loader: don't forget to set meta.isdir

Leftover from #8273

Fixes #8300

.github/CODEOWNERS

@@ -1 +0,0 @@
-*       @iterative/dvc

.github/workflows/packages.yaml

@@ -6,152 +6,92 @@ on:
   release:
     types: [released, prereleased]
   workflow_dispatch:
+
 jobs:
-  deb:
-    runs-on: ubuntu-18.04
+  binary:
+    strategy:
+      matrix:
+        include:
+         - {pkg: "deb",  os: "ubuntu-18.04", asset: "dvc_${{ github.event.release.tag_name }}_amd64.deb"}
+         - {pkg: "rpm", os: "ubuntu-18.04", asset: "dvc-${{ github.event.release.tag_name }}-1.x86_64.rpm"}
+         - {pkg: "osxpkg", os: "macos-10.15", asset: "dvc-${{ github.event.release.tag_name }}.pkg"}
+         - {pkg: "exe",  os: "windows-2019", asset: "dvc-${{ github.event.release.tag_name }}.exe"}
+
+    name: ${{ matrix.pkg }}
+    runs-on: ${{ matrix.os }}
     steps:
     - uses: actions/checkout@v3
       with:
           fetch-depth: 0
+
     - name: Set up Python 3.9
       uses: actions/setup-python@v4
       with:
         python-version: 3.9
-    - name: Set up Ruby 2.4
-      uses: actions/[email protected]
+        cache: pip
+        cache-dependency-path: |
+          setup.cfg
+          scripts/build-requirements.txt
+
+    - name: Set up Ruby 2.6
+      uses: ruby/setup-ruby@v1
+      if: matrix.pkg != 'exe'
       with:
-        ruby-version: 2.4
-    - name: Install deps
-      run: |
-        pip install .[all]
-        pip install -r scripts/build-requirements.txt
-        gem install --no-document fpm
-    - name: Build deb
-      run: python scripts/build.py deb
-    - name: Publish deb
-      if: github.event_name == 'release'
-      uses: actions/[email protected]
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      with:
-        upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/fpm/dvc_${{ github.event.release.tag_name }}_amd64.deb
-        asset_name: dvc_${{ github.event.release.tag_name }}_amd64.deb
-        asset_content_type: binary/octet-stream
-  rpm:
-    runs-on: ubuntu-18.04
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-    - name: Set up Python 3.9
-      uses: actions/setup-python@v4
-      with:
-        python-version: 3.9
-    - name: Set up Ruby 2.4
-      uses: actions/[email protected]
-      with:
-        ruby-version: 2.4
-    - name: Install deps
-      run: |
-        pip install .[all]
-        pip install -r scripts/build-requirements.txt
-        gem install --no-document fpm
-    - name: Build rpm
-      run: python scripts/build.py rpm
-    - name: Publish rpm
-      if: github.event_name == 'release'
-      uses: actions/[email protected]
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      with:
-        upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/fpm/dvc-${{ github.event.release.tag_name }}-1.x86_64.rpm
-        asset_name: dvc-${{ github.event.release.tag_name }}-1.x86_64.rpm
-        asset_content_type: binary/octet-stream
-  osxpkg:
-    runs-on: macos-10.15
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-    - name: Set up Python 3.9
-      uses: actions/setup-python@v4
-      with:
-        python-version: 3.9
-    - name: Set up Ruby 2.4
-      uses: actions/[email protected]
-      with:
-        ruby-version: 2.4
-    - name: Install deps
-      run: |
-        # available lxml (webdav dependency) wheels are built with an old
-        # SDK, which breaks osxpkg notarization. Building from scratch
-        # instead.
-        pip install .[all] --no-binary lxml
-        pip install -r scripts/build-requirements.txt
-        gem install --no-document fpm
-    - name: Build
-      run: python scripts/build.py osxpkg
-    - name: Publish
-      if: github.event_name == 'release'
-      uses: actions/[email protected]
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      with:
-        upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/fpm/dvc-${{ github.event.release.tag_name }}.pkg
-        asset_name: dvc-${{ github.event.release.tag_name }}.pkg
-        asset_content_type: binary/octet-stream
-  exe:
-    runs-on: windows-2019
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-    - name: Set up Python 3.9
-      uses: actions/setup-python@v4
-      with:
-        python-version: 3.9
+        ruby-version: '2.6'
+
+    - name: Install fpm
+      if: matrix.pkg != 'exe'
+      run: gem install --no-document fpm
+
     - name: Install deps
       run: |
+        pip install -U wheel pip setuptools
         pip install .[all]
         pip install -r scripts/build-requirements.txt
+
     - name: Pull images
-      run: |
-        dvc pull
-    - name: Build
-      run: python scripts\build.py exe
-      shell: cmd
-    - name: Publish
+      run: dvc pull
+
+    - name: Build ${{ matrix.pkg }}
+      run: python scripts/build.py ${{ matrix.pkg }}
+
+    - name: Publish ${{ matrix.pkg }}
       if: github.event_name == 'release'
+      uses: actions/[email protected]
       env:
         GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      uses: actions/[email protected]
       with:
         upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/innosetup/dvc-${{ github.event.release.tag_name }}.exe
-        asset_name: dvc-${{ github.event.release.tag_name }}.exe
+        asset_path: scripts/${{ matrix.pkg == 'exe' && 'innosetup' || 'fpm' }}/${{ matrix.asset }}
+        asset_name: ${{ matrix.asset }}
         asset_content_type: binary/octet-stream
+
   pip:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     steps:
     - uses: actions/checkout@v3
       with:
         fetch-depth: 0
+
     - name: Set up Python 3.8
       uses: actions/setup-python@v4
       with:
         python-version: 3.8
+        cache: pip
+        cache-dependency-path: setup.cfg
+
     - name: Install
-      run: pip install ".[all,tests]"
+      run: |
+        pip install -U wheel pip build setuptools
+        pip install ".[all,tests]"
+
     - name: Build packages
       run: |
         pip install -U pip
         ./scripts/build_package.sh
+
     - name: Publish packages to PyPI
       if: github.event_name == 'release'
-      uses: pypa/gh-action-pypi-publish@master
+      uses: pypa/gh-action-pypi-publish@release/v1
       with:
-        user: __token__
         password: ${{ secrets.PYPI_TOKEN }}

.github/workflows/tests.yaml

@@ -43,6 +43,12 @@ jobs:
       matrix:
         os: [ubuntu-20.04, windows-latest, macos-latest]
         pyv: ["3.8", "3.9", "3.10"]
+        pytest-filter: 
+        - "import or plot or live or experiment"
+        - "not (import or plot or live or experiment)"
+        include:
+        - {os: ubuntu-latest, pyv: "3.11-dev"}
+
     steps:
     - uses: actions/checkout@v3
       with:
@@ -68,12 +74,11 @@ jobs:
       run: |
         pip install --upgrade pip setuptools wheel
         pip install -e ".[dev]"
+      env:
+        AIOHTTP_NO_EXTENSIONS: ${{ matrix.pyv == '3.11-dev' && '1' }}
     - name: run tests
       timeout-minutes: 40
-      run: >-
-        python -m tests -n=auto
-        --cov-report=xml --cov-report=term
-        ${{ env.extra_test_args }}
+      run: pytest -nauto -ra --durations 100 --cov=dvc --cov-report=xml --cov-report=term -k "${{ matrix.pytest-filter }}"
     - name: upload coverage report
       uses: codecov/codecov-action@v3
       with:

README.rst

@@ -11,7 +11,7 @@
 • `Contributing`_
 • `Community and Support`_
 
-|CI| |Maintainability| |Coverage| |VS Code| |DOI|
+|CI| |Python Version| |Coverage| |VS Code| |DOI|
 
 |PyPI| |Packages| |Brew| |Conda| |Choco| |Snap|
 
@@ -248,6 +248,10 @@ Barrak, A., Eghan, E.E. and Adams, B. `On the Co-evolution of ML Pipelines and S
    :target: https://codeclimate.com/github/iterative/dvc
    :alt: Code Climate
 
+.. |Python Version| image:: https://img.shields.io/pypi/pyversions/dvc
+   :target: https://pypi.org/project/dvc
+   :alt: Python Version
+
 .. |Coverage| image:: https://codecov.io/gh/iterative/dvc/branch/main/graph/badge.svg
    :target: https://codecov.io/gh/iterative/dvc
    :alt: Codecov

dvc/annotations.py

@@ -0,0 +1,35 @@
+from dataclasses import asdict, dataclass, field, fields
+from typing import Any, ClassVar, Dict, List, Optional
+
+from funcy import compact
+
+
+@dataclass
+class Annotation:
+    PARAM_DESC: ClassVar[str] = "desc"
+    PARAM_TYPE: ClassVar[str] = "type"
+    PARAM_LABELS: ClassVar[str] = "labels"
+    PARAM_META: ClassVar[str] = "meta"
+
+    desc: Optional[str] = None
+    type: Optional[str] = None
+    labels: List[str] = field(default_factory=list)
+    meta: Dict[str, Any] = field(default_factory=dict)
+
+    def update(self, **kwargs) -> "Annotation":
+        for attr, value in kwargs.items():
+            if value and hasattr(self, attr):
+                setattr(self, attr, value)
+        return self
+
+    def to_dict(self) -> Dict[str, str]:
+        return compact(asdict(self))
+
+
+ANNOTATION_FIELDS = [field.name for field in fields(Annotation)]
+ANNOTATION_SCHEMA = {
+    Annotation.PARAM_DESC: str,
+    Annotation.PARAM_TYPE: str,
+    Annotation.PARAM_LABELS: [str],
+    Annotation.PARAM_META: object,
+}

dvc/api/__init__.py

@@ -1,9 +1,15 @@
-from .data import (  # noqa, pylint: disable=redefined-builtin
-    get_url,
-    open,
-    read,
-)
+from dvc.fs.dvc import _DvcFileSystem as DvcFileSystem
+
+from .data import open  # pylint: disable=redefined-builtin
+from .data import get_url, read
 from .experiments import make_checkpoint
 from .params import params_show
 
-__all__ = ["get_url", "make_checkpoint", "open", "params_show", "read"]
+__all__ = [
+    "get_url",
+    "make_checkpoint",
+    "open",
+    "params_show",
+    "read",
+    "DvcFileSystem",
+]

dvc/api/params.py

@@ -37,6 +37,10 @@ def params_show(
             stages to retrieve parameters from.
             Defaults to `None`.
             If `None`, all parameters from all stages will be retrieved.
+            If this method is called from a different location to the one where
+            the `dvc.yaml` is found, the relative path to the `dvc.yaml` must
+            be provided as a prefix with the syntax `{relpath}:{stage}`.
+            For example: `subdir/dvc.yaml:stage-0` or `../dvc.yaml:stage-1`.
         rev (str, optional): Name of the `Git revision`_ to retrieve parameters
             from.
             Defaults to `None`.

dvc/cli/__init__.py

@@ -121,7 +121,7 @@ def _log_exceptions(exc: Exception) -> Optional[int]:
         )
         return None
 
-    from dvc_data.build import IgnoreInCollectedDirError
+    from dvc_data.hashfile.build import IgnoreInCollectedDirError
 
     if isinstance(exc, IgnoreInCollectedDirError):
         logger.exception("")

dvc/cli/actions.py

@@ -0,0 +1,20 @@
+from argparse import _AppendAction
+
+
+class CommaSeparatedArgs(_AppendAction):  # pylint: disable=protected-access
+    def __call__(self, parser, namespace, values, option_string=None):
+        from funcy import ldistinct
+
+        items = getattr(namespace, self.dest) or []
+        items.extend(map(str.strip, values.split(",")))
+        setattr(namespace, self.dest, ldistinct(items))
+
+
+class KeyValueArgs(_AppendAction):
+    def __call__(self, parser, namespace, values, option_string=None):
+        items = getattr(namespace, self.dest) or {}
+        for value in filter(bool, values):
+            key, _, value = value.partition("=")
+            items[key.strip()] = value
+
+        setattr(namespace, self.dest, items)

dvc/commands/add.py

@@ -2,12 +2,43 @@ import argparse
 import logging
 
 from dvc.cli import completion
+from dvc.cli.actions import KeyValueArgs
 from dvc.cli.command import CmdBase
 from dvc.cli.utils import append_doc_link
 
 logger = logging.getLogger(__name__)
 
 
+def _add_annotating_args(parser: argparse.ArgumentParser) -> None:
+    parser.add_argument(
+        "--desc",
+        type=str,
+        metavar="<text>",
+        help="User description of the data.",
+    )
+    parser.add_argument(
+        "--meta",
+        metavar="key=value",
+        nargs=1,
+        action=KeyValueArgs,
+        help="Custom metadata to add to the data.",
+    )
+    parser.add_argument(
+        "--label",
+        dest="labels",
+        type=str,
+        action="append",
+        metavar="<str>",
+        help="Label for the data.",
+    )
+    parser.add_argument(
+        "--type",
+        type=str,
+        metavar="<str>",
+        help="Type of the data.",
+    )
+
+
 class CmdAdd(CmdBase):
     def run(self):
         from dvc.exceptions import (
@@ -28,6 +59,9 @@ class CmdAdd(CmdBase):
                 glob=self.args.glob,
                 desc=self.args.desc,
                 out=self.args.out,
+                type=self.args.type,
+                labels=self.args.labels,
+                meta=self.args.meta,
                 remote=self.args.remote,
                 to_remote=self.args.to_remote,
                 jobs=self.args.jobs,
@@ -109,15 +143,8 @@ def add_parser(subparsers, parent_parser):
         ),
         metavar="<number>",
     )
-    parser.add_argument(
-        "--desc",
-        type=str,
-        metavar="<text>",
-        help=(
-            "User description of the data (optional). "
-            "This doesn't affect any DVC operations."
-        ),
-    )
+
+    _add_annotating_args(parser)
     parser.add_argument(
         "targets", nargs="+", help="Input files/directories to add."
     ).complete = completion.FILE

dvc/commands/data.py

@@ -1,9 +1,12 @@
 import argparse
 import logging
-from typing import TYPE_CHECKING
+from operator import itemgetter
+from typing import TYPE_CHECKING, Any, Dict, Iterable, Set
 
 from funcy import chunks, compact, log_durations
 
+from dvc.cli import completion
+from dvc.cli.actions import CommaSeparatedArgs
 from dvc.cli.command import CmdBase
 from dvc.cli.utils import append_doc_link, fix_subparsers
 from dvc.ui import ui
@@ -31,15 +34,22 @@ class CmdDataStatus(CmdBase):
         "unchanged": "DVC unchanged files",
     }
     HINTS = {
-        "not_in_cache": 'use "dvc pull <file>..." '
-        "to update your local storage",
-        "committed": "git commit the corresponding dvc files "
-        "to update the repo",
-        "uncommitted": 'use "dvc commit <file>..." to track changes',
-        "untracked": 'use "git add <file> ..." or '
-        'dvc add <file>..." to commit to git or to dvc',
-        "git_dirty": "there are {}changes not tracked by dvc, "
-        'use "git status" to see',
+        "not_in_cache": ('use "dvc fetch <file>..." to download files',),
+        "committed": (
+            "git commit the corresponding dvc files to update the repo",
+        ),
+        "uncommitted": (
+            'use "dvc commit <file>..." to track changes',
+            'use "dvc checkout <file>..." to discard changes',
+        ),
+        "untracked": (
+            'use "git add <file> ..." or '
+            '"dvc add <file>..." to commit to git or to dvc',
+        ),
+        "git_dirty": (
+            "there are {}changes not tracked by dvc, "
+            'use "git status" to see',
+        ),
     }
 
     @staticmethod
@@ -76,8 +86,9 @@ class CmdDataStatus(CmdBase):
             color = cls.COLORS.get(stage, None)
 
             ui.write(header)
-            if hint := cls.HINTS.get(stage):
-                ui.write(f"  ({hint})")
+            if hints := cls.HINTS.get(stage):
+                for hint in hints:
+                    ui.write(f"  ({hint})")
 
             if isinstance(stage_status, dict):
                 items = [
@@ -92,9 +103,10 @@ class CmdDataStatus(CmdBase):
                 out = "\n".join(tabs + item for item in chunk)
                 ui.write(colorize(out, color))
 
-        if (hint := cls.HINTS.get("git_dirty")) and git_info.get("is_dirty"):
-            message = hint.format("other " if result else "")
-            ui.write(f"[blue]({message})[/]", styled=True)
+        if (hints := cls.HINTS.get("git_dirty")) and git_info.get("is_dirty"):
+            for hint in hints:
+                message = hint.format("other " if result else "")
+                ui.write(f"[blue]({message})[/]", styled=True)
         return 0
 
     def run(self) -> int:
@@ -102,7 +114,6 @@ class CmdDataStatus(CmdBase):
             status = self.repo.data_status(
                 granular=self.args.granular,
                 untracked_files=self.args.untracked_files,
-                with_dirs=self.args.with_dirs,
             )
 
         if not self.args.unchanged:
@@ -116,6 +127,65 @@ class CmdDataStatus(CmdBase):
         return self._show_status(status)
 
