repo: try to support opening uninitialized/broken repositories (#10736)

This `_get_remote_config()` should use `uninitialized=True` so that it supports
more broader kinds of broken/partially-initialized repositories, including some that have
`.dvc` directory missing, or `.git` directory missing.

This partially reverts #10719. #10608 is also fixed, and no longer requires `core.no_scm` to be passed.
This was already supported by `dvc.api.get_url` as it uses `uninitialized=True`, but this was not respected
in `_get_remote_config()` where it would fail before.

That said, this whole `open_repo`/`_get_remote_config` is terribly broken. For one, it is opening
a local repository, and forcing it's remote config to a repository opened with `Repo(rev=...)`,
where the config may be different.

.github/workflows/benchmarks.yaml

@@ -16,14 +16,14 @@ jobs:
     steps:
       - uses: actions/setup-python@v5
         with:
-            python-version: "3.12"
+            python-version: "3.13"
 
       - uses: actions/checkout@v4
         with:
           ref: ${{ github.event.pull_request.base.sha }}
           fetch-depth: 0
 
-      - uses: astral-sh/setup-uv@v4
+      - uses: astral-sh/setup-uv@v6
         with:
           enable-cache: true
           cache-dependency-glob: pyproject.toml

.github/workflows/build.yaml

@@ -25,25 +25,27 @@ jobs:
 
     - uses: actions/setup-python@v5
       with:
-        python-version: 3.12
+        python-version: 3.13
 
-    - uses: astral-sh/setup-uv@v4
-    - run: uv pip install --upgrade setuptools_scm twine --system
+    - uses: astral-sh/setup-uv@v6
 
-    - name: Force version for Test PyPI upload
-      if: ${{ !startsWith(github.ref, 'refs/tags') }}
-      run: echo version=$(python -m setuptools_scm | awk -F+ '{print $1}' | tail -1) >> $GITHUB_ENV
+    - name: Omit local version for Test PyPI upload
+      if: github.event_name == 'push' && github.ref == 'refs/heads/main'
+      run: echo SETUPTOOLS_SCM_OVERRIDES_FOR_DVC='{local_scheme="no-local-version"}' >> $GITHUB_ENV
 
-    - run: echo 'PKG = "pip"'>dvc/_build.py
-    - run: uv build
-      env:
-        SETUPTOOLS_SCM_PRETEND_VERSION_FOR_DVC: ${{ env.version }}
-    - run: twine check --strict dist/*
+    - name: Build Python Package
+      run: |
+        echo 'PKG = "pip"'>dvc/_build.py
+        uv build
+
+    - name: Check dist
+      run: uv tool run twine check --strict dist/*
 
     - uses: actions/upload-artifact@v4
       with:
         name: Packages
         path: dist/
+        if-no-files-found: error
 
   test-pypi-publish:
     name: Publish dev package to test.pypi.org

.github/workflows/plugin_tests.yaml

@@ -19,7 +19,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest]
-        pyv: ["3.12"]
+        pyv: ["3.13"]
         plugin: ["dvc-s3"]
 
     steps:
@@ -38,7 +38,7 @@ jobs:
       with:
         python-version: ${{ matrix.pyv }}
 
-    - uses: astral-sh/setup-uv@v4
+    - uses: astral-sh/setup-uv@v6
       with:
         enable-cache: true
         cache-dependency-glob: "**/pyproject.toml"

.github/workflows/tests.yaml

@@ -32,8 +32,8 @@ jobs:
     - name: Set up Python
       uses: actions/setup-python@v5
       with:
-        python-version: '3.12'
-    - uses: astral-sh/setup-uv@v4
+        python-version: '3.13'
+    - uses: astral-sh/setup-uv@v6
       with:
         enable-cache: true
         cache-dependency-glob: pyproject.toml
@@ -55,7 +55,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest, macos-latest]
-        pyv: ["3.9", "3.10", "3.11", "3.12"]
+        pyv: ["3.9", "3.10", "3.11", "3.12", "3.13"]
         include:
         - os: windows-latest
           pyv: "3.9"
@@ -69,6 +69,30 @@ jobs:
         - os: windows-latest
           pyv: "3.9"
           pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 4"
+        - os: windows-latest
+          pyv: "3.10"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 1"
+        - os: windows-latest
+          pyv: "3.10"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 2"
+        - os: windows-latest
+          pyv: "3.10"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 3"
+        - os: windows-latest
+          pyv: "3.10"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 4"
+        - os: windows-latest
+          pyv: "3.11"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 1"
+        - os: windows-latest
+          pyv: "3.11"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 2"
+        - os: windows-latest
+          pyv: "3.11"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 3"
+        - os: windows-latest
+          pyv: "3.11"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 4"
         - os: windows-latest
           pyv: "3.12"
           pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 1"
@@ -81,6 +105,18 @@ jobs:
         - os: windows-latest
           pyv: "3.12"
           pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 4"
+        - os: windows-latest
+          pyv: "3.13"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 1"
+        - os: windows-latest
+          pyv: "3.13"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 2"
+        - os: windows-latest
+          pyv: "3.13"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 3"
+        - os: windows-latest
+          pyv: "3.13"
+          pytestargs: "--splitting-algorithm=least_duration --splits 4 --group 4"
     steps:
     - uses: actions/checkout@v4
       with:
@@ -89,7 +125,7 @@ jobs:
       uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.pyv }}
-    - uses: astral-sh/setup-uv@v4
+    - uses: astral-sh/setup-uv@v6
       with:
         enable-cache: true
         cache-dependency-glob: pyproject.toml
@@ -124,7 +160,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - name: Slack Notification
-      uses: rtCamp/[email protected].2
+      uses: rtCamp/[email protected].3
       env:
         SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
         SLACK_COLOR: ${{ job.status }}

