dvc dependency repo: Catch FileNotFoundError (#10501)

- When checkouting non-existent file the _checkout_file recursively
  calls os.stat on a symlinked file which destination doesn't exist.
  This causes a FileNotFoundError but it is not catched and it causes
  an unexpected error instead.

Fixes #10500.

.github/workflows/tests.yaml

@@ -66,7 +66,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - name: Slack Notification
-      uses: rtCamp/action-slack-notify@v2.2.1
+      uses: rtCamp/action-slack-notify@v2.3.0
       env:
         SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
         SLACK_COLOR: ${{ job.status }}

.pre-commit-config.yaml

@@ -6,7 +6,7 @@ ci:
 
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.5.0
+    rev: v4.6.0
     hooks:
       - id: check-added-large-files
       - id: check-case-conflict
@@ -22,13 +22,13 @@ repos:
       - id: sort-simple-yaml
       - id: trailing-whitespace
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: 'v0.2.2'
+    rev: 'v0.5.6'
     hooks:
       - id: ruff
         args: [--fix, --exit-non-zero-on-fix]
       - id: ruff-format
   - repo: https://github.com/codespell-project/codespell
-    rev: v2.2.6
+    rev: v2.3.0
     hooks:
       - id: codespell
         additional_dependencies: ["tomli"]

CITATION.cff

@@ -0,0 +1,25 @@
+cff-version: 1.2.0
+title: 'DVC: Data Version Control - Git for Data & Models"'
+message: >-
+  If you use this software as part of a publication and wish
+  to cite it, please use the metadata from this file.
+type: software
+authors:
+  - name: The DVC team and contributors
+    website: 'https://github.com/iterative'
+repository-code: 'https://github.com/iterative/dvc'
+url: 'https://dvc.org'
+abstract: >-
+  DVC is a tool for data science that takes advantage of existing software engineering toolset.
+  It helps machine learning teams manage large datasets, make projects reproducible, and collaborate better.
+keywords:
+  - ai
+  - collaboration
+  - data-science
+  - data-version-control
+  - developer-tools
+  - git
+  - machine-learning
+  - python
+  - reproducibility
+license: Apache-2.0

README.rst

@@ -1,4 +1,6 @@
-|Banner|
+🚀 Check out our new product `DataChain <https://github.com/iterative/datachain>`_ (and give it a ⭐!) if you need to version and process a large number of files. Contact us at [email protected] to discuss commercial solutions and support for AI reproducibility and data management scenarios.
+
+--------------------------
 
 `Website <https://dvc.org>`_
 • `Docs <https://dvc.org/doc>`_

dvc/__init__.py

@@ -3,6 +3,7 @@ DVC
 ----
 Make your data science projects reproducible and shareable.
 """
+
 import dvc.logger
 from dvc.build import PKG  # noqa: F401
 from dvc.version import __version__, version_tuple  # noqa: F401

dvc/__main__.py

@@ -1,4 +1,5 @@
 """Main entry point for DVC command line tool."""
+
 import sys
 
 from dvc.cli import main

dvc/_debug.py

@@ -162,7 +162,7 @@ def _get_path_func(tool: str, ext: str):
     fmt = f"{tool}.dvc-{{now:%Y%m%d}}_{{now:%H%M%S}}.{ext}"
 
     def func(now: Optional["datetime"] = None) -> str:
-        return fmt.format(now=now or datetime.now())
+        return fmt.format(now=now or datetime.now())  # noqa: DTZ005
 
     return func
 

dvc/api/artifacts.py

@@ -1,3 +1,4 @@
+import os
 from typing import Any, Optional
 
 from dvc.repo import Repo
@@ -36,12 +37,22 @@ def artifacts_show(
     if version and stage:
         raise ValueError("Artifact version and stage are mutually exclusive.")
 
+    from dvc.repo.artifacts import Artifacts
+    from dvc.utils import as_posix
+
     repo_kwargs: dict[str, Any] = {
         "subrepos": True,
         "uninitialized": True,
     }
+
+    dirname, _ = Artifacts.parse_path(name)
     with Repo.open(repo, **repo_kwargs) as _repo:
         rev = _repo.artifacts.get_rev(name, version=version, stage=stage)
         with _repo.switch(rev):
-            path = _repo.artifacts.get_path(name)
-        return {"rev": rev, "path": path}
+            root = _repo.fs.root_marker
+            _dirname = _repo.fs.join(root, dirname) if dirname else root
+            with Repo(_dirname, fs=_repo.fs, scm=_repo.scm) as r:
+                path = r.artifacts.get_path(name)
+                path = _repo.fs.join(_repo.fs.root_marker, as_posix(path))
+                parts = _repo.fs.relparts(path, _repo.root_dir)
+                return {"rev": rev, "path": os.path.join(*parts)}

dvc/api/data.py

@@ -25,7 +25,14 @@ def _wrap_exceptions(repo, url):
         raise PathMissingError(exc.path, url) from exc
 
 
-def get_url(path, repo=None, rev=None, remote=None):
+def get_url(
+    path: str,
+    repo: Optional[str] = None,
+    rev: Optional[str] = None,
+    remote: Optional[str] = None,
+    config: Optional[dict[str, Any]] = None,
+    remote_config: Optional[dict[str, Any]] = None,
+):
     """
     Returns the URL to the storage location of a data file or directory tracked
     in a DVC repo. For Git repos, HEAD is used unless a rev argument is
@@ -35,15 +42,44 @@ def get_url(path, repo=None, rev=None, remote=None):
 
     NOTE: This function does not check for the actual existence of the file or
     directory in the remote storage.
+
+    Args:
+        path (str): location and file name of the target, relative to the root
+            of `repo`.
+        repo (str, optional): location of the DVC project or Git Repo.
+            Defaults to the current DVC project (found by walking up from the
+            current working directory tree).
+            It can be a URL or a file system path.
+            Both HTTP and SSH protocols are supported for online Git repos
+            (e.g. [user@]server:project.git).
+        rev (str, optional): Any `Git revision`_ such as a branch or tag name,
+            a commit hash or a dvc experiment name.
+            Defaults to HEAD.
+            If `repo` is not a Git repo, this option is ignored.
+        remote (str, optional): Name of the `DVC remote`_ used to form the
+            returned URL string.
+            Defaults to the `default remote`_ of `repo`.
+            For local projects, the cache is tried before the default remote.
+        config(dict, optional): config to be passed to the DVC repository.
+            Defaults to None.
+        remote_config(dict, optional): remote config to be passed to the DVC
+            repository.
+            Defaults to None.
+
+    Returns:
+        str: URL to the file or directory.
     """
     from dvc.config import NoRemoteError
     from dvc_data.index import StorageKeyError
 
-    repo_kwargs: dict[str, Any] = {}
-    if remote:
-        repo_kwargs["config"] = {"core": {"remote": remote}}
     with Repo.open(
-        repo, rev=rev, subrepos=True, uninitialized=True, **repo_kwargs
+        repo,
+        rev=rev,
+        subrepos=True,
+        uninitialized=True,
+        remote=remote,
+        config=config,
+        remote_config=remote_config,
     ) as _repo:
         index, entry = _repo.get_data_index_entry(path)
         with reraise(
@@ -249,7 +285,7 @@ def _open(
             if TYPE_CHECKING:
                 from dvc.fs import FileSystem
 
-            fs: Union["FileSystem", DataFileSystem, DVCFileSystem]
+            fs: Union[FileSystem, DataFileSystem, DVCFileSystem]
             if os.path.isabs(path):
                 fs = DataFileSystem(index=_repo.index.data["local"])
                 fs_path = path

dvc/cli/command.py

@@ -18,8 +18,8 @@ class CmdBase(ABC):
 
         os.chdir(args.cd)
 
-        self.repo: "Repo" = Repo(uninitialized=self.UNINITIALIZED)
-        self.config: "Config" = self.repo.config
+        self.repo: Repo = Repo(uninitialized=self.UNINITIALIZED)
+        self.config: Config = self.repo.config
         self.args = args
 
     def do_run(self):

dvc/cli/parser.py

@@ -1,4 +1,5 @@
 """Main parser for the dvc cli."""
+
 import argparse
 import os
 from functools import lru_cache

dvc/commands/add.py

@@ -51,6 +51,7 @@ class CmdAdd(CmdBase):
                 to_remote=self.args.to_remote,
                 remote_jobs=self.args.remote_jobs,
                 force=self.args.force,
+                relink=self.args.relink,
             )
         except FileNotFoundError:
             logger.exception("")
@@ -119,6 +120,13 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Override local file or folder if exists.",
     )
+    parser.add_argument(
+        "--no-relink",
+        dest="relink",
+        action="store_false",
+        help="Don't recreate links from cache to workspace.",
+    )
+    parser.set_defaults(relink=True)
     parser.add_argument(
         "targets", nargs="+", help="Input files/directories to add."
     ).complete = completion.FILE

dvc/commands/artifacts.py

@@ -13,6 +13,9 @@ class CmdArtifactsGet(CmdBaseNoRepo):
         from dvc.scm import CloneError
         from dvc.ui import ui
 
+        if self.args.show_url:
+            return self._show_url()
+
         try:
             count, out = Artifacts.get(
                 self.args.url,
@@ -36,6 +39,28 @@ class CmdArtifactsGet(CmdBaseNoRepo):
             )
             return 1
 
+    def _show_url(self):
+        from dvc.api import artifacts_show, get_url
+        from dvc.ui import ui
+
+        artifact = artifacts_show(
+            self.args.name,
+            version=self.args.rev,
+            stage=self.args.stage,
+            repo=self.args.url,
+        )
+
+        url = get_url(
+            artifact["path"],
+            repo=self.args.url,
+            rev=artifact["rev"],
+            remote=self.args.remote,
+            remote_config=self.args.remote_config,
+        )
+        ui.write(url, force=True)
+
+        return 0
+
 
 def add_parser(subparsers, parent_parser):
     ARTIFACTS_HELP = "DVC model registry artifact commands."
@@ -84,6 +109,14 @@ def add_parser(subparsers, parent_parser):
         help="Destination path to download artifact to",
         metavar="<path>",
     ).complete = completion.DIR
+    get_parser.add_argument(
+        "--show-url",
+        action="store_true",
+        help=(
+            "Print the storage location (URL) the target data would be "
+            "downloaded from, and exit."
+        ),
+    )
     get_parser.add_argument(
         "-j",
         "--jobs",

dvc/commands/commit.py

@@ -20,6 +20,7 @@ class CmdCommit(CmdBase):
                     with_deps=self.args.with_deps,
                     recursive=self.args.recursive,
                     force=self.args.force,
+                    relink=self.args.relink,
                 )
             except DvcException:
                 logger.exception("failed to commit%s", (" " + target) if target else "")
@@ -64,6 +65,13 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Commit cache for subdirectories of the specified directory.",
     )
+    commit_parser.add_argument(
+        "--no-relink",
+        dest="relink",
+        action="store_false",
+        help="Don't recreate links from cache to workspace.",
+    )
+    commit_parser.set_defaults(relink=True)
     commit_parser.add_argument(
         "targets",
         nargs="*",

dvc/commands/data_sync.py

@@ -194,7 +194,7 @@ def add_parser(subparsers, _parent_parser):
     )
     pull_parser.add_argument(
         "--run-cache",
-        action="store_true",
+        action=argparse.BooleanOptionalAction,
         default=False,
         help="Fetch run history for all stages.",
     )
@@ -262,7 +262,7 @@ def add_parser(subparsers, _parent_parser):
     )
     push_parser.add_argument(
         "--run-cache",
-        action="store_true",
+        action=argparse.BooleanOptionalAction,
         default=False,
         help="Push run history for all stages.",
     )
@@ -324,7 +324,7 @@ def add_parser(subparsers, _parent_parser):
     )
     fetch_parser.add_argument(
         "--run-cache",
-        action="store_true",
+        action=argparse.BooleanOptionalAction,
         default=False,
         help="Fetch run history for all stages.",
     )

dvc/commands/dataset.py

@@ -42,7 +42,7 @@ class CmdDatasetAdd(CmdBase):
             if rev := dataset.lock.rev:
                 ver = rev
 
-        ver_part: Optional["Text"] = None
+        ver_part: Optional[Text] = None
         if ver:
             ver_part = ui.rich_text.assemble(" @ ", (ver, "repr.number"))
         text = ui.rich_text.assemble("(", (url, "repr.url"), ver_part or "", ")")

dvc/commands/experiments/pull.py

@@ -1,3 +1,5 @@
+import argparse
+
 from dvc.cli import formatter
 from dvc.cli.command import CmdBase
 from dvc.cli.utils import append_doc_link
@@ -81,7 +83,7 @@ def add_parser(experiments_subparsers, parent_parser):
     )
     experiments_pull_parser.add_argument(
         "--run-cache",
-        action="store_true",
+        action=argparse.BooleanOptionalAction,
         default=False,
         help="Pull run history for all stages.",
     )

dvc/commands/experiments/push.py

@@ -1,3 +1,4 @@
+import argparse
 from typing import Any
 
 from dvc.cli import completion, formatter
@@ -120,7 +121,7 @@ def add_parser(experiments_subparsers, parent_parser):
     )
     experiments_push_parser.add_argument(
         "--run-cache",
-        action="store_true",
+        action=argparse.BooleanOptionalAction,
         default=False,
         help="Push run history for all stages.",
     )

dvc/commands/gc.py

@@ -60,7 +60,7 @@ class CmdGC(CmdBase):
             msg += " of the current and the following repos:"
 
             for repo_path in self.args.repos:
-                msg += "\n  - %s" % os.path.abspath(repo_path)
+                msg += f"\n  - {os.path.abspath(repo_path)}"
         else:
             msg += " of the current repo."
 

dvc/commands/get.py

@@ -12,7 +12,13 @@ class CmdGet(CmdBaseNoRepo):
         from dvc.api import get_url
         from dvc.ui import ui
 
-        url = get_url(self.args.path, repo=self.args.url, rev=self.args.rev)
+        url = get_url(
+            self.args.path,
+            repo=self.args.url,
+            rev=self.args.rev,
+            remote=self.args.remote,
+            remote_config=self.args.remote_config,
+        )
         ui.write(url, force=True)
 
         return 0

dvc/commands/plots.py

@@ -72,7 +72,7 @@ class CmdPlots(CmdBase):
                 html_template_path = os.path.join(self.repo.dvc_dir, html_template_path)
         return html_template_path
 
-    def run(self) -> int:  # noqa: C901, PLR0911, PLR0912
+    def run(self) -> int:  # noqa: C901, PLR0911
         from pathlib import Path
 
         from dvc.render.match import match_defs_renderers

dvc/commands/repro.py

@@ -117,10 +117,7 @@ and then the stage name name.
         "--pull",
         action="store_true",
         default=False,
-        help=(
-            "Try automatically pulling missing cache for outputs restored "
-            "from the run-cache."
-        ),
+        help="Try automatically pulling missing data.",
     )
     repro_parser.add_argument(
         "--allow-missing",

dvc/commands/stage.py

@@ -68,7 +68,7 @@ def prepare_stages_data(
 class CmdStageList(CmdBase):
     def _get_stages(self) -> Iterable["Stage"]:
         if self.args.all:
-            stages: list["Stage"] = self.repo.index.stages
+            stages: list[Stage] = self.repo.index.stages
             logger.trace("%d no. of stages found", len(stages))
             return stages
 

dvc/commands/studio.py

@@ -13,17 +13,21 @@ logger = logger.getChild(__name__)
 
 class CmdStudioLogin(CmdConfig):
     def run(self):
-        from dvc_studio_client.auth import StudioAuthError, get_access_token
-
         from dvc.env import DVC_STUDIO_URL
         from dvc.ui import ui
         from dvc.utils.studio import STUDIO_URL
-
-        name = self.args.name
-        hostname = self.args.hostname or os.environ.get(DVC_STUDIO_URL) or STUDIO_URL
-        scopes = self.args.scopes
+        from dvc_studio_client.auth import StudioAuthError, get_access_token
 
         studio = self.config.get("studio", {})
+        name = self.args.name
+        hostname = (
+            self.args.hostname
+            or os.environ.get(DVC_STUDIO_URL)
+            or studio.get("url")
+            or STUDIO_URL
+        )
+        scopes = self.args.scopes
+
         if studio.get("url", hostname) == hostname and "token" in studio:
             raise DvcException(
                 "Token already exists. "

dvc/compare.py

@@ -88,12 +88,10 @@ class TabularData(MutableSequence[Sequence["CellT"]]):
         return list(it)
 
     @overload
-    def __setitem__(self, item: int, value: Sequence["CellT"]) -> None:
-        ...
+    def __setitem__(self, item: int, value: Sequence["CellT"]) -> None: ...
 