 
+class CmdDataLs(CmdBase):
+    @staticmethod
+    def _show_table(
+        d: Iterable[Dict[str, Any]],
+        filter_types: Set[str],
+        filter_labels: Set[str],
+        markdown: bool = False,
+    ) -> None:
+        from rich.style import Style
+
+        from dvc.compare import TabularData
+
+        td = TabularData(
+            columns=["Path", "Type", "Labels", "Description"], fill_value="-"
+        )
+        for entry in sorted(d, key=itemgetter("path")):
+            typ = entry.get("type", "")
+            desc = entry.get("desc", "-")
+            labels = entry.get("labels", [])
+
+            if filter_types and typ not in filter_types:
+                continue
+            if filter_labels and filter_labels.isdisjoint(labels):
+                continue
+
+            rich_label = ui.rich_text()
+            for index, label in enumerate(labels):
+                if index:
+                    rich_label.append(",")
+                style = Style(bold=label in filter_labels, color="green")
+                rich_label.append(label, style=style)
+
+            if markdown and desc:
+                desc = desc.partition("\n")[0]
+
+            path = ui.rich_text(entry["path"], style="cyan")
+            type_style = Style(bold=typ in filter_types, color="yellow")
+            typ = ui.rich_text(entry.get("type", ""), style=type_style)
+            td.append([path, typ or "-", rich_label or "-", desc])
+
+        td.render(markdown=markdown, rich_table=True)
+
+    def run(self):
+        from dvc.repo.data import ls
+
+        filter_labels = set(self.args.labels)
+        filter_types = set(self.args.type)
+        d = ls(
+            self.repo, targets=self.args.targets, recursive=self.args.recursive
+        )
+        self._show_table(
+            d,
+            filter_labels=filter_labels,
+            filter_types=filter_types,
+            markdown=self.args.markdown,
+        )
+        return 0
+
+
 def add_parser(subparsers, parent_parser):
     data_parser = subparsers.add_parser(
         "data",
@@ -172,10 +242,52 @@ def add_parser(subparsers, parent_parser):
         nargs="?",
         help="Show untracked files.",
     )
-    data_status_parser.add_argument(
-        "--with-dirs",
+    data_status_parser.set_defaults(func=CmdDataStatus)
+
+    DATA_LS_HELP = "List data tracked by DVC with its metadata."
+    data_ls_parser = data_subparsers.add_parser(
+        "ls",
+        aliases=["list"],
+        parents=[parent_parser],
+        description=append_doc_link(DATA_LS_HELP, "data/ls"),
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        help=DATA_LS_HELP,
+    )
+    data_ls_parser.add_argument(
+        "--md",
+        "--show-md",
+        dest="markdown",
         action="store_true",
         default=False,
-        help=argparse.SUPPRESS,
+        help="Show tabulated output in the Markdown format (GFM).",
     )
-    data_status_parser.set_defaults(func=CmdDataStatus)
+    data_ls_parser.add_argument(
+        "--type",
+        action=CommaSeparatedArgs,
+        default=[],
+        help="Comma-separated list of type to filter.",
+    )
+    data_ls_parser.add_argument(
+        "--labels",
+        action=CommaSeparatedArgs,
+        default=[],
+        help="Comma-separated list of labels to filter.",
+    )
+    data_ls_parser.add_argument(
+        "-R",
+        "--recursive",
+        action="store_true",
+        default=False,
+        help="Recursively list from the specified directories.",
+    )
+    data_ls_parser.add_argument(
+        "targets",
+        default=None,
+        nargs="*",
+        help=(
+            "Limit command scope to these tracked files/directories, "
+            ".dvc files, or stage names."
+        ),
+    ).complete = completion.DVCFILES_AND_STAGE
+
+    data_ls_parser.set_defaults(func=CmdDataLs)

dvc/commands/data_sync.py

@@ -43,7 +43,8 @@ class CmdDataPull(CmdDataBase):
             )
             self.log_summary(stats)
         except (CheckoutError, DvcException) as exc:
-            self.log_summary(getattr(exc, "stats", {}))
+            if stats := getattr(exc, "stats", {}):
+                self.log_summary(stats)
             logger.exception("failed to pull data from the cloud")
             return 1
 

dvc/commands/experiments/diff.py

@@ -27,7 +27,7 @@ class CmdExperimentsDiff(CmdBase):
 
         if self.args.json:
             ui.write_json(diff)
-        else:
+        elif diff:
             from dvc.compare import show_diff
 
             precision = self.args.precision or DEFAULT_PRECISION

dvc/commands/imp.py

@@ -19,7 +19,11 @@ class CmdImport(CmdBase):
                 fname=self.args.file,
                 rev=self.args.rev,
                 no_exec=self.args.no_exec,
+                no_download=self.args.no_download,
                 desc=self.args.desc,
+                type=self.args.type,
+                labels=self.args.labels,
+                meta=self.args.meta,
                 jobs=self.args.jobs,
             )
         except DvcException:
@@ -33,6 +37,8 @@ class CmdImport(CmdBase):
 
 
 def add_parser(subparsers, parent_parser):
+    from .add import _add_annotating_args
+
     IMPORT_HELP = (
         "Download file or directory tracked by DVC or by Git "
         "into the workspace, and track it."
@@ -69,20 +75,19 @@ def add_parser(subparsers, parent_parser):
         help="Specify name of the .dvc file this command will generate.",
         metavar="<filename>",
     )
-    import_parser.add_argument(
+    no_download_exec_group = import_parser.add_mutually_exclusive_group()
+    no_download_exec_group.add_argument(
         "--no-exec",
         action="store_true",
         default=False,
-        help="Only create .dvc file without actually downloading it.",
+        help="Only create .dvc file without actually importing target data.",
     )
-    import_parser.add_argument(
-        "--desc",
-        type=str,
-        metavar="<text>",
-        help=(
-            "User description of the data (optional). "
-            "This doesn't affect any DVC operations."
-        ),
+    no_download_exec_group.add_argument(
+        "--no-download",
+        action="store_true",
+        default=False,
+        help="Create .dvc file including target data hash value(s)"
+        " but do not actually download the file(s).",
     )
     import_parser.add_argument(
         "-j",
@@ -94,4 +99,6 @@ def add_parser(subparsers, parent_parser):
         ),
         metavar="<number>",
     )
+
+    _add_annotating_args(import_parser)
     import_parser.set_defaults(func=CmdImport)

dvc/commands/imp_url.py

@@ -17,9 +17,13 @@ class CmdImportUrl(CmdBase):
                 out=self.args.out,
                 fname=self.args.file,
                 no_exec=self.args.no_exec,
+                no_download=self.args.no_download,
                 remote=self.args.remote,
                 to_remote=self.args.to_remote,
                 desc=self.args.desc,
+                type=self.args.type,
+                labels=self.args.labels,
+                meta=self.args.meta,
                 jobs=self.args.jobs,
             )
         except DvcException:
@@ -34,6 +38,8 @@ class CmdImportUrl(CmdBase):
 
 
 def add_parser(subparsers, parent_parser):
+    from .add import _add_annotating_args
+
     IMPORT_HELP = (
         "Download or copy file from URL and take it under DVC control."
     )
@@ -66,11 +72,19 @@ def add_parser(subparsers, parent_parser):
         help="Specify name of the .dvc file this command will generate.",
         metavar="<filename>",
     ).complete = completion.DIR
-    import_parser.add_argument(
+    no_download_exec_group = import_parser.add_mutually_exclusive_group()
+    no_download_exec_group.add_argument(
         "--no-exec",
         action="store_true",
         default=False,
-        help="Only create .dvc file without actually downloading it.",
+        help="Only create .dvc file without actually importing target data.",
+    )
+    no_download_exec_group.add_argument(
+        "--no-download",
+        action="store_true",
+        default=False,
+        help="Create .dvc file including target data hash value(s)"
+        " but do not actually download the file(s).",
     )
     import_parser.add_argument(
         "--to-remote",
@@ -94,13 +108,5 @@ def add_parser(subparsers, parent_parser):
         ),
         metavar="<number>",
     )
-    import_parser.add_argument(
-        "--desc",
-        type=str,
-        metavar="<text>",
-        help=(
-            "User description of the data (optional). "
-            "This doesn't affect any DVC operations."
-        ),
-    )
+    _add_annotating_args(import_parser)
     import_parser.set_defaults(func=CmdImportUrl)

dvc/commands/plots.py

@@ -155,11 +155,6 @@ class CmdPlots(CmdBase):
                 out=renderers_out,
                 templates_dir=self.repo.plots.templates_dir,
             )
-            if self.args.show_vega:
-                renderer = first(filter(lambda r: r.TYPE == "vega", renderers))
-                if renderer:
-                    ui.write_json(json.loads(renderer.get_filled_template()))
-                return 0
             if self.args.json:
                 renderers = _filter_unhandled_renderers(renderers)
                 _show_json(renderers, self.args.split)
@@ -167,6 +162,12 @@ class CmdPlots(CmdBase):
 
             _adjust_vega_renderers(renderers)
 
+            if self.args.show_vega:
+                renderer = first(filter(lambda r: r.TYPE == "vega", renderers))
+                if renderer:
+                    ui.write_json(json.loads(renderer.get_filled_template()))
+                return 0
+
             output_file: Path = (Path.cwd() / out).resolve() / "index.html"
 
             if renderers:
@@ -222,31 +223,22 @@ class CmdPlotsModify(CmdPlots):
 
 
 class CmdPlotsTemplates(CmdBase):
-    TEMPLATES_CHOICES = [
-        "simple",
-        "linear",
-        "confusion",
-        "confusion_normalized",
-        "scatter",
-        "smooth",
-        "bar_horizontal_sorted",
-        "bar_horizontal",
-    ]
-
     def run(self):
-        from dvc_render.vega_templates import dump_templates
+        from dvc.exceptions import InvalidArgumentError
+        from dvc_render.vega_templates import TEMPLATES
 
         try:
-            out = (
-                os.path.join(os.getcwd(), self.args.out)
-                if self.args.out
-                else self.repo.plots.templates_dir
-            )
+            target = self.args.template
+            if target:
+                for template in TEMPLATES:
+                    if target == template.DEFAULT_NAME:
+                        ui.write_json(template.DEFAULT_CONTENT)
+                        return 0
+                raise InvalidArgumentError(f"Unexpected template: {target}.")
 
-            targets = [self.args.target] if self.args.target else None
-            dump_templates(output=out, targets=targets)
-            templates_path = os.path.relpath(out, os.getcwd())
-            ui.write(f"Templates have been written into '{templates_path}'.")
+            else:
+                for template in TEMPLATES:
+                    ui.write(template.DEFAULT_NAME)
 
             return 0
         except DvcException:
@@ -356,8 +348,7 @@ def add_parser(subparsers, parent_parser):
     plots_modify_parser.set_defaults(func=CmdPlotsModify)
 
     TEMPLATES_HELP = (
-        "Write built-in plots templates to a directory "
-        "(.dvc/plots by default)."
+        "List built-in plots templates or show JSON specification for one."
     )
     plots_templates_parser = plots_subparsers.add_parser(
         "templates",
@@ -367,13 +358,14 @@ def add_parser(subparsers, parent_parser):
         formatter_class=argparse.RawDescriptionHelpFormatter,
     )
     plots_templates_parser.add_argument(
-        "target",
+        "template",
         default=None,
         nargs="?",
-        choices=CmdPlotsTemplates.TEMPLATES_CHOICES,
-        help="Template to write. Writes all templates by default.",
+        help=(
+            "Template for which to show JSON specification. "
+            "List all template names by default."
+        ),
     )
-    _add_output_argument(plots_templates_parser, typ="templates")
     plots_templates_parser.set_defaults(func=CmdPlotsTemplates)
 
 

dvc/commands/stage.py

@@ -76,7 +76,7 @@ class CmdStageList(CmdBase):
         # removing duplicates while maintaining order
         collected = chain.from_iterable(
             self.repo.stage.collect(
-                target=target, recursive=self.args.recursive, accept_group=True
+                target=target, recursive=self.args.recursive
             )
             for target in self.args.targets
         )

dvc/commands/update.py

@@ -18,6 +18,7 @@ class CmdUpdate(CmdBase):
                 rev=self.args.rev,
                 recursive=self.args.recursive,
                 to_remote=self.args.to_remote,
+                no_download=self.args.no_download,
                 remote=self.args.remote,
                 jobs=self.args.jobs,
             )
@@ -55,6 +56,13 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Update all stages in the specified directory.",
     )
+    update_parser.add_argument(
+        "--no-download",
+        action="store_true",
+        default=False,
+        help="Update .dvc file hash value(s)"
+        " but do not download the file(s).",
+    )
     update_parser.add_argument(
         "--to-remote",
         action="store_true",

dvc/config_schema.py

@@ -78,7 +78,9 @@ REMOTE_COMMON = {
     "url": str,
     "checksum_jobs": All(Coerce(int), Range(1)),
     "jobs": All(Coerce(int), Range(1)),
+    Optional("worktree"): Bool,
     Optional("no_traverse"): Bool,  # obsoleted
+    Optional("version_aware"): Bool,
 }
 LOCAL_COMMON = {
     "type": supported_cache_type,
@@ -289,4 +291,9 @@ SCHEMA = {
         "bool": All(Lower, Choices("store_true", "boolean_optional")),
         "list": All(Lower, Choices("nargs", "append")),
     },
+    "hydra": {
+        Optional("enabled", default=False): Bool,
+        "config_dir": str,
+        "config_name": str,
+    },
 }

dvc/data_cloud.py

@@ -3,16 +3,37 @@
 import logging
 from typing import TYPE_CHECKING, Iterable, Optional
 
-from dvc_data.db import get_index
+from funcy import cached_property
+
+from dvc_data.hashfile.db import get_index
 
 if TYPE_CHECKING:
     from dvc_data.hashfile.db import HashFileDB
     from dvc_data.hashfile.hash_info import HashInfo
-    from dvc_data.status import CompareStatusResult
+    from dvc_data.hashfile.status import CompareStatusResult
 
 logger = logging.getLogger(__name__)
 
 
+class Remote:
+    def __init__(self, path, fs, **config):
+        self.path = path
+        self.fs = fs
+
+        self.worktree = config.pop("worktree", False)
+        self.config = config
+
+    @cached_property
+    def odb(self):
+        from dvc_data.hashfile.db import get_odb
+
+        path = self.path
+        if self.worktree:
+            path = self.fs.path.join(path, ".dvc", "cache")
+
+        return get_odb(self.fs, path, **self.config)
+
+
 class DataCloud:
     """Class that manages dvc remotes.
 
@@ -27,18 +48,23 @@ class DataCloud:
     def __init__(self, repo):
         self.repo = repo
 
-    def get_remote_odb(
+    def get_remote(
         self,
         name: Optional[str] = None,
         command: str = "<command>",
-    ) -> "HashFileDB":
+    ) -> "Remote":
         from dvc.config import NoRemoteError
 
         if not name:
             name = self.repo.config["core"].get("remote")
 
         if name:
-            return self._init_odb(name)
+            from dvc.fs import get_cloud_fs
+
+            cls, config, fs_path = get_cloud_fs(self.repo, name=name)
+            fs = cls(**config)
+            config["tmp_dir"] = self.repo.index_db_dir
+            return Remote(fs_path, fs, **config)
 
         if bool(self.repo.config["remote"]):
             error_msg = (
@@ -55,13 +81,13 @@ class DataCloud:
 
         raise NoRemoteError(error_msg)
 
-    def _init_odb(self, name):
-        from dvc.fs import get_cloud_fs
-        from dvc_data.db import get_odb
-
-        cls, config, fs_path = get_cloud_fs(self.repo, name=name)
-        config["tmp_dir"] = self.repo.index_db_dir
-        return get_odb(cls(**config), fs_path, **config)
+    def get_remote_odb(
+        self,
+        name: Optional[str] = None,
+        command: str = "<command>",
+    ) -> "HashFileDB":
+        remote = self.get_remote(name=name, command=command)
+        return remote.odb
 
     def _log_missing(self, status: "CompareStatusResult"):
         if status.missing:
@@ -82,7 +108,7 @@ class DataCloud:
         **kwargs,
     ):
         from dvc.exceptions import FileTransferError
-        from dvc_data.transfer import TransferError, transfer
+        from dvc_data.hashfile.transfer import TransferError, transfer
 
         try:
             return transfer(src_odb, dest_odb, objs, **kwargs)
@@ -164,7 +190,7 @@ class DataCloud:
             log_missing: log warning messages if file doesn't exist
                 neither in cache, neither in cloud.
         """
-        from dvc_data.status import compare_status
+        from dvc_data.hashfile.status import compare_status
 
         if not odb:
             odb = self.get_remote_odb(remote, "status")

dvc/dependency/__init__.py

@@ -18,7 +18,7 @@ SCHEMA: Mapping[str, Any] = {
 }
 
 
-def _get(stage, p, info):
+def _get(stage, p, info, fs_config=None):
     if info and info.get(RepoDependency.PARAM_REPO):
         repo = info.pop(RepoDependency.PARAM_REPO)
         return RepoDependency(repo, stage, p, info)
@@ -27,7 +27,7 @@ def _get(stage, p, info):
         params = info.pop(ParamsDependency.PARAM_PARAMS)
         return ParamsDependency(stage, p, params)
 
-    return Dependency(stage, p, info)
+    return Dependency(stage, p, info, fs_config=fs_config)
 
 
 def loadd_from(stage, d_list):
@@ -38,10 +38,18 @@ def loadd_from(stage, d_list):
     return ret
 
 
-def loads_from(stage, s_list, erepo=None):
+def loads_from(stage, s_list, erepo=None, fs_config=None):
     assert isinstance(s_list, list)
     info = {RepoDependency.PARAM_REPO: erepo} if erepo else {}
-    return [_get(stage, s, info.copy()) for s in s_list]
+    return [
+        _get(
+            stage,
+            s,
+            info.copy(),
+            fs_config=fs_config,
+        )
+        for s in s_list
+    ]
 
 
 def _merge_params(s_list):

dvc/dependency/param.py

@@ -39,14 +39,15 @@ class ParamsDependency(Dependency):
 
     def __init__(self, stage, path, params=None, repo=None):
         self.params = list(params) if params else []
-        info = (
-            {self.PARAM_PARAMS: params} if isinstance(params, dict) else None
-        )
+        hash_info = HashInfo()
+        if isinstance(params, dict):
+            hash_info = HashInfo(self.PARAM_PARAMS, params)
         repo = repo or stage.repo
         path = path or os.path.join(repo.root_dir, self.DEFAULT_PARAMS_FILE)
-        super().__init__(stage, path, info=info, repo=repo)
+        super().__init__(stage, path, repo=repo)
+        self.hash_info = hash_info
 
-    def dumpd(self):
+    def dumpd(self, **kwargs):
         ret = super().dumpd()
         if not self.hash_info:
             ret[self.PARAM_PARAMS] = self.params or {}

dvc/dependency/repo.py

@@ -59,11 +59,11 @@ class RepoDependency(Dependency):
     def save(self):
         pass
 
