_link_guard,
-)
-from dvc.objects.diff import ROOT
-from dvc.objects.diff import diff as odiff
-from dvc.objects.stage import stage
 from dvc.types import Optional
 
+from .diff import ROOT
+from .diff import diff as odiff
+from .slow_link_detection import slow_link_guard  # type: ignore[attr-defined]
+from .stage import stage
+
 logger = logging.getLogger(__name__)
 
 

dvc/data/db/__init__.py

@@ -7,7 +7,8 @@ if TYPE_CHECKING:
 
 
 def get_odb(fs, fs_path, **config):
-    from .base import ObjectDB
+    from dvc.objects.db import ObjectDB
+
     from .local import LocalObjectDB
 
     if fs.scheme == Schemes.LOCAL:

dvc/data/db/index.py

@@ -3,7 +3,7 @@ import os
 from abc import ABC, abstractmethod
 from typing import TYPE_CHECKING, Iterable, Set
 
-from ..errors import ObjectDBError
+from dvc.objects.errors import ObjectDBError
 
 if TYPE_CHECKING:
     from dvc.types import StrPath

dvc/data/db/local.py

@@ -6,13 +6,12 @@ from funcy import cached_property
 from shortuuid import uuid
 
 from dvc.hash_info import HashInfo
+from dvc.objects.db import ObjectDB
 from dvc.objects.errors import ObjectFormatError
 from dvc.progress import Tqdm
 from dvc.utils import relpath
 from dvc.utils.fs import copyfile, remove, umask, walk_files
 
-from .base import ObjectDB
-
 logger = logging.getLogger(__name__)
 
 

dvc/data/db/reference.py

@@ -3,11 +3,11 @@ import logging
 import os
 from typing import TYPE_CHECKING, Dict
 
+from dvc.objects.db import ObjectDB
+from dvc.objects.errors import ObjectFormatError
 from dvc.scheme import Schemes
 
-from ..errors import ObjectFormatError
 from ..reference import ReferenceHashFile
-from .base import ObjectDB
 
 if TYPE_CHECKING:
     from dvc.fs.base import FileSystem

dvc/data/diff.py

@@ -3,8 +3,7 @@ from typing import TYPE_CHECKING, List, Optional, Tuple
 
 if TYPE_CHECKING:
     from dvc.hash_info import HashInfo
-
-    from .file import HashFile
+    from dvc.objects.file import HashFile
 
 ADD = "add"
 MODIFY = "modify"
@@ -96,7 +95,7 @@ def diff(
         return entry_obj.hash_info if entry_obj else None
 
     def _in_cache(oid, cache):
-        from .errors import ObjectFormatError
+        from dvc.objects.errors import ObjectFormatError
 
         if not oid:
             return False

dvc/data/reference.py

@@ -4,16 +4,15 @@ import os
 import pickle
 from typing import TYPE_CHECKING, Optional
 
-from .errors import ObjectFormatError
-from .file import HashFile
+from dvc.objects.errors import ObjectFormatError
+from dvc.objects.file import HashFile
 
 if TYPE_CHECKING:
     from dvc.fs.base import FileSystem
     from dvc.hash_info import HashInfo
+    from dvc.objects.db import ObjectDB
     from dvc.types import AnyPath
 
-    from .db.base import ObjectDB
-
 logger = logging.getLogger(__name__)
 
 

dvc/data/stage.py

@@ -8,19 +8,18 @@ from typing import TYPE_CHECKING, Dict, Optional, Tuple
 from dvc.exceptions import DvcIgnoreInCollectedDirError
 from dvc.hash_info import HashInfo
 from dvc.ignore import DvcIgnore
+from dvc.objects.file import HashFile
 from dvc.progress import Tqdm
 from dvc.utils import file_md5
 
 from .db.reference import ReferenceObjectDB
-from .file import HashFile
 from .meta import Meta
 
 if TYPE_CHECKING:
     from dvc.fs.base import FileSystem
+    from dvc.objects.db import ObjectDB
     from dvc.types import AnyPath
 
-    from .db.base import ObjectDB
-
 logger = logging.getLogger(__name__)
 