     @overload
-    def __setitem__(self, item: slice, value: Iterable[Sequence["CellT"]]) -> None:
-        ...
+    def __setitem__(self, item: slice, value: Iterable[Sequence["CellT"]]) -> None: ...
 
     def __setitem__(self, item, value) -> None:
         it = value
@@ -162,7 +160,7 @@ class TabularData(MutableSequence[Sequence["CellT"]]):
             if key not in keys:
                 self.add_column(key)
 
-        row: list["CellT"] = [
+        row: list[CellT] = [
             with_value(d.get(key), self._fill_value) for key in self.keys()
         ]
         self.append(row)

dvc/config.py

@@ -1,4 +1,5 @@
 """DVC config objects."""
+
 import os
 import re
 from contextlib import contextmanager

dvc/config_schema.py

@@ -193,6 +193,7 @@ SCHEMA = {
                     "session_token": str,
                     Optional("listobjects", default=False): Bool,  # obsoleted
                     Optional("use_ssl", default=True): Bool,
+                    Optional("allow_anonymous_login", default=False): Bool,
                     "ssl_verify": Any(Bool, str),
                     "sse": str,
                     "sse_kms_key_id": str,
@@ -232,7 +233,12 @@ SCHEMA = {
                     Optional("verify", default=False): Bool,
                     **REMOTE_COMMON,
                 },
-                "hdfs": {"user": str, "kerb_ticket": str, **REMOTE_COMMON},
+                "hdfs": {
+                    "user": str,
+                    "kerb_ticket": str,
+                    "replication": int,
+                    **REMOTE_COMMON,
+                },
                 "webhdfs": {
                     "kerberos": Bool,
                     "kerberos_principal": str,

dvc/daemon.py

@@ -44,10 +44,10 @@ def _win_detached_subprocess(args: Sequence[str], **kwargs) -> int:
 
     startupinfo = STARTUPINFO()
     startupinfo.dwFlags |= STARTF_USESHOWWINDOW
-    popen = subprocess.Popen(
+    popen = subprocess.Popen(  # noqa: S603
         args,
         close_fds=True,
-        shell=False,  # noqa: S603
+        shell=False,
         startupinfo=startupinfo,
         creationflags=creationflags,
         **kwargs,
@@ -149,7 +149,7 @@ def _spawn(
     env: Optional[Mapping[str, str]] = None,
     output_file: Optional[str] = None,
 ) -> int:
-    file: "AbstractContextManager[Any]" = nullcontext()
+    file: AbstractContextManager[Any] = nullcontext()
     kwargs = {}
     if output_file:
         file = open(output_file, "ab")  # noqa: SIM115

dvc/dependency/db.py

@@ -22,9 +22,12 @@ def download_progress(to: "Output") -> Iterator[Callable[[int], Any]]:
     from dvc.ui import ui
     from dvc.ui._rich_progress import DbDownloadProgress
 
-    with log_durations(logger.debug, f"Saving to {to}"), DbDownloadProgress(
-        console=ui.error_console,
-    ) as progress:
+    with (
+        log_durations(logger.debug, f"Saving to {to}"),
+        DbDownloadProgress(
+            console=ui.error_console,
+        ) as progress,
+    ):
         task = progress.add_task("Saving", total=None, output=to)
         yield lambda n: progress.advance(task, advance=n)
         progress.update(task, description="Saved", total=0)
@@ -34,13 +37,17 @@ class AbstractDependency(Dependency):
     """Dependency without workspace/fs/fs_path"""
 
     def __init__(self, stage: "Stage", info: dict[str, Any], *args, **kwargs):
-        self.repo: "Repo" = stage.repo
+        self.repo: Repo = stage.repo
         self.stage = stage
         self.fs = None
         self.fs_path = None
         self.def_path = None  # type: ignore[assignment]
         self.info = info or {}
 
+    @property
+    def is_in_repo(self):
+        return False
+
 
 class DbDependency(AbstractDependency):
     PARAM_CONNECTION = "connection"
@@ -67,8 +74,8 @@ class DbDependency(AbstractDependency):
         return self.db_info.get(self.PARAM_QUERY) or self.db_info.get(self.PARAM_TABLE)
 
     def __repr__(self):
-        return "{}:{}".format(
-            self.__class__.__name__, "".join(f"{k}=={v}" for k, v in self.info.items())
+        return "{}: {}".format(
+            self.__class__.__name__, "".join(f"{k}={v}" for k, v in self.info.items())
         )
 
     def __str__(self):

dvc/dependency/repo.py

@@ -3,12 +3,14 @@ from typing import TYPE_CHECKING, Any, ClassVar, Optional, Union
 
 import voluptuous as vol
 
+from dvc.prompt import confirm
 from dvc.utils import as_posix
 
 from .base import Dependency
 
 if TYPE_CHECKING:
     from dvc.fs import DVCFileSystem
+    from dvc.output import Output
     from dvc.stage import Stage
 
 
@@ -35,7 +37,7 @@ class RepoDependency(Dependency):
         super().__init__(stage, *args, **kwargs)
 
         self.fs = self._make_fs()
-        self.fs_path = as_posix(self.def_path)
+        self.fs_path = as_posix(self.fs.normpath(self.def_path))
 
     def _parse_path(self, fs, fs_path):  # noqa: ARG002
         return None
@@ -91,6 +93,31 @@ class RepoDependency(Dependency):
             self.PARAM_REPO: self._dump_def_repo(self.def_repo),
         }
 
+    def download(self, to: "Output", jobs: Optional[int] = None):
+        from dvc_data.hashfile.build import build
+        from dvc_data.hashfile.checkout import CheckoutError, checkout
+
+        try:
+            repo = self._make_fs(locked=True).repo
+
+            _, _, obj = build(
+                repo.cache.local,
+                self.fs_path,
+                repo.dvcfs,
+                repo.cache.local.fs.PARAM_CHECKSUM,
+            )
+            checkout(
+                to.fs_path,
+                to.fs,
+                obj,
+                self.repo.cache.local,
+                ignore=None,
+                state=self.repo.state,
+                prompt=confirm,
+            )
+        except (CheckoutError, FileNotFoundError):
+            super().download(to=to, jobs=jobs)
+
     def update(self, rev: Optional[str] = None):
         if rev:
             self.def_repo[self.PARAM_REV] = rev

dvc/dvcfile.py

@@ -68,10 +68,9 @@ def is_git_ignored(repo, path):
 def check_dvcfile_path(repo, path):
     if not is_valid_filename(path):
         raise StageFileBadNameError(
-            "bad DVC file name '{}'. DVC files should be named "
-            "'{}' or have a '.dvc' suffix (e.g. '{}.dvc').".format(
-                relpath(path), PROJECT_FILE, os.path.basename(path)
-            )
+            f"bad DVC file name '{relpath(path)}'. DVC files should be named "
+            f"'{PROJECT_FILE}' or have a '.dvc' suffix "
+            f"(e.g. '{os.path.basename(path)}.dvc')."
         )
 
     if is_git_ignored(repo, path):
@@ -243,16 +242,22 @@ class ProjectFile(FileMixin):
 
     def dump_dataset(self, dataset):
         with modify_yaml(self.path, fs=self.repo.fs) as data:
-            datasets: list[dict] = data.setdefault("datasets", [])
+            parsed = self.datasets if data else []
+            raw = data.setdefault("datasets", [])
             loc = next(
-                (i for i, ds in enumerate(datasets) if ds["name"] == dataset["name"]),
+                (i for i, ds in enumerate(parsed) if ds["name"] == dataset["name"]),
                 None,
             )
             if loc is not None:
-                apply_diff(dataset, datasets[loc])
-                datasets[loc] = dataset
+                if raw[loc] != parsed[loc]:
+                    raise ParametrizedDumpError(
+                        "cannot update a parametrized dataset entry"
+                    )
+
+                apply_diff(dataset, raw[loc])
+                raw[loc] = dataset
             else:
-                datasets.append(dataset)
+                raw.append(dataset)
         self.repo.scm_context.track_file(self.relpath)
 
     def _dump_lockfile(self, stage, **kwargs):
@@ -308,29 +313,29 @@ class ProjectFile(FileMixin):
         return self.LOADER(self, self.contents, self.lockfile_contents)
 
     @property
-    def metrics(self) -> list[str]:
-        return self.contents.get("metrics", [])
+    def artifacts(self) -> dict[str, Optional[dict[str, Any]]]:
+        return self.resolver.resolve_artifacts()
 
     @property
-    def plots(self) -> Any:
-        return self.contents.get("plots", {})
+    def metrics(self) -> list[str]:
+        return self.resolver.resolve_metrics()
 
     @property
     def params(self) -> list[str]:
-        return self.contents.get("params", [])
+        return self.resolver.resolve_params()
+
+    @property
+    def plots(self) -> list[Any]:
+        return self.resolver.resolve_plots()
 
     @property
     def datasets(self) -> list[dict[str, Any]]:
-        return self.contents.get("datasets", [])
+        return self.resolver.resolve_datasets()
 
     @property
     def datasets_lock(self) -> list[dict[str, Any]]:
         return self.lockfile_contents.get("datasets", [])
 
-    @property
-    def artifacts(self) -> dict[str, Optional[dict[str, Any]]]:
-        return self.contents.get("artifacts", {})
-
     def remove(self, force=False):
         if not force:
             logger.warning("Cannot remove pipeline file.")

dvc/env.py

@@ -19,3 +19,4 @@ DVC_STUDIO_URL = "DVC_STUDIO_URL"
 DVC_SQLALCHEMY_ECHO = "DVC_SQLALCHEMY_ECHO"
 DVC_SYSTEM_CONFIG_DIR = "DVC_SYSTEM_CONFIG_DIR"
 DVC_UPDATER_ENDPOINT = "DVC_UPDATER_ENDPOINT"
+DVC_STAGE = "DVC_STAGE"

dvc/exceptions.py

@@ -1,4 +1,5 @@
 """Exceptions raised by the dvc."""
+
 import errno
 from typing import TYPE_CHECKING, Optional
 
@@ -132,11 +133,11 @@ class MoveNotDataSourceError(DvcException):
     def __init__(self, path):
         msg = (
             "move is not permitted for stages that are not data sources. "
-            "You need to either move '{path}' to a new location and edit "
-            "it by hand, or remove '{path}' and create a new one at the "
+            f"You need to either move {path!r} to a new location and edit "
+            f"it by hand, or remove {path!r} and create a new one at the "
             "desired location."
         )
-        super().__init__(msg.format(path=path))
+        super().__init__(msg)
 
 
 class NotDvcRepoError(DvcException):
@@ -270,9 +271,8 @@ class NoOutputInExternalRepoError(DvcException):
         from dvc.utils import relpath
 
         super().__init__(
-            "Output '{}' not found in target repository '{}'".format(
-                relpath(path, external_repo_path), external_repo_url
-            )
+            f"Output {relpath(path, external_repo_path)!r} "
+            f"not found in target repository '{external_repo_url}'"
         )
 
 

dvc/fs/__init__.py

@@ -1,10 +1,10 @@
+import glob
 from typing import Optional
 from urllib.parse import urlparse
 
-from dvc_http import HTTPFileSystem, HTTPSFileSystem  # noqa: F401
-
 from dvc.config import ConfigError as RepoConfigError
 from dvc.config_schema import SCHEMA, Invalid
+from dvc_http import HTTPFileSystem, HTTPSFileSystem  # noqa: F401
 
 # pylint: disable=unused-import
 from dvc_objects.fs import (  # noqa: F401
@@ -70,7 +70,14 @@ def download(
             to_infos = [to]
 
         if isinstance(fs, DVCFileSystem):
-            lfs_prefetch(fs, from_infos)
+            lfs_prefetch(
+                fs,
+                [
+                    f"{fs.normpath(glob.escape(fs_path))}/**"
+                    if fs.isdir(fs_path)
+                    else glob.escape(fs_path)
+                ],
+            )
         cb.set_size(len(from_infos))
         jobs = jobs or fs.jobs
         generic.copy(fs, from_infos, localfs, to_infos, callback=cb, batch_size=jobs)

dvc/fs/dvc.py

@@ -360,7 +360,7 @@ class _DVCFileSystem(AbstractFileSystem):
         except FileNotFoundError:
             return False
 
-    def ls(self, path, detail=True, dvc_only=False, **kwargs):  # noqa: C901
+    def ls(self, path, detail=True, dvc_only=False, **kwargs):  # noqa: C901, PLR0912
         key = self._get_key_from_relative(path)
         repo, dvc_fs, subkey = self._get_subrepo_info(key)
 
@@ -429,7 +429,7 @@ class _DVCFileSystem(AbstractFileSystem):
         ignore_subrepos = kwargs.get("ignore_subrepos", True)
         return self._info(key, path, ignore_subrepos=ignore_subrepos)
 
-    def _info(  # noqa: C901, PLR0912
+    def _info(  # noqa: C901
         self, key, path, ignore_subrepos=True, check_ignored=True
     ):
         repo, dvc_fs, subkey = self._get_subrepo_info(key)

dvc/output.py

@@ -390,7 +390,7 @@ class Output:
         self.can_push = push
 
         self.fs_path = self._parse_path(self.fs, fs_path)
-        self.obj: Optional["HashFile"] = None
+        self.obj: Optional[HashFile] = None
 
         self.remote = remote
 
@@ -902,7 +902,7 @@ class Output:
     def get_obj(
         self, filter_info: Optional[str] = None, **kwargs
     ) -> Optional["HashFile"]:
-        obj: Optional["HashFile"] = None
+        obj: Optional[HashFile] = None
         if self.obj:
             obj = self.obj
         elif self.files:
@@ -1158,7 +1158,7 @@ class Output:
             logger.warning(msg)
             return {}
 