-    def dumpd(self):
+    def dumpd(self, **kwargs):
         return {self.PARAM_PATH: self.def_path, self.PARAM_REPO: self.def_repo}
 
     def download(self, to, jobs=None):
-        from dvc_data.checkout import checkout
+        from dvc_data.hashfile.checkout import checkout
 
         for odb, objs in self.get_used_objs().items():
             self.repo.cloud.pull(objs, jobs=jobs, odb=odb)
@@ -103,8 +103,8 @@ class RepoDependency(Dependency):
         from dvc.config import NoRemoteError
         from dvc.exceptions import NoOutputOrStageError, PathMissingError
         from dvc.utils import as_posix
-        from dvc_data.build import build
-        from dvc_data.objects.tree import Tree, TreeError
+        from dvc_data.hashfile.build import build
+        from dvc_data.hashfile.tree import Tree, TreeError
 
         local_odb = self.repo.odb.local
         locked = kwargs.pop("locked", True)
@@ -153,8 +153,8 @@ class RepoDependency(Dependency):
     def _check_circular_import(self, odb, obj_ids):
         from dvc.exceptions import CircularImportError
         from dvc.fs.dvc import DvcFileSystem
-        from dvc_data.db.reference import ReferenceHashFileDB
-        from dvc_data.objects.tree import Tree
+        from dvc_data.hashfile.db.reference import ReferenceHashFileDB
+        from dvc_data.hashfile.tree import Tree
 
         if not isinstance(odb, ReferenceHashFileDB):
             return

dvc/dvcfile.py

@@ -198,7 +198,7 @@ class SingleStageFile(FileMixin):
         if self.verify:
             check_dvcfile_path(self.repo, self.path)
         logger.debug("Saving information to '%s'.", relpath(self.path))
-        dump_yaml(self.path, serialize.to_single_stage_file(stage))
+        dump_yaml(self.path, serialize.to_single_stage_file(stage, **kwargs))
         self.repo.scm_context.track_file(self.relpath)
 
     def remove_stage(self, stage):  # pylint: disable=unused-argument
@@ -243,9 +243,11 @@ class PipelineFile(FileMixin):
         self._lockfile.dump(stage)
 
     @staticmethod
-    def _check_if_parametrized(stage):
+    def _check_if_parametrized(stage, action: str = "dump") -> None:
         if stage.raw_data.parametrized:
-            raise ParametrizedDumpError(f"cannot dump a parametrized {stage}")
+            raise ParametrizedDumpError(
+                f"cannot {action} a parametrized {stage}"
+            )
 
     def _dump_pipeline_file(self, stage):
         self._check_if_parametrized(stage)
@@ -291,6 +293,7 @@ class PipelineFile(FileMixin):
         self._lockfile.remove()
 
     def remove_stage(self, stage):
+        self._check_if_parametrized(stage, "remove")
         self._lockfile.remove_stage(stage)
         if not self.exists():
             return

dvc/fs/__init__.py

@@ -10,6 +10,8 @@ from dvc_objects.fs import (  # noqa: F401
     Schemes,
     generic,
     get_fs_cls,
+    known_implementations,
+    localfs,
     registry,
     system,
 )
@@ -19,33 +21,26 @@ from dvc_objects.fs.errors import (  # noqa: F401
     ConfigError,
     RemoteMissingDepsError,
 )
-from dvc_objects.fs.implementations.azure import AzureAuthError  # noqa: F401
-from dvc_objects.fs.implementations.azure import AzureFileSystem  # noqa: F401
-from dvc_objects.fs.implementations.gdrive import (  # noqa: F401
-    GDriveFileSystem,
-)
-from dvc_objects.fs.implementations.gs import GSFileSystem  # noqa: F401
-from dvc_objects.fs.implementations.hdfs import HDFSFileSystem  # noqa: F401
-from dvc_objects.fs.implementations.local import localfs  # noqa: F401
-from dvc_objects.fs.implementations.oss import OSSFileSystem  # noqa: F401
-from dvc_objects.fs.implementations.s3 import S3FileSystem  # noqa: F401
-from dvc_objects.fs.implementations.ssh import (  # noqa: F401
-    DEFAULT_PORT as DEFAULT_SSH_PORT,
-)
-from dvc_objects.fs.implementations.ssh import SSHFileSystem  # noqa: F401
-from dvc_objects.fs.implementations.webdav import (  # noqa: F401
-    WebDAVFileSystem,
-    WebDAVSFileSystem,
-)
-from dvc_objects.fs.implementations.webhdfs import (  # noqa: F401
-    WebHDFSFileSystem,
-)
 from dvc_objects.fs.path import Path  # noqa: F401
 
 from .data import DataFileSystem  # noqa: F401
 from .dvc import DvcFileSystem  # noqa: F401
 from .git import GitFileSystem  # noqa: F401
 
+known_implementations.update(
+    {
+        "dvc": {
+            "class": "dvc.fs.dvc.DvcFileSystem",
+            "err": "dvc is supported, but requires 'dvc' to be installed",
+        },
+        "git": {
+            "class": "dvc.fs.git.GitFileSystem",
+            "err": "git is supported, but requires 'dvc' to be installed",
+        },
+    }
+)
+
+
 # pylint: enable=unused-import
 
 
@@ -116,11 +111,8 @@ def get_cloud_fs(repo, **kwargs):
 
     cls = get_fs_cls(remote_conf)
 
-    if cls == GDriveFileSystem and repo:
-        remote_conf["gdrive_credentials_tmp_dir"] = repo.tmp_dir
-
     url = remote_conf.pop("url")
-    if issubclass(cls, WebDAVFileSystem):
+    if cls.protocol in ["webdav", "webdavs"]:
         # For WebDAVFileSystem, provided url is the base path itself, so it
         # should be treated as being a root path.
         fs_path = cls.root_marker

dvc/fs/dvc.py

@@ -4,12 +4,12 @@ import os
 import posixpath
 import threading
 from contextlib import suppress
-from itertools import takewhile
-from typing import TYPE_CHECKING, Callable, Optional, Tuple, Type, Union
+from typing import TYPE_CHECKING, Any, Callable, Optional, Tuple, Type, Union
 
 from fsspec.spec import AbstractFileSystem
 from funcy import cached_property, wrap_prop, wrap_with
 
+from dvc.utils.fs import makedirs
 from dvc_objects.fs.base import FileSystem
 from dvc_objects.fs.callbacks import DEFAULT_CALLBACK
 from dvc_objects.fs.path import Path
@@ -22,29 +22,19 @@ if TYPE_CHECKING:
 logger = logging.getLogger(__name__)
 
 RepoFactory = Union[Callable[[str], "Repo"], Type["Repo"]]
+Key = Tuple[str, ...]
 
 
 def as_posix(path: str) -> str:
     return path.replace(ntpath.sep, posixpath.sep)
 
 
-def _wrap_walk(dvc_fs, *args, **kwargs):
-    for root, dnames, fnames in dvc_fs.walk(*args, **kwargs):
-        yield dvc_fs.path.join(dvc_fs.repo.root_dir, root), dnames, fnames
+# NOT the same as dvc.dvcfile.is_dvc_file()!
+def _is_dvc_file(fname):
+    from dvc.dvcfile import is_valid_filename
+    from dvc.ignore import DvcIgnore
 
-
-def _ls(fs, path):
-    dnames = []
-    fnames = []
-
-    for entry in fs.ls(path, detail=True):
-        name = fs.path.name(entry["name"])
-        if entry["type"] == "directory":
-            dnames.append(name)
-        else:
-            fnames.append(name)
-
-    return dnames, fnames
+    return is_valid_filename(fname) or fname == DvcIgnore.DVCIGNORE_FILE
 
 
 def _merge_info(repo, fs_info, dvc_info):
@@ -70,6 +60,10 @@ def _merge_info(repo, fs_info, dvc_info):
     return ret
 
 
+def _get_dvc_path(dvc_fs, subkey):
+    return dvc_fs.path.join(*subkey) if subkey else ""
+
+
 class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
     """DVC + git-tracked files fs.
 
@@ -82,28 +76,25 @@ class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
 
     root_marker = "/"
 
-    PARAM_REPO_URL = "repo_url"
-    PARAM_REPO_ROOT = "repo_root"
-    PARAM_REV = "rev"
-    PARAM_CACHE_DIR = "cache_dir"
-    PARAM_CACHE_TYPES = "cache_types"
-    PARAM_SUBREPOS = "subrepos"
-
     def __init__(
         self,
+        url: Optional[str] = None,
+        rev: Optional[str] = None,
         repo: Optional["Repo"] = None,
-        subrepos=False,
-        repo_factory: RepoFactory = None,
-        **kwargs,
-    ):
-        super().__init__()
-
+        subrepos: bool = False,
+        repo_factory: Optional[RepoFactory] = None,
+        **repo_kwargs: Any,
+    ) -> None:
         from pygtrie import Trie
 
+        super().__init__()
         if repo is None:
-            repo, repo_factory = self._repo_from_fs_config(
-                subrepos=subrepos, **kwargs
+            repo = self._make_repo(
+                url=url, rev=rev, subrepos=subrepos, **repo_kwargs
             )
+            assert repo
+            # pylint: disable=protected-access
+            repo_factory = repo._fs_conf["repo_factory"]
 
         if not repo_factory:
             from dvc.repo import Repo
@@ -137,78 +128,35 @@ class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         if hasattr(repo, "dvc_dir"):
             self._datafss[key] = DataFileSystem(index=repo.index.data["repo"])
 
-    def _get_key(self, path):
+    def _get_key(self, path) -> Key:
         parts = self.repo.fs.path.relparts(path, self.repo.root_dir)
         if parts == (".",):
             parts = ()
         return parts
 
+    def _get_key_from_relative(self, path) -> Key:
+        parts = self.path.relparts(path, self.root_marker)
+        if parts and parts[0] == os.curdir:
+            parts = parts[1:]
+        return parts
+
+    def _from_key(self, parts: Key) -> str:
+        return self.repo.fs.path.join(self.repo.root_dir, *parts)
+
     @property
     def repo_url(self):
         if self.repo is None:
             return None
         return self.repo.url
 
-    @property
-    def config(self):
-        return {
-            self.PARAM_REPO_URL: self.repo_url,
-            self.PARAM_REPO_ROOT: self.repo.root_dir,
-            self.PARAM_REV: getattr(self.repo.fs, "rev", None),
-            self.PARAM_CACHE_DIR: os.path.abspath(self.repo.odb.local.path),
-            self.PARAM_CACHE_TYPES: self.repo.odb.local.cache_types,
-            self.PARAM_SUBREPOS: self._traverse_subrepos,
-        }
-
     @classmethod
-    def _repo_from_fs_config(
-        cls, **config
-    ) -> Tuple["Repo", Optional["RepoFactory"]]:
-        from dvc.external_repo import erepo_factory, external_repo
+    def _make_repo(cls, **kwargs) -> "Repo":
         from dvc.repo import Repo
 
-        url = config.get(cls.PARAM_REPO_URL)
-        root = config.get(cls.PARAM_REPO_ROOT)
-        assert url or root
+        with Repo.open(uninitialized=True, **kwargs) as repo:
+            return repo
 
-        def _open(*args, **kwargs):
-            # NOTE: if original repo was an erepo (and has a URL),
-            # we cannot use Repo.open() since it will skip erepo
-            # cache/remote setup for local URLs
-            if url is None:
-                return Repo.open(*args, **kwargs)
-            return external_repo(*args, **kwargs)
-
-        cache_dir = config.get(cls.PARAM_CACHE_DIR)
-        cache_config = (
-            {}
-            if not cache_dir
-            else {
-                "cache": {
-                    "dir": cache_dir,
-                    "type": config.get(cls.PARAM_CACHE_TYPES),
-                }
-            }
-        )
-        repo_kwargs: dict = {
-            "rev": config.get(cls.PARAM_REV),
-            "subrepos": config.get(cls.PARAM_SUBREPOS, False),
-            "uninitialized": True,
-        }
-        factory: Optional["RepoFactory"] = None
-        if url is None:
-            repo_kwargs["config"] = cache_config
-        else:
-            repo_kwargs["cache_dir"] = cache_dir
-            factory = erepo_factory(url, root, cache_config)
-
-        with _open(
-            url if url else root,
-            **repo_kwargs,
-        ) as repo:
-            return repo, factory
-
-    def _get_repo(self, path: str) -> "Repo":
+    def _get_repo(self, key: Key) -> "Repo":
         """Returns repo that the path falls in, using prefix.
 
         If the path is already tracked/collected, it just returns the repo.
@@ -216,33 +164,21 @@ class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         Otherwise, it collects the repos that might be in the path's parents
         and then returns the appropriate one.
         """
-        if not self.repo.fs.path.isin_or_eq(path, self.repo.root_dir):
-            # outside of repo
-            return self.repo
-
-        key = self._get_key(path)
         repo = self._subrepos_trie.get(key)
         if repo:
             return repo
 
         prefix_key, repo = self._subrepos_trie.longest_prefix(key)
-        prefix = self.repo.fs.path.join(
-            self.repo.root_dir,
-            *prefix_key,  # pylint: disable=not-an-iterable
-        )
-
-        parents = (parent for parent in self.repo.fs.path.parents(path))
-        dirs = [path] + list(takewhile(lambda p: p != prefix, parents))
-        dirs.reverse()
-        self._update(dirs, starting_repo=repo)
+        dir_keys = (key[:i] for i in range(len(prefix_key) + 1, len(key) + 1))
+        self._update(dir_keys, starting_repo=repo)
         return self._subrepos_trie.get(key) or self.repo
 
     @wrap_with(threading.Lock())
-    def _update(self, dirs, starting_repo):
+    def _update(self, dir_keys, starting_repo):
         """Checks for subrepo in directories and updates them."""
         repo = starting_repo
-        for d in dirs:
-            key = self._get_key(d)
+        for key in dir_keys:
+            d = self._from_key(key)
             if self._is_dvc_repo(d):
                 repo = self.repo_factory(
                     d,
@@ -265,79 +201,64 @@ class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         repo_path = self.repo.fs.path.join(dir_path, Repo.DVC_DIR)
         return self.repo.fs.isdir(repo_path)
 
-    def _get_fs_pair(
-        self, path
-    ) -> Tuple[
-        Optional["Repo"],
-        Optional[FileSystem],
-        Optional[str],
-        Optional[DataFileSystem],
-        Optional[str],
-    ]:
+    def _get_subrepo_info(
+        self, key: Key
+    ) -> Tuple["Repo", Optional[DataFileSystem], Key]:
         """
-        Returns a pair of fss based on repo the path falls in, using prefix.
+        Returns information about the subrepo the key is part of.
         """
-        parts = self.path.relparts(path, self.root_marker)
-        if parts and parts[0] == os.curdir:
-            parts = parts[1:]
-
-        fs_path = self.repo.fs.path.join(self.repo.root_dir, *parts)
-        repo = self._get_repo(fs_path)
-        fs = repo.fs
-
-        repo_parts = fs.path.relparts(repo.root_dir, self.repo.root_dir)
-        if repo_parts[0] == os.curdir:
-            repo_parts = repo_parts[1:]
-
-        dvc_parts = parts[len(repo_parts) :]
-        if dvc_parts and dvc_parts[0] == os.curdir:
-            dvc_parts = dvc_parts[1:]
-
-        key = self._get_key(repo.root_dir)
-        dvc_fs = self._datafss.get(key)
-        if dvc_fs:
-            dvc_path = dvc_fs.path.join(*dvc_parts) if dvc_parts else ""
+        repo = self._get_repo(key)
+        repo_key: Key
+        if repo is self.repo:
+            repo_key = ()
+            subkey = key
         else:
-            dvc_path = None
+            repo_key = self._get_key(repo.root_dir)
+            subkey = key[len(repo_key) :]
 
-        return repo, fs, fs_path, dvc_fs, dvc_path
+        dvc_fs = self._datafss.get(repo_key)
+        return repo, dvc_fs, subkey
 
-    def open(
-        self, path, mode="r", encoding="utf-8", **kwargs
+    def _open(
+        self, path, mode="rb", **kwargs
     ):  # pylint: disable=arguments-renamed, arguments-differ
-        if "b" in mode:
-            encoding = None
-
-        _, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
+        key = self._get_key_from_relative(path)
+        fs_path = self._from_key(key)
         try:
-            return fs.open(fs_path, mode=mode, encoding=encoding)
+            return self.repo.fs.open(fs_path, mode=mode)
         except FileNotFoundError:
+            _, dvc_fs, subkey = self._get_subrepo_info(key)
             if not dvc_fs:
                 raise
 
-        return dvc_fs.open(dvc_path, mode=mode, encoding=encoding, **kwargs)
+        dvc_path = _get_dvc_path(dvc_fs, subkey)
+        return dvc_fs.open(dvc_path, mode=mode)
 
     def isdvc(self, path, **kwargs):
-        _, _, _, dvc_fs, dvc_path = self._get_fs_pair(path)
+        key = self._get_key_from_relative(path)
+        _, dvc_fs, subkey = self._get_subrepo_info(key)
+        dvc_path = _get_dvc_path(dvc_fs, subkey)
         return dvc_fs is not None and dvc_fs.isdvc(dvc_path, **kwargs)
 
     def ls(  # pylint: disable=arguments-differ
         self, path, detail=True, dvc_only=False, **kwargs
     ):
-        repo, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
-
-        dvcignore = repo.dvcignore
-        ignore_subrepos = kwargs.get("ignore_subrepos", True)
+        key = self._get_key_from_relative(path)
+        repo, dvc_fs, subkey = self._get_subrepo_info(key)
 
         names = set()
         if dvc_fs:
             with suppress(FileNotFoundError):
+                dvc_path = _get_dvc_path(dvc_fs, subkey)
                 for entry in dvc_fs.ls(dvc_path, detail=False):
                     names.add(dvc_fs.path.name(entry))
 