 
@@ -245,8 +244,9 @@ def _get_staging(odb: "ObjectDB") -> "ObjectDB":
 
 
 def _load_from_state(odb, staging, fs_path, fs, name):
+    from dvc.objects.errors import ObjectFormatError
+
     from . import check, load
-    from .errors import ObjectFormatError
     from .tree import Tree
 
     state = odb.state

dvc/data/status.py

@@ -7,9 +7,10 @@ from dvc.scheme import Schemes
 from .tree import Tree
 
 if TYPE_CHECKING:
-    from .db.base import ObjectDB
+    from dvc.objects.db import ObjectDB
+    from dvc.objects.file import HashFile
+
     from .db.index import ObjectDBIndexBase
-    from .file import HashFile
 
 logger = logging.getLogger(__name__)
 

dvc/data/transfer.py

@@ -10,8 +10,8 @@ from dvc.utils.threadpool import ThreadPoolExecutor
 
 if TYPE_CHECKING:
     from dvc.hash_info import HashInfo
+    from dvc.objects.db import ObjectDB
 
-    from .db.base import ObjectDB
     from .db.index import ObjectDBIndexBase
     from .tree import Tree
 
@@ -40,7 +40,8 @@ def _log_exceptions(func):
 def find_tree_by_obj_id(
     odbs: Iterable[Optional["ObjectDB"]], obj_id: "HashInfo"
 ) -> Optional["Tree"]:
-    from .errors import ObjectFormatError
+    from dvc.objects.errors import ObjectFormatError
+
     from .tree import Tree
 
     for odb in odbs:

dvc/data/tree.py

@@ -5,8 +5,9 @@ from typing import TYPE_CHECKING, Dict, Optional, Tuple
 
 from funcy import cached_property
 
-from .errors import ObjectFormatError
-from .file import HashFile
+from dvc.objects.errors import ObjectFormatError
+from dvc.objects.file import HashFile
+
 from .stage import get_file_hash
 
 if TYPE_CHECKING:

dvc/data_cloud.py

@@ -3,11 +3,11 @@
 import logging
 from typing import TYPE_CHECKING, Iterable, Optional
 
-from dvc.objects.db import get_index
+from dvc.data.db import get_index
 
 if TYPE_CHECKING:
     from dvc.hash_info import HashInfo
-    from dvc.objects.db.base import ObjectDB
+    from dvc.objects.db import ObjectDB
 
 logger = logging.getLogger(__name__)
 
@@ -55,8 +55,8 @@ class DataCloud:
         raise NoRemoteError(error_msg)
 
     def _init_odb(self, name):
+        from dvc.data.db import get_odb
         from dvc.fs import get_cloud_fs
-        from dvc.objects.db import get_odb
 
         cls, config, fs_path = get_cloud_fs(self.repo, name=name)
         config["tmp_dir"] = self.repo.index_db_dir
@@ -78,7 +78,7 @@ class DataCloud:
                 By default remote from core.remote config option is used.
             odb: optional ODB to push to. Overrides remote.
         """
-        from dvc.objects.transfer import transfer
+        from dvc.data.transfer import transfer
 
         if not odb:
             odb = self.get_remote_odb(remote, "push")
@@ -107,7 +107,7 @@ class DataCloud:
                 By default remote from core.remote config option is used.
             odb: optional ODB to pull from. Overrides remote.
         """
-        from dvc.objects.transfer import transfer
+        from dvc.data.transfer import transfer
 
         if not odb:
             odb = self.get_remote_odb(remote, "pull")
@@ -141,7 +141,7 @@ class DataCloud:
             log_missing: log warning messages if file doesn't exist
                 neither in cache, neither in cloud.
         """
-        from dvc.objects.status import compare_status
+        from dvc.data.status import compare_status
 
         if not odb:
             odb = self.get_remote_odb(remote, "status")

dvc/dependency/repo.py

@@ -9,7 +9,7 @@ from .base import Dependency
 
 if TYPE_CHECKING:
     from dvc.hash_info import HashInfo
-    from dvc.objects.db.base import ObjectDB
+    from dvc.objects.db import ObjectDB
     from dvc.objects.file import HashFile
 
 
@@ -61,7 +61,7 @@ class RepoDependency(Dependency):
         return {self.PARAM_PATH: self.def_path, self.PARAM_REPO: self.def_repo}
 
     def download(self, to, jobs=None):
-        from dvc.objects.checkout import checkout
+        from dvc.data.checkout import checkout
 
         for odb, objs in self.get_used_objs().items():
             self.repo.cloud.pull(objs, jobs=jobs, odb=odb)
@@ -98,9 +98,9 @@ class RepoDependency(Dependency):
         self, obj_only=False, **kwargs
     ) -> Tuple[Dict[Optional["ObjectDB"], Set["HashInfo"]], "HashFile"]:
         from dvc.config import NoRemoteError
+        from dvc.data.stage import stage
+        from dvc.data.tree import Tree
         from dvc.exceptions import NoOutputOrStageError, PathMissingError
-        from dvc.objects.stage import stage
-        from dvc.objects.tree import Tree
 
