[pre-commit.ci] pre-commit autoupdate (#10908)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

.github/workflows/benchmarks.yaml

@@ -16,14 +16,14 @@ jobs:
     steps:
       - uses: actions/setup-python@v6
         with:
-            python-version: "3.13"
+            python-version: "3.14"
 
       - uses: actions/checkout@v5
         with:
           ref: ${{ github.event.pull_request.base.sha }}
           fetch-depth: 0
 
-      - uses: astral-sh/setup-uv@v6
+      - uses: astral-sh/setup-uv@v7
         with:
           enable-cache: true
           cache-dependency-glob: pyproject.toml

.github/workflows/build.yaml

@@ -25,9 +25,9 @@ jobs:
 
     - uses: actions/setup-python@v6
       with:
-        python-version: 3.13
+        python-version: "3.14"
 
-    - uses: astral-sh/setup-uv@v6
+    - uses: astral-sh/setup-uv@v7
 
     - name: Omit local version for Test PyPI upload
       if: github.event_name == 'push' && github.ref == 'refs/heads/main'
@@ -41,7 +41,7 @@ jobs:
     - name: Check dist
       run: uv tool run twine check --strict dist/*
 
-    - uses: actions/upload-artifact@v4
+    - uses: actions/upload-artifact@v5
       with:
         name: Packages
         path: dist/
@@ -61,7 +61,7 @@ jobs:
       id-token: write
 
     steps:
-    - uses: actions/download-artifact@v5
+    - uses: actions/download-artifact@v6
       with:
         name: Packages
         path: dist
@@ -86,7 +86,7 @@ jobs:
       id-token: write
 
     steps:
-    - uses: actions/download-artifact@v5
+    - uses: actions/download-artifact@v6
       with:
         name: Packages
         path: dist

.github/workflows/codeql.yml

@@ -31,10 +31,10 @@ jobs:
       uses: actions/checkout@v5
 
     - name: Initialize CodeQL
-      uses: github/codeql-action/init@v3
+      uses: github/codeql-action/init@v4
       with:
         languages: ${{ matrix.language }}
         queries: security-extended
 
     - name: Perform CodeQL Analysis
-      uses: github/codeql-action/analyze@v3
+      uses: github/codeql-action/analyze@v4

.github/workflows/plugin_tests.yaml

@@ -19,7 +19,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest]
-        pyv: ["3.13"]
+        pyv: ["3.14"]
         plugin: ["dvc-s3"]
 
     steps:
@@ -38,7 +38,7 @@ jobs:
       with:
         python-version: ${{ matrix.pyv }}
 
-    - uses: astral-sh/setup-uv@v6
+    - uses: astral-sh/setup-uv@v7
       with:
         enable-cache: true
         cache-dependency-glob: "**/pyproject.toml"

.github/workflows/tests.yaml

@@ -32,8 +32,8 @@ jobs:
     - name: Set up Python
       uses: actions/setup-python@v6
       with:
-        python-version: '3.13'
-    - uses: astral-sh/setup-uv@v6
+        python-version: '3.14'
+    - uses: astral-sh/setup-uv@v7
       with:
         enable-cache: true
         cache-dependency-glob: pyproject.toml
@@ -55,7 +55,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest, macos-latest]
-        pyv: ["3.9", "3.10", "3.11", "3.12", "3.13"]
+        pyv: ["3.9", "3.10", "3.11", "3.12", "3.13", "3.14"]
         include:
         - os: windows-latest
           pyv: "3.9"
@@ -117,6 +117,18 @@ jobs:
         - os: windows-latest
           pyv: "3.13"
           pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 4"
+        - os: windows-latest
+          pyv: "3.14"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 1"
+        - os: windows-latest
+          pyv: "3.14"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 2"
+        - os: windows-latest
+          pyv: "3.14"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 3"
+        - os: windows-latest
+          pyv: "3.14"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 4"
     steps:
     - uses: actions/checkout@v5
       with:
@@ -125,7 +137,7 @@ jobs:
       uses: actions/setup-python@v6
       with:
         python-version: ${{ matrix.pyv }}
-    - uses: astral-sh/setup-uv@v6
+    - uses: astral-sh/setup-uv@v7
       with:
         enable-cache: true
         cache-dependency-glob: pyproject.toml