-        if not dvc_only and fs:
+        ignore_subrepos = kwargs.get("ignore_subrepos", True)
+        if not dvc_only:
+            fs = self.repo.fs
+            fs_path = self._from_key(key)
             try:
-                for entry in dvcignore.ls(
+                for entry in repo.dvcignore.ls(
                     fs, fs_path, detail=False, ignore_subrepos=ignore_subrepos
                 ):
                     names.add(fs.path.name(entry))
@@ -345,26 +266,21 @@ class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
                 pass
 
         dvcfiles = kwargs.get("dvcfiles", False)
-
-        def _func(fname):
-            from dvc.dvcfile import is_valid_filename
-            from dvc.ignore import DvcIgnore
-
-            if dvcfiles:
-                return True
-
-            return not (
-                is_valid_filename(fname) or fname == DvcIgnore.DVCIGNORE_FILE
-            )
-
-        names = filter(_func, names)
+        if not dvcfiles:
+            names = (name for name in names if not _is_dvc_file(name))
 
         infos = []
         paths = []
         for name in names:
             entry_path = self.path.join(path, name)
+            entry_key = key + (name,)
             try:
-                info = self.info(entry_path, ignore_subrepos=ignore_subrepos)
+                info = self._info(
+                    entry_key,
+                    entry_path,
+                    ignore_subrepos=ignore_subrepos,
+                    check_ignored=False,
+                )
             except FileNotFoundError:
                 continue
             infos.append(info)
@@ -378,45 +294,57 @@ class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
     def get_file(  # pylint: disable=arguments-differ
         self, rpath, lpath, callback=DEFAULT_CALLBACK, **kwargs
     ):
-        _, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(rpath)
+        key = self._get_key_from_relative(rpath)
+        fs_path = self._from_key(key)
+        fs = self.repo.fs
 
-        if fs:
-            try:
-                fs.get_file(fs_path, lpath, callback=callback, **kwargs)
-                return
-            except FileNotFoundError:
-                if not dvc_fs:
-                    raise
+        if self.isdir(rpath):
+            makedirs(lpath, exist_ok=True)
+            return None
+
+        try:
+            fs.get_file(fs_path, lpath, callback=callback, **kwargs)
+            return
+        except FileNotFoundError:
+            _, dvc_fs, subkey = self._get_subrepo_info(key)
+            if not dvc_fs:
+                raise
+        dvc_path = _get_dvc_path(dvc_fs, subkey)
         dvc_fs.get_file(dvc_path, lpath, callback=callback, **kwargs)
 
     def info(self, path, **kwargs):
-        repo, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
-
-        dvcignore = repo.dvcignore
+        key = self._get_key_from_relative(path)
         ignore_subrepos = kwargs.get("ignore_subrepos", True)
+        return self._info(key, path, ignore_subrepos=ignore_subrepos)
+
+    def _info(self, key, path, ignore_subrepos=True, check_ignored=True):
+        repo, dvc_fs, subkey = self._get_subrepo_info(key)
 
         dvc_info = None
         if dvc_fs:
             try:
-                dvc_info = dvc_fs.info(dvc_path)
+                dvc_info = dvc_fs.fs.index.info(subkey)
+                dvc_path = _get_dvc_path(dvc_fs, subkey)
+                dvc_info["name"] = dvc_path
             except FileNotFoundError:
                 pass
 
         fs_info = None
-        if fs:
-            try:
-                fs_info = fs.info(fs_path)
-                if dvcignore.is_ignored(
-                    fs, fs_path, ignore_subrepos=ignore_subrepos
-                ):
-                    fs_info = None
-            except (FileNotFoundError, NotADirectoryError):
-                if not dvc_info:
-                    raise
+        fs = self.repo.fs
+        fs_path = self._from_key(key)
+        try:
+            fs_info = fs.info(fs_path)
+            if check_ignored and repo.dvcignore.is_ignored(
+                fs, fs_path, ignore_subrepos=ignore_subrepos
+            ):
+                fs_info = None
+        except (FileNotFoundError, NotADirectoryError):
+            if not dvc_info:
+                raise
 
         # NOTE: if some parent in fs_path turns out to be a file, it means
         # that the whole repofs branch doesn't exist.
-        if fs and not fs_info and dvc_info:
+        if dvc_info and not fs_info:
             for parent in fs.path.parents(fs_path):
                 try:
                     if fs.info(parent)["type"] != "directory":
@@ -432,14 +360,6 @@ class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         info["name"] = path
         return info
 
-    def checksum(self, path):
-        _, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
-
-        try:
-            return fs.checksum(fs_path)
-        except FileNotFoundError:
-            return dvc_fs.checksum(dvc_path)
-
 
 class DvcFileSystem(FileSystem):
     protocol = "local"
@@ -468,10 +388,6 @@ class DvcFileSystem(FileSystem):
     def repo_url(self):
         return self.fs.repo_url
 
-    @property
-    def config(self):
-        return self.fs.config
-
     def from_os_path(self, path):
         if os.path.isabs(path):
             path = os.path.relpath(path, self.repo.root_dir)

dvc/fs/git.py

@@ -15,6 +15,7 @@ class GitFileSystem(FileSystem):  # pylint:disable=abstract-method
     """Proxies the repo file access methods to Git objects"""
 
     protocol = "local"
+    PARAM_CHECKSUM = "md5"
 
     def __init__(
         self,

dvc/info.py

@@ -14,6 +14,14 @@ from dvc.scm import SCMError
 from dvc.utils import error_link
 from dvc.utils.pkg import PKG
 
+SUBPROJECTS = (
+    "dvc_data",
+    "dvc_objects",
+    "dvc_render",
+    "dvc_task",
+    "dvclive",
+    "scmrepo",
+)
 package = "" if PKG is None else f"({PKG})"
 
 
@@ -23,6 +31,7 @@ def get_dvc_info():
         "---------------------------------",
         f"Platform: Python {platform.python_version()} on "
         f"{platform.platform()}",
+        f"Subprojects:{_get_subprojects()}",
         f"Supports:{_get_supported_remotes()}",
     ]
 
@@ -88,13 +97,29 @@ def _get_linktype_support_info(repo):
     return ", ".join(links)
 
 
+def _get_subprojects():
+    subprojects = []
+    for subproject in SUBPROJECTS:
+        try:
+            version = importlib_metadata.version(subproject)
+            subprojects.append(f"{subproject} = {version}")
+        except ImportError:
+            pass
+
+    return "\n\t" + "\n\t".join(subprojects)
+
+
 def _get_supported_remotes():
     supported_remotes = []
     for scheme in registry:
-        if scheme in [Schemes.LOCAL, Schemes.MEMORY]:
+        if scheme in [Schemes.LOCAL, Schemes.MEMORY, "dvc", "git"]:
+            continue
+
+        try:
+            fs_cls = registry[scheme]
+        except ImportError:
             continue
 
-        fs_cls = registry[scheme]
         if not fs_cls.get_missing_deps():
             dependencies = []
             for requirement in fs_cls.REQUIRES:

dvc/machine/backend/base.py

@@ -3,7 +3,8 @@ from contextlib import contextmanager
 from typing import TYPE_CHECKING, Iterator, Optional
 
 if TYPE_CHECKING:
-    from dvc.fs import SSHFileSystem
+    from dvc_ssh import SSHFileSystem
+
     from dvc.types import StrPath
 
 

dvc/machine/backend/terraform.py

@@ -3,8 +3,9 @@ from contextlib import contextmanager
 from functools import partial, partialmethod
 from typing import TYPE_CHECKING, Iterator, Optional
 
+from dvc_ssh import DEFAULT_PORT, SSHFileSystem
+
 from dvc.exceptions import DvcException
-from dvc.fs import DEFAULT_SSH_PORT, SSHFileSystem
 from dvc.utils.fs import makedirs
 
 from .base import BaseMachineBackend
@@ -68,7 +69,7 @@ class TerraformBackend(BaseMachineBackend):
             resource = tpi.default_resource(name)
         return {
             "host": resource["instance_ip"],
-            "port": DEFAULT_SSH_PORT,
+            "port": DEFAULT_PORT,
             "username": "ubuntu",
             "fs_factory": partial(_sshfs, dict(resource)),
         }

dvc/odbmgr.py

@@ -1,5 +1,5 @@
 from dvc.fs import GitFileSystem, Schemes
-from dvc_data.db import get_odb
+from dvc_data.hashfile.db import get_odb
 
 
 def _get_odb(repo, settings, fs=None):

dvc/output.py

@@ -1,10 +1,10 @@
 import logging
 import os
 from collections import defaultdict
-from typing import TYPE_CHECKING, Dict, Optional, Set, Type
+from typing import TYPE_CHECKING, Dict, List, Optional, Set, Tuple, Type
 from urllib.parse import urlparse
 
-from funcy import collecting, project
+from funcy import cached_property, collecting, project
 from voluptuous import And, Any, Coerce, Length, Lower, Required, SetTo
 
 from dvc import prompt
@@ -17,29 +17,25 @@ from dvc.exceptions import (
     MergeError,
     RemoteCacheRequiredError,
 )
-from dvc_data import Tree
-from dvc_data import check as ocheck
-from dvc_data import load as oload
-from dvc_data.build import build
-from dvc_data.checkout import checkout
+from dvc_data.hashfile import Tree
+from dvc_data.hashfile import check as ocheck
+from dvc_data.hashfile import load as oload
+from dvc_data.hashfile.build import build
+from dvc_data.hashfile.checkout import checkout
 from dvc_data.hashfile.hash_info import HashInfo
 from dvc_data.hashfile.istextfile import istextfile
 from dvc_data.hashfile.meta import Meta
-from dvc_data.transfer import transfer as otransfer
+from dvc_data.hashfile.transfer import transfer as otransfer
 from dvc_objects.errors import ObjectFormatError
 
-from .fs import (
-    HDFSFileSystem,
-    LocalFileSystem,
-    RemoteMissingDepsError,
-    S3FileSystem,
-    Schemes,
-    get_cloud_fs,
-)
+from .annotations import ANNOTATION_FIELDS, ANNOTATION_SCHEMA, Annotation
+from .fs import LocalFileSystem, RemoteMissingDepsError, Schemes, get_cloud_fs
 from .utils import relpath
 from .utils.fs import path_isin
 
 if TYPE_CHECKING:
+    from dvc_data.hashfile.obj import HashFile
+    from dvc_data.index import DataIndexKey
     from dvc_objects.db import ObjectDB
 
 logger = logging.getLogger(__name__)
@@ -65,10 +61,12 @@ CASE_SENSITIVE_CHECKSUM_SCHEMA = Any(
 #
 # so when a few types of outputs share the same name, we only need
 # specify it once.
+HDFS_PARAM_CHECKSUM = "checksum"
+S3_PARAM_CHECKSUM = "etag"
 CHECKSUMS_SCHEMA = {
     LocalFileSystem.PARAM_CHECKSUM: CHECKSUM_SCHEMA,
-    HDFSFileSystem.PARAM_CHECKSUM: CHECKSUM_SCHEMA,
-    S3FileSystem.PARAM_CHECKSUM: CASE_SENSITIVE_CHECKSUM_SCHEMA,
+    HDFS_PARAM_CHECKSUM: CHECKSUM_SCHEMA,
+    S3_PARAM_CHECKSUM: CASE_SENSITIVE_CHECKSUM_SCHEMA,
 }
 
 
@@ -85,9 +83,10 @@ def loadd_from(stage, d_list):
         plot = d.pop(Output.PARAM_PLOT, False)
         persist = d.pop(Output.PARAM_PERSIST, False)
         checkpoint = d.pop(Output.PARAM_CHECKPOINT, False)
-        desc = d.pop(Output.PARAM_DESC, False)
         live = d.pop(Output.PARAM_LIVE, False)
         remote = d.pop(Output.PARAM_REMOTE, None)
+        annot = {field: d.pop(field, None) for field in ANNOTATION_FIELDS}
+        files = d.pop(Output.PARAM_FILES, None)
         ret.append(
             _get(
                 stage,
@@ -98,9 +97,10 @@ def loadd_from(stage, d_list):
                 plot=plot,
                 persist=persist,
                 checkpoint=checkpoint,
-                desc=desc,
                 live=live,
                 remote=remote,
+                **annot,
+                files=files,
             )
         )
     return ret
@@ -192,6 +192,7 @@ def load_from_pipeline(stage, data, typ="outs"):
                 Output.PARAM_PERSIST,
                 Output.PARAM_CHECKPOINT,
                 Output.PARAM_REMOTE,
+                *ANNOTATION_FIELDS,
             ],
         )
 
@@ -245,6 +246,7 @@ class Output:
     PARAM_PATH = "path"
     PARAM_CACHE = "cache"
     PARAM_CHECKPOINT = "checkpoint"
+    PARAM_FILES = "files"
     PARAM_METRIC = "metric"
     PARAM_METRIC_TYPE = "type"
     PARAM_METRIC_XPATH = "xpath"
@@ -257,7 +259,6 @@ class Output:
     PARAM_PLOT_TITLE = "title"
     PARAM_PLOT_HEADER = "header"
     PARAM_PERSIST = "persist"
-    PARAM_DESC = "desc"
     PARAM_LIVE = "live"
     PARAM_LIVE_SUMMARY = "summary"
     PARAM_LIVE_HTML = "html"
@@ -289,11 +290,29 @@ class Output:
         checkpoint=False,
         live=False,
         desc=None,
+        type=None,  # pylint: disable=redefined-builtin
+        labels=None,
+        meta=None,
         remote=None,
         repo=None,
+        fs_config=None,
+        files: List[Dict[str, Any]] = None,
     ):
+        self.annot = Annotation(
+            desc=desc, type=type, labels=labels or [], meta=meta or {}
+        )
         self.repo = stage.repo if not repo and stage else repo
-        fs_cls, fs_config, fs_path = get_cloud_fs(self.repo, url=path)
+        meta = Meta.from_dict(info or {})
+        # NOTE: when version_aware is not passed into get_cloud_fs, it will be
+        # set based on whether or not path is versioned
+        fs_kwargs = {"version_aware": True} if meta.version_id else {}
+
+        if fs_config is not None:
+            fs_kwargs.update(**fs_config)
+
+        fs_cls, fs_config, fs_path = get_cloud_fs(
+            self.repo, url=path, **fs_kwargs
+        )
         self.fs = fs_cls(**fs_config)
 
         if (
@@ -325,21 +344,38 @@ class Output:
         # By resolved path, which contains actual location,
         # should be absolute and don't contain remote:// refs.
         self.stage = stage
-        self.meta = Meta.from_dict(info)
-        self.hash_info = HashInfo.from_dict(info)
+        self.meta = meta
+        self.files = files
         self.use_cache = False if self.IS_DEPENDENCY else cache
         self.metric = False if self.IS_DEPENDENCY else metric
         self.plot = False if self.IS_DEPENDENCY else plot
         self.persist = persist
         self.checkpoint = checkpoint
         self.live = live
-        self.desc = desc
 
         self.fs_path = self._parse_path(self.fs, fs_path)
-        self.obj = None
+        self.obj: Optional["HashFile"] = None
 
         self.remote = remote
 
+        if self.fs.version_aware:
+            _, version_id = self.fs.path.coalesce_version(
+                self.def_path, self.meta.version_id
+            )
+            self.meta.version_id = version_id
+
+        if self.is_in_repo:
+            self.hash_name = "md5"
+        else:
+            self.hash_name = self.fs.PARAM_CHECKSUM
+
+        self.hash_info = HashInfo(
+            name=self.hash_name,
+            value=getattr(self.meta, self.hash_name, None),
+        )
+        if self.hash_info and self.hash_info.isdir:
+            self.meta.isdir = True
+
     def _parse_path(self, fs, fs_path):
         parsed = urlparse(self.def_path)
         if (
@@ -380,6 +416,12 @@ class Output:
 
         return self.fs.path.relpath(self.fs_path, self.repo.root_dir)
 
+    def clear(self):
+        self.hash_info = HashInfo.from_dict({})
+        self.meta = Meta.from_dict({})
+        self.obj = None
+        self.files = None
+
     @property
     def protocol(self):
         return self.fs.protocol
@@ -419,21 +461,27 @@ class Output:
         )
 
     def get_hash(self):
+        _, hash_info = self._get_hash_meta()
+        return hash_info
+
+    def _get_hash_meta(self):
         if self.use_cache:
             odb = self.odb
-            name = self.odb.fs.PARAM_CHECKSUM
         else:
             odb = self.repo.odb.local
-            name = self.fs.PARAM_CHECKSUM
-        _, _, obj = build(
+        _, meta, obj = build(
             odb,
             self.fs_path,
             self.fs,
-            name,
+            self.hash_name,
             ignore=self.dvcignore,
             dry_run=not self.use_cache,
         )
-        return obj.hash_info
+        return meta, obj.hash_info
+
+    def get_meta(self) -> Meta:
+        meta, _ = self._get_hash_meta()
+        return meta
 
     @property
     def is_dir_checksum(self):
@@ -452,6 +500,17 @@ class Output:
 
         return self.fs.exists(self.fs_path)
 
+    @cached_property
+    def index_key(self) -> Tuple[str, "DataIndexKey"]:
+        if self.is_in_repo:
+            workspace = "repo"
+            key = self.repo.fs.path.relparts(self.fs_path, self.repo.root_dir)
+        else:
+            workspace = self.fs.protocol
+            no_drive = self.fs.path.flavour.splitdrive(self.fs_path)[1]
+            key = self.fs.path.parts(no_drive)[1:]
+        return workspace, key
+
     def changed_checksum(self):
         return self.hash_info != self.get_hash()
 
@@ -469,6 +528,11 @@ class Output:
         except (FileNotFoundError, ObjectFormatError):
             return True
 
+    def changed_meta(self) -> bool:
+        if self.fs.version_aware and self.meta.version_id:
+            return self.meta.version_id == self.get_meta().version_id
+        return False
+
     def workspace_status(self):
         if not self.exists:
             return {str(self): "deleted"}
@@ -551,11 +615,12 @@ class Output:
                 self.repo.odb.local,
                 self.fs_path,
                 self.fs,
-                self.fs.PARAM_CHECKSUM,
+                self.hash_name,
                 ignore=self.dvcignore,
                 dry_run=True,
             )
             self.hash_info = obj.hash_info
+            self.files = None
             if not self.IS_DEPENDENCY:
                 logger.debug(
                     "Output '%s' doesn't use cache. Skipping saving.", self
@@ -568,18 +633,19 @@ class Output:
             self.odb,
             self.fs_path,
             self.fs,
-            self.odb.fs.PARAM_CHECKSUM,
+            self.hash_name,
             ignore=self.dvcignore,
         )
         self.hash_info = self.obj.hash_info
+        self.files = None
 
     def set_exec(self):
         if self.isfile() and self.meta.isexec:
             self.odb.set_exec(self.fs_path)
 
     def _checkout(self, *args, **kwargs):
-        from dvc_data.checkout import CheckoutError as _CheckoutError
-        from dvc_data.checkout import LinkError, PromptError
+        from dvc_data.hashfile.checkout import CheckoutError as _CheckoutError
+        from dvc_data.hashfile.checkout import LinkError, PromptError
 
         kwargs.setdefault("ignore", self.dvcignore)
         try:
@@ -610,7 +676,7 @@ class Output:
                     self.odb,
                     filter_info or self.fs_path,
                     self.fs,
-                    self.odb.fs.PARAM_CHECKSUM,
+                    self.hash_name,
                     ignore=self.dvcignore,
                 )
                 otransfer(
@@ -639,7 +705,7 @@ class Output:
             self.odb,
             self.fs_path,
             self.fs,
-            self.odb.fs.PARAM_CHECKSUM,
+            self.hash_name,
             ignore=self.dvcignore,
         )
         save_obj = save_obj.filter(prefix)
@@ -653,8 +719,10 @@ class Output:
         )
         return checkout_obj
 