         local_odb = self.repo.odb.local
         locked = kwargs.pop("locked", True)
@@ -150,10 +150,10 @@ class RepoDependency(Dependency):
             return used_obj_ids, staged_obj
 
     def _check_circular_import(self, odb, obj_ids):
+        from dvc.data.db.reference import ReferenceObjectDB
+        from dvc.data.tree import Tree
         from dvc.exceptions import CircularImportError
         from dvc.fs.repo import RepoFileSystem
-        from dvc.objects.db.reference import ReferenceObjectDB
-        from dvc.objects.tree import Tree
 
         if not isinstance(odb, ReferenceObjectDB):
             return

dvc/objects/db.py

@@ -357,7 +357,7 @@ class ObjectDB:
         pass
 
     def gc(self, used, jobs=None, cache_odb=None, shallow=True):
-        from ..tree import Tree
+        from dvc.data.tree import Tree
 
         if self.read_only:
             raise ObjectDBPermissionError("Cannot gc read-only ODB")

dvc/objects/file.py

@@ -3,14 +3,12 @@ import logging
 import os
 from typing import TYPE_CHECKING, Optional
 
-from .errors import ObjectFormatError
-
 if TYPE_CHECKING:
     from dvc.fs.base import FileSystem
     from dvc.hash_info import HashInfo
     from dvc.types import AnyPath
 
-    from .db.base import ObjectDB
+    from .db import ObjectDB
 
 logger = logging.getLogger(__name__)
 
@@ -68,7 +66,8 @@ class HashFile:
         self._check_hash(odb)
 
     def _check_hash(self, odb):
-        from .stage import get_file_hash
+        from dvc.data.stage import get_file_hash
+        from dvc.objects.errors import ObjectFormatError
 
         _, actual = get_file_hash(
             self.fs_path, self.fs, self.hash_info.name, odb.state

dvc/output.py

@@ -7,7 +7,7 @@ from urllib.parse import urlparse
 from funcy import collecting, project
 from voluptuous import And, Any, Coerce, Length, Lower, Required, SetTo
 
-from dvc import objects, prompt
+from dvc import prompt
 from dvc.exceptions import (
     CheckoutError,
     CollectCacheError,
@@ -15,25 +15,27 @@ from dvc.exceptions import (
     MergeError,
     RemoteCacheRequiredError,
 )
-from dvc.objects.checkout import checkout
 
+from .data import check as ocheck
+from .data import load as oload
+from .data.checkout import checkout
+from .data.meta import Meta
+from .data.stage import stage as ostage
+from .data.transfer import transfer as otransfer
+from .data.tree import Tree
 from .fs import get_cloud_fs
 from .fs.hdfs import HDFSFileSystem
 from .fs.local import LocalFileSystem
 from .fs.s3 import S3FileSystem
 from .hash_info import HashInfo
 from .istextfile import istextfile
-from .objects import Tree
 from .objects.errors import ObjectFormatError
-from .objects.meta import Meta
-from .objects.stage import stage as ostage
-from .objects.transfer import transfer as otransfer
 from .scheme import Schemes
 from .utils import relpath
 from .utils.fs import path_isin
 
 if TYPE_CHECKING:
-    from .objects.db.base import ObjectDB
+    from .objects.db import ObjectDB
 
 logger = logging.getLogger(__name__)
 
@@ -456,7 +458,7 @@ class Output:
             return True
 
         try:
-            objects.check(self.odb, obj)
+            ocheck(self.odb, obj)
             return False
         except (FileNotFoundError, ObjectFormatError):
             return True
@@ -710,7 +712,7 @@ class Output:
             obj = self.obj
         elif self.hash_info:
             try:
-                obj = objects.load(self.odb, self.hash_info)
+                obj = oload(self.odb, self.hash_info)
             except FileNotFoundError:
                 return None
         else:
@@ -857,7 +859,7 @@ class Output:
 
         obj = self.odb.get(self.hash_info)
         try:
-            objects.check(self.odb, obj)
+            ocheck(self.odb, obj)
         except FileNotFoundError:
             if self.remote:
                 kwargs["remote"] = self.remote
@@ -867,7 +869,7 @@ class Output:
             return self.obj
 
         try:
-            self.obj = objects.load(self.odb, self.hash_info)
+            self.obj = oload(self.odb, self.hash_info)
         except (FileNotFoundError, ObjectFormatError):
             self.obj = None
 