.pre-commit-config.yaml

@@ -22,7 +22,7 @@ repos:
       - id: sort-simple-yaml
       - id: trailing-whitespace
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: 'v0.13.2'
+    rev: 'v0.14.4'
     hooks:
       - id: ruff-check
         args: [--fix, --exit-non-zero-on-fix]

dvc/dependency/__init__.py

@@ -41,7 +41,7 @@ def _get(stage, p, info, **kwargs):
     return Dependency(stage, p, info, **kwargs)
 
 
-def loadd_from(stage, d_list):
+def loadd_from(stage, d_list) -> list[Dependency]:
     ret = []
     for d in d_list:
         p = d.pop(Output.PARAM_PATH, None)

dvc/dependency/param.py

@@ -110,6 +110,7 @@ class ParamsDependency(Dependency):
         except MissingParamsFile:
             return {}
 
+        assert self.repo
         try:
             return read_param_file(
                 self.repo.fs,

dvc/dependency/repo.py

@@ -150,6 +150,7 @@ class RepoDependency(Dependency):
         # local cache instead of the remote's cache. This avoids re-streaming of data,
         # but messes up the call to `_get_remote_config()` downstream, which will need
         # to ignore cache parameters.
+        assert self.repo
         config["cache"] = self.repo.config["cache"]
         config["cache"]["dir"] = self.repo.cache.local_cache_dir
 

dvc/logger.py

@@ -1,8 +1,6 @@
 """Manages logging configuration for DVC repo."""
 
 import logging
-import logging.config
-import logging.handlers
 import os
 import sys
 from collections.abc import Iterator

dvc/output.py

@@ -42,6 +42,7 @@ from .utils import relpath
 from .utils.fs import path_isin
 
 if TYPE_CHECKING:
+    from dvc.repo import Repo
     from dvc_data.hashfile.obj import HashFile
     from dvc_data.index import DataIndexKey
 
@@ -83,7 +84,7 @@ def _get(stage, path, **kwargs):
     return Output(stage, path, **kwargs)
 
 
-def loadd_from(stage, d_list):
+def loadd_from(stage, d_list) -> list["Output"]:
     ret = []
     for d in d_list:
         p = d.pop(Output.PARAM_PATH)
@@ -329,7 +330,7 @@ class Output:
         self.annot = Annotation(
             desc=desc, type=type, labels=labels or [], meta=meta or {}
         )
-        self.repo = stage.repo if not repo and stage else repo
+        self.repo: Optional[Repo] = stage.repo if not repo and stage else repo
         meta_d = merge_file_meta_from_cloud(info or {})
         meta = Meta.from_dict(meta_d)
         # NOTE: when version_aware is not passed into get_cloud_fs, it will be
@@ -511,6 +512,7 @@ class Output:
         from dvc.cachemgr import LEGACY_HASH_NAMES
 
         assert self.is_in_repo
+        assert self.repo
         odb_name = "legacy" if self.hash_name in LEGACY_HASH_NAMES else "repo"
         return getattr(self.repo.cache, odb_name)
 
@@ -518,6 +520,7 @@ class Output:
     def local_cache(self):
         from dvc.cachemgr import LEGACY_HASH_NAMES
 