.pre-commit-config.yaml

@@ -22,13 +22,13 @@ repos:
       - id: sort-simple-yaml
       - id: trailing-whitespace
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: 'v0.8.0'
+    rev: 'v0.11.8'
     hooks:
       - id: ruff
         args: [--fix, --exit-non-zero-on-fix]
       - id: ruff-format
   - repo: https://github.com/codespell-project/codespell
-    rev: v2.3.0
+    rev: v2.4.1
     hooks:
       - id: codespell
         additional_dependencies: ["tomli"]

README.rst

@@ -1,6 +1,4 @@
-🚀 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.
-
---------------------------
+|Banner|
 
 `Website <https://dvc.org>`_
 • `Docs <https://dvc.org/doc>`_

dvc/api/show.py

@@ -27,8 +27,7 @@ def _postprocess(results):
             }
             processed[rev] = processed[rev] | to_merge
 
-    if "workspace" in processed:
-        del processed["workspace"]
+    processed.pop("workspace", None)
 
     return processed
 

dvc/cli/utils.py

@@ -40,7 +40,7 @@ def hide_subparsers_from_help(subparsers):
     # from the 'positional arguments' choices list
     # see: https://bugs.python.org/issue22848
     # Need to set `add_help=False`, but avoid setting `help`
-    # (not even to `argparse.SUPPPRESS`).
+    # (not even to `argparse.SUPPRESS`).
     # NOTE: The argument is the parent subparser, not the subcommand parser.
     cmds = [cmd for cmd, parser in subparsers.choices.items() if parser.add_help]
     subparsers.metavar = "{{{}}}".format(",".join(cmds))

dvc/commands/ls/__init__.py

@@ -28,7 +28,7 @@ def _format_entry(entry, name, with_size=True, with_hash=False):
     ret = []
     if with_size:
         size = entry.get("size")
-        if size is None:
+        if size is None or (size <= 0 and entry.get("isdir")):
             size = ""
         else:
             size = naturalsize(size)
@@ -77,10 +77,7 @@ def _build_tree_structure(
 
     num_entries = len(entries)
     for i, (name, entry) in enumerate(entries.items()):
-        # show full path for root, otherwise only show the name
-        if _depth > 0:
-            entry["path"] = name
-
+        entry["path"] = name
         is_last = i >= num_entries - 1
         tree_part = ""
         if _depth > 0:

dvc/commands/ls_url.py

@@ -3,26 +3,45 @@ from dvc.cli.command import CmdBaseNoRepo
 from dvc.cli.utils import DictAction, append_doc_link
 from dvc.log import logger
 
-from .ls import show_entries
+from .ls import show_entries, show_tree
 
 logger = logger.getChild(__name__)
 
 
 class CmdListUrl(CmdBaseNoRepo):
-    def run(self):
-        from dvc.config import Config
+    def _show_tree(self, config):
+        from dvc.fs import parse_external_url
+        from dvc.repo.ls import _ls_tree
+
+        fs, fs_path = parse_external_url(
+            self.args.url, fs_config=self.args.fs_config, config=config
+        )
+        entries = _ls_tree(fs, fs_path, maxdepth=self.args.level)
+        show_tree(entries, with_color=True, with_size=self.args.size)
+        return 0
+
+    def _show_list(self, config):
         from dvc.repo import Repo
 
         entries = Repo.ls_url(
             self.args.url,
             recursive=self.args.recursive,
+            maxdepth=self.args.level,
             fs_config=self.args.fs_config,
-            config=Config.from_cwd(),
+            config=config,
         )
         if entries:
             show_entries(entries, with_color=True, with_size=self.args.size)
         return 0
 
+    def run(self):
+        from dvc.config import Config
+
+        config = Config.from_cwd()
+        if self.args.tree:
+            return self._show_tree(config=config)
+        return self._show_list(config=config)
+
 
 def add_parser(subparsers, parent_parser):
     LS_HELP = "List directory contents from URL."
@@ -40,6 +59,19 @@ def add_parser(subparsers, parent_parser):
     lsurl_parser.add_argument(
         "-R", "--recursive", action="store_true", help="Recursively list files."
     )
+    lsurl_parser.add_argument(
+        "-T",
+        "--tree",
+        action="store_true",
+        help="Recurse into directories as a tree.",
+    )
+    lsurl_parser.add_argument(
+        "-L",
+        "--level",
+        metavar="depth",
+        type=int,
+        help="Limit the depth of recursion.",
+    )
     lsurl_parser.add_argument("--size", action="store_true", help="Show sizes.")
     lsurl_parser.add_argument(
         "--fs-config",

dvc/commands/remote.py

@@ -110,8 +110,16 @@ class CmdRemoteDefault(CmdRemote):
 class CmdRemoteList(CmdRemote):
     def run(self):
         conf = self.config.read(self.args.level)
+        default_remote = conf["core"].get("remote")
+
         for name, remote_conf in conf["remote"].items():
-            ui.write(name, remote_conf["url"], sep="\t")
+            if name == default_remote:
+                text = f"{name}\t{remote_conf['url']}\t(default)"
+                color = "green"
+            else:
+                text = f"{name}\t{remote_conf['url']}"
+                color = ""
+            ui.write(ui.rich_text(text, style=color), styled=True)
         return 0
 
 

dvc/config.py

@@ -1,6 +1,8 @@
 """DVC config objects."""
 