@@ -884,7 +886,7 @@ class Output:
             logger.debug(f"failed to pull cache for '{self}'")
 
         try:
-            objects.check(self.odb, self.odb.get(self.hash_info))
+            ocheck(self.odb, self.odb.get(self.hash_info))
         except FileNotFoundError:
             msg = (
                 "Missing cache for directory '{}'. "
@@ -1012,7 +1014,7 @@ class Output:
             )
 
     def merge(self, ancestor, other):
-        from dvc.objects.tree import du, merge
+        from dvc.data.tree import du, merge
 
         assert other
 

dvc/repo/__init__.py

@@ -150,11 +150,11 @@ class Repo:
         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.objects.db import ODBManager
         from dvc.repo.live import Live
         from dvc.repo.metrics import Metrics
         from dvc.repo.params import Params
@@ -406,7 +406,7 @@ class Repo:
         def _add_suffix(objs: Set["HashFile"], suffix: str) -> None:
             from itertools import chain
 
-            from dvc.objects import iterobjs
+            from dvc.data import iterobjs
 
             for obj in chain.from_iterable(map(iterobjs, objs)):
                 if obj.name is not None:

dvc/repo/diff.py

@@ -113,8 +113,8 @@ def _paths_checksums(repo, targets):
 
 
 def _output_paths(repo, targets):
+    from dvc.data.stage import stage as ostage
     from dvc.fs.local import LocalFileSystem
-    from dvc.objects.stage import stage as ostage
 
     on_working_fs = isinstance(repo.fs, LocalFileSystem)
 

dvc/repo/experiments/executor/ssh.py

@@ -156,7 +156,7 @@ class SSHExecutor(BaseExecutor):
         return sshfs.fs.execute(f"cd {working_dir};{cmd}", **kwargs)
 
     def init_cache(self, dvc: "Repo", rev: str, run_cache: bool = True):
-        from dvc.objects.db import ODBManager, get_odb
+        from dvc.data.db import ODBManager, get_odb
         from dvc.repo import Repo
         from dvc.repo.push import push
 

dvc/repo/gc.py

@@ -44,7 +44,7 @@ def gc(
 
     from contextlib import ExitStack
 
-    from dvc.objects.db import get_index
+    from dvc.data.db import get_index
     from dvc.repo import Repo
 
     if not repos:

dvc/schema.py

@@ -3,7 +3,7 @@ from collections.abc import Mapping
 from voluptuous import Any, Optional, Required, Schema
 
 from dvc import dependency, output
-from dvc.objects.meta import Meta
+from dvc.data.meta import Meta
 from dvc.output import CHECKSUMS_SCHEMA, Output
 from dvc.parsing import DO_KWD, FOREACH_KWD, VARS_KWD
 from dvc.parsing.versions import SCHEMA_KWD, lockfile_version_schema

dvc/stage/cache.py

@@ -21,7 +21,7 @@ class RunCacheNotFoundError(DvcException):
 
 
 def _get_cache_hash(cache, key=False):
-    from dvc.objects.meta import Meta
+    from dvc.data.meta import Meta
 
     if key:
         cache["outs"] = [out["path"] for out in cache.get("outs", [])]
@@ -236,7 +236,7 @@ class StageCache:
         return ret
 
     def push(self, remote: Optional[str], odb: Optional["ObjectDB"] = None):
-        from dvc.objects.transfer import _log_exceptions
+        from dvc.data.transfer import _log_exceptions
 
         if odb is None:
             odb = self.repo.cloud.get_remote_odb(remote)
@@ -247,7 +247,7 @@ class StageCache:
         )
 
     def pull(self, remote: Optional[str]):
-        from dvc.objects.transfer import _log_exceptions
+        from dvc.data.transfer import _log_exceptions
 
         odb = self.repo.cloud.get_remote_odb(remote)
         return self._transfer(

dvc/stage/loader.py

@@ -6,8 +6,8 @@ from itertools import chain
 from funcy import cached_property, get_in, lcat, once, project
 
 from dvc import dependency, output
+from dvc.data.meta import Meta
 from dvc.hash_info import HashInfo
-from dvc.objects.meta import Meta
 from dvc.parsing import FOREACH_KWD, JOIN, DataResolver, EntryNotFound
 from dvc.parsing.versions import LOCKFILE_VERSION
 