+        assert self.repo
         if self.hash_name in LEGACY_HASH_NAMES:
             return self.repo.cache.legacy
         return self.repo.cache.local
@@ -585,6 +588,7 @@ class Output:
     def index_key(self) -> tuple[str, "DataIndexKey"]:
         if self.is_in_repo:
             workspace = "repo"
+            assert self.repo
             key = self.repo.fs.relparts(self.fs_path, self.repo.root_dir)
         else:
             workspace = self.fs.protocol
@@ -640,6 +644,7 @@ class Output:
     @property
     def dvcignore(self) -> Optional["DvcIgnoreFilter"]:
         if self.fs.protocol == "local":
+            assert self.repo
             return self.repo.dvcignore
         return None
 
@@ -661,6 +666,7 @@ class Output:
         if not self.use_scm_ignore:
             return
 
+        assert self.repo
         if self.repo.scm.is_tracked(self.fspath):
             raise OutputAlreadyTrackedError(self)
 
@@ -670,6 +676,7 @@ class Output:
         if not self.use_scm_ignore:
             return
 
+        assert self.repo
         self.repo.scm_context.ignore_remove(self.fspath)
 
     def save(self) -> None:
@@ -776,6 +783,7 @@ class Output:
                         callback=cb,
                     )
             if relink:
+                assert self.repo
                 rel = self.fs.relpath(filter_info or self.fs_path)
                 with CheckoutCallback(desc=f"Checking out {rel}", unit="files") as cb:
                     self._checkout(
@@ -957,6 +965,7 @@ class Output:
         added = not self.exists
 
         try:
+            assert self.repo
             modified = self._checkout(
                 filter_info or self.fs_path,
                 self.fs,
@@ -987,18 +996,24 @@ class Output:
         if ignore_remove:
             self.ignore_remove()
 
-    def move(self, out):
+    def move(self, out: "Output") -> None:
+        src_exists = self.exists
+        if src_exists:
+            self.fs.move(self.fs_path, out.fs_path)
+        else:
+            logger.warning("%r missing", self.fspath)
+
         if self.protocol == "local" and self.use_scm_ignore:
+            assert self.repo
             self.repo.scm_context.ignore_remove(self.fspath)
 
-        self.fs.move(self.fs_path, out.fs_path)
         self.def_path = out.def_path
         self.fs_path = out.fs_path
-        self.save()
-        self.commit()
-
-        if self.protocol == "local" and self.use_scm_ignore:
-            self.repo.scm_context.ignore(self.fspath)
+        try:
+            self.save()
+            self.commit()
+        except self.DoesNotExistError:
+            self.ignore()
 
     def transfer(
         self, source, odb=None, jobs=None, update=False, no_progress_bar=False
@@ -1079,6 +1094,7 @@ class Output:
             if self.remote:
                 kwargs["remote"] = self.remote
             with suppress(Exception):
+                assert self.repo
                 self.repo.cloud.pull([obj.hash_info], **kwargs)
 
         if self.obj:
@@ -1172,6 +1188,7 @@ class Output:
             return {}
 
         if self.remote:
+            assert self.repo
             remote_odb = self.repo.cloud.get_remote_odb(
                 name=self.remote, hash_name=self.hash_name
             )

dvc/render/match.py

@@ -2,7 +2,6 @@ import os
 from collections import defaultdict
 from typing import TYPE_CHECKING, NamedTuple, Optional
 
-import dpath
 import dpath.options
 from funcy import get_in, last
 

dvc/repo/__init__.py

@@ -20,6 +20,7 @@ if TYPE_CHECKING:
     from dvc.fs.data import DataFileSystem
     from dvc.fs.dvc import DVCFileSystem
     from dvc.lock import LockBase
+    from dvc.output import Output
     from dvc.scm import Git, NoSCM
     from dvc.stage import Stage
     from dvc.types import DictStrAny
@@ -544,7 +545,9 @@ class Repo:
 
         return used
 
-    def find_outs_by_path(self, path, outs=None, recursive=False, strict=True):
+    def find_outs_by_path(
+        self, path, outs=None, recursive=False, strict=True
+    ) -> list["Output"]:
         # using `outs_graph` to ensure graph checks are run
         outs = outs or self.index.outs_graph
 

dvc/repo/data.py

@@ -1,6 +1,6 @@
 import os
 import posixpath
-from collections import defaultdict
+from collections import defaultdict, deque
 from collections.abc import Iterable, Iterator, Mapping
 from typing import TYPE_CHECKING, Optional, TypedDict, Union
 