-        obj: Optional["HashFile"]
+        obj: Optional[HashFile]
         if self.is_dir_checksum:
             obj = self._collect_used_dir_cache(**kwargs)
         else:
@@ -1353,7 +1353,7 @@ class Output:
         cache = self.cache if self.use_cache else self.local_cache
         assert isinstance(cache, HashFileDB)
 
-        new: "HashFile"
+        new: HashFile
         try:
             assert self.hash_name
             staging, meta, obj = self._build(
@@ -1365,7 +1365,7 @@ class Output:
                 dry_run=not self.use_cache,
             )
         except FileNotFoundError as exc:
-            if self.fs_path == path:
+            if not self.exists:
                 raise self.DoesNotExistError(self) from exc
             if not self.is_dir_checksum:
                 raise

dvc/parsing/__init__.py

@@ -21,6 +21,7 @@ from .context import (
     VarsAlreadyLoaded,
 )
 from .interpolate import (
+    check_expression,
     check_recursive_parse_errors,
     is_interpolated_string,
     recurse,
@@ -38,10 +39,16 @@ if TYPE_CHECKING:
 
 logger = logger.getChild(__name__)
 
-STAGES_KWD = "stages"
 VARS_KWD = "vars"
 WDIR_KWD = "wdir"
+
+ARTIFACTS_KWD = "artifacts"
+DATASETS_KWD = "datasets"
+METRICS_KWD = "metrics"
 PARAMS_KWD = "params"
+PLOTS_KWD = "plots"
+STAGES_KWD = "stages"
+
 FOREACH_KWD = "foreach"
 MATRIX_KWD = "matrix"
 DO_KWD = "do"
@@ -163,6 +170,27 @@ class DataResolver:
             for name, definition in stages_data.items()
         }
 
+        self.artifacts = [
+            ArtifactDefinition(self, self.context, name, definition, ARTIFACTS_KWD)
+            for name, definition in d.get(ARTIFACTS_KWD, {}).items()
+        ]
+        self.datasets = [
+            TopDefinition(self, self.context, str(i), definition, DATASETS_KWD)
+            for i, definition in enumerate(d.get(DATASETS_KWD, []))
+        ]
+        self.metrics = [
+            TopDefinition(self, self.context, str(i), definition, METRICS_KWD)
+            for i, definition in enumerate(d.get(METRICS_KWD, []))
+        ]
+        self.params = [
+            TopDefinition(self, self.context, str(i), definition, PARAMS_KWD)
+            for i, definition in enumerate(d.get(PARAMS_KWD, []))
+        ]
+        self.plots = [
+            TopDefinition(self, self.context, str(i), definition, PLOTS_KWD)
+            for i, definition in enumerate(d.get(PLOTS_KWD, []))
+        ]
+
     def resolve_one(self, name: str):
         group, key = split_group_name(name)
 
@@ -186,6 +214,27 @@ class DataResolver:
         logger.trace("Resolved dvc.yaml:\n%s", data)
         return {STAGES_KWD: data}
 
+    # Top-level sections are eagerly evaluated, whereas stages are lazily evaluated,
+    # one-by-one.
+
+    def resolve_artifacts(self) -> dict[str, Optional[dict[str, Any]]]:
+        d: dict[str, Optional[dict[str, Any]]] = {}
+        for item in self.artifacts:
+            d.update(item.resolve())
+        return d
+
+    def resolve_datasets(self) -> list[dict[str, Any]]:
+        return [item.resolve() for item in self.datasets]
+
+    def resolve_metrics(self) -> list[str]:
+        return [item.resolve() for item in self.metrics]
+
+    def resolve_params(self) -> list[str]:
+        return [item.resolve() for item in self.params]
+
+    def resolve_plots(self) -> list[Any]:
+        return [item.resolve() for item in self.plots]
+
     def has_key(self, key: str):
         return self._has_group_and_key(*split_group_name(key))
 
@@ -514,10 +563,10 @@ class MatrixDefinition:
         iterable = self.resolved_iterable
         assert isinstance(iterable, Mapping)
 
-        ret: dict[str, "DictStrAny"] = {}
+        ret: dict[str, DictStrAny] = {}
         matrix = {key: enumerate(v) for key, v in iterable.items()}
         for combination in product(*matrix.values()):
-            d: "DictStrAny" = {}
+            d: DictStrAny = {}
             fragments: list[str] = []
             for k, (i, v) in zip(matrix.keys(), combination):
                 d[k] = v
@@ -565,3 +614,43 @@ class MatrixDefinition:
                 return entry.resolve_stage(skip_checks=True)
             except ContextError as exc:
                 format_and_raise(exc, f"stage '{generated}'", self.relpath)
+
+
+class TopDefinition:
+    def __init__(
+        self,
+        resolver: DataResolver,
+        context: Context,
+        name: str,
+        definition: "Any",
+        where: str,
+    ):
+        self.resolver = resolver
+        self.context = context
+        self.name = name
+        self.definition = definition
+        self.where = where
+        self.relpath = self.resolver.relpath
+
+    def resolve(self):
+        try:
+            check_recursive_parse_errors(self.definition)
+            return self.context.resolve(self.definition)
+        except (ParseError, ContextError) as exc:
+            format_and_raise(exc, f"'{self.where}.{self.name}'", self.relpath)
+
+
+class ArtifactDefinition(TopDefinition):
+    def resolve(self) -> dict[str, Optional[dict[str, Any]]]:
+        try:
+            check_expression(self.name)
+            name = self.context.resolve(self.name)
+            if not isinstance(name, str):
+                typ = type(name).__name__
+                raise ResolveError(
+                    f"failed to resolve '{self.where}.{self.name}'"
+                    f" in '{self.relpath}': expected str, got " + typ
+                )
+        except (ParseError, ContextError) as exc:
+            format_and_raise(exc, f"'{self.where}.{self.name}'", self.relpath)
+        return {name: super().resolve()}

dvc/progress.py

@@ -1,4 +1,5 @@
 """Manages progress bars for DVC repo."""
+
 import logging
 import sys
 from threading import RLock
@@ -40,7 +41,7 @@ class Tqdm(tqdm):
         "miniters": 1,
     }
 
-    def __init__(  # noqa: PLR0913
+    def __init__(
         self,
         iterable=None,
         disable=None,

dvc/repo/__init__.py

@@ -184,12 +184,12 @@ class Repo:
         # used by DVCFileSystem to determine if it should traverse subrepos
         self.subrepos = subrepos
 
-        self.cloud: "DataCloud" = DataCloud(self)
-        self.stage: "StageLoad" = StageLoad(self)
+        self.cloud: DataCloud = DataCloud(self)
+        self.stage: StageLoad = StageLoad(self)
 
-        self.lock: "LockBase"
+        self.lock: LockBase
         self.cache: CacheManager
-        self.state: "StateBase"
+        self.state: StateBase
         if isinstance(self.fs, GitFileSystem) or not self.dvc_dir:
             self.lock = LockNoop()
             self.state = StateNoop()
@@ -551,11 +551,7 @@ class Repo:
 
             if out.protocol == "local" and match(fs_path, out.fs_path):
                 return True
-
-            if recursive and out.fs.isin(out.fs_path, fs_path):
-                return True
-
-            return False
+            return recursive and out.fs.isin(out.fs_path, fs_path)
 
         matched = list(filter(func, outs))
         if not matched:

dvc/repo/add.py

@@ -171,11 +171,16 @@ def _add_transfer(
     stage.dump()
 
 
-def _add(stage: "Stage", source: Optional[str] = None, no_commit: bool = False) -> None:
+def _add(
+    stage: "Stage",
+    source: Optional[str] = None,
+    no_commit: bool = False,
+    relink: bool = True,
+) -> None:
     out = stage.outs[0]
     path = out.fs.abspath(source) if source else None
     try:
-        stage.add_outs(path, no_commit=no_commit)
+        stage.add_outs(path, no_commit=no_commit, relink=relink)
     except CacheLinkError:
         stage.dump()
         raise
@@ -194,6 +199,7 @@ def add(
     to_remote: bool = False,
     remote_jobs: Optional[int] = None,
     force: bool = False,
+    relink: bool = True,
 ) -> list["Stage"]:
     add_targets = find_targets(targets, glob=glob)
     if not add_targets:
@@ -224,7 +230,12 @@ def add(
     with warn_link_failures() as link_failures:
         for source, (stage, output_exists) in progress_iter(stages_with_targets):
             try:
-                _add(stage, source if output_exists else None, no_commit=no_commit)
+                _add(
+                    stage,
+                    source if output_exists else None,
+                    no_commit=no_commit,
+                    relink=relink,
+                )
             except CacheLinkError:
                 link_failures.append(stage.relpath)
     return stages

dvc/repo/artifacts.py

@@ -12,7 +12,7 @@ from dvc.exceptions import (
     InvalidArgumentError,
 )
 from dvc.log import logger
-from dvc.utils import relpath, resolve_output
+from dvc.utils import as_posix, relpath, resolve_output
 from dvc.utils.objects import cached_property
 from dvc.utils.serialize import modify_yaml
 
@@ -99,7 +99,7 @@ class Artifacts:
         """Read artifacts from dvc.yaml."""
         artifacts: dict[str, dict[str, Artifact]] = {}
         for dvcfile, dvcfile_artifacts in self.repo.index._artifacts.items():
-            dvcyaml = relpath(dvcfile, self.repo.root_dir)
+            dvcyaml = self.repo.fs.relpath(dvcfile, self.repo.root_dir)
             artifacts[dvcyaml] = {}
             for name, value in dvcfile_artifacts.items():
                 try:
@@ -137,18 +137,16 @@ class Artifacts:
 
         assert not (version and stage)
         name = _reformat_name(name)
-        tags: list["GitTag"] = find(
-            name=name, version=version, stage=stage, scm=self.scm
-        )
+        tags: list[GitTag] = find(name=name, version=version, stage=stage, scm=self.scm)
         if not tags:
             raise ArtifactNotFoundError(name, version=version, stage=stage)
         if version or stage:
             return tags[-1].target
-        gto_tags: list["GTOTag"] = sort_versions(parse_tag(tag) for tag in tags)
+        gto_tags: list[GTOTag] = sort_versions(parse_tag(tag) for tag in tags)
         return gto_tags[0].tag.target
 
-    def get_path(self, name: str):
-        """Return repo fspath for the given artifact."""
+    @classmethod
+    def parse_path(cls, name: str) -> tuple[Optional[str], str]:
         from gto.constants import SEPARATOR_IN_NAME, fullname_re
 
         name = _reformat_name(name)
@@ -158,13 +156,37 @@ class Artifacts:
         dirname = m.group("dirname")
         if dirname:
             dirname = dirname.rstrip(SEPARATOR_IN_NAME)
-        dvcyaml = os.path.join(dirname, PROJECT_FILE) if dirname else PROJECT_FILE
-        artifact_name = m.group("name")
+
+        return dirname, m.group("name")
+
+    def get_path(self, name: str):
+        """Return fspath for the given artifact relative to the git root."""
+        from dvc.fs import GitFileSystem
+
+        dirname, artifact_name = self.parse_path(name)
+        # `name`/`dirname` are expected to be a git root relative.
+        # We convert it to dvc-root relative path so that we can read artifacts
+        # from dvc.yaml file.
+        # But we return dirname intact, as we want to return a git-root relative path.
+        # This is useful when reading from `dvcfs` from remote.
+        fs = self.repo.fs
+        assert self.scm
+        if isinstance(fs, GitFileSystem):
+            scm_root = fs.root_marker
+        else:
+            scm_root = self.scm.root_dir
+
+        dirparts = posixpath.normpath(dirname).split(posixpath.sep) if dirname else ()
+        abspath = fs.join(scm_root, *dirparts, PROJECT_FILE)
+        rela = fs.relpath(abspath, self.repo.root_dir)
         try:
-            artifact = self.read()[dvcyaml][artifact_name]
+            artifact = self.read()[rela][artifact_name]
         except KeyError as exc:
             raise ArtifactNotFoundError(name) from exc
-        return os.path.join(dirname, artifact.path) if dirname else artifact.path
+
+        path = posixpath.join(dirname or "", artifact.path)
+        parts = posixpath.normpath(path).split(posixpath.sep)
+        return os.path.join(*parts)
 
     def download(
         self,
@@ -177,15 +199,28 @@ class Artifacts:
     ) -> tuple[int, str]:
         """Download the specified artifact."""
         from dvc.fs import download as fs_download
+        from dvc.repo import Repo
 
         logger.debug("Trying to download artifact '%s' via DVC", name)
         rev = self.get_rev(name, version=version, stage=stage)
+
+        dirname, _ = self.parse_path(name)
         with self.repo.switch(rev):
-            path = self.get_path(name)
+            root = self.repo.fs.root_marker
+            _dirname = self.repo.fs.join(root, dirname) if dirname else root
+            with Repo(_dirname, fs=self.repo.fs, scm=self.repo.scm) as r:
+                path = r.artifacts.get_path(name)
+                path = self.repo.fs.join(root, as_posix(path))
+                path = self.repo.fs.relpath(path, self.repo.root_dir)
+                # when the `repo` is a subrepo, the path `/subrepo/myart.pkl` for dvcfs
+                # should be translated as `/myart.pkl`,
+                # i.e. relative to the root of the subrepo
+                path = self.repo.fs.join(root, path)
+                path = self.repo.fs.normpath(path)
+
             out = resolve_output(path, out, force=force)
             fs = self.repo.dvcfs
-            fs_path = fs.from_os_path(path)
-            count = fs_download(fs, fs_path, os.path.abspath(out), jobs=jobs)
+            count = fs_download(fs, path, os.path.abspath(out), jobs=jobs)
         return count, out
 
     @staticmethod
@@ -200,10 +235,9 @@ class Artifacts:
         dvc_studio_config: Optional[dict[str, Any]] = None,
         **kwargs,
     ) -> tuple[int, str]:
-        from dvc_studio_client.model_registry import get_download_uris
-
         from dvc.fs import HTTPFileSystem, generic, localfs
         from dvc.fs.callbacks import TqdmCallback
+        from dvc_studio_client.model_registry import get_download_uris
 
         logger.debug("Trying to download artifact '%s' via studio", name)
         out = out or os.getcwd()
@@ -231,7 +265,7 @@ class Artifacts:
                 from_infos.append(url)
         except DvcException:
             raise
-        except Exception as exc:  # noqa: BLE001
+        except Exception as exc:
             raise DvcException(
                 f"Failed to download artifact '{name}' via Studio"
             ) from exc
@@ -249,7 +283,7 @@ class Artifacts:
         return len(to_infos), relpath(localfs.commonpath(to_infos))
 