+import ntpath
 import os
+import posixpath
 import re
 from contextlib import contextmanager
 from functools import partial
@@ -268,12 +270,17 @@ class Config(dict):
         if re.match(r"\w+://", path):
             return path
 
+        if os.name == "nt" and posixpath.isabs(path) and ntpath.sep not in path:
+            return path
+
         if os.path.isabs(path):
             return path
 
         # on windows convert slashes to backslashes
         # to have path compatible with abs_conf_dir
         if os.path.sep == "\\" and "/" in path:
+            if path.startswith("/"):
+                path = path.replace("/", "\\\\", 1)
             path = path.replace("/", "\\")
 
         expanded = os.path.expanduser(path)
@@ -305,6 +312,9 @@ class Config(dict):
         if os.path.expanduser(path) != path:
             return localfs.as_posix(path)
 
+        if os.name == "nt" and posixpath.isabs(path) and ntpath.sep not in path:
+            return path
+
         if isinstance(path, RelPath) or not os.path.isabs(path):
             path = relpath(path, conf_dir)
             return localfs.as_posix(path)

dvc/dagascii.py

@@ -104,21 +104,21 @@ class AsciiCanvas:
                 if dx == 0:
                     y = y0
                 else:
-                    y = y0 + int(round((x - x0) * dy / float(dx)))
+                    y = y0 + round((x - x0) * dy / float(dx))
                 self.point(x, y, char)
         elif y0 < y1:
             for y in range(y0, y1 + 1):
                 if dy == 0:
                     x = x0
                 else:
-                    x = x0 + int(round((y - y0) * dx / float(dy)))
+                    x = x0 + round((y - y0) * dx / float(dy))
                 self.point(x, y, char)
         else:
             for y in range(y1, y0 + 1):
                 if dy == 0:
                     x = x0
                 else:
-                    x = x1 + int(round((y - y1) * dx / float(dy)))
+                    x = x1 + round((y - y1) * dx / float(dy))
                 self.point(x, y, char)
 
     def text(self, x, y, text):
@@ -258,8 +258,8 @@ def draw(vertices, edges):
     maxx = max(Xs)
     maxy = max(Ys)
 
-    canvas_cols = int(math.ceil(math.ceil(maxx) - math.floor(minx))) + 1
-    canvas_lines = int(round(maxy - miny))
+    canvas_cols = math.ceil(math.ceil(maxx) - math.floor(minx)) + 1
+    canvas_lines = round(maxy - miny)
 
     canvas = AsciiCanvas(canvas_cols, canvas_lines)
 
@@ -270,10 +270,10 @@ def draw(vertices, edges):
             start = edge.view._pts[index - 1]
             end = edge.view._pts[index]
 
-            start_x = int(round(start[0] - minx))
-            start_y = int(round(start[1] - miny))
-            end_x = int(round(end[0] - minx))
-            end_y = int(round(end[1] - miny))
+            start_x = round(start[0] - minx)
+            start_y = round(start[1] - miny)
+            end_x = round(end[0] - minx)
+            end_y = round(end[1] - miny)
 
             assert start_x >= 0
             assert start_y >= 0
@@ -288,12 +288,12 @@ def draw(vertices, edges):
         y = vertex.view.xy[1]
 
         canvas.box(
-            int(round(x - minx)),
-            int(round(y - miny)),
+            round(x - minx),
+            round(y - miny),
             vertex.view.w,
             vertex.view.h,
         )
 
-        canvas.text(int(round(x - minx)) + 1, int(round(y - miny)) + 1, vertex.data)
+        canvas.text(round(x - minx) + 1, round(y - miny) + 1, vertex.data)
 
     return canvas.draw()

dvc/dependency/repo.py

@@ -146,6 +146,10 @@ class RepoDependency(Dependency):
         else:
             config = Config.load_file(conf)
 
+        # Setup config to the new DVCFileSystem to use the remote repo, but rely on the
+        # local cache instead of the remote's cache. This avoids re-streaming of data,
+        # but messes up the call to `_get_remote_config()` downstream, which will need
+        # to ignore cache parameters.
         config["cache"] = self.repo.config["cache"]
         config["cache"]["dir"] = self.repo.cache.local_cache_dir
 

dvc/fs/dvc.py

@@ -500,7 +500,7 @@ class _DVCFileSystem(AbstractFileSystem):
             **kwargs,
         )
 
-    def _get(  # noqa: C901, PLR0912
+    def _get(  # noqa: C901, PLR0912, PLR0915
         self,
         rpath,
         lpath,
@@ -578,7 +578,8 @@ class _DVCFileSystem(AbstractFileSystem):
             kw = kwargs
             if isinstance(fs, DataFileSystem):
                 kw = kw | {"info": info}
-            return fs.get_file(src, dest, callback=callback, **kw)
+            with callback.branched(src, dest) as child:
+                fs.get_file(src, dest, callback=child, **kw)
 
         if batch_size == 1:
             ctx: AbstractContextManager = nullcontext()
@@ -589,7 +590,7 @@ class _DVCFileSystem(AbstractFileSystem):
 
         with ctx:
             it = ((fs, f) for fs, files in _files.items() for f in files)
-            deque(map_fn(get_file, it), maxlen=0)
+            deque(callback.wrap(map_fn(get_file, it)), maxlen=0)
         return result
 
     def get_file(self, rpath, lpath, **kwargs):
@@ -732,9 +733,10 @@ class DVCFileSystem(FileSystem):
         return self.fs.repo_url
 
     def from_os_path(self, path: str) -> str:
-        if os.path.isabs(path):
+        if os.path.isabs(path) or (
+            os.name == "nt" and posixpath.isabs(path) and ntpath.sep not in path
+        ):
             path = os.path.relpath(path, self.repo.root_dir)
-
         return as_posix(path)
 
     def close(self):

dvc/logger.py

@@ -36,7 +36,7 @@ def add_logging_level(level_name, level_num, method_name=None):
             self._log(level_num, message, args, **kwargs)
 
     def log_to_root(message, *args, **kwargs):
-        logging.log(level_num, message, *args, **kwargs)
+        logging.log(level_num, message, *args, **kwargs)  # noqa: LOG015
 
     # getLevelName resolves the numeric log level if already defined,
     # otherwise returns a string

dvc/parsing/context.py

@@ -105,7 +105,7 @@ class Meta:
 