@@ -385,6 +385,26 @@ def _transform_git_paths_to_dvc(repo: "Repo", files: Iterable[str]) -> list[str]
     return [repo.fs.relpath(file, start) for file in files]
 
 
+def iter_index(
+    index: Union["BaseDataIndex", "DataIndexView"], shallow: bool = False
+) -> Iterator[tuple["DataIndexKey", Optional["DataIndexEntry"]]]:
+    if not shallow:
+        yield from index.iteritems(shallow=shallow)
+        return
+
+    # only iterate until we find entries with hash_info in shallow mode
+    todo: deque[tuple[DataIndexKey, dict]] = deque([((), index.info(()))])
+    while todo:
+        key, info = todo.popleft()
+        entry = info.get("entry")
+        if info.get("type") == "directory" and not (entry and entry.hash_info):
+            try:
+                todo.extend(index.ls(key, detail=True))
+            except (KeyError, DataIndexDirError):
+                pass
+        yield key, entry
+
+
 def _get_entries_not_in_remote(
     repo: "Repo",
     filter_keys: Optional[Iterable["DataIndexKey"]] = None,
@@ -418,7 +438,7 @@ def _get_entries_not_in_remote(
 
     n = 0
     with TqdmCallback(size=n, desc="Checking remote", unit="entry") as cb:
-        for key, entry in view.iteritems(shallow=not granular):
+        for key, entry in iter_index(view, shallow=not granular):
             if not (entry and entry.hash_info):
                 continue
 

dvc/repo/experiments/executor/base.py

@@ -9,6 +9,7 @@ from enum import IntEnum
 from itertools import chain
 from typing import TYPE_CHECKING, Any, Callable, NamedTuple, Optional, Union
 
+import funcy
 from funcy import nullcontext
 from scmrepo.exceptions import SCMError
 
@@ -815,11 +816,22 @@ class BaseExecutor(ABC):
 
     @contextmanager
     def set_temp_refs(self, scm: "Git", temp_dict: dict[str, str]):
+        # Retry ref set, get, and remove operations to handle transient issues during
+        # concurrent Git access.
+        # Dulwich deletes parent directories of refs if they happen to be empty after
+        # removing a ref, which can interfere with `set_ref` in other processes.
+        # `remove_ref` may also fail with a `FileLocked` error when refs are packed,
+        # since multiple processes might attempt to write to the same file.
+        retry = funcy.retry(10, errors=Exception, timeout=0.1)
+        set_ref = retry(scm.set_ref)
+        get_ref = retry(scm.get_ref)
+        remove_ref = retry(scm.remove_ref)
+
         try:
             for ref, rev in temp_dict.items():
-                scm.set_ref(ref, rev)
+                set_ref(ref, rev)
             yield
         finally:
             for ref in temp_dict:
-                if scm.get_ref(ref):
-                    scm.remove_ref(ref)
+                if get_ref(ref):
+                    remove_ref(ref)

dvc/repo/move.py

@@ -1,10 +1,13 @@
 import os
+from typing import TYPE_CHECKING
 
-from dvc.exceptions import MoveNotDataSourceError
 from dvc.repo.scm_context import scm_context
 
 from . import locked
 
+if TYPE_CHECKING:
+    from . import Repo
+
 
 def _expand_target_path(from_path, to_path):
     if os.path.isdir(to_path):
@@ -14,7 +17,7 @@ def _expand_target_path(from_path, to_path):
 