     @classmethod
-    def get(  # noqa: PLR0913
+    def get(
         cls,
         url: str,
         name: str,
@@ -278,6 +312,13 @@ class Artifacts:
 
         name = _reformat_name(name)
         saved_exc: Optional[Exception] = None
+
+        local_dvc_studio_config = Config().get("studio", {})
+        args_dvc_studio_config = {}
+        if config and not isinstance(config, dict):
+            config = Config.load_file(config)
+            args_dvc_studio_config = config.get("studio", {})
+
         try:
             logger.trace("Trying studio-only config")
             return cls._download_studio(
@@ -288,14 +329,13 @@ class Artifacts:
                 out=out,
                 force=force,
                 jobs=jobs,
+                dvc_studio_config=local_dvc_studio_config | args_dvc_studio_config,
             )
         except FileExistsLocallyError:
             raise
         except Exception as exc:  # noqa: BLE001
             saved_exc = exc
 
-        if config and not isinstance(config, dict):
-            config = Config.load_file(config)
         with Repo.open(
             url=url,
             subrepos=True,
@@ -305,7 +345,7 @@ class Artifacts:
             remote_config=remote_config,
         ) as repo:
             logger.trace("Trying repo [studio] config")
-            dvc_studio_config = dict(repo.config.get("studio"))
+            repo_dvc_studio_config = repo.config.get("studio", {})
             try:
                 return cls._download_studio(
                     url,
@@ -315,7 +355,9 @@ class Artifacts:
                     out=out,
                     force=force,
                     jobs=jobs,
-                    dvc_studio_config=dvc_studio_config,
+                    dvc_studio_config=local_dvc_studio_config
+                    | repo_dvc_studio_config
+                    | args_dvc_studio_config,
                 )
             except FileExistsLocallyError:
                 raise
@@ -333,7 +375,7 @@ class Artifacts:
                 )
             except FileExistsLocallyError:
                 raise
-            except Exception as exc:  # noqa: BLE001
+            except Exception as exc:
                 if saved_exc:
                     logger.exception(str(saved_exc), exc_info=saved_exc.__cause__)
                 raise DvcException(

dvc/repo/checkout.py

@@ -125,7 +125,7 @@ def checkout(  # noqa: C901
             (StageFileDoesNotExistError, StageFileBadNameError, NoOutputOrStageError),
         ):
             raise CheckoutErrorSuggestGit(target) from exc
-        raise
+        raise  # noqa: PLE0704
 
     view = self.index.targets_view(
         targets, recursive=recursive, with_deps=with_deps, onerror=onerror

dvc/repo/datasets.py

@@ -16,8 +16,7 @@ from dvc.types import StrPath
 from dvc_data.hashfile.meta import Meta
 
 if TYPE_CHECKING:
-    from dql.dataset import DatasetRecord  # type: ignore[import]
-    from dvcx.dataset import DatasetVersion  # type: ignore[import]
+    from dvcx.dataset import DatasetRecord, DatasetVersion  # type: ignore[import]
     from typing_extensions import Self
 
     from dvc.repo import Repo
@@ -243,7 +242,7 @@ class DatasetNotFoundError(DvcException, KeyError):
 
 class Datasets(Mapping[str, Dataset]):
     def __init__(self, repo: "Repo") -> None:
-        self.repo: "Repo" = repo
+        self.repo: Repo = repo
 
     def __repr__(self):
         return repr(dict(self))

dvc/repo/diff.py

@@ -26,7 +26,7 @@ def _diff(old, new, data_keys, with_missing=False):
     from dvc_data.index.diff import ADD, DELETE, MODIFY, RENAME
     from dvc_data.index.diff import diff as idiff
 
-    ret: "dict[str, list[dict]]" = {
+    ret: dict[str, list[dict]] = {
         "added": [],
         "deleted": [],
         "modified": [],

dvc/repo/experiments/__init__.py

@@ -119,7 +119,7 @@ class Experiments:
         **kwargs,
     ):
         """Reproduce and checkout a single (standalone) experiment."""
-        exp_queue: "BaseStashQueue" = (
+        exp_queue: BaseStashQueue = (
             self.tempdir_queue if tmp_dir else self.workspace_queue
         )
         self.queue_one(exp_queue, **kwargs)

dvc/repo/experiments/apply.py

@@ -25,7 +25,7 @@ def apply(repo: "Repo", rev: str, **kwargs):
     from dvc.repo.checkout import checkout as dvc_checkout
     from dvc.scm import RevError, resolve_rev
 
-    exps: "Experiments" = repo.experiments
+    exps: Experiments = repo.experiments
 
     is_stash: bool = False
 
@@ -57,7 +57,7 @@ def apply(repo: "Repo", rev: str, **kwargs):
 
 
 def _apply(repo: "Repo", rev: str, name: Optional[str] = None, is_stash: bool = False):
-    exps: "Experiments" = repo.experiments
+    exps: Experiments = repo.experiments
 
     with exps.apply_stash.preserve_workspace(rev, name=name):
         with repo.scm.detach_head(rev, force=True):

dvc/repo/experiments/collect.py

@@ -69,16 +69,16 @@ def collect_rev(
 
 def _collect_rev(
     repo: "Repo",
-    rev: str,
+    revision: str,
     param_deps: bool = False,
     **kwargs,
 ) -> SerializableExp:
-    with repo.switch(rev) as rev:
+    with repo.switch(revision) as rev:
         if rev == "workspace":
             timestamp: Optional[datetime] = None
         else:
             commit = repo.scm.resolve_commit(rev)
-            timestamp = datetime.fromtimestamp(commit.commit_time)
+            timestamp = datetime.fromtimestamp(commit.commit_time)  # noqa: DTZ006
 
         return SerializableExp.from_repo(
             repo,
@@ -174,7 +174,7 @@ def collect_active(
     """
     if not baseline_revs:
         return {}
-    result: dict[str, list["ExpRange"]] = {}
+    result: dict[str, list[ExpRange]] = {}
     exps = repo.experiments
     for queue in (exps.workspace_queue, exps.tempdir_queue, exps.celery_queue):
         for baseline, active_exps in queue.collect_active_data(
@@ -220,7 +220,7 @@ def collect_successful(
     Returns:
         Dict mapping baseline revision to successful experiments.
     """
-    result: dict[str, list["ExpRange"]] = {}
+    result: dict[str, list[ExpRange]] = {}
     for baseline_rev in baseline_revs:
         result[baseline_rev] = list(_collect_baseline(repo, baseline_rev, **kwargs))
     return result
@@ -304,7 +304,7 @@ def collect(
         )
 
     workspace_data = collect_rev(repo, "workspace", **kwargs)
-    result: list["ExpState"] = [workspace_data]
+    result: list[ExpState] = [workspace_data]
     queued = collect_queued(repo, baseline_revs, **kwargs) if not hide_queued else {}
     active = collect_active(repo, baseline_revs, **kwargs)
     failed = collect_failed(repo, baseline_revs, **kwargs) if not hide_failed else {}
@@ -338,6 +338,6 @@ def _sorted_ranges(exp_ranges: Iterable["ExpRange"]) -> list["ExpRange"]:
         if head_exp and head_exp.data and head_exp.data.timestamp:
             return head_exp.data.timestamp, head_exp.rev
 
-        return datetime.fromtimestamp(0), ""
+        return datetime.fromtimestamp(0), ""  # noqa: DTZ006
 
     return sorted(exp_ranges, key=_head_timestamp, reverse=True)

dvc/repo/experiments/executor/base.py

@@ -25,6 +25,7 @@ from dvc.utils import dict_sha256, env2bool, relpath
 from dvc.utils.fs import remove
 from dvc.utils.studio import (
     env_to_config,
+    get_repo_url,
     get_subrepo_relpath,
 )
 
@@ -198,7 +199,7 @@ class BaseExecutor(ABC):
     @classmethod
     def from_info(cls, info: "ExecutorInfo") -> "Self":
         if info.result_hash:
-            result: Optional["ExecutorResult"] = ExecutorResult(
+            result: Optional[ExecutorResult] = ExecutorResult(
                 info.result_hash,
                 (ExpRefInfo.from_ref(info.result_ref) if info.result_ref else None),
                 info.result_force,
@@ -469,7 +470,7 @@ class BaseExecutor(ABC):
             cls._set_log_level(log_level)
 
         exp_hash: Optional[str] = None
-        exp_ref: Optional["ExpRefInfo"] = None
+        exp_ref: Optional[ExpRefInfo] = None
         repro_force: bool = False
 
         if info.name:
@@ -558,7 +559,7 @@ class BaseExecutor(ABC):
             )
 
         ref: Optional[str] = dvc.scm.get_ref(EXEC_BRANCH, follow=False)
-        exp_ref: Optional["ExpRefInfo"] = ExpRefInfo.from_ref(ref) if ref else None
+        exp_ref: Optional[ExpRefInfo] = ExpRefInfo.from_ref(ref) if ref else None
         if cls.WARN_UNTRACKED:
             untracked = dvc.scm.untracked_files()
             if untracked:
@@ -584,9 +585,8 @@ class BaseExecutor(ABC):
         message: Optional[str] = None,
         **kwargs,
     ) -> Iterator["Repo"]:
-        from dvc_studio_client.post_live_metrics import post_live_metrics
-
         from dvc.repo import Repo
+        from dvc_studio_client.post_live_metrics import post_live_metrics
 
         with Repo(os.path.join(info.root_dir, info.dvc_dir)) as dvc:
             info.status = TaskStatus.RUNNING
@@ -611,6 +611,8 @@ class BaseExecutor(ABC):
             # inferring repo url will fail if not set here
             run_env_config = env_to_config(kwargs.get("run_env", {}))
             dvc_studio_config = run_env_config | dvc_studio_config
+            # override studio repo url if exp git remote set
+            repo_url = get_repo_url(dvc)
             try:
                 post_live_metrics(
                     "start",
@@ -621,6 +623,7 @@ class BaseExecutor(ABC):
                     dvc_studio_config=dvc_studio_config,
                     message=message,
                     subdir=get_subrepo_relpath(dvc),
+                    studio_repo_url=repo_url,
                 )
                 logger.debug("Running repro in '%s'", os.getcwd())
                 yield dvc
@@ -646,6 +649,7 @@ class BaseExecutor(ABC):
                     experiment_rev=dvc.experiments.scm.get_ref(EXEC_BRANCH),
                     metrics=_gather_metrics(dvc, on_error="return"),
                     dvc_studio_config=dvc_studio_config,
+                    studio_repo_url=repo_url,
                 )
 
                 if infofile is not None:

dvc/repo/experiments/executor/local.py

@@ -96,9 +96,10 @@ class TempDirExecutor(BaseLocalExecutor):
             temp_merge: stash_rev,
             temp_baseline: entry.baseline_rev,
         }
-        with get_exp_rwlock(
-            repo, writes=[temp_head, temp_merge, temp_baseline]
-        ), self.set_temp_refs(scm, temp_ref_dict):
+        with (
+            get_exp_rwlock(repo, writes=[temp_head, temp_merge, temp_baseline]),
+            self.set_temp_refs(scm, temp_ref_dict),
+        ):
             # Executor will be initialized with an empty git repo that
             # we populate by pushing:
             #   EXEC_HEAD - the base commit for this experiment
@@ -200,7 +201,7 @@ class WorkspaceExecutor(BaseLocalExecutor):
     @classmethod
     def from_stash_entry(cls, repo: "Repo", entry: "ExpStashEntry", **kwargs):
         root_dir = repo.scm.root_dir
-        executor: "WorkspaceExecutor" = cls._from_stash_entry(
+        executor: WorkspaceExecutor = cls._from_stash_entry(
             repo, entry, root_dir, **kwargs
         )
         logger.debug("Init workspace executor in '%s'", root_dir)

dvc/repo/experiments/pull.py

@@ -30,7 +30,7 @@ def pull(  # noqa: C901
     pull_cache: bool = False,
     **kwargs,
 ) -> Iterable[str]:
-    exp_ref_set: set["ExpRefInfo"] = set()
+    exp_ref_set: set[ExpRefInfo] = set()
     if all_commits:
         exp_ref_set.update(exp_refs(repo.scm, git_remote))
     elif exp_names:
@@ -97,7 +97,7 @@ def _pull(
     def group_result(refspec):
         return results[str(refspec)]
 
-    pull_result: Mapping[SyncStatus, list["ExpRefInfo"]] = group_by(group_result, refs)
+    pull_result: Mapping[SyncStatus, list[ExpRefInfo]] = group_by(group_result, refs)
 
     return pull_result
 

dvc/repo/experiments/push.py

@@ -96,7 +96,7 @@ def push(
     push_cache: bool = False,
     **kwargs: Any,
 ) -> dict[str, Any]:
-    exp_ref_set: set["ExpRefInfo"] = set()
+    exp_ref_set: set[ExpRefInfo] = set()
     assert isinstance(repo.scm, Git)
     if all_commits:
         exp_ref_set.update(exp_refs(repo.scm))
@@ -162,7 +162,7 @@ def _push(
     def group_result(refspec):
         return results[str(refspec)]
 
-    pull_result: Mapping[SyncStatus, list["ExpRefInfo"]] = group_by(group_result, refs)
+    pull_result: Mapping[SyncStatus, list[ExpRefInfo]] = group_by(group_result, refs)
 
     return pull_result
 

dvc/repo/experiments/queue/base.py

@@ -4,7 +4,6 @@ from collections.abc import Collection, Generator, Iterable, Mapping
 from dataclasses import asdict, dataclass
 from typing import TYPE_CHECKING, Any, NamedTuple, Optional, Union
 
-from dvc_studio_client.post_live_metrics import get_studio_config
 from funcy import retry
 
 from dvc.dependency import ParamsDependency
@@ -24,6 +23,7 @@ from dvc.repo.experiments.utils import (
 )
 from dvc.utils.objects import cached_property
 from dvc.utils.studio import config_to_env
+from dvc_studio_client.post_live_metrics import get_studio_config
 
 from .utils import get_remote_executor_refs
 
@@ -180,7 +180,7 @@ class BaseStashQueue(ABC):
 
         def _get_timestamp(rev: str) -> datetime:
             commit = self.scm.resolve_commit(rev)
-            return datetime.fromtimestamp(commit.commit_time)
+            return datetime.fromtimestamp(commit.commit_time)  # noqa: DTZ006
 
         def _format_entry(
             entry: QueueEntry,

dvc/repo/experiments/queue/celery.py

@@ -470,14 +470,14 @@ class LocalCeleryQueue(BaseStashQueue):
             exp_names = [exp_names]
         results: dict[str, ExpRefAndQueueEntry] = {}
 
-        exp_ref_match: dict[str, Optional["ExpRefInfo"]] = resolve_name(
+        exp_ref_match: dict[str, Optional[ExpRefInfo]] = resolve_name(
             self.scm, exp_names, git_remote
         )
         if not git_remote:
-            queue_entry_match: dict[
-                str, Optional["QueueEntry"]
-            ] = self.match_queue_entry_by_name(
-                exp_names, self.iter_queued(), self.iter_done()
+            queue_entry_match: dict[str, Optional[QueueEntry]] = (
+                self.match_queue_entry_by_name(
+                    exp_names, self.iter_queued(), self.iter_done()
+                )
             )
 
         for exp_name in exp_names:
@@ -613,7 +613,7 @@ class LocalCeleryQueue(BaseStashQueue):
         """Map exp refs to any available successful executors."""
         from dvc.repo.experiments.serialize import ExpExecutor, LocalExpExecutor
 