     @staticmethod
     def update_path(meta: "Meta", path: Union[str, int]):
-        dpaths = meta.dpaths[:] + [str(path)]
+        dpaths = [*meta.dpaths, str(path)]
         return replace(meta, dpaths=dpaths)
 
     def __str__(self):

dvc/pathspec_math.py

@@ -27,11 +27,9 @@ def _remove_slash(rule):
 
 def _match_all_level(rule):
     if rule[:-1].find("/") >= 0 and not rule.startswith("**/"):
-        if rule.startswith("/"):
-            rule = rule[1:]
+        rule = rule.removeprefix("/")
         return False, rule
-    if rule.startswith("**/"):
-        rule = rule[3:]
+    rule = rule.removeprefix("**/")
     return True, rule
 
 

dvc/repo/experiments/queue/base.py

@@ -372,7 +372,7 @@ class BaseStashQueue(ABC):
         )
 
     def _stash_commit_deps(self, *args, **kwargs):
-        if len(args):
+        if args:
             targets = args[0]
         else:
             targets = kwargs.get("targets")

dvc/repo/experiments/run.py

@@ -93,7 +93,7 @@ def run(  # noqa: C901, PLR0912
 
     for idx, sweep_overrides in enumerate(sweeps):
         if hydra_sweep and name_prefix is not None:
-            kwargs["name"] = f"{name_prefix}-{idx+1}"
+            kwargs["name"] = f"{name_prefix}-{idx + 1}"
         queue_entry = repo.experiments.queue_one(
             repo.experiments.celery_queue,
             targets=targets,

dvc/repo/ls.py

@@ -99,7 +99,9 @@ def ls_tree(
         path = path or ""
         fs: DVCFileSystem = repo.dvcfs
         fs_path = fs.from_os_path(path)
-        return _ls_tree(fs, fs_path, dvc_only, maxdepth)
+        return _ls_tree(
+            fs, fs_path, maxdepth=maxdepth, dvc_only=dvc_only, dvcfiles=True
+        )
 
 
 def _ls(
@@ -145,27 +147,34 @@ def _ls(
     return ret_list
 
 
-def _ls_tree(
-    fs, path, dvc_only: bool = False, maxdepth: Optional[int] = None, _info=None
-):
-    ret = {}
+def _ls_tree(fs, path, maxdepth=None, _info=None, **fs_kwargs):
     info = _info or fs.info(path)
+    if _info is None:
+        # preserve the original path name
+        name = path
+        if not name:
+            name = os.curdir if fs.protocol == "local" else fs.root_marker
+        path = info["name"]
+    else:
+        name = path.rsplit(fs.sep, 1)[-1]
 
-    path = info["name"].rstrip(fs.sep) or os.curdir
-    name = path.rsplit("/", 1)[-1]
+    ret = {}
     ls_info = _adapt_info(info)
     ls_info["path"] = path
 
     recurse = maxdepth is None or maxdepth > 0
     if recurse and info["type"] == "directory":
-        infos = fs.ls(path, dvcfiles=True, dvc_only=dvc_only, detail=True)
+        try:
+            infos = fs.ls(path, detail=True, **fs_kwargs)
+        except FileNotFoundError:
+            # broken symlink?
+            infos = []
+
         infos.sort(key=lambda f: f["name"])
         maxdepth = maxdepth - 1 if maxdepth is not None else None
         contents = {}
         for info in infos:
-            d = _ls_tree(
-                fs, info["name"], dvc_only=dvc_only, maxdepth=maxdepth, _info=info
-            )
+            d = _ls_tree(fs, info["name"], maxdepth=maxdepth, _info=info, **fs_kwargs)
             contents.update(d)
         ls_info["contents"] = contents
 

dvc/repo/ls_url.py

@@ -1,19 +1,30 @@
+from fsspec.implementations.local import LocalFileSystem as _LocalFileSystem
+
 from dvc.exceptions import URLMissingError
-from dvc.fs import parse_external_url
+from dvc.fs import LocalFileSystem, parse_external_url
 
 
-def ls_url(url, *, fs_config=None, recursive=False, config=None):
+def ls_url(url, *, fs_config=None, recursive=False, maxdepth=None, config=None):
     fs, fs_path = parse_external_url(url, fs_config=fs_config, config=config)
     try:
         info = fs.info(fs_path)
     except FileNotFoundError as exc:
         raise URLMissingError(url) from exc
-    if info["type"] != "directory":
+    if maxdepth == 0 or info["type"] != "directory":
         return [{"path": info["name"], "isdir": False}]
 
+    if isinstance(fs, LocalFileSystem):
+        # dvc's LocalFileSystem does not support maxdepth yet
+        walk = _LocalFileSystem().walk
+    else:
+        walk = fs.walk
+
     ret = []
-    for _, dirs, files in fs.walk(fs_path, detail=True):
-        if not recursive:
+    for root, dirs, files in walk(fs_path, detail=True, maxdepth=maxdepth):
+        parts = fs.relparts(root, fs_path)
+        if parts == (".",):
+            parts = ()
+        if not recursive or (maxdepth and len(parts) >= maxdepth - 1):
             files.update(dirs)
 
         for info in files.values():

dvc/repo/open_repo.py

@@ -99,7 +99,7 @@ def clean_repos():
 
 def _get_remote_config(url):
     try:
-        repo = Repo(url)
+        repo = Repo(url, uninitialized=True)
     except NotDvcRepoError:
         return {}
 
@@ -109,10 +109,16 @@ def _get_remote_config(url):
             # Fill the empty upstream entry with a new remote pointing to the
             # original repo's cache location.
             name = "auto-generated-upstream"
-            return {
-                "core": {"remote": name},
-                "remote": {name: {"url": repo.cache.local_cache_dir}},
-            }
+            try:
+                local_cache_dir = repo.cache.local_cache_dir
+            except AttributeError:
+                # if the `.dvc` dir is missing, we get an AttributeError
+                return {}
+            else:
+                return {
+                    "core": {"remote": name},
+                    "remote": {name: {"url": local_cache_dir}},
+                }
 