 @locked
 @scm_context
-def move(self, from_path, to_path):
+def move(self: "Repo", from_path, to_path):
     """
     Renames an output file and modifies the stage associated
     to reflect the change on the pipeline.
@@ -32,9 +35,12 @@ def move(self, from_path, to_path):
     It only works with outputs generated by `add` or `import`,
     also known as data sources.
     """
-    from dvc import output
+    from dvc import dependency, output
     from dvc.dvcfile import DVC_FILE_SUFFIX
+    from dvc.exceptions import MoveNotDataSourceError
     from dvc.stage import Stage
+    from dvc.stage.exceptions import StageFileAlreadyExistsError
+    from dvc_objects.fs.local import LocalFileSystem
 
     from_out = output.loads_from(Stage(self), [from_path])[0]
     assert from_out.protocol == "local"
@@ -45,6 +51,7 @@ def move(self, from_path, to_path):
     assert len(outs) == 1
     out = outs[0]
     stage = out.stage
+    deps = stage.deps
 
     if not stage.is_data_source:
         raise MoveNotDataSourceError(stage.addressing)
@@ -58,21 +65,38 @@ def move(self, from_path, to_path):
         )
         new_wdir = os.path.abspath(os.path.join(os.curdir, os.path.dirname(to_path)))
         to_path = os.path.relpath(to_path, new_wdir)
-        new_stage = self.stage.create(
-            single_stage=True,
-            fname=new_fname,
-            wdir=new_wdir,
-            outs=[to_path],
-            meta=stage.meta,
-        )
-
-        os.unlink(stage.path)
-        stage = new_stage
+        try:
+            new_stage = self.stage.create(
+                single_stage=True,
+                fname=new_fname,
+                wdir=new_wdir,
+                outs=[to_path],
+                meta=stage.meta,
+                frozen=stage.frozen,
+                always_changed=stage.always_changed,
+                desc=stage.desc,
+            )
+        except StageFileAlreadyExistsError:
+            # reraise to remove `--force` hint
+            raise StageFileAlreadyExistsError(f"{new_fname!r} already exists") from None
     else:
+        new_stage = stage
         to_path = os.path.relpath(to_path, stage.wdir)
 
-    to_out = output.loads_from(stage, [to_path], out.use_cache, out.metric)[0]
+    def with_dep_path_adjusted(dep: dependency.Dependency):
+        d = dep.dumpd()
+        if isinstance(dep.fs, LocalFileSystem) and not os.path.isabs(dep.def_path):
+            return d | {"path": os.path.relpath(dep.fspath, new_stage.wdir)}
+        return d
 
-    out.move(to_out)
-    stage.save()
-    stage.dump()
+    new_stage.outs = output.loadd_from(new_stage, [out.dumpd() | {"path": to_path}])
+    new_stage.deps = dependency.loadd_from(
+        new_stage, [with_dep_path_adjusted(dep) for dep in deps]
+    )
+    out.move(new_stage.outs[0])
+    new_stage.md5 = new_stage.compute_md5()
+    new_stage.dump()
+    if stage != new_stage:
+        stage.dvcfile.remove()
+        self.scm_context.track_file(stage.dvcfile.relpath)
+    return stage, new_stage

dvc/stage/__init__.py

@@ -32,7 +32,7 @@ from .utils import (
 )
 
 if TYPE_CHECKING:
-    from dvc.dependency import ParamsDependency
+    from dvc.dependency import Dependency, ParamsDependency
     from dvc.dvcfile import ProjectFile, SingleStageFile
     from dvc.output import Output
     from dvc.repo import Repo
@@ -153,8 +153,8 @@ class Stage(params.StageParams):
         self._path = path
         self.cmd = cmd
         self.wdir = wdir
-        self.outs = outs
-        self.deps = deps
+        self.outs: list[Output] = outs
+        self.deps: list[Dependency] = deps
         self.md5 = md5
         self.frozen = locked or frozen
         self.always_changed = always_changed