dvc/stage/utils.py

@@ -4,8 +4,8 @@ from typing import TYPE_CHECKING, Union
 
 from funcy import concat, first, lsplit, rpartial, without
 
+from dvc.data.meta import Meta
 from dvc.exceptions import InvalidArgumentError
-from dvc.objects.meta import Meta
 
 from .exceptions import (
     MissingDataSource,

dvc/state.py

@@ -103,7 +103,7 @@ class State(StateBase):  # pylint: disable=too-many-instance-attributes
             HashInfo or None: hash for the specified path info or None if it
             doesn't exist in the state database.
         """
-        from .objects.meta import Meta
+        from .data.meta import Meta
 
         if not isinstance(fs, LocalFileSystem):
             return None, None

dvc/testing/fixtures.py

@@ -149,7 +149,7 @@ def local_remote(make_remote):
 @pytest.fixture
 def make_workspace(tmp_dir, dvc, make_cloud):
     def _make_workspace(name, typ="local"):
-        from dvc.objects.db import ODBManager
+        from dvc.data.db import ODBManager
 
         cloud = make_cloud(typ)  # pylint: disable=W0621
 

dvc/testing/test_workspace.py

@@ -20,7 +20,7 @@ class TestImport:
         pytest.skip()
 
     def test_import_dir(self, tmp_dir, dvc, workspace, stage_md5, dir_md5):
-        from dvc.objects.db import ODBManager
+        from dvc.data.db import ODBManager
 
         workspace.gen(
             {"dir": {"file": "file", "subdir": {"subfile": "subfile"}}}

tests/func/objects/db/test_index.py

@@ -1,8 +1,8 @@
 import pytest
 
+from dvc.data.db import get_index
 from dvc.exceptions import DownloadError, UploadError
 from dvc.fs.local import LocalFileSystem
-from dvc.objects.db import get_index
 from dvc.utils.fs import remove
 from tests.utils import clean_staging
 

tests/func/test_add.py

@@ -11,6 +11,7 @@ import colorama
 import pytest
 
 import dvc as dvc_module
+from dvc.data.db import ODBManager
 from dvc.dvcfile import DVC_FILE_SUFFIX
 from dvc.exceptions import (
     DvcException,
@@ -22,7 +23,6 @@ from dvc.exceptions import (
 from dvc.fs.local import LocalFileSystem
 from dvc.hash_info import HashInfo
 from dvc.main import main
-from dvc.objects.db import ODBManager
 from dvc.output import (
     OutputAlreadyTrackedError,
     OutputDoesNotExistError,
@@ -102,7 +102,7 @@ def test_add_unsupported_file(dvc):
 
 
 def test_add_directory(tmp_dir, dvc):
-    from dvc.objects import load
+    from dvc.data import load
 
     (stage,) = tmp_dir.dvc_gen({"dir": {"file": "file"}})
 
@@ -378,7 +378,7 @@ class TestDoubleAddUnchanged(TestDvc):
 
 
 def test_should_update_state_entry_for_file_after_add(mocker, dvc, tmp_dir):
-    file_md5_counter = mocker.spy(dvc_module.objects.stage, "file_md5")
+    file_md5_counter = mocker.spy(dvc_module.data.stage, "file_md5")
     tmp_dir.gen("foo", "foo")
 
     ret = main(["config", "cache.type", "copy"])
@@ -409,7 +409,7 @@ def test_should_update_state_entry_for_file_after_add(mocker, dvc, tmp_dir):
 def test_should_update_state_entry_for_directory_after_add(
     mocker, dvc, tmp_dir
 ):
-    file_md5_counter = mocker.spy(dvc_module.objects.stage, "file_md5")
+    file_md5_counter = mocker.spy(dvc_module.data.stage, "file_md5")
 
     tmp_dir.gen({"data/data": "foo", "data/data_sub/sub_data": "foo"})
 
@@ -456,7 +456,7 @@ class TestAddCommit(TestDvc):
 
 def test_should_collect_dir_cache_only_once(mocker, tmp_dir, dvc):
     tmp_dir.gen({"data/data": "foo"})
-    counter = mocker.spy(dvc_module.objects.stage, "_stage_tree")
+    counter = mocker.spy(dvc_module.data.stage, "_stage_tree")
     ret = main(["add", "data"])
     assert ret == 0
     assert counter.mock.call_count == 1
@@ -880,7 +880,7 @@ def test_add_with_cache_link_error(tmp_dir, dvc, mocker, capsys):
     tmp_dir.gen("foo", "foo")
 
     mocker.patch(
-        "dvc.objects.checkout.test_links",
+        "dvc.data.checkout.test_links",
         return_value=[],
     )
     dvc.add("foo")

tests/func/test_checkout.py

@@ -85,7 +85,7 @@ class TestCheckoutCorruptedCacheFile(TestRepro):
 
 class TestCheckoutCorruptedCacheDir(TestDvc):
     def test(self):
-        from dvc.objects import load
+        from dvc.data import load
 