-        result: dict[str, "ExpExecutor"] = {}
+        result: dict[str, ExpExecutor] = {}
         for entry, exec_result in self.iter_success():
             if baseline_revs and entry.baseline_rev not in baseline_revs:
                 continue

dvc/repo/experiments/queue/remove.py

@@ -21,11 +21,11 @@ def remove_tasks(  # noqa: C901, PLR0912
     """
     from celery.result import AsyncResult
 
-    stash_revs: dict[str, "ExpStashEntry"] = {}
-    failed_stash_revs: list["ExpStashEntry"] = []
-    done_entry_set: set["QueueEntry"] = set()
+    stash_revs: dict[str, ExpStashEntry] = {}
+    failed_stash_revs: list[ExpStashEntry] = []
+    done_entry_set: set[QueueEntry] = set()
     stash_rev_all = celery_queue.stash.stash_revs
-    failed_rev_all: dict[str, "ExpStashEntry"] = {}
+    failed_rev_all: dict[str, ExpStashEntry] = {}
     if celery_queue.failed_stash:
         failed_rev_all = celery_queue.failed_stash.stash_revs
     for entry in queue_entries:
@@ -90,17 +90,17 @@ def celery_clear(
     """
 
     removed: list[str] = []
-    entry_list: list["QueueEntry"] = []
+    entry_list: list[QueueEntry] = []
     if queued:
-        queue_entries: list["QueueEntry"] = list(self.iter_queued())
+        queue_entries: list[QueueEntry] = list(self.iter_queued())
         entry_list.extend(queue_entries)
         removed.extend(_get_names(queue_entries))
     if failed:
-        failed_tasks: list["QueueDoneResult"] = list(self.iter_failed())
+        failed_tasks: list[QueueDoneResult] = list(self.iter_failed())
         entry_list.extend([result.entry for result in failed_tasks])
         removed.extend(_get_names(failed_tasks))
     if success:
-        success_tasks: list["QueueDoneResult"] = list(self.iter_success())
+        success_tasks: list[QueueDoneResult] = list(self.iter_success())
         entry_list.extend([result.entry for result in success_tasks])
         removed.extend(_get_names(success_tasks))
 
@@ -125,7 +125,7 @@ def celery_remove(self: "LocalCeleryQueue", revs: Collection[str]) -> list[str]:
 
     remained: list[str] = []
     removed: list[str] = []
-    entry_to_remove: list["QueueEntry"] = []
+    entry_to_remove: list[QueueEntry] = []
     for name, entry in match_results.items():
         if entry:
             entry_to_remove.append(entry)

dvc/repo/experiments/queue/tempdir.py

@@ -125,7 +125,7 @@ class TempDirQueue(WorkspaceQueue):
                 )
         except DvcException:
             raise
-        except Exception as exc:  # noqa: BLE001
+        except Exception as exc:
             raise DvcException(f"Failed to reproduce experiment '{rev[:7]}'") from exc
         finally:
             executor.cleanup(infofile)

dvc/repo/experiments/queue/workspace.py

@@ -131,7 +131,7 @@ class WorkspaceQueue(BaseStashQueue):
                 )
         except DvcException:
             raise
-        except Exception as exc:  # noqa: BLE001
+        except Exception as exc:
             raise DvcException(f"Failed to reproduce experiment '{rev[:7]}'") from exc
         finally:
             executor.cleanup(infofile)

dvc/repo/experiments/remove.py

@@ -34,19 +34,19 @@ def remove(  # noqa: C901, PLR0912
     removed: list[str] = []
     if not any([exp_names, queue, all_commits, rev]):
         return removed
-    celery_queue: "LocalCeleryQueue" = repo.experiments.celery_queue
+    celery_queue: LocalCeleryQueue = repo.experiments.celery_queue
 
     if queue:
         removed.extend(celery_queue.clear(queued=True))
 
     assert isinstance(repo.scm, Git)
 
-    exp_ref_list: list["ExpRefInfo"] = []
-    queue_entry_list: list["QueueEntry"] = []
+    exp_ref_list: list[ExpRefInfo] = []
+    queue_entry_list: list[QueueEntry] = []
     if exp_names:
-        results: dict[
-            str, "ExpRefAndQueueEntry"
-        ] = celery_queue.get_ref_and_entry_by_names(exp_names, git_remote)
+        results: dict[str, ExpRefAndQueueEntry] = (
+            celery_queue.get_ref_and_entry_by_names(exp_names, git_remote)
+        )
         remained: list[str] = []
         for name, result in results.items():
             if not result.exp_ref_info and not result.queue_entry:
@@ -94,7 +94,7 @@ def _resolve_exp_by_baseline(
 ) -> dict[str, "ExpRefInfo"]:
     assert isinstance(repo.scm, Git)
 
-    commit_ref_dict: dict[str, "ExpRefInfo"] = {}
+    commit_ref_dict: dict[str, ExpRefInfo] = {}
     rev_dict = iter_revs(repo.scm, rev, num)
     rev_set = set(rev_dict.keys())
     ref_info_dict = exp_refs_by_baseline(repo.scm, rev_set, git_remote)

dvc/repo/experiments/serialize.py

@@ -50,7 +50,12 @@ class SerializableExp:
         Params, metrics, deps, outs are filled via repo fs/index, all other fields
         should be passed via kwargs.
         """
-        from dvc.dependency import ParamsDependency, RepoDependency
+        from dvc.dependency import (
+            DatasetDependency,
+            DbDependency,
+            ParamsDependency,
+            RepoDependency,
+        )
 
         rev = rev or repo.get_rev()
         assert rev
@@ -68,7 +73,10 @@ class SerializableExp:
                     nfiles=dep.meta.nfiles if dep.meta else None,
                 )
                 for dep in repo.index.deps
-                if not isinstance(dep, (ParamsDependency, RepoDependency))
+                if not isinstance(
+                    dep,
+                    (ParamsDependency, RepoDependency, DatasetDependency, DbDependency),
+                )
             },
             outs={
                 relpath(out.fs_path, repo.root_dir): ExpOut(

dvc/repo/experiments/show.py

@@ -117,7 +117,7 @@ def _build_rows(
     **kwargs,
 ) -> Iterator[tuple["CellT", ...]]:
     for baseline in baseline_states:
-        row: dict[str, "CellT"] = dict.fromkeys(all_headers, fill_value)
+        row: dict[str, CellT] = dict.fromkeys(all_headers, fill_value)
         row["Experiment"] = ""
         if baseline.name:
             row["rev"] = baseline.name
@@ -232,7 +232,7 @@ def _exp_range_rows(
         logger.debug("Returning tip commit for legacy checkpoint exp")
     exp = first(exp_range.revs)
     if exp:
-        row: dict[str, "CellT"] = dict.fromkeys(all_headers, fill_value)
+        row: dict[str, CellT] = dict.fromkeys(all_headers, fill_value)
         row["Experiment"] = exp.name or ""
         row["rev"] = exp.rev[:7] if Git.is_sha(exp.rev) else exp.rev
         row["typ"] = "branch_base" if is_base else "branch_commit"
@@ -306,7 +306,7 @@ def format_time(
         return fill_value
     if iso:
         return timestamp.isoformat()
-    if timestamp.date() == date.today():
+    if timestamp.date() == date.today():  # noqa: DTZ011
         fmt = "%I:%M %p"
     else:
         fmt = "%b %d, %Y"

dvc/repo/experiments/utils.py

@@ -199,7 +199,7 @@ def exp_commits(
     scm: "Git", ref_infos: Optional[Iterable[ExpRefInfo]] = None
 ) -> Iterable[str]:
     """Iterate over all experiment commits."""
-    shas: set["str"] = set()
+    shas: set[str] = set()
     refs = ref_infos if ref_infos else exp_refs(scm)
     for ref_info in refs:
         shas.update(scm.branch_revs(str(ref_info), ref_info.baseline_sha))
@@ -287,7 +287,7 @@ def gen_random_name():
     # fmt: on
     # Use custom random generator to make sure that names are random even if
     # global random seed is set (common for ML pipelines).
-    random_generator = random.Random()
+    random_generator = random.Random()  # noqa: S311
     adjective = random_generator.choice(ADJECTIVES)
     noun = random_generator.choice(NOUNS)
     return f"{adjective}-{noun}"

dvc/repo/fetch.py

@@ -2,6 +2,7 @@ from typing import TYPE_CHECKING
 
 from dvc.exceptions import DownloadError
 from dvc.log import logger
+from dvc.stage.cache import RunCacheNotSupported
 from dvc.ui import ui
 from dvc_data.index import DataIndex, FileStorage
 
@@ -49,14 +50,12 @@ def _collect_indexes(  # noqa: PLR0913
         config["core"] = core
 
     def stage_filter(stage: "Stage") -> bool:
-        if push and stage.is_repo_import:
-            return False
-        return True
+        return not (push and stage.is_repo_import)
 
     def outs_filter(out: "Output") -> bool:
         if push and not out.can_push:
             return False
-        return True
+        return not (remote and out.remote and remote != out.remote)
 
     for rev in repo.brancher(
         revs=revs,
@@ -138,6 +137,8 @@ def fetch(  # noqa: PLR0913
     try:
         if run_cache:
             self.stage_cache.pull(remote)
+    except RunCacheNotSupported as e:
+        logger.debug("failed to pull run cache: %s", e)
     except DownloadError as exc:
         failed_count += exc.amount
 
@@ -202,7 +203,7 @@ def fetch(  # noqa: PLR0913
 
 
 def _log_unversioned(data: list["DataIndex"]) -> tuple[list["DataIndex"], int]:
-    ret: list["DataIndex"] = []
+    ret: list[DataIndex] = []
     unversioned: list[str] = []
     for fs_index in data:
         remote = fs_index.storage_map[()].remote

dvc/repo/gc.py

@@ -50,7 +50,7 @@ def _used_obj_ids_not_in_remote(
 
 
 @locked
-def gc(  # noqa: PLR0913, C901
+def gc(  # noqa: C901, PLR0912, PLR0913
     self: "Repo",
     all_branches: bool = False,
     cloud: bool = False,
@@ -96,7 +96,7 @@ def gc(  # noqa: PLR0913, C901
         repos = []
     all_repos = [Repo(path) for path in repos]
 
-    odb_to_obj_ids: "ObjectContainer" = {}
+    odb_to_obj_ids: ObjectContainer = {}
     with ExitStack() as stack:
         for repo in all_repos:
             stack.enter_context(repo.lock)

dvc/repo/get.py

@@ -54,7 +54,7 @@ def get(
         from dvc.fs import download
         from dvc.fs.data import DataFileSystem
 
-        fs: Union[DataFileSystem, "DVCFileSystem"]
+        fs: Union[DataFileSystem, DVCFileSystem]
         if os.path.isabs(path):
             fs = DataFileSystem(index=repo.index.data["local"])
             fs_path = fs.from_os_path(path)

dvc/repo/graph.py

@@ -22,7 +22,7 @@ def check_acyclic(graph: "DiGraph") -> None:
     except nx.NetworkXNoCycle:
         return
 
-    stages: set["Stage"] = set()
+    stages: set[Stage] = set()
     for from_node, to_node, _ in edges:
         stages.add(from_node)
         stages.add(to_node)

dvc/repo/imp_db.py

@@ -4,7 +4,6 @@ from funcy import compact
 
 from dvc.exceptions import OutputDuplicationError
 from dvc.repo.scm_context import scm_context
-from dvc.ui import ui
 from dvc.utils import resolve_output, resolve_paths
 
 if TYPE_CHECKING:
@@ -25,11 +24,6 @@ def imp_db(
     force: bool = False,
     connection: Optional[str] = None,
 ):
-    ui.warn("WARNING: import-db is an experimental feature.")
-    ui.warn(
-        "The functionality may change or break without notice, "
-        "which could lead to unexpected behavior."
-    )
     assert sql or table
     assert output_format in ("csv", "json")
 

dvc/repo/index.py

@@ -300,7 +300,7 @@ class Index:
         self._artifacts = artifacts or {}
         self._datasets: dict[str, list[dict[str, Any]]] = datasets or {}
         self._datasets_lock: dict[str, list[dict[str, Any]]] = datasets_lock or {}
-        self._collected_targets: dict[int, list["StageInfo"]] = {}
+        self._collected_targets: dict[int, list[StageInfo]] = {}
 
     @cached_property
     def rev(self) -> Optional[str]:
@@ -416,7 +416,7 @@ class Index:
 
     @cached_property
     def out_data_keys(self) -> dict[str, set["DataIndexKey"]]:
-        by_workspace: dict[str, set["DataIndexKey"]] = defaultdict(set)
+        by_workspace: dict[str, set[DataIndexKey]] = defaultdict(set)
 
         by_workspace["repo"] = set()
         by_workspace["local"] = set()
@@ -470,7 +470,7 @@ class Index:
 
     @cached_property
     def data_keys(self) -> dict[str, set["DataIndexKey"]]:
-        by_workspace: dict[str, set["DataIndexKey"]] = defaultdict(set)
+        by_workspace: dict[str, set[DataIndexKey]] = defaultdict(set)
 
         by_workspace["repo"] = set()
         by_workspace["local"] = set()
@@ -488,7 +488,7 @@ class Index:
     def metric_keys(self) -> dict[str, set["DataIndexKey"]]:
         from .metrics.show import _collect_top_level_metrics
 
-        by_workspace: dict[str, set["DataIndexKey"]] = defaultdict(set)
+        by_workspace: dict[str, set[DataIndexKey]] = defaultdict(set)
 
         by_workspace["repo"] = set()
 
@@ -509,7 +509,7 @@ class Index:
     def param_keys(self) -> dict[str, set["DataIndexKey"]]:
         from .params.show import _collect_top_level_params
 
-        by_workspace: dict[str, set["DataIndexKey"]] = defaultdict(set)
+        by_workspace: dict[str, set[DataIndexKey]] = defaultdict(set)
         by_workspace["repo"] = set()
 
         param_paths = _collect_top_level_params(self.repo)
@@ -525,7 +525,7 @@ class Index:
 
     @cached_property
     def plot_keys(self) -> dict[str, set["DataIndexKey"]]:
-        by_workspace: dict[str, set["DataIndexKey"]] = defaultdict(set)
+        by_workspace: dict[str, set[DataIndexKey]] = defaultdict(set)
 
         by_workspace["repo"] = set()
 
@@ -548,7 +548,7 @@ class Index:
 
     @cached_property
     def data(self) -> "dict[str, DataIndex]":
-        prefix: "DataIndexKey"
+        prefix: DataIndexKey
         loaded = False
 
         index = self.repo.data_index
@@ -600,7 +600,7 @@ class Index:
         if not onerror:
 
             def onerror(_target, _exc):
-                raise
+                raise  # noqa: PLE0704
 
         targets = ensure_list(targets)
         if not targets:
@@ -627,7 +627,7 @@ class Index:
         jobs: Optional[int] = None,
         push: bool = False,
     ) -> "ObjectContainer":
-        used: "ObjectContainer" = defaultdict(set)
+        used: ObjectContainer = defaultdict(set)
         pairs = self.collect_targets(targets, recursive=recursive, with_deps=with_deps)
         for stage, filter_info in pairs:
             for odb, objs in stage.get_used_objs(
@@ -748,7 +748,7 @@ class IndexView:
 