-    def dumpd(self):
-        ret = {**self.hash_info.to_dict(), **self.meta.to_dict()}
+    def dumpd(self, **kwargs):
+        meta = self.meta.to_dict()
+        meta.pop("isdir", None)
+        ret = {**self.hash_info.to_dict(), **meta}
 
         if self.is_in_repo:
             path = self.fs.path.as_posix(
@@ -668,9 +736,7 @@ class Output:
         if self.IS_DEPENDENCY:
             return ret
 
-        if self.desc:
-            ret[self.PARAM_DESC] = self.desc
-
+        ret.update(self.annot.to_dict())
         if not self.use_cache:
             ret[self.PARAM_CACHE] = self.use_cache
 
@@ -699,6 +765,18 @@ class Output:
         if self.remote:
             ret[self.PARAM_REMOTE] = self.remote
 
+        if (
+            self.use_cache
+            and self.is_in_repo
+            and self.hash_info.isdir
+            and (kwargs.get("with_files") or self.files is not None)
+        ):
+            if self.obj:
+                obj = self.obj
+            else:
+                obj = self.get_obj()
+            ret[self.PARAM_FILES] = obj.as_list(with_meta=True)
+
         return ret
 
     def verify_metric(self):
@@ -731,14 +809,22 @@ class Output:
         ) as cb:
             self.fs.get(self.fs_path, to.fs_path, batch_size=jobs, callback=cb)
 
-    def get_obj(self, filter_info=None, **kwargs):
+    def get_obj(
+        self, filter_info: Optional[str] = None, **kwargs
+    ) -> Optional["HashFile"]:
+        obj: Optional["HashFile"] = None
         if self.obj:
             obj = self.obj
         elif self.hash_info:
-            try:
-                obj = oload(self.odb, self.hash_info)
-            except FileNotFoundError:
-                return None
+            if self.files:
+                tree = Tree.from_list(self.files, hash_name=self.hash_name)
+                tree.digest()
+                obj = tree
+            else:
+                try:
+                    obj = oload(self.odb, self.hash_info)
+                except (FileNotFoundError, ObjectFormatError):
+                    return None
         else:
             return None
 
@@ -858,6 +944,7 @@ class Output:
         )
 
         self.hash_info = obj.hash_info
+        self.files = None
         return obj
 
     def get_files_number(self, filter_info=None):
@@ -928,6 +1015,7 @@ class Output:
 
         obj = self.get_obj()
         if filter_info and filter_info != self.fs_path:
+            assert obj
             prefix = self.fs.path.parts(
                 self.fs.path.relpath(filter_info, self.fs_path)
             )
@@ -1020,7 +1108,7 @@ class Output:
         other = out.dumpd()
 
         ignored = [
-            self.fs.PARAM_CHECKSUM,
+            self.hash_name,
             Meta.PARAM_SIZE,
             Meta.PARAM_NFILES,
         ]
@@ -1040,8 +1128,8 @@ class Output:
             )
 
     def merge(self, ancestor, other):
-        from dvc_data.objects.tree import MergeError as TreeMergeError
-        from dvc_data.objects.tree import du, merge
+        from dvc_data.hashfile.tree import MergeError as TreeMergeError
+        from dvc_data.hashfile.tree import du, merge
 
         assert other
 
@@ -1064,6 +1152,7 @@ class Output:
         self.odb.add(merged.path, merged.fs, merged.oid)
 
         self.hash_info = merged.hash_info
+        self.files = None
         self.meta = Meta(
             size=du(self.odb, merged),
             nfiles=len(merged),
@@ -1086,21 +1175,33 @@ class Output:
         return bool(self.plot) or bool(self.live)
 
 
+META_SCHEMA = {
+    Meta.PARAM_SIZE: int,
+    Meta.PARAM_NFILES: int,
+    Meta.PARAM_ISEXEC: bool,
+    Meta.PARAM_VERSION_ID: str,
+}
+
 ARTIFACT_SCHEMA = {
     **CHECKSUMS_SCHEMA,
+    **META_SCHEMA,
     Required(Output.PARAM_PATH): str,
     Output.PARAM_PLOT: bool,
     Output.PARAM_PERSIST: bool,
     Output.PARAM_CHECKPOINT: bool,
-    Meta.PARAM_SIZE: int,
-    Meta.PARAM_NFILES: int,
-    Meta.PARAM_ISEXEC: bool,
+}
+
+DIR_FILES_SCHEMA: Dict[str, Any] = {
+    **CHECKSUMS_SCHEMA,
+    **META_SCHEMA,
+    Required(Tree.PARAM_RELPATH): str,
 }
 
 SCHEMA = {
     **ARTIFACT_SCHEMA,
+    **ANNOTATION_SCHEMA,
     Output.PARAM_CACHE: bool,
     Output.PARAM_METRIC: Output.METRIC_SCHEMA,
-    Output.PARAM_DESC: str,
     Output.PARAM_REMOTE: str,
+    Output.PARAM_FILES: [DIR_FILES_SCHEMA],
 }

dvc/parsing/context.py

@@ -159,7 +159,7 @@ class Value(Node):
 PRIMITIVES = (int, float, str, bytes, bool)
 
 
-class Container(Node, ABC):
+class Container(Node, ABC):  # noqa: B024
     meta: Meta
     data: Union[list, dict]
     _key_transform = staticmethod(identity)

dvc/repo/__init__.py

@@ -3,7 +3,7 @@ import os
 from collections import defaultdict
 from contextlib import contextmanager
 from functools import wraps
-from typing import TYPE_CHECKING, Callable, Optional
+from typing import TYPE_CHECKING, Callable, List, Optional
 
 from funcy import cached_property
 
@@ -18,6 +18,7 @@ if TYPE_CHECKING:
     from dvc.fs import FileSystem
     from dvc.repo.scm_context import SCMContext
     from dvc.scm import Base
+    from dvc.stage import Stage
 
 logger = logging.getLogger(__name__)
 
@@ -444,6 +445,44 @@ class Repo:
 
         return used
 
+    def partial_imports(
+        self,
+        targets=None,
+        all_branches=False,
+        all_tags=False,
+        all_commits=False,
+        all_experiments=False,
+        commit_date: Optional[str] = None,
+        recursive=False,
+        revs=None,
+        num=1,
+    ) -> List["Stage"]:
+        """Get the stages related to the given target and collect dependencies
+        which are missing outputs.
+
+        This is useful to retrieve files which have been imported to the repo
+        using --no-download.
+
+        Returns:
+            A list of partially imported stages
+        """
+        from itertools import chain
+
+        partial_imports = chain.from_iterable(
+            self.index.partial_imports(targets, recursive=recursive)
+            for _ in self.brancher(
+                revs=revs,
+                all_branches=all_branches,
+                all_tags=all_tags,
+                all_commits=all_commits,
+                all_experiments=all_experiments,
+                commit_date=commit_date,
+                num=num,
+            )
+        )
+
+        return list(partial_imports)
+
     @property
     def stages(self):  # obsolete, only for backward-compatibility
         return self.index.stages

dvc/repo/add.py

@@ -264,6 +264,7 @@ def create_stages(
             outs=[out],
             external=external,
         )
-        if kwargs.get("desc"):
-            stage.outs[0].desc = kwargs["desc"]
+
+        out_obj = stage.outs[0]
+        out_obj.annot.update(**kwargs)
         yield stage

dvc/repo/data.py

@@ -1,7 +1,21 @@
 import os
+import posixpath
 from collections import defaultdict
-from typing import TYPE_CHECKING, Any, Dict, List, Optional, TypedDict, cast
+from itertools import chain
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    Iterable,
+    Iterator,
+    List,
+    Optional,
+    Tuple,
+    TypedDict,
+    cast,
+)
 
+from dvc.fs.git import GitFileSystem
 from dvc.ui import ui
 
 if TYPE_CHECKING:
@@ -10,19 +24,20 @@ if TYPE_CHECKING:
     from dvc.output import Output
     from dvc.repo import Repo
     from dvc_data.hashfile.db import HashFileDB
+    from dvc_data.hashfile.hash_info import HashInfo
     from dvc_data.hashfile.obj import HashFile
 
 
-def _in_cache(obj: Optional["HashFile"], cache: "HashFileDB") -> bool:
+def posixpath_to_os_path(path: str) -> str:
+    return path.replace(posixpath.sep, os.path.sep)
+
+
+def _in_cache(obj: "HashInfo", cache: "HashFileDB") -> bool:
     from dvc_objects.errors import ObjectFormatError
 
-    if not obj:
-        return False
-    if not obj.hash_info.value:
-        return False
-
+    assert obj.value
     try:
-        cache.check(obj.hash_info.value)
+        cache.check(obj.value)
         return True
     except (FileNotFoundError, ObjectFormatError):
         return False
@@ -30,31 +45,30 @@ def _in_cache(obj: Optional["HashFile"], cache: "HashFileDB") -> bool:
 
 def _shallow_diff(
     root: str,
-    old_obj: Optional["HashFile"],
-    new_obj: Optional["HashFile"],
+    old: Optional["HashInfo"],
+    new: Optional["HashInfo"],
     cache: "HashFileDB",
 ) -> Dict[str, List[str]]:
-    # TODO: add support for shallow diff in dvc-data
-    # TODO: we may want to recursively do in_cache check
     d = {}
 
-    from dvc_data.objects.tree import Tree
+    old_root = new_root = root
+    if old and old.isdir:
+        old_root = os.path.sep.join([root, ""])
+    if new and new.isdir:
+        new_root = os.path.sep.join([root, ""])
 
-    if isinstance(new_obj, Tree):
-        root = os.path.sep.join([root, ""])
+    if old and not _in_cache(old, cache):
+        d["not_in_cache"] = [old_root]
 
-    if not _in_cache(old_obj, cache):
-        d["not_in_cache"] = [root]
-
-    if old_obj is None and new_obj is None:
+    if not old and not new:
         return d
-    if old_obj is None:
-        return {"added": [root], **d}
-    if new_obj is None:
-        return {"deleted": [root], **d}
-    if old_obj.hash_info != new_obj.hash_info:
-        return {"modified": [root], **d}
-    return {"unchanged": [root], **d}
+    if not new:
+        return {"deleted": [old_root], **d}
+    if not old:
+        return {"added": [new_root], **d}
+    if old != new:
+        return {"modified": [new_root], **d}
+    return {"unchanged": [new_root], **d}
 
 
 def _granular_diff(
@@ -62,19 +76,12 @@ def _granular_diff(
     old_obj: Optional["HashFile"],
     new_obj: Optional["HashFile"],
     cache: "HashFileDB",
-    with_dirs: bool = False,
 ) -> Dict[str, List[str]]:
-    from dvc_data.diff import ROOT
-    from dvc_data.diff import diff as odiff
-    from dvc_data.objects.tree import Tree
+    from dvc_data.hashfile.diff import ROOT
+    from dvc_data.hashfile.diff import diff as odiff
 
-    drop_root = False
-    trees = isinstance(old_obj, Tree) or isinstance(new_obj, Tree)
-    if trees:
-        drop_root = not with_dirs
-
-    def path_join(root: str, *paths: str) -> str:
-        if not trees and paths == ROOT:
+    def path_join(root: str, *paths: str, isdir: bool = False) -> str:
+        if not isdir and paths == ROOT:
             return root
         return os.path.sep.join([root, *paths])
 
@@ -83,38 +90,46 @@ def _granular_diff(
     output: Dict[str, List[str]] = defaultdict(list)
     for state in ("added", "deleted", "modified", "unchanged"):
         items = getattr(diff_data, state)
-        output[state].extend(
-            path_join(root, *item.new.key)
-            for item in items
-            if not (drop_root and item.new.key == ROOT)
-        )
-        # TODO: PERF: diff is checking not_in_cache for each even if we only
-        # need it for the index.
-        # BUG: not_in_cache file also shows up as modified in staged and
-        # unstaged. We currently don't know if it is really modified.
-        output["not_in_cache"].extend(
-            path_join(root, *item.new.key)
-            for item in items
-            if not item.old.in_cache
-            and not (drop_root and item.new.key == ROOT)
-            and state != "added"
-        )
-    return output
+        for item in items:  # pylint: disable=not-an-iterable
+            entry = item.old if state == "deleted" else item.new
+            isdir = entry.oid.isdir if entry.oid else False
+
+            path = path_join(root, *entry.key, isdir=isdir)
+            output[state].append(path)
+            if not item.old.in_cache and state != "added":
+                output["not_in_cache"].append(path)
+    return dict(output)
+
+
+def _get_obj_items(root: str, obj: Optional["HashFile"]) -> List[str]:
+    if not obj:
+        return []
+
+    from dvc_data.hashfile.tree import Tree
+
+    sep = os.path.sep
+    if isinstance(obj, Tree):
+        return [sep.join([root, *key]) for key, _, _ in obj]
+    return [root]
 
 
 def _diff(
     root: str,
+    old_oid: Optional["HashInfo"],
     old_obj: Optional["HashFile"],
+    new_oid: Optional["HashInfo"],
     new_obj: Optional["HashFile"],
-    cache: "HashFileDB",
+    odb: "HashFileDB",
     granular: bool = False,
-    with_dirs: bool = False,
 ) -> Dict[str, List[str]]:
-    if granular:
-        return _granular_diff(
-            root, old_obj, new_obj, cache, with_dirs=with_dirs
-        )
-    return _shallow_diff(root, old_obj, new_obj, cache)
+    if not granular:
+        return _shallow_diff(root, old_oid, new_oid, odb)
+    if (old_oid and not old_obj) or (new_oid and not new_obj):
+        # we don't have enough information to give full details
+        unknown = _get_obj_items(root, new_obj)
+        shallow_diff = _shallow_diff(root, old_oid, new_oid, odb)
+        return {**shallow_diff, "unknown": unknown}
+    return _granular_diff(root, old_obj, new_obj, odb)
 
 
 class GitInfo(TypedDict, total=False):
@@ -141,6 +156,8 @@ def _git_info(scm: "Base", untracked_files: str = "all") -> GitInfo:
         empty_repo = False
 
     staged, unstaged, untracked = scm.status(untracked_files=untracked_files)
+    if os.name == "nt":
+        untracked = [posixpath_to_os_path(path) for path in untracked]
     # NOTE: order is important here.
     return GitInfo(
         staged=staged,
@@ -152,7 +169,7 @@ def _git_info(scm: "Base", untracked_files: str = "all") -> GitInfo:
 
 
 def _diff_index_to_wtree(repo: "Repo", **kwargs: Any) -> Dict[str, List[str]]:
-    from dvc_data.build import build
+    from dvc_data.hashfile.build import build
 
     unstaged_diff = defaultdict(list)
     for out in repo.index.outs:
@@ -175,8 +192,18 @@ def _diff_index_to_wtree(repo: "Repo", **kwargs: Any) -> Dict[str, List[str]]:
         cache = repo.odb.repo
         root = str(out)
         old = out.get_obj()
+
         with ui.status(f"Calculating diff for {root} between index/workspace"):
-            d = _diff(root, old, new, cache, **kwargs)
+            d = _diff(
+                root,
+                out.hash_info,
+                old,
+                new.hash_info if new else None,
+                new,
+                cache,
+                **kwargs,
+            )
+
         for state, items in d.items():
             if not items:
                 continue
@@ -188,7 +215,9 @@ def _diff_head_to_index(
     repo: "Repo", head: str = "HEAD", **kwargs: Any
 ) -> Dict[str, List[str]]:
     # we need to store objects from index and the HEAD to diff later
-    objs: Dict[str, Dict[str, "HashFile"]] = defaultdict(dict)
+    objs: Dict[str, Dict[str, Tuple["HashFile", "HashInfo"]]]
+    objs = defaultdict(dict)
+
     staged_diff = defaultdict(list)
     for rev in repo.brancher(revs=[head]):
         for out in repo.index.outs:
@@ -197,15 +226,20 @@ def _diff_head_to_index(
                 continue
 
             root = str(out)
+            if isinstance(out.fs, GitFileSystem):
+                root = posixpath_to_os_path(root)
             typ = "index" if rev == "workspace" else head
-            objs[root][typ] = out.get_obj()
+            objs[root][typ] = (out.get_obj(), out.hash_info)
 
     cache = repo.odb.repo
     for root, obj_d in objs.items():
-        old = obj_d.get(head, None)
-        new = obj_d.get("index", None)
+        old_obj, old_oid = obj_d.get(head, (None, None))
+        new_obj, new_oid = obj_d.get("index", (None, None))
         with ui.status(f"Calculating diff for {root} between head/index"):
-            d = _diff(root, old, new, cache, **kwargs)
+            d = _diff(
+                root, old_oid, old_obj, new_oid, new_obj, cache, **kwargs
+            )
+
         for state, items in d.items():
             if not items:
                 continue
@@ -216,28 +250,36 @@ def _diff_head_to_index(
 
 class Status(TypedDict):
     not_in_cache: List[str]
-    committed: Dict[str, Any]
-    uncommitted: Dict[str, Any]
+    committed: Dict[str, List[str]]
+    uncommitted: Dict[str, List[str]]
     untracked: List[str]
     unchanged: List[str]
     git: GitInfo
 
 
-def _transform_git_paths_to_dvc(repo: "Repo", files: List[str]):
+def _transform_git_paths_to_dvc(
+    repo: "Repo", files: Iterable[str]
+) -> List[str]:
     """Transform files rel. to Git root to DVC root, and drop outside files."""
     rel = repo.fs.path.relpath(repo.root_dir, repo.scm.root_dir).rstrip("/")
-    if rel in (os.curdir, ""):
-        return files
 
-    prefix = rel + os.sep
-    length = len(prefix)
-    return [file[length:] for file in files if file.startswith(prefix)]
+    # if we have repo root in a different location than scm's root,
+    # i.e. subdir repo, all git_paths need to be transformed rel. to the DVC
+    # repo root and anything outside need to be filtered out.
+    if rel not in (os.curdir, ""):
+        prefix = rel + os.sep
+        length = len(prefix)
+        files = (file[length:] for file in files if file.startswith(prefix))
+
+    start = repo.fs.path.relpath(repo.fs.path.getcwd(), repo.root_dir)
+    if start in (os.curdir, ""):
+        return list(files)
+    # we need to convert repo relative paths to curdir relative.
+    return [repo.fs.path.relpath(file, start) for file in files]
 
 
 def status(repo: "Repo", untracked_files: str = "no", **kwargs: Any) -> Status:
-    from scmrepo.exceptions import SCMError
-
-    from dvc.scm import NoSCMError
+    from dvc.scm import NoSCMError, SCMError
 
     head = kwargs.pop("head", "HEAD")
     uncommitted_diff = _diff_index_to_wtree(repo, **kwargs)
@@ -265,3 +307,18 @@ def status(repo: "Repo", untracked_files: str = "no", **kwargs: Any) -> Status:
         unchanged=list(unchanged),
         git=git_info,
     )
+
+
+def ls(
+    repo: "Repo",
+    targets: List[Optional[str]] = None,
+    recursive: bool = False,
+) -> Iterator[Dict[str, Any]]:
+    targets = targets or [None]
+    pairs = chain.from_iterable(
+        repo.stage.collect_granular(target, recursive=recursive)
+        for target in targets
+    )
+    for stage, filter_info in pairs:
+        for out in stage.filter_outs(filter_info):
+            yield {"path": str(out), **out.annot.to_dict()}

dvc/repo/diff.py

@@ -115,7 +115,7 @@ def _paths_checksums(repo, targets):
 
 def _output_paths(repo, targets):
     from dvc.fs import LocalFileSystem
-    from dvc_data.build import build
+    from dvc_data.hashfile.build import build
 
     on_working_fs = isinstance(repo.fs, LocalFileSystem)
 

dvc/repo/experiments/executor/manager/base.py

@@ -21,7 +21,7 @@ if TYPE_CHECKING:
 logger = logging.getLogger(__name__)
 
 
-class BaseExecutorManager(ABC, Mapping):
+class BaseExecutorManager(ABC, Mapping):  # noqa: B024
     """Manages executors for a collection of experiments to be run."""
 