dvc/stage/serialize.py

@@ -178,7 +178,7 @@ def to_single_stage_lockfile(stage: "Stage", **kwargs) -> dict:
     res = OrderedDict([("cmd", stage.cmd)])
     params, deps = split_params_deps(stage)
     deps, outs = (
-        [_dumpd(item) for item in sorted(items, key=attrgetter("def_path"))]
+        [_dumpd(item) for item in sorted(items, key=attrgetter("def_path"))]  # type: ignore[call-overload]
         for items in [deps, stage.outs]
     )
     params = _serialize_params_values(params)

dvc/testing/benchmarks/fixtures.py

@@ -158,7 +158,7 @@ def make_bench(request):
             # pytest >= 8.4.0
             wrapped_func = fixture_function._get_wrapped_function()
         except AttributeError:
-            wrapped_func = fixture_function.__pytest_wrapped__.obj  # ty: ignore[unresolved-attribute]
+            wrapped_func = fixture_function.__pytest_wrapped__.obj  # type: ignore[attr-defined]  # ty: ignore[unresolved-attribute]
         assert inspect.isgeneratorfunction(wrapped_func)
 
         generator = wrapped_func(request)

pyproject.toml

@@ -29,6 +29,7 @@ classifiers = [
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
     "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
 ]
 dynamic = ["version"]
 

tests/conftest.py

@@ -194,6 +194,7 @@ def isolate(tmp_path_factory):
             monkeypatch.setenv(env_var, os.fspath(path))
     else:
         monkeypatch.setenv("HOME", str(home_dir))
+        monkeypatch.setenv("XDG_CONFIG_HOME", str(home_dir / ".config"))
 
     monkeypatch.setenv("GIT_CONFIG_NOSYSTEM", "1")
     contents = b"""

tests/func/test_data_status.py

@@ -1211,3 +1211,40 @@ def test_renames(
         dvc.data_status(granular=True, with_renames=True, targets=targets)
         == EMPTY_STATUS | {"git": M.dict()} | expected_granular
     )
+
+
+def test_shallow_should_iterate_upto_tracked_directory(
+    M, tmp_dir, dvc, scm, local_remote
+):
+    """Testing regression for https://github.com/iterative/dvc/issues/10899."""
+
+    tmp_dir.scm_gen({"dir": {".gitkeep": ""}}, commit="mk dir")
+    dataset_root = join("dir", "data", "")
+    tmp_dir.dvc_gen({dataset_root: {"foo": "foo", "bar": "bar"}})
+    files = [dataset_root, join(dataset_root, "foo"), join(dataset_root, "bar")]
+
+    assert dvc.data_status(not_in_remote=True) == {
+        **EMPTY_STATUS,
+        "committed": {"added": [dataset_root]},
+        "not_in_remote": [dataset_root],
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True, not_in_remote=True) == {
+        **EMPTY_STATUS,
+        "committed": {"added": M.unordered(*files)},
+        "not_in_remote": M.unordered(*files),
+        "git": M.dict(),
+    }
+
+    assert dvc.push() == 3
+
+    assert dvc.data_status(not_in_remote=True, remote_refresh=True) == {
+        **EMPTY_STATUS,
+        "committed": {"added": [dataset_root]},
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True, not_in_remote=True, remote_refresh=True) == {
+        **EMPTY_STATUS,
+        "committed": {"added": M.unordered(*files)},
+        "git": M.dict(),
+    }

tests/func/test_move.py

@@ -6,14 +6,20 @@ import pytest
 
 from dvc.cli import main
 from dvc.exceptions import MoveNotDataSourceError, OutputNotFoundError
+from dvc.stage.exceptions import StageFileAlreadyExistsError
 
 
-def test_move(tmp_dir, dvc):
-    tmp_dir.dvc_gen("foo", "foo")
-    dvc.move("foo", "foo1")
+def test_move(tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen("foo", "bar")
+    assert (tmp_dir / "foo.dvc").exists()
+    dvc.move("foo", "bar")
 