     @cached_property
     def out_data_keys(self) -> dict[str, set["DataIndexKey"]]:
-        by_workspace: dict[str, set["DataIndexKey"]] = defaultdict(set)
+        by_workspace: dict[str, set[DataIndexKey]] = defaultdict(set)
 
         by_workspace["repo"] = set()
         by_workspace["local"] = set()
@@ -764,7 +764,7 @@ class IndexView:
 
     @cached_property
     def _data_prefixes(self) -> dict[str, "_DataPrefixes"]:
-        prefixes: dict[str, "_DataPrefixes"] = defaultdict(
+        prefixes: dict[str, _DataPrefixes] = defaultdict(
             lambda: _DataPrefixes(set(), set())
         )
         for out, filter_info in self._filtered_outs:
@@ -781,7 +781,7 @@ class IndexView:
 
     @cached_property
     def data_keys(self) -> dict[str, set["DataIndexKey"]]:
-        ret: dict[str, set["DataIndexKey"]] = defaultdict(set)
+        ret: dict[str, set[DataIndexKey]] = defaultdict(set)
 
         for out, filter_info in self._filtered_outs:
             if not out.use_cache:
@@ -811,7 +811,7 @@ class IndexView:
             except KeyError:
                 return False
 
-        data: dict[str, Union["DataIndex", "DataIndexView"]] = {}
+        data: dict[str, Union[DataIndex, DataIndexView]] = {}
         for workspace, data_index in self._index.data.items():
             if self.stages:
                 data[workspace] = view(data_index, partial(key_filter, workspace))

dvc/repo/ls.py

@@ -77,7 +77,7 @@ def _ls(
     recursive: Optional[bool] = None,
     dvc_only: bool = False,
 ):
-    fs: "DVCFileSystem" = repo.dvcfs
+    fs: DVCFileSystem = repo.dvcfs
     fs_path = fs.from_os_path(path)
 
     fs_path = fs.info(fs_path)["name"]

dvc/repo/metrics/show.py

@@ -1,3 +1,4 @@
+import logging
 import os
 from collections.abc import Iterable, Iterator
 from itertools import chain
@@ -8,7 +9,7 @@ from scmrepo.exceptions import SCMError
 
 from dvc.log import logger
 from dvc.scm import NoSCMError
-from dvc.utils import as_posix, expand_paths
+from dvc.utils import as_posix
 from dvc.utils.collections import ensure_list
 from dvc.utils.serialize import load_path
 
@@ -103,7 +104,7 @@ def _collect_metrics(
     paths = (fs.from_os_path(metric) for metric in metrics)
     # make paths absolute for DVCFileSystem
     repo_paths = (f"{fs.root_marker}{path}" for path in paths)
-    return ldistinct(expand_paths(fs, repo_paths))
+    return ldistinct(try_expand_paths(fs, repo_paths))
 
 
 class FileResult(TypedDict, total=False):
@@ -116,6 +117,22 @@ class Result(TypedDict, total=False):
     error: Exception
 
 
+def try_expand_paths(fs: "FileSystem", paths: Iterable[str]) -> Iterator[str]:
+    for path in paths:
+        try:
+            if fs.isdir(path):
+                yield from fs.find(path)
+                continue
+        except Exception as e:  # noqa: BLE001
+            logger.debug(
+                "failed to expand %r: %s",
+                path,
+                e,
+                exc_info=logger.isEnabledFor(logging.TRACE),  # type: ignore[attr-defined]
+            )
+        yield path
+
+
 def to_relpath(fs: "FileSystem", root_dir: str, d: Result) -> Result:
     relpath = fs.relpath
     cwd = fs.getcwd()
@@ -207,7 +224,7 @@ def show(
                 on_error=on_error,
             )
             res[rev] = Result(data=result)
-        except Exception as exc:  # noqa: BLE001
+        except Exception as exc:
             if on_error == "raise":
                 raise
 

dvc/repo/open_repo.py

@@ -146,7 +146,7 @@ def _cached_clone(url, rev):
 
 
 @wrap_with(threading.Lock())
-def _clone_default_branch(url, rev):  # noqa: PLR0912
+def _clone_default_branch(url, rev):
     """Get or create a clean clone of the url.
 
     The cloned is reactualized with git pull unless rev is a known sha.

dvc/repo/params/show.py

@@ -6,9 +6,9 @@ from typing import TYPE_CHECKING, Any, Optional, Union
 
 from dvc.dependency.param import ParamsDependency, read_param_file
 from dvc.log import logger
-from dvc.repo.metrics.show import FileResult, Result
+from dvc.repo.metrics.show import FileResult, Result, try_expand_paths
 from dvc.stage import PipelineStage
-from dvc.utils import as_posix, expand_paths
+from dvc.utils import as_posix
 from dvc.utils.collections import ensure_list
 
 if TYPE_CHECKING:
@@ -77,7 +77,7 @@ def _collect_params(
         path = fs.from_os_path(param)
         # make paths absolute for DVCFileSystem
         repo_path = f"{fs.root_marker}{path}"
-        ret.update(dict.fromkeys(expand_paths(fs, [repo_path]), _params))
+        ret.update(dict.fromkeys(try_expand_paths(fs, [repo_path]), _params))
     return ret
 
 
@@ -191,7 +191,7 @@ def show(
                 on_error=on_error,
             )
             res[rev] = Result(data=params)
-        except Exception as exc:  # noqa: BLE001
+        except Exception as exc:
             if on_error == "raise":
                 raise
             logger.warning("failed to load params in revision %r, %s", rev, str(exc))

dvc/repo/plots/__init__.py

@@ -369,11 +369,9 @@ def _matches(targets, config_file, plot_id):
         return True
 
     full_id = get_plot_id(plot_id, config_file)
-    if any(
+    return any(
         (re.match(target, plot_id) or re.match(target, full_id)) for target in targets
-    ):
-        return True
-    return False
+    )
 
 
 def _normpath(path):

dvc/repo/push.py

@@ -1,10 +1,14 @@
 from contextlib import suppress
 
 from dvc.exceptions import InvalidArgumentError, UploadError
+from dvc.log import logger
+from dvc.stage.cache import RunCacheNotSupported
 from dvc.ui import ui
 
 from . import locked
 
+logger = logger.getChild(__name__)
+
 
 def _rebuild(idx, path, fs, cb):
     from dvc_data.index import DataIndex, DataIndexEntry, Meta
@@ -95,8 +99,11 @@ def push(  # noqa: PLR0913
                 "Multiple rev push is unsupported for cloud versioned remotes"
             )
 
-    used_run_cache = self.stage_cache.push(remote) if run_cache else []
-    transferred_count += len(used_run_cache)
+    try:
+        used_run_cache = self.stage_cache.push(remote) if run_cache else []
+        transferred_count += len(used_run_cache)
+    except RunCacheNotSupported as e:
+        logger.debug("failed to push run cache: %s", e)
 
     if isinstance(targets, str):
         targets = [targets]

dvc/repo/reproduce.py

@@ -29,7 +29,7 @@ def collect_stages(
     recursive: bool = False,
     glob: bool = False,
 ) -> list["Stage"]:
-    stages: list["Stage"] = []
+    stages: list[Stage] = []
     for target in targets:
         stages.extend(repo.stage.collect(target, recursive=recursive, glob=glob))
     return ldistinct(stages)
@@ -166,10 +166,10 @@ def _reproduce(
 ) -> list["Stage"]:
     assert on_error in ("fail", "keep-going", "ignore")
 
-    result: list["Stage"] = []
-    failed: list["Stage"] = []
-    to_skip: dict["Stage", "Stage"] = {}
-    ret: Optional["Stage"] = None
+    result: list[Stage] = []
+    failed: list[Stage] = []
+    to_skip: dict[Stage, Stage] = {}
+    ret: Optional[Stage] = None
 
     force_state = dict.fromkeys(stages, force)
 
@@ -228,7 +228,7 @@ def reproduce(
     if not kwargs.get("interactive", False):
         kwargs["interactive"] = self.config["core"].get("interactive", False)
 
-    stages: list["Stage"] = []
+    stages: list[Stage] = []
     if not all_pipelines:
         targets_list = ensure_list(targets or PROJECT_FILE)
         stages = collect_stages(self, targets_list, recursive=recursive, glob=glob)

dvc/repo/scm_context.py

@@ -21,7 +21,7 @@ class SCMContext:
     def __init__(self, scm: "Base", config: Optional[dict[str, Any]] = None) -> None:
         from funcy import get_in
 
-        self.scm: "Base" = scm
+        self.scm: Base = scm
         self.autostage: bool = get_in(
             config or {}, ["core", "autostage"], default=False
         )

dvc/repo/stage.py

@@ -108,7 +108,7 @@ def locked(f):
 
 class StageLoad:
     def __init__(self, repo: "Repo") -> None:
-        self.repo: "Repo" = repo
+        self.repo: Repo = repo
 
     @property
     def fs(self):

dvc/repo/trie.py

@@ -25,15 +25,11 @@ def build_outs_trie(stages):
                 overlapping = out
             if parent and overlapping:
                 msg = (
-                    "The output paths:\n'{}'('{}')\n'{}'('{}')\n"
+                    f"The output paths:\n'{parent!s}'('{parent.stage.addressing}')\n"
+                    f"'{overlapping!s}'('{overlapping.stage.addressing}')\n"
                     "overlap and are thus in the same tracked directory.\n"
                     "To keep reproducibility, outputs should be in separate "
                     "tracked directories or tracked individually."
-                ).format(
-                    str(parent),
-                    parent.stage.addressing,
-                    str(overlapping),
-                    overlapping.stage.addressing,
                 )
                 raise OverlappingOutputPathsError(parent, overlapping, msg)
 

dvc/repo/update.py

@@ -34,7 +34,7 @@ def update(  # noqa: C901
         raise InvalidArgumentError("--remote can't be used without --to-remote")
 
     import_stages = set()
-    other_stage_infos: list["StageInfo"] = []
+    other_stage_infos: list[StageInfo] = []
 
     for stage_info in self.index.collect_targets(targets, recursive=recursive):
         if stage_info.stage.is_import:

dvc/repo/worktree.py

@@ -76,14 +76,10 @@ def worktree_view(
     """
 
     def stage_filter(stage: "Stage") -> bool:
-        if push and stage.is_repo_import:
-            return False
-        return True
+        return not (push and stage.is_repo_import)
 
     def outs_filter(out: "Output") -> bool:
-        if not out.is_in_repo or not out.use_cache or (push and not out.can_push):
-            return False
-        return True
+        return out.is_in_repo and out.use_cache and (not push or out.can_push)
 
     return index.targets_view(
         targets,
@@ -170,7 +166,7 @@ def update_worktree_stages(repo: "Repo", stage_infos: Iterable["StageInfo"]):
 
     view = IndexView(repo.index, stage_infos, outs_filter=outs_filter)
     local_index = view.data["repo"]
-    remote_indexes: dict[str, tuple["Remote", "DataIndex"]] = {}
+    remote_indexes: dict[str, tuple[Remote, DataIndex]] = {}
     for stage in view.stages:
         for out in stage.outs:
             _update_worktree_out(repo, out, local_index, remote_indexes)

dvc/scm.py

@@ -1,4 +1,5 @@
 """Manages source control systems (e.g. Git)."""
+
 import os
 from collections.abc import Iterator, Mapping
 from contextlib import contextmanager
@@ -70,8 +71,7 @@ def SCM(
     *,
     search_parent_directories: bool = ...,
     no_scm: Literal[False] = ...,
-) -> "Git":
-    ...
+) -> "Git": ...
 
 
 @overload
@@ -80,8 +80,7 @@ def SCM(
     *,
     search_parent_directories: bool = ...,
     no_scm: Literal[True],
-) -> "NoSCM":
-    ...
+) -> "NoSCM": ...
 
 
 @overload
@@ -90,8 +89,7 @@ def SCM(
     *,
     search_parent_directories: bool = ...,
     no_scm: bool = ...,
-) -> Union["Git", "NoSCM"]:
-    ...
+) -> Union["Git", "NoSCM"]: ...
 
 
 def SCM(root_dir, *, search_parent_directories=True, no_scm=False):
@@ -241,7 +239,9 @@ def iter_revs(
         if commit_date:
             from datetime import datetime
 
-            commit_datestamp = datetime.strptime(commit_date, "%Y-%m-%d").timestamp()
+            commit_datestamp = (
+                datetime.strptime(commit_date, "%Y-%m-%d").timestamp()  # noqa: DTZ007
+            )
 
             def _time_filter(rev):
                 try:

dvc/stage/__init__.py

@@ -290,7 +290,7 @@ class Stage(params.StageParams):
         )
 
     def short_description(self) -> Optional["str"]:
-        desc: Optional["str"] = None
+        desc: Optional[str] = None
         if self.desc:
             with suppress(ValueError):
                 # try to use first non-empty line as a description
@@ -339,7 +339,7 @@ class Stage(params.StageParams):
         for out in self.outs:
             status = out.status()
             if status:
-                if allow_missing and status[str(out)] == "not in cache":
+                if allow_missing and status[str(out)] in ["not in cache", "deleted"]:
                     continue
                 logger.debug(
                     "Output '%s' of %s changed because it is '%s'.",
@@ -418,17 +418,28 @@ class Stage(params.StageParams):
 
     @rwlocked(read=["deps"], write=["outs"])
     def reproduce(self, interactive=False, **kwargs) -> Optional["Stage"]:
-        if not (
-            kwargs.get("force", False)
-            or self.changed(
-                kwargs.get("allow_missing", False), kwargs.pop("upstream", None)
-            )
-        ):
+        force = kwargs.get("force", False)
+        allow_missing = kwargs.get("allow_missing", False)
+        pull = kwargs.get("pull", False)
+        upstream = kwargs.pop("upstream", None)
+        if force:
+            pass
+        # Skip stages with missing data if otherwise unchanged
+        elif not self.changed(allow_missing, upstream):
             if not isinstance(self, PipelineStage) and self.is_data_source:
                 logger.info("'%s' didn't change, skipping", self.addressing)
             else:
                 logger.info("Stage '%s' didn't change, skipping", self.addressing)
             return None
+        # Pull stages with missing data if otherwise unchanged
+        elif not self.changed(True, upstream) and pull:
+            try:
+                logger.info("Pulling data for %s", self)
+                self.repo.pull(self.addressing, jobs=kwargs.get("jobs", None))
+                self.checkout()
+                return None
+            except CheckoutError:
+                logger.info("Unable to pull data for %s", self)
 
         msg = f"Going to reproduce {self}. Are you sure you want to continue?"
         if interactive and not prompt.confirm(msg):
@@ -597,16 +608,10 @@ class Stage(params.StageParams):
         if (self.cmd or self.is_import) and not self.frozen and not dry:
             self.remove_outs(ignore_remove=False, force=False)
 