         # Use original remote to make sure that we are using correct url,
         # credential paths, etc if they are relative to the config location.

dvc/testing/benchmarks/fixtures.py

@@ -138,8 +138,7 @@ def dvc_bin(request, make_dvc_bin):
         version = Version(parse(dvc_version).base_version)
         if version not in spec:
             pytest.skip(
-                f"Version {dvc_version} "
-                f"does not satisfy requirement {spec!r}: {reason}"
+                f"Version {dvc_version} does not satisfy requirement {spec!r}: {reason}"
             )
     return make_dvc_bin
 

dvc/testing/workspace_tests.py

@@ -225,6 +225,39 @@ class TestLsUrl:
             ],
         )
 
+        result = ls_url(
+            str(cloud / "dir"), fs_config=cloud.config, recursive=True, maxdepth=0
+        )
+        match_files(
+            fs,
+            result,
+            [{"path": (cloud / "dir").fs_path, "isdir": False}],
+        )
+
+        result = ls_url(
+            str(cloud / "dir"), fs_config=cloud.config, recursive=True, maxdepth=1
+        )
+        match_files(
+            fs,
+            result,
+            [
+                {"path": "foo", "isdir": False},
+                {"path": "subdir", "isdir": True},
+            ],
+        )
+
+        result = ls_url(
+            str(cloud / "dir"), fs_config=cloud.config, recursive=True, maxdepth=2
+        )
+        match_files(
+            fs,
+            result,
+            [
+                {"path": "foo", "isdir": False},
+                {"path": "subdir/bar", "isdir": False},
+            ],
+        )
+
     def test_nonexistent(self, cloud):
         with pytest.raises(URLMissingError):
             ls_url(str(cloud / "dir"), fs_config=cloud.config)

dvc/utils/studio.py

@@ -23,7 +23,7 @@ if TYPE_CHECKING:
 
 logger = logger.getChild(__name__)
 
-STUDIO_URL = "https://studio.dvc.ai"
+STUDIO_URL = "https://studio.datachain.ai"
 
 
 def post(

pyproject.toml

@@ -1,6 +1,6 @@
 [build-system]
 build-backend = "setuptools.build_meta"
-requires = ["setuptools>=61", "setuptools_scm[toml]>=7"]
+requires = ["setuptools>=77", "setuptools_scm[toml]>=8"]
 
 [project]
 name = "dvc"
@@ -16,7 +16,8 @@ keywords = [
     "machine-learning",
     "reproducibility",
 ]
-license = { text = "Apache License 2.0" }
+license = "Apache-2.0"
+license-files = ["LICENSE"]
 maintainers = [{ name = "Iterative", email = "[email protected]" }]
 authors = [{ name = "Dmitry Petrov", email = "[email protected]" }]
 requires-python = ">=3.9"
@@ -27,6 +28,7 @@ classifiers = [
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
 ]
 dynamic = ["version"]
 
@@ -83,7 +85,7 @@ gdrive = ["dvc-gdrive>=3,<4"]
 gs = ["dvc-gs>=3,<4"]
 hdfs = ["dvc-hdfs>=3,<4"]
 lint = [
-    "mypy==1.13.0",
+    "mypy==1.15.0",
     "pandas-stubs",
     "types-colorama",
     "types-psutil",
@@ -148,9 +150,6 @@ dvc = "dvc.fs.dvc_path:DVCPath"
 hook-dirs = "dvc.__pyinstaller:get_hook_dirs"
 tests = "dvc.__pyinstaller:get_PyInstaller_tests"
 
-[tool.setuptools]
-license-files = ["LICENSE"]
-
 [tool.setuptools.packages.find]
 exclude = ["tests", "tests.*"]
 namespaces = false
@@ -172,6 +171,8 @@ filterwarnings = [
     "ignore:Deprecated call to `pkg_resources.declare_namespace:DeprecationWarning",
     # see https://github.com/networkx/networkx/issues/5723.
     "ignore:nx.nx_pydot.* depends on the pydot package, which has.*known issues and is not actively maintained:DeprecationWarning",
+    # TODO: investigate where we are not closing sqlite3.Connection
+    "ignore:unclosed database.*sqlite3.Connection:ResourceWarning",
 ]
 log_level = "debug"
 markers = [
@@ -264,14 +265,14 @@ show-fixes = true
 
 [tool.ruff.lint]
 ignore = [
-    "N818", "S101", "ISC001", "PT007", "RET502", "RET503", "SIM105", "SIM108", "SIM117",
+    "N818", "S101", "A005", "PT007", "RET502", "RET503", "SIM105", "SIM108", "SIM117",
     "TRY003", "TRY300", "PERF203", "PLR2004", "PLW2901", "LOG007",
 ]
 select = [
     "F", "E", "W", "C90", "I", "N", "UP", "YTT", "ASYNC", "S", "BLE", "B", "A", "C4", "DTZ", "T10",
     "EXE", "ISC", "ICN", "LOG", "G", "INP", "PIE", "T20", "PYI", "PT", "Q", "RSE", "RET",
     "SLOT", "SIM", "TID", "TC", "ARG", "PGH", "PLC", "PLE", "PLR", "PLW", "TRY",
-    "FLY", "PERF", "FURB", "RUF", "RUF027", "RUF029", "RUF034", "RUF036",
+    "FLY", "PERF", "FURB", "RUF", "RUF027", "RUF029", "RUF036",
 ]
 preview = true
 explicit-preview-rules = true

tests/func/api/test_data.py

@@ -5,6 +5,7 @@ from funcy import first, get_in
 
 from dvc import api
 from dvc.exceptions import OutputNotFoundError, PathMissingError
+from dvc.scm import CloneError
 from dvc.testing.api_tests import TestAPI  # noqa: F401
 from dvc.testing.tmp_dir import make_subrepo
 from dvc.utils.fs import remove
@@ -65,6 +66,34 @@ def test_get_url_from_remote(tmp_dir, erepo_dir, cloud, local_cloud):
     )
 