         # NOTE: using 'copy' so that cache and link don't have same inode
         ret = main(["config", "cache.type", "copy"])

tests/func/test_data_cloud.py

@@ -107,7 +107,7 @@ def test_warn_on_outdated_stage(tmp_dir, dvc, local_remote, caplog):
 
 def test_hash_recalculation(mocker, dvc, tmp_dir, local_remote):
     tmp_dir.gen({"foo": "foo"})
-    test_file_md5 = mocker.spy(dvc_module.objects.stage, "file_md5")
+    test_file_md5 = mocker.spy(dvc_module.data.stage, "file_md5")
     ret = main(["config", "cache.type", "hardlink"])
     assert ret == 0
     ret = main(["add", "foo"])
@@ -169,7 +169,7 @@ def test_verify_hashes(
     remove("dir")
     remove(dvc.odb.local.cache_dir)
 
-    hash_spy = mocker.spy(dvc_module.objects.stage, "file_md5")
+    hash_spy = mocker.spy(dvc_module.data.stage, "file_md5")
 
     dvc.pull()
     assert hash_spy.call_count == 0

tests/func/test_external_repo.py

@@ -3,9 +3,9 @@ from unittest.mock import ANY, patch
 
 from scmrepo.git import Git
 
+from dvc.data.stage import stage
+from dvc.data.transfer import transfer
 from dvc.external_repo import CLONES, external_repo
-from dvc.objects.stage import stage
-from dvc.objects.transfer import transfer
 from dvc.utils import relpath
 from dvc.utils.fs import makedirs, remove
 from tests.unit.fs.test_repo import make_subrepo

tests/func/test_gc.py

@@ -6,10 +6,10 @@ import configobj
 import pytest
 from git import Repo
 
+from dvc.data.db.local import LocalObjectDB
 from dvc.exceptions import CollectCacheError
 from dvc.fs.local import LocalFileSystem
 from dvc.main import main
-from dvc.objects.db.local import LocalObjectDB
 from dvc.repo import Repo as DvcRepo
 from dvc.utils.fs import remove
 from tests.basic_env import TestDir, TestDvcGit

tests/func/test_get.py

@@ -3,8 +3,8 @@ import os
 
 import pytest
 
+from dvc.data.db import ODBManager
 from dvc.main import main
-from dvc.objects.db import ODBManager
 from dvc.repo import Repo
 from dvc.repo.get import GetDVCFileError
 from dvc.system import System

tests/func/test_import.py

@@ -7,9 +7,9 @@ from funcy import first
 from scmrepo.git import Git
 
 from dvc.config import NoRemoteError
+from dvc.data.db import ODBManager
 from dvc.dvcfile import Dvcfile
 from dvc.exceptions import DownloadError, PathMissingError
-from dvc.objects.db import ODBManager
 from dvc.stage.exceptions import StagePathNotFoundError
 from dvc.system import System
 from dvc.utils.fs import makedirs, remove
@@ -529,7 +529,7 @@ def test_import_with_no_exec(tmp_dir, dvc, erepo_dir):
 
 
 def test_import_with_jobs(mocker, dvc, erepo_dir):
-    import dvc.objects.transfer as otransfer
+    import dvc.data.transfer as otransfer
 
     with erepo_dir.chdir():
         erepo_dir.dvc_gen(

tests/func/test_odb.py

@@ -4,9 +4,9 @@ import stat
 import configobj
 import pytest
 
+from dvc.data.db import ODBManager
 from dvc.hash_info import HashInfo
 from dvc.main import main
-from dvc.objects.db import ODBManager
 from dvc.objects.errors import ObjectFormatError
 from dvc.utils import relpath
 from tests.basic_env import TestDir, TestDvc
@@ -47,7 +47,7 @@ class TestCacheLoadBadDirCache(TestDvc):
         self.assertEqual(len(ret), 0)
 
     def test(self):
-        from dvc.objects import load
+        from dvc.data import load
 
         dir_hash = "123.dir"
         fname = os.fspath(self.dvc.odb.local.hash_to_path(dir_hash))

tests/func/test_remote.py

@@ -139,8 +139,8 @@ def test_upper_case_remote(tmp_dir, dvc, local_cloud):
 
 
 def test_dir_hash_should_be_key_order_agnostic(tmp_dir, dvc):
-    from dvc.objects.stage import stage
-    from dvc.objects.tree import Tree
+    from dvc.data.stage import stage
+    from dvc.data.tree import Tree
 
     tmp_dir.gen({"data": {"1": "1 content", "2": "2 content"}})
 