-        if (
-            self.is_import and (not self.frozen or kwargs.get("pull"))
-        ) or self.is_partial_import:
+        if (self.is_import and not self.frozen) or self.is_partial_import:
             self._sync_import(dry, force, kwargs.get("jobs", None), no_download)
         elif not self.frozen and self.cmd:
             self._run_stage(dry, force, **kwargs)
-        elif kwargs.get("pull"):
-            logger.info("Pulling data for %s", self)
-            self.repo.pull(self.addressing, jobs=kwargs.get("jobs", None))
-            self.checkout()
         elif not dry:
             args = ("outputs", "frozen ") if self.frozen else ("data sources", "")
             logger.info("Verifying %s in %s%s", *args, self)
@@ -653,7 +658,7 @@ class Stage(params.StageParams):
     def checkout(
         self, allow_missing: bool = False, **kwargs
     ) -> dict[str, list["StrPath"]]:
-        stats: dict[str, list["StrPath"]] = defaultdict(list)
+        stats: dict[str, list[StrPath]] = defaultdict(list)
         if self.is_partial_import:
             return stats
 

dvc/stage/cache.py

@@ -229,7 +229,7 @@ class StageCache:
         if not dry:
             cached_stage.checkout()
 
-    def transfer(self, from_odb, to_odb):
+    def transfer(self, from_odb, to_odb, force=True):
         from dvc.fs import HTTPFileSystem, LocalFileSystem
         from dvc.fs.callbacks import TqdmCallback
 
@@ -258,10 +258,11 @@ class StageCache:
 
             dst = to_fs.join(to_odb.path, rel)
             key = to_fs.parent(dst)
+
             # check if any build cache already exists for this key
             # TODO: check if MaxKeys=1 or something like that applies
             # or otherwise this will take a lot of time!
-            if to_fs.exists(key) and first(to_fs.find(key)):
+            if not force and to_fs.exists(key) and first(to_fs.find(key)):
                 continue
 
             src_name = from_fs.name(src)

dvc/stage/loader.py

@@ -50,7 +50,7 @@ class StageLoader(Mapping):
         from dvc.output import Output, merge_file_meta_from_cloud
 
         assert isinstance(lock_data, dict)
-        items: Iterable[tuple[str, "Output"]] = chain(
+        items: Iterable[tuple[str, Output]] = chain(
             ((StageParams.PARAM_DEPS, dep) for dep in stage.deps),
             ((StageParams.PARAM_OUTS, out) for out in stage.outs),
         )

dvc/stage/run.py

@@ -2,6 +2,9 @@ import os
 import signal
 import subprocess
 import threading
+from functools import cache
+
+from packaging.version import InvalidVersion, Version
 
 from dvc.log import logger
 from dvc.utils import fix_env
@@ -12,25 +15,59 @@ from .exceptions import StageCmdFailedError
 logger = logger.getChild(__name__)
 
 
+@cache
+def _fish_supports_no_config(executable) -> bool:
+    """
+    Check if the fish shell supports the --no-config option.
+
+    Parameters:
+    executable (str): The path to the fish shell executable.
+
+    Returns:
+    bool: True if the fish version is greater than 3.3.0, False otherwise.
+    """
+    try:
+        output = subprocess.check_output(  # noqa: S603
+            [executable, "--version"],
+            text=True,
+        )
+        version = Version(output.split(" ")[-1].strip())
+        version_to_check = Version("3.3.0")
+        return version >= version_to_check
+    except (subprocess.CalledProcessError, IndexError, InvalidVersion):
+        logger.trace("could not check fish version, defaulting to False")
+        return False
+
+
+def _warn_if_fish(executable):
+    if (
+        executable is None
+        or os.path.basename(executable) != "fish"
+        or _fish_supports_no_config(executable)
+    ):
+        return
+    logger.warning(
+        "DVC detected that you are using a version of fish shell below 3.3.0 "
+        "Be aware that it might cause problems by overwriting "
+        "your current environment variables with values defined "
+        "in 'config.fish', which might affect your command. See "
+        "https://github.com/iterative/dvc/issues/1307. "
+    )
+
+
 def _make_cmd(executable, cmd):
     if executable is None:
         return cmd
-    opts = {"zsh": ["--no-rcs"], "bash": ["--noprofile", "--norc"]}
+    opts = {
+        "zsh": ["--no-rcs"],
+        "bash": ["--noprofile", "--norc"],
+        "fish": [],
+    }
     name = os.path.basename(executable).lower()
-    return [executable, *opts.get(name, []), "-c", cmd]
-
-
-def warn_if_fish(executable):
-    if executable is None or os.path.basename(os.path.realpath(executable)) != "fish":
-        return
-
-    logger.warning(
-        "DVC detected that you are using fish as your default "
-        "shell. Be aware that it might cause problems by overwriting "
-        "your current environment variables with values defined "
-        "in '.fishrc', which might affect your command. See "
-        "https://github.com/iterative/dvc/issues/1307. "
-    )
+    opt = opts.get(name, [])
+    if name == "fish" and _fish_supports_no_config(executable):
+        opt.append("--no-config")
+    return [executable, *opt, "-c", cmd]
 
 
 def _enforce_cmd_list(cmd):
@@ -39,7 +76,7 @@ def _enforce_cmd_list(cmd):
 
 
 def prepare_kwargs(stage, run_env=None):
-    from dvc.env import DVC_ROOT
+    from dvc.env import DVC_ROOT, DVC_STAGE
 
     kwargs = {"cwd": stage.wdir, "env": fix_env(None), "close_fds": True}
 
@@ -48,6 +85,9 @@ def prepare_kwargs(stage, run_env=None):
     if DVC_ROOT not in kwargs["env"]:
         kwargs["env"][DVC_ROOT] = stage.repo.root_dir
 
+    # Create DVC_STAGE env variable for every command
+    kwargs["env"][DVC_STAGE] = stage.addressing
+
     # NOTE: when you specify `shell=True`, `Popen` [1] will default to
     # `/bin/sh` on *nix and will add ["/bin/sh", "-c"] to your command.
     # But we actually want to run the same shell that we are running
@@ -103,7 +143,7 @@ def cmd_run(stage, dry=False, run_env=None):
     executable = get_executable()
 
     if not dry:
-        warn_if_fish(executable)
+        _warn_if_fish(executable)
 
     for cmd in commands:
         display_command(cmd)
@@ -114,7 +154,11 @@ def cmd_run(stage, dry=False, run_env=None):
 
 
 def _pull_missing_deps(stage):
+    from dvc.dependency import DatasetDependency, DbDependency
+
     for dep in stage.deps:
+        if isinstance(dep, (DatasetDependency, DbDependency)):
+            continue
         if not dep.exists:
             stage.repo.pull(dep.def_path)
 

dvc/stage/utils.py

@@ -92,9 +92,7 @@ def check_no_externals(stage):
     from dvc.utils import format_link
 
     def _is_cached_external(out):
-        if out.is_in_repo or not out.use_cache:
-            return False
-        return True
+        return not out.is_in_repo and out.use_cache
 
     outs = [str(out) for out in stage.outs if _is_cached_external(out)]
     if not outs:

dvc/testing/benchmarks/cli/stories/test_modify_data.py

@@ -1,6 +1,7 @@
 """
 Tests for modifications to an existing dataset.
 """
+
 import glob
 import os
 import random

dvc/testing/benchmarks/fixtures.py

@@ -104,6 +104,11 @@ def make_dvc_bin(
             venv.run("pip", "install", f"git+file://{dvc_git_repo}@{dvc_rev}#egg={egg}")
             if dvc_rev in ["2.18.1", "2.11.0", "2.6.3"]:
                 venv.run("pip", "install", "fsspec==2022.11.0")
+            try:
+                if version.Version(dvc_rev) < version.Version("3.50.3"):
+                    venv.run("pip", "install", "pygit2==1.14.1")
+            except version.InvalidVersion:
+                pass
             dvc_venvs[dvc_rev] = venv
         dvc_bin = venv.which("dvc")
     else:

dvc/testing/fixtures.py

@@ -228,10 +228,10 @@ def docker_services(tmp_path_factory, request):
         pytest.skip("disabled for Windows on CI")
 
     try:
-        subprocess.check_output(
+        subprocess.check_output(  # noqa: S602
             "docker ps",  # noqa: S607
             stderr=subprocess.STDOUT,
-            shell=True,  # noqa: S602
+            shell=True,
         )
     except subprocess.CalledProcessError as err:
         out = (err.output or b"").decode("utf-8")

dvc/testing/remote_tests.py

@@ -3,7 +3,6 @@ import shutil
 
 import pytest
 
-from dvc.stage.cache import RunCacheNotSupported
 from dvc.utils.fs import remove
 from dvc_data.hashfile.tree import Tree
 
@@ -97,8 +96,10 @@ class TestRemote:
         status_dir = dvc.cloud.status(dir_hashes)
         _check_status(status_dir, ok=dir_hashes)
 
-    @pytest.mark.xfail(raises=RunCacheNotSupported, strict=False)
     def test_stage_cache_push_pull(self, tmp_dir, dvc, remote):
+        if remote.scheme in ("http", "https"):
+            pytest.skip("HTTP remote does not support stage cache")
+
         tmp_dir.gen("foo", "foo")
         stage = dvc.stage.add(
             deps=["foo"], outs=["bar"], name="copy-foo-bar", cmd="cp foo bar"

dvc/ui/__init__.py

@@ -93,7 +93,7 @@ class Console:
             styled=styled,
         )
 
-    def write_json(  # noqa: PLR0913
+    def write_json(
         self,
         data: Any,
         indent: Optional[int] = None,

dvc/utils/__init__.py

@@ -5,7 +5,6 @@ import json
 import os
 import re
 import sys
-from collections.abc import Iterable, Iterator
 from typing import TYPE_CHECKING, Optional
 
 import colorama
@@ -13,8 +12,6 @@ import colorama
 if TYPE_CHECKING:
     from typing import TextIO
 
-    from dvc.fs import FileSystem
-
 
 LARGE_DIR_SIZE = 100
 TARGET_REGEX = re.compile(r"(?P<path>.*?)(:(?P<name>[^\\/:]*))??$")
@@ -410,14 +407,6 @@ def errored_revisions(rev_data: dict) -> list:
     return result
 
 
-def expand_paths(fs: "FileSystem", paths: Iterable[str]) -> Iterator[str]:
-    for path in paths:
-        if fs.isdir(path):
-            yield from fs.find(path)
-        else:
-            yield path
-
-
 def isatty(stream: "Optional[TextIO]") -> bool:
     if stream is None:
         return False

dvc/utils/collections.py

@@ -39,9 +39,7 @@ def apply_diff(src, dest):  # noqa: C901
                     dest[i] = value
     else:
         raise AssertionError(  # noqa: TRY004
-            "Can't apply diff from {} to {}".format(
-                src.__class__.__name__, dest.__class__.__name__
-            )
+            f"Can't apply diff from {type(src).__name__} to {type(dest).__name__}"
         )
 
 

dvc/utils/serialize/_common.py

@@ -1,4 +1,5 @@
 """Common utilities for serialize."""
+
 import os
 from contextlib import AbstractContextManager, contextmanager
 from typing import TYPE_CHECKING, Any, Callable, Optional, Protocol, TextIO, Union
@@ -15,25 +16,21 @@ if TYPE_CHECKING:
 class DumperFn(Protocol):
     def __call__(
         self, path: "StrPath", data: Any, fs: Optional["FileSystem"] = None
-    ) -> Any:
-        ...
+    ) -> Any: ...
 
 
 class DumpersFn(Protocol):
-    def __call__(self, data: Any, stream: TextIO) -> Any:
-        ...
+    def __call__(self, data: Any, stream: TextIO) -> Any: ...
 
 
 class ModifierFn(Protocol):
     def __call__(
         self, path: "StrPath", fs: Optional["FileSystem"] = None
-    ) -> AbstractContextManager[dict]:
-        ...
+    ) -> AbstractContextManager[dict]: ...
 
 
 class LoaderFn(Protocol):
-    def __call__(self, path: "StrPath", fs: Optional["FileSystem"] = None) -> Any:
-        ...
+    def __call__(self, path: "StrPath", fs: Optional["FileSystem"] = None) -> Any: ...
 
 
 ReadType = Union[bytes, None, str]

dvc/utils/strictyaml.py

@@ -6,6 +6,7 @@ Used for parsing dvc.yaml, dvc.lock and .dvc files.
 
 Not to be confused with strictyaml, a python library with similar motivations.
 """
+
 import re
 import typing
 from contextlib import suppress

dvc/utils/studio.py

@@ -1,3 +1,4 @@
+import os
 from typing import TYPE_CHECKING, Any, Optional
 from urllib.parse import urljoin
 
@@ -22,7 +23,7 @@ if TYPE_CHECKING:
 
 logger = logger.getChild(__name__)
 
-STUDIO_URL = "https://studio.iterative.ai"
+STUDIO_URL = "https://studio.dvc.ai"
 
 
 def post(
@@ -125,3 +126,19 @@ def get_subrepo_relpath(repo: "Repo") -> str:
     relpath = as_posix(repo.fs.relpath(repo.root_dir, scm_root_dir))
 
     return "" if relpath == "." else relpath
+
+
+def get_repo_url(repo: "Repo") -> str:
+    from dulwich.porcelain import get_remote_repo
+
+    from dvc.env import DVC_EXP_GIT_REMOTE
+
+    repo_url = os.getenv(
+        DVC_EXP_GIT_REMOTE, repo.config.get("exp", {}).get("git_remote")
+    )
+    if repo_url:
+        try:
+            _, repo_url = get_remote_repo(repo.scm.dulwich.repo, repo_url)
+        except IndexError:
+            pass
+    return repo_url

pyproject.toml

@@ -31,24 +31,30 @@ classifiers = [
 dynamic = ["version"]
 
 dependencies = [
+    "attrs>=22.2.0",
+    "celery",
     "colorama>=0.3.9",
     "configobj>=5.0.6",
     "distro>=1.3",
     "dpath<3,>=2.1.0",
-    "dvc-data>=3.13,<3.15",
+    "dulwich",
+    "dvc-data>=3.15,<3.16",
     "dvc-http>=2.29.0",
+    "dvc-objects",
     "dvc-render>=1.0.1,<2",
-    "dvc-studio-client>=0.20,<1",
+    "dvc-studio-client>=0.21,<1",
     "dvc-task>=0.3.0,<1",
     "flatten_dict<1,>=0.4.1",
-    # https://github.com/iterative/dvc/issues/9654
-    "flufl.lock>=5,<8",
+    "flufl.lock>=8.1.0,<9", # https://github.com/iterative/dvc/issues/9654
+    "fsspec>=2024.2.0",
     "funcy>=1.14",
     "grandalf<1,>=0.7",
     "gto>=1.6.0,<2",
     "hydra-core>=1.1",
     "iterative-telemetry>=0.0.7",
+    "kombu",
     "networkx>=2.5",
+    "omegaconf",
     "packaging>=19",
     "pathspec>=0.10.3",
     "platformdirs<4,>=3.1.1",
@@ -59,7 +65,7 @@ dependencies = [
     "requests>=2.22",
     "rich>=12",
     "ruamel.yaml>=0.17.11",
-    "scmrepo>=3,<4",
+    "scmrepo>=3.3.4,<4",
     "shortuuid>=0.5",
     "shtab<2,>=1.3.4",
     "tabulate>=0.8.7",
@@ -77,7 +83,7 @@ gdrive = ["dvc-gdrive>=3,<4"]
 gs = ["dvc-gs>=3,<4"]
 hdfs = ["dvc-hdfs>=3,<4"]
 lint = [
-    "mypy==1.8.0",
+    "mypy==1.10.1",
     "pandas-stubs",
     "types-colorama",
     "types-psutil",
@@ -89,7 +95,7 @@ lint = [
     "typing-extensions",
 ]
 oss = ["dvc-oss>=3,<4"]
-s3 = ["dvc-s3>=3,<4"]
+s3 = ["dvc-s3>=3.2.0,<4"]
 ssh = ["dvc-ssh>=4,<5"]
 ssh_gssapi = ["dvc-ssh[gssapi]>=4,<5"]
 testing = [
@@ -250,6 +256,7 @@ module = [
 