 
+def test_get_url_ignore_scm(tmp_dir, dvc, cloud, scm):
+    tmp_dir.add_remote(config=cloud.config)
+    tmp_dir.dvc_gen("foo", "foo", commit="add foo")
+
+    repo_posix = tmp_dir.as_posix()
+    expected_url = (cloud / "files" / "md5" / "ac/bd18db4cc2f85cedef654fccc4a4d8").url
+
+    # Test baseline with scm
+    assert api.get_url("foo", repo=repo_posix) == expected_url
+
+    # Simulate gitless environment (e.g. deployed container)
+    (tmp_dir / ".git").rename(tmp_dir / "gitless_environment")
+
+    assert api.get_url("foo", repo=repo_posix) == expected_url
+    assert (
+        api.get_url("foo", repo=repo_posix, config={"core": {"no_scm": True}})
+        == expected_url
+    )
+
+    # Addressing repos with `file://` triggers git, so it fails in a gitless environment
+    repo_url = f"file://{repo_posix}"
+    with pytest.raises(
+        CloneError,
+        match="SCM error",
+    ):
+        api.get_url("foo", repo=repo_url, config={"core": {"no_scm": True}})
+
+
 def test_open_external(tmp_dir, erepo_dir, cloud):
     erepo_dir.add_remote(config=cloud.config)
 

tests/func/experiments/test_remove.py

@@ -18,7 +18,7 @@ def test_remove_experiments_by_ref(tmp_dir, scm, dvc, exp_stage, caplog):
         ref_name_list.append(str(ref_info))
 
     with pytest.raises(InvalidArgumentError):
-        dvc.experiments.remove(ref_name_list[:2] + ["non-exist"])
+        dvc.experiments.remove([*ref_name_list[:2], "non-exist"])
     assert scm.get_ref(ref_name_list[0]) is not None
     assert scm.get_ref(ref_name_list[1]) is not None
     assert scm.get_ref(ref_name_list[2]) is not None

tests/func/experiments/test_set_params.py

@@ -133,7 +133,7 @@ def test_hydra_sweep_prefix_name(tmp_dir, params_repo, dvc):
     db_values = ["mysql", "postgresql"]
     param = "+db=" + ",".join(db_values)
     dvc.experiments.run(params=[param], queue=True, name=prefix)
-    expected_names = [f"{prefix}-{i+1}" for i, _ in enumerate(db_values)]
+    expected_names = [f"{prefix}-{i + 1}" for i, _ in enumerate(db_values)]
     exp_names = [entry.name for entry in dvc.experiments.celery_queue.iter_queued()]
     for name, expected in zip(exp_names, expected_names):
         assert name == expected

tests/func/plots/test_show.py

@@ -205,11 +205,8 @@ def test_plots_show_nested_x_dict(tmp_dir, dvc, scm):
 
     dvclive_metrics_dir.gen(
         {
-            "Error.tsv": "step\tError\n" "0\t0.11\n" "1\t0.22\n" "2\t0.44\n",
-            "Max_Leaf_Nodes.tsv": "step\tMax_Leaf_Nodes\n"
-            "0\t5\n"
-            "1\t50\n"
-            "2\t500\n",
+            "Error.tsv": "step\tError\n0\t0.11\n1\t0.22\n2\t0.44\n",
+            "Max_Leaf_Nodes.tsv": "step\tMax_Leaf_Nodes\n0\t5\n1\t50\n2\t500\n",
         }
     )
 

tests/func/repro/test_repro.py

@@ -759,7 +759,7 @@ def test_repro_dry(tmp_dir, dvc, copy_script):
 
     stages = dvc.reproduce(stage.addressing, dry=True)
 
-    assert len(stages), 2
+    assert len(stages) == 2
     assert not filecmp.cmp("file1", "bar", shallow=False)
 
     ret = main(["repro", "--dry", stage.addressing])

tests/func/test_checkout.py

@@ -744,3 +744,12 @@ def test_checkout_dir_compat(tmp_dir, dvc):
     remove("data")
     dvc.checkout()
     assert (tmp_dir / "data").read_text() == {"foo": "foo"}
+
+
+def test_checkout_cleanup_properly_on_untracked_nested_directories(tmp_dir, scm, dvc):
+    tmp_dir.dvc_gen({"datasets": {"dir1": {"file1": "file1"}}})
+    tmp_dir.gen({"datasets": {"dir2": {"dir3": {"file2": "file2"}}}})
+
+    dvc.checkout(force=True)
+
+    assert (tmp_dir / "datasets").read_text() == {"dir1": {"file1": "file1"}}

tests/func/test_ls.py

@@ -6,8 +6,9 @@ from os.path import join
 
 import pytest
 
+from dvc.fs import MemoryFileSystem
 from dvc.repo import Repo
-from dvc.repo.ls import ls_tree
+from dvc.repo.ls import _ls_tree, ls_tree
 from dvc.scm import CloneError
 