@@ -150,7 +150,7 @@ def test_dir_hash_should_be_key_order_agnostic(tmp_dir, dvc):
         [{"relpath": "1", "md5": "1"}, {"relpath": "2", "md5": "2"}]
     )
     tree.digest()
-    with patch("dvc.objects.stage._stage_tree", return_value=(None, tree)):
+    with patch("dvc.data.stage._stage_tree", return_value=(None, tree)):
         _, _, obj = stage(dvc.odb.local, path, dvc.odb.local.fs, "md5")
         hash1 = obj.hash_info
 
@@ -158,7 +158,7 @@ def test_dir_hash_should_be_key_order_agnostic(tmp_dir, dvc):
         [{"md5": "1", "relpath": "1"}, {"md5": "2", "relpath": "2"}]
     )
     tree.digest()
-    with patch("dvc.objects.stage._stage_tree", return_value=(None, tree)):
+    with patch("dvc.data.stage._stage_tree", return_value=(None, tree)):
         _, _, obj = stage(dvc.odb.local, path, dvc.odb.local.fs, "md5")
         hash2 = obj.hash_info
 

tests/func/test_repo.py

@@ -1,5 +1,5 @@
+from dvc.data.db import ODBManager
 from dvc.dvcfile import PIPELINE_FILE, PIPELINE_LOCK
-from dvc.objects.db import ODBManager
 from dvc.system import System
 
 

tests/unit/fs/test_dvc.py

@@ -4,9 +4,9 @@ import shutil
 import pytest
 
 from dvc.config import NoRemoteError
+from dvc.data.stage import stage
 from dvc.fs.dvc import DvcFileSystem
 from dvc.hash_info import HashInfo
-from dvc.objects.stage import stage
 from dvc.utils.fs import remove
 
 
@@ -227,7 +227,7 @@ def test_get_hash_dir(tmp_dir, dvc, mocker):
         {"dir": {"foo": "foo", "bar": "bar", "subdir": {"data": "data"}}}
     )
     fs = DvcFileSystem(repo=dvc)
-    get_file_hash_spy = mocker.spy(dvc_module.objects.stage, "get_file_hash")
+    get_file_hash_spy = mocker.spy(dvc_module.data.stage, "get_file_hash")
     assert (
         fs.info((tmp_dir / "dir").fs_path)["md5"]
         == "8761c4e9acad696bee718615e23e22db.dir"

tests/unit/fs/test_repo.py

@@ -4,9 +4,9 @@ from unittest import mock
 
 import pytest
 
+from dvc.data.stage import stage
 from dvc.fs.repo import RepoFileSystem
 from dvc.hash_info import HashInfo
-from dvc.objects.stage import stage
 from tests.utils import clean_staging
 
 
@@ -584,9 +584,9 @@ def test_get_hash_mixed_dir(tmp_dir, scm, dvc):
 
 
 def test_get_hash_dirty_file(tmp_dir, dvc):
-    from dvc.objects import check
+    from dvc.data import check
+    from dvc.data.stage import get_file_hash
     from dvc.objects.errors import ObjectFormatError
-    from dvc.objects.stage import get_file_hash
 
     tmp_dir.dvc_gen("file", "file")
     file_hash_info = HashInfo("md5", "8c7dd922ad47494fc02c388e12c00eac")

tests/unit/objects/db/test_local.py

@@ -3,9 +3,9 @@ import os
 
 import pytest
 
+from dvc.data.db.local import LocalObjectDB
 from dvc.fs.local import LocalFileSystem
 from dvc.hash_info import HashInfo
-from dvc.objects.db.local import LocalObjectDB
 
 
 def test_status_download_optimization(mocker, dvc):
@@ -13,7 +13,7 @@ def test_status_download_optimization(mocker, dvc):
     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
     """
-    from dvc.objects.status import compare_status
+    from dvc.data.status import compare_status
 
     odb = LocalObjectDB(LocalFileSystem(), os.getcwd())
     obj_ids = {
@@ -85,9 +85,9 @@ def test_set_exec_ignore_errors(tmp_dir, dvc, mocker, err):
 