+    assert not (tmp_dir / "foo.dvc").exists()
+    assert (tmp_dir / "bar.dvc").exists()
     assert not (tmp_dir / "foo").is_file()
-    assert (tmp_dir / "foo1").is_file()
+    assert (tmp_dir / "bar").is_file()
+    # should only have the new path in the .gitignore, and only once
+    assert (tmp_dir / ".gitignore").read_text().splitlines() == ["/bar"]
 
 
 def test_move_non_existent_file(dvc):
@@ -21,6 +27,24 @@ def test_move_non_existent_file(dvc):
         dvc.move("non_existent_file", "dst")
 
 
+def test_move_missing_file(tmp_dir, dvc, scm, caplog):
+    tmp_dir.dvc_gen("foo", "foo")
+    (tmp_dir / "foo").unlink()
+    contents = (tmp_dir / "foo.dvc").parse()
+    dvc.move("foo", "bar")
+
+    assert not (tmp_dir / "foo.dvc").exists()
+    # only the path should be changed in the dvc file
+    contents["outs"][0]["path"] = "bar"
+    assert contents == (tmp_dir / "bar.dvc").parse()
+
+    # file should not be checked out
+    assert not (tmp_dir / "foo").is_file()
+    assert not (tmp_dir / "bar").is_file()
+    # should only have the new path in the .gitignore, and only once
+    assert (tmp_dir / ".gitignore").read_text().splitlines() == ["/bar"]
+
+
 def test_move_directory(tmp_dir, dvc):
     tmp_dir.dvc_gen("data", {"foo": "foo", "bar": "bar"})
     dvc.move("data", "dst")
@@ -195,3 +219,176 @@ def test_move_meta(tmp_dir, dvc):
           custom_key: 42
     """
     )
+
+
+def test_import(tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen("foo", "foo", commit="add foo")
+    imp_stage = dvc.imp(os.curdir, "foo", "foo_imported")
+
+    dvc.move("foo_imported", "foo_moved")
+
+    (stage,) = dvc.stage.collect("foo_moved.dvc")
+    assert imp_stage.md5 != stage.md5
+    res = (tmp_dir / "foo_moved.dvc").read_text()
+    assert res == textwrap.dedent(
+        f"""\
+        md5: {stage.md5}
+        frozen: true
+        deps:
+        - path: foo
+          repo:
+            url: {os.curdir}
+            rev_lock: {scm.get_rev()}
+        outs:
+        - md5: acbd18db4cc2f85cedef654fccc4a4d8
+          size: 3
+          hash: md5
+          path: foo_moved
+    """
+    )
+
+
+@pytest.mark.parametrize(
+    "path_func",
+    [pytest.param(os.path.abspath, id="abs"), pytest.param(os.path.relpath, id="rel")],
+)
+def test_import_url_in_repo(tmp_dir, dvc, path_func):
+    tmp_dir.gen("foo", "foo")
+    imp_stage = dvc.imp_url(path_func(tmp_dir / "foo"), "foo_imported")
+    (tmp_dir / "data").mkdir()
+
+    dvc.move("foo_imported", os.path.join("data", "foo_moved"))
+
+    (stage,) = dvc.stage.collect(os.path.join("data", "foo_moved.dvc"))
+    assert imp_stage.md5 != stage.md5
+    res = (tmp_dir / "data" / "foo_moved.dvc").read_text()
+    assert res == textwrap.dedent(
+        f"""\
+        md5: {stage.md5}
+        frozen: true
+        deps:
+        - md5: acbd18db4cc2f85cedef654fccc4a4d8
+          size: 3
+          hash: md5
+          path: ../foo
+        outs:
+        - md5: acbd18db4cc2f85cedef654fccc4a4d8
+          size: 3
+          hash: md5
+          path: foo_moved
+    """
+    )
+
+
+@pytest.mark.parametrize(
+    "path_func",
+    [pytest.param(os.path.abspath, id="abs"), pytest.param(os.path.relpath, id="rel")],
+)
+def test_import_url_out_of_repo(tmp_dir, dvc, scm, path_func, make_tmp_dir):