 FS_STRUCTURE = {
@@ -998,3 +999,64 @@ def test_ls_tree_maxdepth(M, tmp_dir, scm, dvc):
             "structure.xml.dvc": None,
         }
     }
+
+
+def test_fs_ls_tree():
+    fs = MemoryFileSystem(global_store=False)
+    fs.pipe({f: content.encode() for f, content in FS_STRUCTURE.items()})
+    root = fs.root_marker
+
+    files = _ls_tree(fs, "README.md")
+    assert _simplify_tree(files) == {"README.md": None}
+    files = _ls_tree(fs, root)
+    expected = {
+        root: {
+            ".gitignore": None,
+            "README.md": None,
+            "model": {
+                "script.py": None,
+                "train.py": None,
+            },
+        }
+    }
+    assert _simplify_tree(files) == expected
+
+    files = _ls_tree(fs, "model")
+    assert _simplify_tree(files) == {
+        "model": {
+            "script.py": None,
+            "train.py": None,
+        }
+    }
+
+
+def test_fs_ls_tree_maxdepth():
+    fs = MemoryFileSystem(global_store=False)
+    fs.pipe({f: content.encode() for f, content in FS_STRUCTURE.items()})
+
+    files = _ls_tree(fs, "/", maxdepth=0)
+    assert _simplify_tree(files) == {"/": None}
+
+    files = _ls_tree(fs, "/", maxdepth=1)
+    assert _simplify_tree(files) == {
+        "/": {
+            ".gitignore": None,
+            "README.md": None,
+            "model": None,
+        }
+    }
+
+    files = _ls_tree(fs, "/", maxdepth=2)
+    assert _simplify_tree(files) == {
+        "/": {
+            ".gitignore": None,
+            "README.md": None,
+            "model": {
+                "script.py": None,
+                "train.py": None,
+            },
+        }
+    }
+
+    files = _ls_tree(fs, "README.md", maxdepth=3)
+    assert _simplify_tree(files) == {"README.md": None}

tests/func/test_merge_driver.py

@@ -132,7 +132,8 @@ def test_merge_different_output_options(tmp_dir, dvc, caplog):
     )
 
     (tmp_dir / "their").write_text(
-        "outs:\n"
+        r"outs:"
+        "\n"
         "- md5: f987654321.dir\n"
         "  hash: md5\n"
         "  path: path\n"

tests/func/test_remote.py

@@ -115,6 +115,20 @@ def test_show_default(dvc, capsys):
     assert out == "foo\n"
 
 
+def test_list_shows_default(dvc, capsys):
+    default_remote = "foo"
+    other_remote = "bar"
+    bucket_url = "s3://bucket/name"
+    assert main(["remote", "add", default_remote, bucket_url]) == 0
+    assert main(["remote", "add", other_remote, bucket_url]) == 0
+    assert main(["remote", "default", default_remote]) == 0
+    assert main(["remote", "list"]) == 0
+    out, _ = capsys.readouterr()
+    out_lines = out.splitlines()
+    assert out_lines[0].split() == [default_remote, bucket_url, "(default)"]
+    assert out_lines[1].split() == [other_remote, bucket_url]
+
+
 def test_upper_case_remote(tmp_dir, dvc, local_cloud):
     remote_name = "UPPERCASEREMOTE"
 

tests/func/test_repo_index.py

@@ -114,13 +114,13 @@ def assert_index_equal(first, second, strict=True, ordered=True):
     assert len(first) == len(second), "Index have different no. of stages"
     assert set(first) == set(second), "Index does not have same stages"
     if ordered:
-        assert list(first) == list(
-            second
-        ), "Index does not have same sequence of stages"
+        assert list(first) == list(second), (
+            "Index does not have same sequence of stages"
+        )
     if strict:
-        assert set(map(id, first)) == set(
-            map(id, second)
-        ), "Index is not strictly equal"
+        assert set(map(id, first)) == set(map(id, second)), (
+            "Index is not strictly equal"
+        )
 
 
 def test_skip_graph_checks(dvc, mocker):

tests/func/test_state.py

@@ -65,7 +65,7 @@ def test_get_unused_links(tmp_dir, dvc):
     assert set(dvc.state.get_unused_links(links, dvc.fs)) == set()
     assert set(
         dvc.state.get_unused_links(
-            (links[:1] + [os.path.join(dvc.root_dir, "not-existing-file")]),
+            ([*links[:1], os.path.join(dvc.root_dir, "not-existing-file")]),
             dvc.fs,
         )
     ) == {"bar"}

tests/integration/plots/test_plots.py

@@ -553,11 +553,8 @@ def test_nested_x_defn_collection(tmp_dir, dvc, scm, capsys):
     )
     dvclive_metrics_dir.gen(
         {
-            "Error.tsv": "step\tError\n" "0\t0.11\n" "1\t0.22\n" "2\t0.44\n",
-            "Max_Leaf_Nodes.tsv": "step\tMax_Leaf_Nodes\n"
-            "0\t5\n"
-            "1\t50\n"
-            "2\t500\n",
+            "Error.tsv": "step\tError\n0\t0.11\n1\t0.22\n2\t0.44\n",
+            "Max_Leaf_Nodes.tsv": "step\tMax_Leaf_Nodes\n0\t5\n1\t50\n2\t500\n",
         }
     )
     (other_logger_dir / "multiple_metrics.json").dump(

tests/unit/command/test_dataset.py

@@ -106,7 +106,8 @@ def test_add_already_exists(dvc, caplog, mocker):
                 "missing": "Updating mydataset (s3://bucket/path)\n",
                 "unchanged": "Nothing to update\n",
                 "updated": (
-                    "Updating mydataset (s3://bucket/path)\n"
+                    r"Updating mydataset (s3://bucket/path)"
+                    "\n"
                     "M\tbaz\n"
                     "A\tfoobar\n"
                     "D\tfoo\n"

tests/unit/command/test_ls_url.py

@@ -1,6 +1,7 @@
 from dvc.cli import parse_args
 from dvc.commands.ls_url import CmdListUrl
 from dvc.config import Config
+from dvc.fs import LocalFileSystem
 
 
 def test_ls_url(mocker, M):
@@ -12,12 +13,16 @@ def test_ls_url(mocker, M):
     assert cmd.run() == 0
 
     m.assert_called_once_with(
-        "src", recursive=False, fs_config=None, config=M.instance_of(Config)
+        "src",
+        recursive=False,
+        maxdepth=None,
+        fs_config=None,
+        config=M.instance_of(Config),
     )
 