     EXECUTOR_CLS: Type = BaseExecutor

dvc/repo/experiments/executor/ssh.py

@@ -5,10 +5,9 @@ import sys
 from contextlib import contextmanager
 from typing import TYPE_CHECKING, Callable, Iterable, Optional
 
+from dvc_ssh import SSHFileSystem
 from funcy import first
 
-from dvc.fs import SSHFileSystem
-
 from ..refs import (
     EXEC_BRANCH,
     EXEC_CHECKPOINT,

dvc/repo/experiments/queue/base.py

@@ -1,5 +1,6 @@
 import logging
 import os
+import sys
 from abc import ABC, abstractmethod
 from dataclasses import asdict, dataclass
 from typing import (
@@ -19,6 +20,7 @@ from typing import (
 
 from funcy import cached_property
 
+from dvc.dependency import ParamsDependency
 from dvc.env import DVCLIVE_RESUME
 from dvc.exceptions import DvcException
 from dvc.ui import ui
@@ -520,14 +522,36 @@ class BaseStashQueue(ABC):
                 provided via `exp run --set-param`.
 
         .. _Hydra Override:
-            https://hydra.cc/docs/next/advanced/override_grammar/basic/
+            https://hydra.cc/docs/advanced/override_grammar/basic/
         """
         logger.debug("Using experiment params '%s'", params)
 
-        from dvc.utils.hydra import apply_overrides
+        try:
+            from dvc.utils.hydra import apply_overrides, compose_and_dump
+        except ValueError:
+            if sys.version_info >= (3, 11):
+                raise DvcException(
+                    "--set-param is not supported in Python >= 3.11"
+                )
+            raise
 
+        hydra_config = self.repo.config.get("hydra", {})
+        hydra_enabled = hydra_config.get("enabled", False)
+        hydra_output_file = ParamsDependency.DEFAULT_PARAMS_FILE
         for path, overrides in params.items():
-            apply_overrides(path, overrides)
+            if hydra_enabled and path == hydra_output_file:
+                config_dir = os.path.join(
+                    self.repo.root_dir, hydra_config.get("config_dir", "conf")
+                )
+                config_name = hydra_config.get("config_name", "config")
+                compose_and_dump(
+                    path,
+                    config_dir,
+                    config_name,
+                    overrides,
+                )
+            else:
+                apply_overrides(path, overrides)
 
         # Force params file changes to be staged in git
         # Otherwise in certain situations the changes to params file may be

dvc/repo/experiments/run.py

@@ -1,6 +1,8 @@
 import logging
 from typing import Dict, Iterable, Optional
 
+from dvc.dependency.param import ParamsDependency
+from dvc.exceptions import InvalidArgumentError
 from dvc.repo import locked
 from dvc.ui import ui
 from dvc.utils.cli_parse import to_path_overrides
@@ -26,26 +28,59 @@ def run(
     Returns a dict mapping new experiment SHAs to the results
     of `repro` for that experiment.
     """
+    from dvc.utils.hydra import to_hydra_overrides
+
     if run_all:
         entries = list(repo.experiments.celery_queue.iter_queued())
         return repo.experiments.reproduce_celery(entries, jobs=jobs)
 
     if params:
-        params = to_path_overrides(params)
+        path_overrides = to_path_overrides(params)
+    else:
+        path_overrides = {}
 
-    if queue:
-        if not kwargs.get("checkpoint_resume", None):
-            kwargs["reset"] = True
+    hydra_enabled = repo.config.get("hydra", {}).get("enabled", False)
+    hydra_output_file = ParamsDependency.DEFAULT_PARAMS_FILE
+    if hydra_enabled and hydra_output_file not in path_overrides:
+        # Force `_update_params` even if `--set-param` was not used
+        path_overrides[hydra_output_file] = []
+
+    hydra_sweep = any(
+        x.is_sweep_override()
+        for param_file in path_overrides
+        for x in to_hydra_overrides(path_overrides[param_file])
+    )
+
+    if hydra_sweep and not queue:
+        raise InvalidArgumentError(
+            "Sweep overrides can't be used without `--queue`"
+        )
+
+    if not queue:
+        return repo.experiments.reproduce_one(
+            targets=targets, params=path_overrides, tmp_dir=tmp_dir, **kwargs
+        )
+
+    if hydra_sweep:
+        from dvc.utils.hydra import get_hydra_sweeps
+
+        sweeps = get_hydra_sweeps(path_overrides)
+    else:
+        sweeps = [path_overrides]
+
+    if not kwargs.get("checkpoint_resume", None):
+        kwargs["reset"] = True
+
+    for sweep_overrides in sweeps:
         queue_entry = repo.experiments.queue_one(
             repo.experiments.celery_queue,
             targets=targets,
-            params=params,
+            params=sweep_overrides,
             **kwargs,
         )
+        if sweep_overrides:
+            ui.write(f"Queueing with overrides '{sweep_overrides}'.")
         name = queue_entry.name or queue_entry.stash_rev[:7]
         ui.write(f"Queued experiment '{name}' for future execution.")
-        return {}
 
-    return repo.experiments.reproduce_one(
-        targets=targets, params=params, tmp_dir=tmp_dir, **kwargs
-    )
+    return {}

dvc/repo/experiments/show.py

@@ -142,6 +142,9 @@ def show(
     fetch_running: bool = True,
 ):
 
+    if repo.scm.no_commits:
+        return {}
+
     if onerror is None:
         onerror = onerror_collect
 

dvc/repo/fetch.py

@@ -1,6 +1,8 @@
 import logging
+from contextlib import suppress
 from typing import TYPE_CHECKING, Optional
 
+from dvc.config import NoRemoteError
 from dvc.exceptions import DownloadError, FileTransferError
 from dvc.fs import Schemes
 
@@ -12,6 +14,36 @@ if TYPE_CHECKING:
 logger = logging.getLogger(__name__)
 
 
+def _fetch_worktree(repo, remote):
+    from dvc_data.index import save
+
+    index = repo.index.data["repo"]
+    for key, entry in index.iteritems():
+        entry.fs = remote.fs
+        entry.path = remote.fs.path.join(
+            remote.path,
+            *key,
+        )
+    save(index)
+
+    for stage in repo.index.stages:
+        for out in stage.outs:
+            if not out.use_cache:
+                continue
+
+            if not out.is_in_repo:
+                continue
+
+            workspace, key = out.index_key
+            entry = repo.index.data[workspace][key]
+            out.hash_info = entry.hash_info
+            out.meta = entry.meta
+
+        stage.dvcfile.dump(stage)
+
+    return len(index)
+
+
 @locked
 def fetch(
     self,
@@ -42,6 +74,11 @@ def fetch(
     if isinstance(targets, str):
         targets = [targets]
 
+    with suppress(NoRemoteError):
+        _remote = self.cloud.get_remote(name=remote)
+        if _remote.worktree:
+            return _fetch_worktree(self, _remote)
+
     used = self.used_objs(
         targets,
         all_branches=all_branches,
@@ -93,6 +130,18 @@ def fetch(
             downloaded += d
             failed += f
 
+    d, f = _fetch_partial_imports(
+        self,
+        targets,
+        all_branches=all_branches,
+        all_tags=all_tags,
+        all_commits=all_commits,
+        recursive=recursive,
+        revs=revs,
+    )
+    downloaded += d
+    failed += f
+
     if failed:
         raise DownloadError(failed)
 
@@ -107,3 +156,33 @@ def _fetch(repo, obj_ids, **kwargs):
     except FileTransferError as exc:
         failed += exc.amount
     return downloaded, failed
+
+
+def _fetch_partial_imports(repo, targets, **kwargs):
+    from dvc.stage.exceptions import DataSourceChanged
+
+    downloaded = 0
+    failed = 0
+    for stage in repo.partial_imports(targets, **kwargs):
+        dep = stage.deps[0]
+        out = stage.outs[0]
+        try:
+            if dep.changed_checksum():
+                raise DataSourceChanged(f"{stage} ({dep})")
+
+            logger.info("Importing '%s' -> '%s'", dep, out)
+            out.transfer(
+                dep.def_path, jobs=kwargs.get("jobs"), odb=repo.odb.local
+            )
+        except DataSourceChanged as exc:
+            logger.warning(f"{exc}")
+            failed += 1
+            continue
+        if not any(
+            kwargs.get(kw, None)
+            for kw in ("all_branches", "all_tags", "all_commits", "revs")
+        ):
+            stage.dump()
+
+        downloaded += 1
+    return downloaded, failed

dvc/repo/gc.py

@@ -60,8 +60,8 @@ def gc(
     from contextlib import ExitStack
 
     from dvc.repo import Repo
-    from dvc_data.db import get_index
-    from dvc_data.gc import gc as ogc
+    from dvc_data.hashfile.db import get_index
+    from dvc_data.hashfile.gc import gc as ogc
 
     if not repos:
         repos = []

dvc/repo/imp_url.py

@@ -17,11 +17,16 @@ def imp_url(
     fname=None,
     erepo=None,
     frozen=True,
+    no_download=False,
     no_exec=False,
     remote=None,
     to_remote=False,
     desc=None,
+    type=None,  # pylint: disable=redefined-builtin
+    labels=None,
+    meta=None,
     jobs=None,
+    fs_config=None,
 ):
     from dvc.dvcfile import Dvcfile
     from dvc.stage import Stage, create_stage, restore_fields
@@ -31,9 +36,9 @@ def imp_url(
         self, out, always_local=to_remote and not out
     )
 
-    if to_remote and no_exec:
+    if to_remote and (no_exec or no_download):
         raise InvalidArgumentError(
-            "--no-exec can't be combined with --to-remote"
+            "--no-exec/--no-download cannot be combined with --to-remote"
         )
 
     if not to_remote and remote:
@@ -57,12 +62,12 @@ def imp_url(
         deps=[url],
         outs=[out],
         erepo=erepo,
+        fs_config=fs_config,
     )
     restore_fields(stage)
 
-    if desc:
-        stage.outs[0].desc = desc
-
+    out_obj = stage.outs[0]
+    out_obj.annot.update(desc=desc, type=type, labels=labels, meta=meta)
     dvcfile = Dvcfile(self, stage.path)
     dvcfile.remove()
 
@@ -80,7 +85,7 @@ def imp_url(
         stage.save_deps()
         stage.md5 = stage.compute_md5()
     else:
-        stage.run(jobs=jobs)
+        stage.run(jobs=jobs, no_download=no_download)
 
     stage.frozen = frozen
 

dvc/repo/index.py

@@ -183,15 +183,7 @@ class Index:
             if not out.use_cache:
                 continue
 
-            if out.is_in_repo:
-                workspace = "repo"
-                key = self.repo.fs.path.relparts(
-                    out.fs_path, self.repo.root_dir
-                )
-            else:
-                workspace = out.fs.protocol
-                no_drive = out.fs.path.flavour.splitdrive(out.fs_path)[1]
-                key = out.fs.path.parts(no_drive)[1:]
+            workspace, key = out.index_key
 
             try:
                 remote = self.repo.cloud.get_remote_odb(out.remote)
@@ -202,9 +194,10 @@ class Index:
 
             data_index[key] = DataIndexEntry(
                 meta=out.meta,
-                obj=out.obj,
+                obj=out.get_obj() if out.files else out.obj,
                 hash_info=out.hash_info,
                 odb=out.odb,
+                cache=out.odb,
                 remote=remote,
             )
 
@@ -246,6 +239,28 @@ class Index:
                 used[odb].update(objs)
         return used
 
+    def partial_imports(
+        self,
+        targets: "TargetType" = None,
+        recursive: bool = False,
+    ) -> List["Stage"]:
+        from itertools import chain
+
+        from dvc.utils.collections import ensure_list
+
+        collect_targets: Sequence[Optional[str]] = (None,)
+        if targets:
+            collect_targets = ensure_list(targets)
+
+        pairs = chain.from_iterable(
+            self.stage_collector.collect_granular(
+                target, recursive=recursive, with_deps=True
+            )
+            for target in collect_targets
+        )
+
+        return [stage for stage, _ in pairs if stage.is_partial_import]
+
     # Following methods help us treat the collection as a set-like structure
     # and provides faux-immutability.
     # These methods do not preserve stages order.

dvc/repo/ls.py

@@ -1,5 +1,4 @@
 import os
-from itertools import chain
 from typing import TYPE_CHECKING, Optional
 
 from dvc.exceptions import PathMissingError
@@ -66,15 +65,16 @@ def _ls(
 
     infos = {}
     for root, dirs, files in fs.walk(
-        fs_path, dvcfiles=True, dvc_only=dvc_only
+        fs_path, dvcfiles=True, dvc_only=dvc_only, detail=True
     ):
-        entries = chain(files, dirs) if not recursive else files
+        if not recursive:
+            files.update(dirs)
 
-        for entry in entries:
-            entry_fs_path = fs.path.join(root, entry)
+        for name, entry in files.items():
+            entry_fs_path = fs.path.join(root, name)
             relparts = fs.path.relparts(entry_fs_path, fs_path)
             name = os.path.join(*relparts)
-            infos[name] = fs.info(entry_fs_path)
+            infos[name] = entry
 
         if not recursive:
             break
@@ -85,11 +85,10 @@ def _ls(
     ret = {}
     for name, info in infos.items():
         dvc_info = info.get("dvc_info", {})
-        if fs.isdvc(info["name"], recursive=True) or not dvc_only:
-            ret[name] = {
-                "isout": dvc_info.get("isout", False),
-                "isdir": info["type"] == "directory",
-                "isexec": info.get("isexec", False),
-            }
+        ret[name] = {
+            "isout": dvc_info.get("isout", False),
+            "isdir": info["type"] == "directory",
+            "isexec": info.get("isexec", False),
+        }
 
     return ret

dvc/repo/metrics/show.py

@@ -85,12 +85,14 @@ def _read_metrics(repo, metrics, rev, onerror=None):
 
     res = {}
     for metric in metrics:
+        rel_metric_path = os.path.join(relpath, *fs.path.parts(metric))
         if not fs.isfile(metric):
-            continue
+            if fs.isfile(rel_metric_path):
+                metric = rel_metric_path
+            else:
+                continue
 
-        res[os.path.join(relpath, *fs.path.parts(metric))] = _read_metric(
-            metric, fs, rev, onerror=onerror
-        )
+        res[rel_metric_path] = _read_metric(metric, fs, rev, onerror=onerror)
 
     return res
 

dvc/repo/params/show.py

@@ -56,7 +56,9 @@ def _collect_configs(
             default_params
         ):
             fs_paths.append(default_params)
-
+    if targets and (deps or stages) and not params:
+        # A target has been provided but it is not used in the stages
+        fs_paths = []
     return params, fs_paths
 
 

dvc/repo/plots/__init__.py

@@ -386,30 +386,56 @@ def _collect_output_plots(
     return result
 
 
-def _adjust_definitions_to_cwd(fs, config_relpath, plots_definitions):
-    result = defaultdict(dict)
+def _id_is_path(plot_props=None):
+    if not plot_props:
+        return True
 
-    config_dirname = fs.path.dirname(config_relpath)
+    y_def = plot_props.get("y")
+    return not isinstance(y_def, dict)
 
-    for plot_id, plot_def in plots_definitions.items():
 