+    external = make_tmp_dir("external")
+    external.gen("foo", "foo")
+
+    imp_stage = dvc.imp_url(path_func(external / "foo"), "foo_imported")
+
+    data_dir = tmp_dir / "data"
+    data_dir.mkdir()
+
+    new_path = data_dir / "foo_moved"
+    new_dvcfile = new_path.with_suffix(".dvc")
+    dvc.move("foo_imported", os.fspath(new_path))
+
+    (stage,) = dvc.stage.collect(os.fspath(new_dvcfile))
+    assert imp_stage.md5 != stage.md5
+
+    with data_dir.chdir():
+        expected_path = path_func(external / "foo")
+
+    assert new_dvcfile.parse() == {
+        "md5": stage.md5,
+        "frozen": True,
+        "deps": [
+            {
+                "md5": "acbd18db4cc2f85cedef654fccc4a4d8",
+                "size": 3,
+                "hash": "md5",
+                "path": expected_path,
+            }
+        ],
+        "outs": [
+            {
+                "md5": "acbd18db4cc2f85cedef654fccc4a4d8",
+                "size": 3,
+                "hash": "md5",
+                "path": "foo_moved",
+            }
+        ],
+    }
+
+
+@pytest.mark.parametrize(
+    "path_func",
+    [pytest.param(os.path.abspath, id="abs"), pytest.param(os.path.relpath, id="rel")],
+)
+def test_all_metadata_are_preserved(tmp_dir, dvc, make_tmp_dir, path_func):
+    external = make_tmp_dir("external")
+    external.gen("foo", "foo")
+
+    contents = {
+        "md5": "bad",  # placeholder, does not matter for the test
+        "frozen": True,
+        "desc": "this is a stage description",
+        "always_changed": True,
+        "meta": {"custom_key": 42},
+        "deps": [
+            {
+                "md5": "acbd18db4cc2f85cedef654fccc4a4d8",
+                "size": 3,
+                "hash": "md5",
+                "path": path_func(external / "foo"),
+            }
+        ],
+        "outs": [
+            {
+                "md5": "acbd18db4cc2f85cedef654fccc4a4d8",
+                "path": "foo_imported",
+                "persist": True,
+                "hash": "md5",
+                "size": 3,
+                "desc": "this is a description",
+                "type": "model",
+                "labels": ["label1", "label2"],
+                "meta": {"custom_key": 42},
+                "cache": False,
+                "remote": "myremote",
+                "push": False,
+            }
+        ],
+    }
+    (tmp_dir / "foo_imported.dvc").dump(contents)
+    (tmp_dir / "foo_imported").write_text("foo")
+
+    data_dir = tmp_dir / "data"
+    data_dir.mkdir()
+
+    new_path = data_dir / "foo_moved"
+    new_dvcfile = new_path.with_suffix(".dvc")
+    dvc.move("foo_imported", os.fspath(new_path))
+
+    (stage,) = dvc.stage.collect(os.fspath(new_dvcfile))
+
+    with data_dir.chdir():
+        expected_path = path_func(external / "foo")
+
+    contents["outs"][0] |= {"path": "foo_moved"}
+    contents["deps"][0] |= {"path": expected_path}
+    contents |= {"md5": stage.md5}
+    assert new_dvcfile.parse() == contents
+
+
+def test_move_dst_stage_file_already_exists(tmp_dir, dvc):
+    tmp_dir.dvc_gen({"foo": "foo", "bar": "bar"})
+
+    with pytest.raises(StageFileAlreadyExistsError) as exc_info:
+        dvc.move("foo", "bar")
+    assert str(exc_info.value) == "'bar.dvc' already exists"
+    assert exc_info.value.__cause__ is None