 
 def test_recursive(mocker, M):
-    cli_args = parse_args(["ls-url", "-R", "src"])
+    cli_args = parse_args(["ls-url", "-R", "-L", "2", "src"])
     assert cli_args.func == CmdListUrl
     cmd = cli_args.func(cli_args)
     m = mocker.patch("dvc.repo.Repo.ls_url", autospec=True)
@@ -25,5 +30,16 @@ def test_recursive(mocker, M):
     assert cmd.run() == 0
 
     m.assert_called_once_with(
-        "src", recursive=True, fs_config=None, config=M.instance_of(Config)
+        "src", recursive=True, maxdepth=2, fs_config=None, config=M.instance_of(Config)
     )
+
+
+def test_tree(mocker, M):
+    cli_args = parse_args(["ls-url", "--tree", "--level", "2", "src"])
+    assert cli_args.func == CmdListUrl
+    cmd = cli_args.func(cli_args)
+    m = mocker.patch("dvc.repo.ls._ls_tree", autospec=True)
+
+    assert cmd.run() == 0
+
+    m.assert_called_once_with(M.instance_of(LocalFileSystem), "src", maxdepth=2)

tests/unit/fs/test_dvc_info.py

@@ -123,6 +123,6 @@ def test_info_on_subrepos(make_tmp_dir, tmp_dir, dvc, scm, dvcfs):
 
     for path in ["subrepo", "subrepo/foo", "subrepo/foobar"]:
         info = dvcfs.info(path)
-        assert info["repo"].root_dir == str(
-            subrepo
-        ), f"repo root didn't match for {path}"
+        assert info["repo"].root_dir == str(subrepo), (
+            f"repo root didn't match for {path}"
+        )

tests/unit/output/test_load.py

@@ -99,7 +99,7 @@ def test_load_from_pipeline_error_on_typ(dvc, typ):
         output.load_from_pipeline(Stage(dvc), ["file1"], typ)
 
 
-@pytest.mark.parametrize("key", [3, "list".split()])
+@pytest.mark.parametrize("key", [3, ["list"]])
 def test_load_from_pipeline_illegal_type(dvc, key):
     stage = Stage(dvc)
     with pytest.raises(ValueError, match=f"'{type(key).__name__}' not supported."):

tests/unit/test_config.py

@@ -130,6 +130,5 @@ def test_feature_section_supports_arbitrary_values(caplog):
     assert "random_key_1" not in data
     assert "random_key_2" not in data
     assert (
-        "'feature.random_key_1', 'feature.random_key_2' "
-        "config options are unsupported"
+        "'feature.random_key_1', 'feature.random_key_2' config options are unsupported"
     ) in caplog.text

tests/unit/test_context.py

@@ -60,7 +60,7 @@ def test_context_list():
         _ = context["lst"][3]
 
     context["lst"].insert(0, "baz")
-    assert context["lst"] == CtxList(["baz"] + lst[:2])
+    assert context["lst"] == CtxList(["baz", *lst[:2]])
 
 
 def test_context_setitem_getitem():

tests/unit/test_ignore.py

@@ -168,8 +168,9 @@ def test_match_ignore_from_file(
 ):
     from dvc.fs import localfs
 
+    root = r"\\" if os.name == "nt" else "/"
     dvcignore_path = os.path.join(
-        os.path.sep, "full", "path", "to", "ignore", "file", ".dvcignore"
+        root, "full", "path", "to", "ignore", "file", ".dvcignore"
     )
     dvcignore_dirname = os.path.dirname(dvcignore_path)
 

tests/unit/test_logger.py

@@ -1,7 +1,5 @@
 import logging
-import time
 import traceback
-from datetime import datetime
 
 import colorama
 import pytest
@@ -22,11 +20,9 @@ colors = {
 
 @pytest.fixture
 def dt(mocker):
-    mocker.patch(
-        "time.time",
-        return_value=time.mktime(datetime(2020, 2, 2).timetuple()),  # noqa: DTZ001
-    )
-    return "2020-02-02 00:00:00,000"
+    dt_str = "2020-02-02 00:00:00,000"
+    mocker.patch.object(formatter, "formatTime", return_value=dt_str)
+    return dt_str
 
 
 class TestColorFormatter:
@@ -104,9 +100,7 @@ class TestColorFormatter:
                 logger.exception("")
 
             expected = (
-                "{red}{datetime}{nc} "
-                "{red}ERROR{nc}: description\n"
-                "{stack_trace}".format(
+                "{red}{datetime}{nc} {red}ERROR{nc}: description\n{stack_trace}".format(
                     stack_trace=stack_trace,
                     **colors,
                     datetime=dt,
@@ -144,9 +138,7 @@ class TestColorFormatter:
                 logger.exception("something", extra={"tb_only": True})
 
             expected = (
-                "{red}{datetime}{nc} "
-                "{red}ERROR{nc}: something\n"
-                "{stack_trace}".format(
+                "{red}{datetime}{nc} {red}ERROR{nc}: something\n{stack_trace}".format(
                     stack_trace=stack_trace,
                     **colors,
                     datetime=dt,

tests/unit/ui/test_pager.py

@@ -1,3 +1,5 @@
+from pydoc import plainpager
+
 import pytest
 
 from dvc.env import DVC_PAGER
@@ -70,8 +72,8 @@ def test_dvc_sets_some_options_on_less_if_less_env_defined(mocker, monkeypatch,
     assert find_pager() == "less --RAW-CONTROL-CHARS --chop-long-lines"
 
 
-def test_make_pager_when_no_pager_found(mocker, monkeypatch):
-    assert make_pager(None).__name__ == "plainpager"
+def test_make_pager_when_no_pager_found():
+    assert make_pager(None) is plainpager
 
 
 def test_pager(mocker, monkeypatch):