-        y_def = plot_def.get("y", None) if plot_def else None
-        if y_def is None or not isinstance(y_def, dict):
-            # plot_id is filename
-            new_plot_id = _normpath(fs.path.join(config_dirname, plot_id))
-            result[new_plot_id] = plot_def or {}
+def _adjust_sources(fs, plot_props, config_dir):
+    new_plot_props = deepcopy(plot_props)
+    old_y = new_plot_props.pop("y", {})
+    new_y = {}
+    for filepath, val in old_y.items():
+        new_y[_normpath(fs.path.join(config_dir, filepath))] = val
+    new_plot_props["y"] = new_y
+    return new_plot_props
+
+
+def _resolve_definitions(
+    fs, targets, props, config_path, definitions, onerror=None
+):
+    config_dir = fs.path.dirname(config_path)
+    result: Dict[str, Dict] = {}
+    for plot_id, plot_props in definitions.items():
+        if plot_props is None:
+            plot_props = {}
+        if _id_is_path(plot_props):
+            data_path = _normpath(fs.path.join(config_dir, plot_id))
+            if _matches(targets, config_path, plot_id):
+                unpacked = unpack_if_dir(
+                    fs,
+                    data_path,
+                    props={**plot_props, **props},
+                    onerror=onerror,
+                )
+                dpath.util.merge(
+                    result,
+                    unpacked,
+                )
         else:
-            new_plot_def = deepcopy(plot_def)
-            old_y = new_plot_def.pop("y")
-            new_y = {}
-            for filepath, val in old_y.items():
-                new_y[_normpath(fs.path.join(config_dirname, filepath))] = val
-            new_plot_def["y"] = new_y
-            result[plot_id] = new_plot_def
-    return dict(result)
+            if _matches(targets, config_path, plot_id):
+                adjusted_props = _adjust_sources(fs, plot_props, config_dir)
+                dpath.util.merge(
+                    result, {"data": {plot_id: {**adjusted_props, **props}}}
+                )
+
+    return result
 
 
-def _collect_pipeline_files(repo, targets: List[str], props):
+def _collect_pipeline_files(repo, targets: List[str], props, onerror=None):
     from dvc.dvcfile import PipelineFile
 
     result: Dict[str, Dict] = {}
@@ -417,21 +443,19 @@ def _collect_pipeline_files(repo, targets: List[str], props):
     for dvcfile in dvcfiles:
         if isinstance(dvcfile, PipelineFile):
             dvcfile_path = _relpath(repo.dvcfs, dvcfile.path)
-            dvcfile_defs = _adjust_definitions_to_cwd(
-                repo.fs, dvcfile_path, dvcfile.load().get("plots", {})
+            dvcfile_defs = dvcfile.load().get("plots", {})
+            resolved = _resolve_definitions(
+                repo.dvcfs,
+                targets,
+                props,
+                dvcfile_path,
+                dvcfile_defs,
+                onerror=onerror,
+            )
+            dpath.util.merge(
+                result,
+                {dvcfile_path: resolved},
             )
-            for plot_id, plot_props in dvcfile_defs.items():
-                if plot_props is None:
-                    plot_props = {}
-                if _matches(targets, dvcfile_path, plot_id):
-                    dpath.util.merge(
-                        result,
-                        {
-                            dvcfile_path: {
-                                "data": {plot_id: {**plot_props, **props}}
-                            }
-                        },
-                    )
     return result
 
 
@@ -453,7 +477,10 @@ def _collect_definitions(
     fs = DvcFileSystem(repo=repo)
 
     if not config_files:
-        dpath.util.merge(result, _collect_pipeline_files(repo, targets, props))
+        dpath.util.merge(
+            result,
+            _collect_pipeline_files(repo, targets, props, onerror=onerror),
+        )
 
     if targets or (not targets and not config_files):
         dpath.util.merge(
@@ -463,12 +490,20 @@ def _collect_definitions(
 
     if config_files:
         for path in config_files:
-            definitions = parse(fs, path)
-            definitions = _adjust_definitions_to_cwd(
-                repo.fs, path, definitions
-            )
+            definitions = parse(fs, path).get("data", {})
             if definitions:
-                dpath.util.merge(result, {path: definitions})
+                resolved = _resolve_definitions(
+                    repo.dvcfs,
+                    targets,
+                    props,
+                    path,
+                    definitions,
+                    onerror=onerror,
+                )
+                dpath.util.merge(
+                    result,
+                    {path: resolved},
+                )
 
     for target in targets:
         if not result or fs.exists(target):

dvc/repo/plots/diff.py

@@ -12,6 +12,8 @@ def _revisions(repo, revs, experiment):
 
 
 def diff(repo, *args, revs=None, experiment=False, **kwargs):
+    if repo.scm.no_commits:
+        return {}
     return repo.plots.show(
         *args, revs=_revisions(repo, revs, experiment), **kwargs
     )

dvc/repo/push.py

@@ -9,6 +9,52 @@ if TYPE_CHECKING:
     from dvc_objects.db.base import ObjectDB
 
 
+def _push_worktree(repo, remote):
+    from dvc_data.hashfile.tree import tree_from_index
+    from dvc_data.index import build, checkout, collect
+
+    index = repo.index.data["repo"]
+    checkout(index, remote.path, remote.fs, force=True)
+    build(index, remote.path, remote.fs)
+    if any(out.isdir() for out in repo.index.outs):
+        collect(index, remote.path, remote.fs)
+
+    for stage in repo.index.stages:
+        for out in stage.outs:
+            if not out.use_cache:
+                continue
+
+            if not out.is_in_repo:
+                continue
+
+            workspace, key = out.index_key
+            index = repo.index.data[workspace]
+            entry = index[key]
+            if out.isdir():
+                old_tree = out.get_obj()
+                entry.hash_info = old_tree.hash_info
+                entry.meta = out.meta
+                for subkey, entry in index.iteritems(key):
+                    if entry.meta.isdir:
+                        continue
+                    fs_path = repo.fs.path.join(repo.root_dir, *subkey)
+                    _, hash_info = old_tree.get(
+                        repo.fs.path.relparts(fs_path, out.fs_path)
+                    )
+                    entry.hash_info = hash_info
+                tree_meta, new_tree = tree_from_index(index, key)
+                new_tree.digest()
+                out.obj = new_tree
+                out.hash_info = new_tree.hash_info
+                out.meta = tree_meta
+            else:
+                out.hash_info = entry.hash_info
+                out.meta = entry.meta
+        stage.dvcfile.dump(stage, with_files=True)
+
+    return len(index)
+
+
 @locked
 def push(
     self,
@@ -26,6 +72,10 @@ def push(
     odb: Optional["ObjectDB"] = None,
     include_imports=False,
 ):
+    _remote = self.cloud.get_remote(name=remote)
+    if _remote.worktree:
+        return _push_worktree(self, _remote)
+
     used_run_cache = (
         self.stage_cache.push(remote, odb=odb) if run_cache else []
     )

dvc/repo/remove.py

@@ -1,6 +1,13 @@
 import logging
 import typing
 
+from dvc.dvcfile import DVC_FILE_SUFFIX
+from dvc.stage.exceptions import (
+    StageFileDoesNotExistError,
+    StageFileIsNotDvcFileError,
+    StageNotFound,
+)
+
 from . import locked
 
 if typing.TYPE_CHECKING:
@@ -11,7 +18,16 @@ logger = logging.getLogger(__name__)
 
 @locked
 def remove(self: "Repo", target: str, outs: bool = False):
-    stages = self.stage.from_target(target)
+    try:
+        stages = self.stage.from_target(target, accept_group=False)
+    except (StageNotFound, StageFileDoesNotExistError) as e:
+        # If the user specified a tracked file as a target instead of a stage,
+        # e.g. `data.csv` instead of `data.csv.dvc`,
+        # give a more helpful error message.
+        if self.fs.exists(target + DVC_FILE_SUFFIX):
+            raise StageFileIsNotDvcFileError(target) from e
+        else:
+            raise
 
     for stage in stages:
         stage.remove(remove_outs=outs, force=outs)

dvc/repo/reproduce.py

@@ -102,7 +102,6 @@ def reproduce(
     from .graph import get_pipeline, get_pipelines
 
     glob = kwargs.pop("glob", False)
-    accept_group = not glob
 
     if isinstance(targets, str):
         targets = [targets]
@@ -137,7 +136,6 @@ def reproduce(
                 self.stage.collect(
                     target,
                     recursive=recursive,
-                    accept_group=accept_group,
                     glob=glob,
                 )
             )

dvc/repo/stage.py

@@ -82,7 +82,6 @@ def _collect_specific_target(
     target: str,
     with_deps: bool,
     recursive: bool,
-    accept_group: bool,
 ) -> Tuple[StageIter, "OptStr", "OptStr"]:
     from dvc.dvcfile import is_valid_filename
 
@@ -98,13 +97,11 @@ def _collect_specific_target(
         msg = "Checking if stage '%s' is in '%s'"
         logger.debug(msg, target, PIPELINE_FILE)
         if not (recursive and loader.fs.isdir(target)):
-            stages = _maybe_collect_from_dvc_yaml(
-                loader, target, with_deps, accept_group=accept_group
-            )
+            stages = _maybe_collect_from_dvc_yaml(loader, target, with_deps)
             if stages:
                 return stages, file, name
     elif not with_deps and is_valid_filename(file):
-        stages = loader.load_all(file, name, accept_group=accept_group)
+        stages = loader.load_all(file, name)
         return stages, file, name
     return [], file, name
 
@@ -209,7 +206,7 @@ class StageLoad:
         return stage
 
     def from_target(
-        self, target: str, accept_group: bool = False, glob: bool = False
+        self, target: str, accept_group: bool = True, glob: bool = False
     ) -> StageList:
         """
         Returns a list of stage from the provided target.
@@ -250,15 +247,12 @@ class StageLoad:
         self,
         stages: "StageLoader",
         name: str = None,
-        accept_group: bool = False,
+        accept_group: bool = True,
         glob: bool = False,
     ) -> Iterable[str]:
 
-        assert not (accept_group and glob)
-
         if not name:
             return stages.keys()
-
         if accept_group and stages.is_foreach_generated(name):
             return self._get_group_keys(stages, name)
         if glob:
@@ -269,7 +263,7 @@ class StageLoad:
         self,
         path: str = None,
         name: str = None,
-        accept_group: bool = False,
+        accept_group: bool = True,
         glob: bool = False,
     ) -> StageList:
         """Load a list of stages from a file.
@@ -338,7 +332,6 @@ class StageLoad:
         with_deps: bool = False,
         recursive: bool = False,
         graph: "DiGraph" = None,
-        accept_group: bool = False,
         glob: bool = False,
     ) -> StageIter:
         """Collect list of stages from the provided target.
@@ -347,15 +340,14 @@ class StageLoad:
             target: if not provided, all of the stages in the graph are
                 returned.
                 Target can be:
-                - a stage name in the `dvc.yaml` file.
+                - a foreach group name or a stage name in the `dvc.yaml` file.
+                - a generated stage name from a foreach group.
                 - a path to `dvc.yaml` or `.dvc` file.
                 - in case of a stage to a dvc.yaml file in a different
                   directory than current working directory, it can be a path
                   to dvc.yaml file, followed by a colon `:`, followed by stage
                   name (eg: `../dvc.yaml:build`).
                 - in case of `recursive`, it can be a path to a directory.
-                - in case of `accept_group`, it can be a group name of
-                    `foreach` generated stage.
                 - in case of `glob`, it can be a wildcard pattern to match
                   stages. Example: `build*` for stages in `dvc.yaml` file, or
                   `../dvc.yaml:build*` for stages in dvc.yaml in a different
@@ -367,8 +359,6 @@ class StageLoad:
             recursive: if true and if `target` is a directory, all of the
                 stages inside that directory is returned.
             graph: graph to use. Defaults to `repo.graph`.
-            accept_group: if true, all of the `foreach` generated stages of
-                the specified target is returned.
             glob: Use `target` as a pattern to match stages in a file.
         """
         if not target:
@@ -380,7 +370,7 @@ class StageLoad:
             path = self.fs.path.abspath(target)
             return collect_inside_path(path, graph or self.graph)
 
-        stages = self.from_target(target, accept_group=accept_group, glob=glob)
+        stages = self.from_target(target, glob=glob)
         if not with_deps:
             return stages
 
@@ -392,14 +382,14 @@ class StageLoad:
         with_deps: bool = False,
         recursive: bool = False,
         graph: "DiGraph" = None,
-        accept_group: bool = False,
     ) -> List[StageInfo]:
         """Collects a list of (stage, filter_info) from the given target.
 
         Priority is in the order of following in case of ambiguity:
         - .dvc file or .yaml file
         - dir if recursive and directory exists
-        - stage_name
+        - foreach_group_name or stage_name
+        - generated stage name from a foreach group
         - output file
 
         Args:
@@ -418,7 +408,7 @@ class StageLoad:
         target = as_posix(target)
 
         stages, file, _ = _collect_specific_target(
-            self, target, with_deps, recursive, accept_group
+            self, target, with_deps, recursive
         )
         if not stages:
             if not (recursive and self.fs.isdir(target)):
@@ -440,7 +430,6 @@ class StageLoad:
                     with_deps,
                     recursive,
                     graph,
-                    accept_group=accept_group,
                 )
             except StageFileDoesNotExistError as exc:
                 # collect() might try to use `target` as a stage name

dvc/repo/update.py

@@ -10,6 +10,7 @@ def update(
     rev=None,
     recursive=False,
     to_remote=False,
+    no_download=False,
     remote=None,
     jobs=None,
 ):
@@ -21,6 +22,11 @@ def update(
     if isinstance(targets, str):
         targets = [targets]
 
+    if to_remote and no_download:
+        raise InvalidArgumentError(
+            "--to-remote can't be used with --no-download"
+        )
+
     if not to_remote and remote:
         raise InvalidArgumentError(
             "--remote can't be used without --to-remote"
@@ -31,7 +37,13 @@ def update(
         stages.update(self.stage.collect(target, recursive=recursive))
 
     for stage in stages:
-        stage.update(rev, to_remote=to_remote, remote=remote, jobs=jobs)
+        stage.update(
+            rev,
+            to_remote=to_remote,
+            remote=remote,
+            no_download=no_download,
+            jobs=jobs,
+        )
         dvcfile = Dvcfile(self, stage.path)
         dvcfile.dump(stage)
 

dvc/schema.py

@@ -3,11 +3,11 @@ from collections.abc import Mapping
 from voluptuous import Any, Optional, Required, Schema
 
 from dvc import dependency, output
-from dvc.output import CHECKSUMS_SCHEMA, Output
+from dvc.annotations import ANNOTATION_SCHEMA
+from dvc.output import CHECKSUMS_SCHEMA, DIR_FILES_SCHEMA, META_SCHEMA, Output
 from dvc.parsing import DO_KWD, FOREACH_KWD, VARS_KWD
 from dvc.parsing.versions import SCHEMA_KWD, lockfile_version_schema
 from dvc.stage.params import StageParams
-from dvc_data.hashfile.meta import Meta
 
 STAGES = "stages"
 SINGLE_STAGE_SCHEMA = {
@@ -25,10 +25,9 @@ SINGLE_STAGE_SCHEMA = {
 
 DATA_SCHEMA = {
     **CHECKSUMS_SCHEMA,
+    **META_SCHEMA,
     Required("path"): str,
-    Meta.PARAM_SIZE: int,
-    Meta.PARAM_NFILES: int,
-    Meta.PARAM_ISEXEC: bool,
+    Output.PARAM_FILES: [DIR_FILES_SCHEMA],
 }
 LOCK_FILE_STAGE_SCHEMA = {
     Required(StageParams.PARAM_CMD): Any(str, list),
@@ -47,16 +46,15 @@ LOCKFILE_V2_SCHEMA = {
 
 OUT_PSTAGE_DETAILED_SCHEMA = {
     str: {
+        **ANNOTATION_SCHEMA,  # type: ignore
         Output.PARAM_CACHE: bool,
         Output.PARAM_PERSIST: bool,
         Output.PARAM_CHECKPOINT: bool,
-        Output.PARAM_DESC: str,
         Output.PARAM_REMOTE: str,
     }
 }
 
 PLOTS = "plots"
-PLOTS_SCHEMA = dict
 PLOT_PROPS = {
     Output.PARAM_PLOT_TEMPLATE: str,
     Output.PARAM_PLOT_X: str,
@@ -112,6 +110,15 @@ def either_or(primary, fallback, fallback_includes=None):
     return validator
 
 
+PLOT_DEFINITION = {
+    Output.PARAM_PLOT_X: str,
+    Output.PARAM_PLOT_Y: Any(str, [str], {str: Any(str, [str])}),
+    Output.PARAM_PLOT_X_LABEL: str,
+    Output.PARAM_PLOT_Y_LABEL: str,
+    Output.PARAM_PLOT_TITLE: str,
+    Output.PARAM_PLOT_TEMPLATE: str,
+}
+PLOTS_SCHEMA = {str: Any(PLOT_DEFINITION, None)}
 FOREACH_IN = {
     Required(FOREACH_KWD): Any(dict, list, str),
     Required(DO_KWD): STAGE_DEFINITION,

dvc/scm.py

@@ -160,7 +160,7 @@ def resolve_rev(scm: "Git", rev: str) -> str:
     except InternalRevError as exc:
         # `scm` will only resolve git branch and tag names,
         # if rev is not a sha it may be an abbreviated experiment name
-        if not rev.startswith("refs/"):
+        if not (rev == "HEAD" or rev.startswith("refs/")):
             from dvc.repo.experiments.utils import (
                 AmbiguousExpRefInfo,
                 resolve_name,

dvc/stage/__init__.py

@@ -78,6 +78,7 @@ def create_stage(cls, repo, path, external=False, **kwargs):
 
     wdir = os.path.abspath(kwargs.get("wdir", None) or os.curdir)
     path = os.path.abspath(path)
+
     check_dvcfile_path(repo, path)
     check_stage_path(repo, wdir, is_wdir=kwargs.get("wdir"))
     check_stage_path(repo, os.path.dirname(path))
@@ -87,7 +88,7 @@ def create_stage(cls, repo, path, external=False, **kwargs):
     if not external:
         check_no_externals(stage)
     fill_stage_dependencies(
-        stage, **project(kwargs, ["deps", "erepo", "params"])
+        stage, **project(kwargs, ["deps", "erepo", "params", "fs_config"])
     )
     check_circular_dependency(stage)
     check_duplicated_arguments(stage)
@@ -109,16 +110,13 @@ def restore_fields(stage):
     # will be used to restore comments later
     # noqa, pylint: disable=protected-access
     stage._stage_text = old._stage_text
-
     stage.meta = old.meta
     stage.desc = old.desc
 