 [tool.codespell]
 ignore-words-list = "ba,datas,fo,uptodate,cachable,falsy"
+skip = "CODE_OF_CONDUCT.md"
 
 [tool.ruff]
 output-format = "full"
@@ -261,10 +268,10 @@ ignore = [
     "TRY003", "TRY300", "PLR2004", "PLW2901", "LOG007",
 ]
 select = [
-    "F", "E", "W", "C90", "I", "N", "UP", "YTT", "ASYNC", "S", "BLE", "B", "A", "C4", "T10",
+    "F", "E", "W", "C90", "I", "N", "UP", "YTT", "ASYNC", "S", "BLE", "B", "A", "C4", "DTZ", "T10",
     "EXE", "ISC", "ICN", "G", "INP", "PIE", "T20", "PYI", "PT", "Q", "RSE", "RET",
     "SLOT", "SIM", "TID", "TCH", "ARG", "PGH", "PLC", "PLE", "PLR", "PLW", "TRY",
-    "FLY", "PERF101", "LOG", "RUF", "RUF022", "RUF023", "RUF024", "RUF025", "RUF026",
+    "FLY", "PERF101", "LOG", "RUF", "RUF022", "RUF023", "RUF024", "RUF025", "RUF026", "RUF027"
 ]
 preview = true
 explicit-preview-rules = true
@@ -286,7 +293,7 @@ strict = true
 ignore-variadic-names = true
 
 [tool.ruff.lint.isort]
-known-first-party = ["dvc", "dvc_data", "dvc_objects", "dvc_render", "dvc_task", "tests"]
+known-first-party = ["dvc", "dvc_*", "tests"]
 
 [tool.ruff.lint.pep8-naming]
 extend-ignore-names = ["M", "SCM"]

tests/dir_helpers.py

@@ -43,7 +43,6 @@ from global repo template to creating everything inplace, which:
     - does not create unnecessary files
 """
 
-
 import os
 
 import pytest

tests/func/api/test_artifacts.py

@@ -0,0 +1,74 @@
+from os.path import join, normpath
+
+import pytest
+
+from dvc.api import artifacts_show
+from dvc.testing.tmp_dir import make_subrepo
+from dvc.utils import as_posix
+from tests.func.artifacts.test_artifacts import get_tag_and_name, make_artifact
+
+
+@pytest.mark.parametrize("sub", ["sub", ""])
+def test_artifacts_show(tmp_dir, dvc, scm, sub):
+    subdir = tmp_dir / sub
+
+    dirname = str(subdir.relative_to(tmp_dir))
+    tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
+    make_artifact(tmp_dir, "myart", tag, subdir / "myart.pkl")
+
+    assert artifacts_show(name) == {
+        "path": normpath(join(dirname, "myart.pkl")),
+        "rev": scm.get_rev(),
+    }
+    assert artifacts_show(name, repo=tmp_dir.fs_path) == {
+        "path": normpath(join(dirname, "myart.pkl")),
+        "rev": scm.get_rev(),
+    }
+    assert artifacts_show(name, repo=f"file://{tmp_dir.as_posix()}") == {
+        "path": normpath(join(dirname, "myart.pkl")),
+        "rev": scm.get_rev(),
+    }
+
+    assert artifacts_show(name, repo=subdir.fs_path) == {
+        "path": normpath(join(dirname, "myart.pkl")),
+        "rev": scm.get_rev(),
+    }
+    with subdir.chdir():
+        assert artifacts_show(name) == {
+            "path": normpath(join(dirname, "myart.pkl")),
+            "rev": scm.get_rev(),
+        }
+
+
+@pytest.mark.parametrize("sub", ["sub", ""])
+def test_artifacts_show_subrepo(tmp_dir, scm, sub):
+    subrepo = tmp_dir / "subrepo"
+    make_subrepo(subrepo, scm)
+    subdir = subrepo / sub
+
+    dirname = str(subdir.relative_to(tmp_dir))
+    tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
+    make_artifact(subrepo, "myart", tag, subdir / "myart.pkl")
+
+    assert artifacts_show(name) == {
+        "path": join(dirname, "myart.pkl"),
+        "rev": scm.get_rev(),
+    }
+    assert artifacts_show(name, repo=tmp_dir.fs_path) == {
+        "path": join(dirname, "myart.pkl"),
+        "rev": scm.get_rev(),
+    }
+    assert artifacts_show(name, repo=f"file://{tmp_dir.as_posix()}") == {
+        "path": join(dirname, "myart.pkl"),
+        "rev": scm.get_rev(),
+    }
+
+    assert artifacts_show(name, repo=subdir.fs_path) == {
+        "path": str((subdir / "myart.pkl").relative_to(subrepo)),
+        "rev": scm.get_rev(),
+    }
+    with subdir.chdir():
+        assert artifacts_show(name) == {
+            "path": str((subdir / "myart.pkl").relative_to(subrepo)),
+            "rev": scm.get_rev(),
+        }

tests/func/api/test_data.py

@@ -31,6 +31,40 @@ def test_get_url_requires_dvc(tmp_dir, scm):
         api.get_url("foo", repo=f"file://{tmp_dir.as_posix()}")
 
 
+def test_get_url_from_remote(tmp_dir, erepo_dir, cloud, local_cloud):
+    erepo_dir.add_remote(config=cloud.config, name="other")
+    erepo_dir.add_remote(config=local_cloud.config, default=True)
+    with erepo_dir.chdir():
+        erepo_dir.dvc_gen("foo", "foo", commit="add foo")
+
+    # Using file url to force clone to tmp repo
+    repo_url = f"file://{erepo_dir.as_posix()}"
+    expected_rel_path = os.path.join(
+        "files", "md5", "ac/bd18db4cc2f85cedef654fccc4a4d8"
+    )
+
+    # Test default remote
+    assert api.get_url("foo", repo=repo_url) == (local_cloud / expected_rel_path).url
+
+    # Test remote arg
+    assert (
+        api.get_url("foo", repo=repo_url, remote="other")
+        == (cloud / expected_rel_path).url
+    )
+
+    # Test config arg
+    assert (
+        api.get_url("foo", repo=repo_url, config={"core": {"remote": "other"}})
+        == (cloud / expected_rel_path).url
+    )
+
+    # Test remote_config arg
+    assert (
+        api.get_url("foo", repo=repo_url, remote_config={"url": cloud.url})
+        == (cloud / expected_rel_path).url
+    )
+
+
 def test_open_external(tmp_dir, erepo_dir, cloud):
     erepo_dir.add_remote(config=cloud.config)
 
@@ -253,3 +287,12 @@ def test_read_from_remote(tmp_dir, erepo_dir, cloud, local_cloud):
         )
         == "foo content"
     )
+
+    assert (
+        api.read(
+            os.path.join("dir", "foo"),
+            repo=f"file://{erepo_dir.as_posix()}",
+            remote_config={"url": cloud.url},
+        )
+        == "foo content"
+    )

tests/func/artifacts/test_artifacts.py

@@ -6,7 +6,9 @@ import pytest
 
 from dvc.annotations import Artifact
 from dvc.exceptions import ArtifactNotFoundError, InvalidArgumentError
-from dvc.repo.artifacts import check_name_format
+from dvc.repo.artifacts import Artifacts, check_name_format
+from dvc.testing.tmp_dir import make_subrepo
+from dvc.utils import as_posix
 from dvc.utils.strictyaml import YAMLSyntaxError, YAMLValidationError
 
 dvcyaml = {
@@ -185,7 +187,7 @@ def test_get_rev(tmp_dir, dvc, scm):
         dvc.artifacts.get_rev("myart", stage="prod")
 
 
-def test_get_path(tmp_dir, dvc):
+def test_get_path(tmp_dir, dvc, scm):
     (tmp_dir / "dvc.yaml").dump(dvcyaml)
     subdir = tmp_dir / "subdir"
     subdir.mkdir()
@@ -196,3 +198,95 @@ def test_get_path(tmp_dir, dvc):
     assert dvc.artifacts.get_path("subdir/dvc.yaml:myart") == os.path.join(
         "subdir", "myart.pkl"
     )
+
+
+def test_parametrized(tmp_dir, dvc):
+    (tmp_dir / "params.yaml").dump({"path": "myart.pkl"})
+    (tmp_dir / "dvc.yaml").dump(
+        {"artifacts": {"myart": {"type": "model", "path": "${path}"}}}
+    )
+    assert tmp_dir.dvc.artifacts.read() == {
+        "dvc.yaml": {"myart": Artifact(path="myart.pkl", type="model")}
+    }
+
+
+def test_get_path_subrepo(tmp_dir, scm, dvc):
+    subrepo = tmp_dir / "subrepo"
+    make_subrepo(subrepo, scm)
+    (subrepo / "dvc.yaml").dump(dvcyaml)
+
+    assert dvc.artifacts.get_path("subrepo:myart") == os.path.join(
+        "subrepo", "myart.pkl"
+    )
+    assert dvc.artifacts.get_path("subrepo/dvc.yaml:myart") == os.path.join(
+        "subrepo", "myart.pkl"
+    )
+
+    assert subrepo.dvc.artifacts.get_path("subrepo:myart") == os.path.join(
+        "subrepo", "myart.pkl"
+    )
+    assert subrepo.dvc.artifacts.get_path("subrepo/dvc.yaml:myart") == os.path.join(
+        "subrepo", "myart.pkl"
+    )
+
+
+def get_tag_and_name(dirname, name, version):
+    tagname = f"{name}@{version}"
+    if dirname in (os.curdir, ""):
+        return tagname, name
+    return f"{dirname}={tagname}", f"{dirname}:{name}"
+
+
+def make_artifact(tmp_dir, name, tag, path) -> Artifact:
+    artifact = Artifact(path=path.name, type="model")
+    dvcfile = path.with_name("dvc.yaml")
+
+    tmp_dir.scm_gen(path, "hello_world", commit="add myart.pkl")
+    tmp_dir.dvc.artifacts.add(name, artifact, dvcfile=os.fspath(dvcfile))
+    tmp_dir.scm.add_commit([dvcfile], message="add dvc.yaml")
+    tmp_dir.scm.tag(tag, annotated=True, message="foo")
+    return artifact
+
+
+@pytest.mark.parametrize("sub", ["sub", ""])
+def test_artifacts_download(tmp_dir, dvc, scm, sub):
+    subdir = tmp_dir / sub
+    dirname = str(subdir.relative_to(tmp_dir))
+    tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
+    make_artifact(tmp_dir, "myart", tag, subdir / "myart.pkl")
+
+    result = (1, "myart.pkl")
+    assert Artifacts.get(".", name, force=True) == result
+    assert Artifacts.get(tmp_dir.fs_path, name, force=True) == result
+    assert Artifacts.get(f"file://{tmp_dir.as_posix()}", name, force=True) == result
+    assert Artifacts.get(subdir.fs_path, name, force=True) == result
+    with subdir.chdir():
+        assert Artifacts.get(".", name, force=True) == result
+
+
+@pytest.mark.parametrize("sub", ["sub", ""])
+def test_artifacts_download_subrepo(tmp_dir, scm, sub):
+    subrepo = tmp_dir / "subrepo"
+    make_subrepo(subrepo, scm)
+    subdir = subrepo / sub
+
+    dirname = str(subdir.relative_to(tmp_dir))
+    tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
+    make_artifact(subrepo, "myart", tag, subdir / "myart.pkl")
+
+    result = (1, "myart.pkl")
+    assert Artifacts.get(".", name) == result
+    assert Artifacts.get(tmp_dir.fs_path, name, force=True) == result
+    assert Artifacts.get(f"file://{tmp_dir.as_posix()}", name, force=True) == result
+    assert Artifacts.get(subdir.fs_path, name, force=True) == result
+    with subdir.chdir():
+        assert Artifacts.get(".", name, force=True) == result
+
+
+def test_artifacts_download_studio(tmp_dir, dvc, mocker):
+    with dvc.config.edit("global") as conf:
+        conf["studio"]["token"] = "mytoken"
+
+    download_studio = mocker.patch("dvc.repo.artifacts.Artifacts._download_studio")
+    Artifacts.get("myart.pkl", "myart.pkl")
+    assert download_studio.call_args.kwargs["dvc_studio_config"]["token"] == "mytoken"

tests/func/experiments/test_show.py

@@ -101,7 +101,7 @@ def test_show_simple(tmp_dir, scm, dvc, exp_stage):
 @pytest.mark.parametrize("workspace", [True, False])
 def test_show_experiment(tmp_dir, scm, dvc, exp_stage, workspace):
     baseline_rev = scm.get_rev()
-    timestamp = datetime.fromtimestamp(
+    timestamp = datetime.fromtimestamp(  # noqa: DTZ006
         scm.gitpython.repo.rev_parse(baseline_rev).committed_date
     )
 
@@ -423,7 +423,7 @@ def test_show_csv(tmp_dir, scm, dvc, exp_stage, capsys):
     baseline_rev = scm.get_rev()
 
     def _get_rev_isotimestamp(rev):
-        return datetime.fromtimestamp(
+        return datetime.fromtimestamp(  # noqa: DTZ006
             scm.gitpython.repo.rev_parse(rev).committed_date
         ).isoformat()
 
@@ -453,15 +453,11 @@ def test_show_csv(tmp_dir, scm, dvc, exp_stage, capsys):
         in cap.out
     )
     assert (
-        "{},{},branch_base,{},,2,2,{}".format(
-            ref_info1.name, rev1[:7], _get_rev_isotimestamp(rev1), data_hash
-        )
+        f"{ref_info1.name},{rev1[:7]},branch_base,{_get_rev_isotimestamp(rev1)},,2,2,{data_hash}"
         in cap.out
     )
     assert (
-        "{},{},branch_commit,{},,3,3,{}".format(
-            ref_info2.name, rev2[:7], _get_rev_isotimestamp(rev2), data_hash
-        )
+        f"{ref_info2.name},{rev2[:7]},branch_commit,{_get_rev_isotimestamp(rev2)},,3,3,{data_hash}"
         in cap.out
     )
 
@@ -656,7 +652,7 @@ def test_metrics_renaming(tmp_dir, dvc, scm, capsys, copy_script):
     cap = capsys.readouterr()
 
     def _get_rev_isotimestamp(rev):
-        return datetime.fromtimestamp(
+        return datetime.fromtimestamp(  # noqa: DTZ006
             scm.gitpython.repo.rev_parse(rev).committed_date
         ).isoformat()