 
 def test_staging_file(tmp_dir, dvc):
-    from dvc.objects import check
-    from dvc.objects.stage import stage
-    from dvc.objects.transfer import transfer
+    from dvc.data import check
+    from dvc.data.stage import stage
+    from dvc.data.transfer import transfer
 
     tmp_dir.gen("foo", "foo")
     fs = LocalFileSystem()
@@ -113,9 +113,9 @@ def test_staging_file(tmp_dir, dvc):
 
 
 def test_staging_dir(tmp_dir, dvc):
-    from dvc.objects import check
-    from dvc.objects.stage import stage
-    from dvc.objects.transfer import transfer
+    from dvc.data import check
+    from dvc.data.stage import stage
+    from dvc.data.transfer import transfer
 
     tmp_dir.gen({"dir": {"foo": "foo", "bar": "bar"}})
     fs = LocalFileSystem()

tests/unit/objects/test_tree.py

@@ -2,9 +2,9 @@ from operator import itemgetter
 
 import pytest
 
+from dvc.data.meta import Meta
+from dvc.data.tree import Tree, _merge
 from dvc.hash_info import HashInfo
-from dvc.objects.meta import Meta
-from dvc.objects.tree import Tree, _merge
 
 
 @pytest.mark.parametrize(

tests/unit/output/test_local.py

@@ -1,8 +1,8 @@
 import os
 from unittest.mock import patch
 
+from dvc.data.meta import Meta
 from dvc.hash_info import HashInfo
-from dvc.objects.meta import Meta
 from dvc.output import Output
 from dvc.stage import Stage
 from dvc.utils import relpath

tests/unit/remote/test_base.py

@@ -3,7 +3,7 @@ import posixpath
 from unittest import mock
 
 from dvc.fs.base import FileSystem
-from dvc.objects.db.base import ObjectDB
+from dvc.objects.db import ObjectDB
 
 
 class _CallableOrNone:

tests/unit/remote/test_index.py

@@ -3,7 +3,7 @@ import os
 import pytest
 from funcy import first
 
-from dvc.objects.db.index import ObjectDBIndex
+from dvc.data.db.index import ObjectDBIndex
 
 
 @pytest.fixture

tests/unit/remote/test_remote.py

@@ -2,10 +2,10 @@ import os
 
 import pytest
 
+from dvc.data.db import get_index
 from dvc.fs import get_cloud_fs
 from dvc.fs.gs import GSFileSystem
 from dvc.fs.s3 import S3FileSystem
-from dvc.objects.db import get_index
 
 
 def test_remote_with_hash_jobs(dvc):

tests/unit/remote/test_slow_link_detection.py

@@ -1,14 +1,12 @@
 import pytest
 
-import dvc.objects.db.slow_link_detection
-from dvc.objects.db.slow_link_detection import slow_link_guard
+import dvc.data.slow_link_detection
+from dvc.data.slow_link_detection import slow_link_guard
 
 
 @pytest.fixture(autouse=True)
 def timeout_immediately(monkeypatch):
-    monkeypatch.setattr(
-        dvc.objects.db.slow_link_detection, "timeout_seconds", 0.0
-    )
+    monkeypatch.setattr(dvc.data.slow_link_detection, "timeout_seconds", 0.0)
 
 
 @pytest.fixture
@@ -27,7 +25,7 @@ def test_show_warning_once(caplog, make_remote):
     slow_link_guard(lambda x, y: None)(None, remote)
     slow_link_guard(lambda x, y: None)(None, remote)
 
-    slow_link_detection = dvc.objects.db.slow_link_detection
+    slow_link_detection = dvc.data.slow_link_detection
     message = slow_link_detection.message  # noqa, pylint: disable=no-member
     assert len(caplog.records) == 1
     assert caplog.records[0].message == message

tests/unit/stage/test_cache.py

@@ -167,7 +167,7 @@ def test_stage_cache_wdir(tmp_dir, dvc, mocker):
 def test_shared_stage_cache(tmp_dir, dvc, run_copy):
     import stat
 
-    from dvc.objects.db import ODBManager
+    from dvc.data.db import ODBManager
 
     tmp_dir.gen("foo", "foo")
 

tests/utils/__init__.py

@@ -45,8 +45,8 @@ def dump_sv(stream, metrics, delimiter=",", header=True):
 
 
 def clean_staging():
+    from dvc.data.stage import _STAGING_MEMFS_PATH
     from dvc.fs.memory import MemoryFileSystem
-    from dvc.objects.stage import _STAGING_MEMFS_PATH
 
     try:
         MemoryFileSystem().fs.rm(