-    old_fields = {out.def_path: (out.desc, out.remote) for out in old.outs}
-
+    old_fields = {out.def_path: (out.annot, out.remote) for out in old.outs}
     for out in stage.outs:
-        out_fields = old_fields.get(out.def_path, None)
-        if out_fields:
-            out.desc, out.remote = out_fields
+        if out_fields := old_fields.get(out.def_path, None):
+            out.annot, out.remote = out_fields
 
 
 class Stage(params.StageParams):
@@ -242,6 +240,14 @@ class Stage(params.StageParams):
         """Whether the DVC file was created with `dvc import`."""
         return not self.cmd and len(self.deps) == 1 and len(self.outs) == 1
 
+    @property
+    def is_partial_import(self) -> bool:
+        """
+        Whether the DVC file was created using `dvc import --no-download`
+        or `dvc import-url --no-download`.
+        """
+        return self.is_import and (not self.outs[0].hash_info)
+
     @property
     def is_repo_import(self):
         if not self.is_import:
@@ -433,18 +439,30 @@ class Stage(params.StageParams):
 
         return self
 
-    def update(self, rev=None, to_remote=False, remote=None, jobs=None):
+    def update(
+        self,
+        rev=None,
+        to_remote=False,
+        remote=None,
+        no_download=None,
+        jobs=None,
+    ):
         if not (self.is_repo_import or self.is_import):
             raise StageUpdateError(self.relpath)
         update_import(
-            self, rev=rev, to_remote=to_remote, remote=remote, jobs=jobs
+            self,
+            rev=rev,
+            to_remote=to_remote,
+            remote=remote,
+            no_download=no_download,
+            jobs=jobs,
         )
 
     def reload(self):
         return self.dvcfile.stage
 
-    def dumpd(self):
-        return get_dump(self)
+    def dumpd(self, **kwargs):
+        return get_dump(self, **kwargs)
 
     def compute_md5(self):
         # `dvc add`ed files don't need stage md5
@@ -457,6 +475,7 @@ class Stage(params.StageParams):
 
     def save(self, allow_missing=False):
         self.save_deps(allow_missing=allow_missing)
+
         self.save_outs(allow_missing=allow_missing)
         self.md5 = self.compute_md5()
 
@@ -525,14 +544,16 @@ class Stage(params.StageParams):
         no_commit=False,
         force=False,
         allow_missing=False,
+        no_download=False,
         **kwargs,
     ):
         if (self.cmd or self.is_import) and not self.frozen and not dry:
             self.remove_outs(ignore_remove=False, force=False)
 
-        if not self.frozen and self.is_import:
-            jobs = kwargs.get("jobs", None)
-            self._sync_import(dry, force, jobs)
+        if (not self.frozen and self.is_import) 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)
         else:
@@ -544,7 +565,7 @@ class Stage(params.StageParams):
                 self._check_missing_outputs()
 
         if not dry:
-            if kwargs.get("checkpoint_func", None):
+            if kwargs.get("checkpoint_func", None) or no_download:
                 allow_missing = True
             self.save(allow_missing=allow_missing)
             if not no_commit:
@@ -555,8 +576,8 @@ class Stage(params.StageParams):
         return run_stage(self, dry, force, **kwargs)
 
     @rwlocked(read=["deps"], write=["outs"])
-    def _sync_import(self, dry, force, jobs):
-        sync_import(self, dry, force, jobs)
+    def _sync_import(self, dry, force, jobs, no_download):
+        sync_import(self, dry, force, jobs, no_download)
 
     @rwlocked(read=["outs"])
     def _check_missing_outputs(self):
@@ -571,6 +592,8 @@ class Stage(params.StageParams):
     @rwlocked(write=["outs"])
     def checkout(self, allow_missing=False, **kwargs):
         stats = defaultdict(list)
+        if self.is_partial_import:
+            return {}
         for out in self.filter_outs(kwargs.get("filter_info")):
             key, outs = self._checkout(
                 out,
@@ -658,6 +681,9 @@ class Stage(params.StageParams):
         self, *args, **kwargs
     ) -> Dict[Optional["ObjectDB"], Set["HashInfo"]]:
         """Return set of object IDs used by this stage."""
+        if self.is_partial_import:
+            return {}
+
         used_objs = defaultdict(set)
         for out in self.filter_outs(kwargs.get("filter_info")):
             for odb, objs in out.get_used_objs(*args, **kwargs).items():

dvc/stage/cache.py

@@ -9,7 +9,7 @@ from funcy import cached_property, first
 from dvc import fs
 from dvc.exceptions import DvcException
 from dvc.utils import dict_sha256, relpath
-from dvc_data.transfer import _log_exceptions
+from dvc_data.hashfile.transfer import _log_exceptions
 
 if TYPE_CHECKING:
     from dvc_objects.db import ObjectDB

dvc/stage/exceptions.py

@@ -31,7 +31,7 @@ class StageFileIsNotDvcFileError(DvcException):
 
         sname = fname + DVC_FILE_SUFFIX
         if is_dvc_file(sname):
-            msg += f" Do you mean '{sname}'?"
+            msg += f". Do you mean '{sname}'?"
 
         super().__init__(msg)
 
@@ -80,6 +80,11 @@ class MissingDataSource(DvcException):
         super().__init__(msg)
 
 
+class DataSourceChanged(DvcException):
+    def __init__(self, path: str):
+        super().__init__(f"data source changed: {path}")
+
+
 class StageNotFound(DvcException, KeyError):
     def __init__(self, file, name):
         self.file = file.relpath

dvc/stage/imports.py

@@ -17,20 +17,36 @@ def _update_import_on_remote(stage, remote, jobs):
     stage.outs[0].transfer(url, odb=odb, jobs=jobs, update=True)
 
 
-def update_import(stage, rev=None, to_remote=False, remote=None, jobs=None):
+def update_import(
+    stage, rev=None, to_remote=False, remote=None, no_download=None, jobs=None
+):
     stage.deps[0].update(rev=rev)
+    outs = stage.outs
+    deps = stage.deps
+
     frozen = stage.frozen
     stage.frozen = False
+
+    if stage.outs:
+        stage.outs[0].clear()
     try:
         if to_remote:
             _update_import_on_remote(stage, remote, jobs)
         else:
-            stage.reproduce(jobs=jobs)
+            stage.reproduce(no_download=no_download, jobs=jobs)
     finally:
+        if deps == stage.deps:
+            stage.outs = outs
         stage.frozen = frozen
 
 
-def sync_import(stage, dry=False, force=False, jobs=None):
+def sync_import(
+    stage,
+    dry=False,
+    force=False,
+    jobs=None,
+    no_download=False,
+):
     """Synchronize import's outs to the workspace."""
     logger.info("Importing '%s' -> '%s'", stage.deps[0], stage.outs[0])
     if dry:
@@ -40,4 +56,8 @@ def sync_import(stage, dry=False, force=False, jobs=None):
         stage.outs[0].checkout()
     else:
         stage.save_deps()
-        stage.deps[0].download(stage.outs[0], jobs=jobs)
+        if not no_download:
+            stage.deps[0].download(
+                stage.outs[0],
+                jobs=jobs,
+            )

dvc/stage/loader.py

@@ -69,7 +69,10 @@ class StageLoader(Mapping):
             info = info.copy()
             info.pop("path", None)
             item.meta = Meta.from_dict(info)
-            item.hash_info = HashInfo.from_dict(info)
+            hash_value = getattr(item.meta, item.hash_name, None)
+            item.hash_info = HashInfo(item.hash_name, hash_value)
+            if item.hash_info and item.hash_info.isdir:
+                item.meta.isdir = True
 
     @classmethod
     def load_stage(cls, dvcfile, name, stage_data, lock_data=None):

dvc/stage/serialize.py

@@ -6,7 +6,7 @@ from typing import TYPE_CHECKING, List, no_type_check
 from funcy import post_processing
 
 from dvc.dependency import ParamsDependency
-from dvc.output import Output
+from dvc.output import Annotation, Output
 from dvc.utils.collections import apply_diff
 from dvc.utils.serialize import parse_yaml_for_update
 
@@ -27,7 +27,7 @@ PARAM_METRIC = Output.PARAM_METRIC
 PARAM_PLOT = Output.PARAM_PLOT
 PARAM_PERSIST = Output.PARAM_PERSIST
 PARAM_CHECKPOINT = Output.PARAM_CHECKPOINT
-PARAM_DESC = Output.PARAM_DESC
+PARAM_DESC = Annotation.PARAM_DESC
 PARAM_REMOTE = Output.PARAM_REMOTE
 
 DEFAULT_PARAMS_FILE = ParamsDependency.DEFAULT_PARAMS_FILE
@@ -38,8 +38,9 @@ sort_by_path = partial(sorted, key=attrgetter("def_path"))
 
 @post_processing(OrderedDict)
 def _get_flags(out):
-    if out.desc:
-        yield PARAM_DESC, out.desc
+    annot = out.annot.to_dict()
+    yield from annot.items()
+
     if not out.use_cache:
         yield PARAM_CACHE, False
     if out.checkpoint:
@@ -155,10 +156,12 @@ def to_single_stage_lockfile(stage: "Stage") -> dict:
     assert stage.cmd
 
     def _dumpd(item):
+        meta_d = item.meta.to_dict()
+        meta_d.pop("isdir", None)
         ret = [
             (item.PARAM_PATH, item.def_path),
             *item.hash_info.to_dict().items(),
-            *item.meta.to_dict().items(),
+            *meta_d.items(),
         ]
 
         return OrderedDict(ret)
@@ -185,8 +188,8 @@ def to_lockfile(stage: "PipelineStage") -> dict:
     return {stage.name: to_single_stage_lockfile(stage)}
 
 
-def to_single_stage_file(stage: "Stage"):
-    state = stage.dumpd()
+def to_single_stage_file(stage: "Stage", **kwargs):
+    state = stage.dumpd(**kwargs)
 
     # When we load a stage we parse yaml with a fast parser, which strips
     # off all the comments and formatting. To retain those on update we do

dvc/stage/utils.py

@@ -4,6 +4,7 @@ from typing import TYPE_CHECKING, Union
 
 from funcy import concat, first, lsplit, rpartial, without
 
+from dvc.annotations import ANNOTATION_FIELDS
 from dvc.exceptions import InvalidArgumentError
 from dvc_data.hashfile.meta import Meta
 
@@ -107,12 +108,16 @@ def _load_live_output(
     return outs
 
 
-def fill_stage_dependencies(stage, deps=None, erepo=None, params=None):
+def fill_stage_dependencies(
+    stage, deps=None, erepo=None, params=None, fs_config=None
+):
     from dvc.dependency import loads_from, loads_params
 
     assert not stage.deps
     stage.deps = []
-    stage.deps += loads_from(stage, deps or [], erepo=erepo)
+    stage.deps += loads_from(
+        stage, deps or [], erepo=erepo, fs_config=fs_config
+    )
     stage.deps += loads_params(stage, params or [])
 
 
@@ -194,9 +199,9 @@ def compute_md5(stage):
     return dict_md5(
         d,
         exclude=[
+            *ANNOTATION_FIELDS,
             stage.PARAM_LOCKED,  # backward compatibility
             stage.PARAM_FROZEN,
-            Output.PARAM_DESC,
             Output.PARAM_METRIC,
             Output.PARAM_PERSIST,
             Output.PARAM_CHECKPOINT,
@@ -221,7 +226,7 @@ def resolve_paths(fs, path, wdir=None):
     return path, wdir
 
 
-def get_dump(stage):
+def get_dump(stage, **kwargs):
     return {
         key: value
         for key, value in {
@@ -230,8 +235,8 @@ def get_dump(stage):
             stage.PARAM_CMD: stage.cmd,
             stage.PARAM_WDIR: resolve_wdir(stage.wdir, stage.path),
             stage.PARAM_FROZEN: stage.frozen,
-            stage.PARAM_DEPS: [d.dumpd() for d in stage.deps],
-            stage.PARAM_OUTS: [o.dumpd() for o in stage.outs],
+            stage.PARAM_DEPS: [d.dumpd(**kwargs) for d in stage.deps],
+            stage.PARAM_OUTS: [o.dumpd(**kwargs) for o in stage.outs],
             stage.PARAM_ALWAYS_CHANGED: stage.always_changed,
             stage.PARAM_META: stage.meta,
         }.items()

dvc/testing/test_api.py

@@ -1,4 +1,7 @@
+import pytest
+
 from dvc import api
+from dvc.api import DvcFileSystem
 from dvc.utils.fs import remove
 
 
@@ -18,3 +21,103 @@ class TestAPI:
 
         with api.open("foo") as fd:
             assert fd.read() == "foo-text"
+
+    @pytest.mark.parametrize(
+        "clear_cache", [True, False], ids=["cache", "no_cache"]
+    )
+    @pytest.mark.parametrize(
+        "fs_kwargs",
+        [
+            {},
+            {"url": "{path}"},
+            {"url": "{path}", "rev": "master"},
+            {"url": "file://{posixpath}"},
+            {"url": "file://{posixpath}", "rev": "master"},
+        ],
+        ids=["current", "local", "local_rev", "git", "git_rev"],
+    )
+    def test_filesystem(
+        self,
+        M,
+        tmp_dir,
+        make_tmp_dir,
+        scm,
+        dvc,
+        remote,
+        fs_kwargs,
+        clear_cache,
+    ):
+        tmp_dir.scm_gen({"scripts": {"script1": "script1"}}, commit="scripts")
+        tmp_dir.dvc_gen({"data": {"foo": "foo", "bar": "bar"}}, commit="data")
+        dvc.push()
+
+        if clear_cache:
+            remove(dvc.odb.repo.path)
+
+        if url := fs_kwargs.get("url"):
+            fs_kwargs["url"] = url.format(
+                path=tmp_dir, posixpath=tmp_dir.as_posix()
+            )
+
+        fs = DvcFileSystem(**fs_kwargs)
+
+        assert fs.ls("/", detail=False) == M.unordered(
+            "/.gitignore", "/scripts", "/data"
+        )
+        assert fs.ls("scripts", detail=False) == ["scripts/script1"]
+        assert fs.ls("data", detail=False) == M.unordered(
+            "data/foo", "data/bar"
+        )
+
+        data_info = M.dict(
+            name="/data",
+            type="directory",
+            dvc_info=M.dict(isdvc=True, isout=True),
+        )
+        scripts_info = M.dict(name="/scripts", type="directory", isexec=False)
+
+        assert sorted(fs.ls("/"), key=lambda i: i["name"]) == [
+            M.dict(name="/.gitignore", type="file", isexec=False),
+            data_info,
+            scripts_info,
+        ]
+
+        with pytest.raises(FileNotFoundError):
+            fs.info("/not-existing-path")
+
+        assert fs.info("/") == M.dict(name="/", isexec=False, type="directory")
+        assert fs.info("/data") == data_info
+        assert fs.info("/scripts") == scripts_info
+        assert fs.info("/data/foo") == M.dict(name="/data/foo", type="file")
+        assert fs.info("/scripts/script1") == M.dict(
+            name="/scripts/script1", type="file"
+        )
+
+        assert not fs.isdvc("/")
+        assert fs.isdvc("/data")
+        assert fs.isdvc("/data/foo")
+        assert not fs.isdvc("/scripts")
+        assert not fs.isdvc("/scripts/script1")
+
+        with pytest.raises((IsADirectoryError, PermissionError)):
+            fs.open("data")
+        with pytest.raises((IsADirectoryError, PermissionError)):
+            fs.open("scripts")
+        with fs.open("/data/foo") as fobj:
+            assert fobj.read() == b"foo"
+        with fs.open("/scripts/script1") as fobj:
+            assert fobj.read() == b"script1"
+
+        tmp = make_tmp_dir("temp-download")
+        fs.get_file("data/foo", tmp / "foo")
+        assert (tmp / "foo").read_text() == "foo"
+
+        fs.get_file("scripts/script1", tmp / "script1")
+        assert (tmp / "script1").read_text() == "script1"
+
+        fs.get("/", (tmp / "all").fs_path, recursive=True)
+        assert (tmp / "all").read_text() == {
+            ".gitignore": "/data\n",
+            "data": {"bar": "bar", "foo": "foo"},
+            "scripts": {"script1": "script1"},
+        }

dvc/utils/hydra.py

@@ -1,21 +1,50 @@
+from collections import defaultdict
 from pathlib import Path
 from typing import TYPE_CHECKING, List
 
+from hydra import compose, initialize_config_dir
 from hydra._internal.config_loader_impl import ConfigLoaderImpl
+from hydra._internal.core_plugins.basic_sweeper import BasicSweeper
 from hydra.core.override_parser.overrides_parser import OverridesParser
+from hydra.core.override_parser.types import ValueType
 from hydra.errors import ConfigCompositionException, OverrideParseException
-from hydra.types import RunMode
 from omegaconf import OmegaConf
 
 from dvc.exceptions import InvalidArgumentError
 
 from .collections import merge_dicts, remove_missing_keys, to_omegaconf
-from .serialize import MODIFIERS
+from .serialize import DUMPERS, MODIFIERS
 
 if TYPE_CHECKING:
     from dvc.types import StrPath
 
 
+def compose_and_dump(
+    output_file: "StrPath",
+    config_dir: str,
+    config_name: str,
+    overrides: List[str],
+) -> None:
+    """Compose Hydra config and dumpt it to `output_file`.
+
+    Args:
+        output_file: File where the composed config will be dumped.
+        config_dir: Folder containing the Hydra config files.
+            Must be absolute file system path.
+        config_name: Name of the config file containing defaults,
+            without the .yaml extension.
+        overrides: List of `Hydra Override`_ patterns.
+
+    .. _Hydra Override:
+        https://hydra.cc/docs/advanced/override_grammar/basic/
+    """
+    with initialize_config_dir(config_dir, version_base=None):
+        cfg = compose(config_name=config_name, overrides=overrides)
+
+    dumper = DUMPERS[Path(output_file).suffix.lower()]
+    dumper(output_file, OmegaConf.to_object(cfg))
+
+
 def apply_overrides(path: "StrPath", overrides: List[str]) -> None:
     """Update `path` params with the provided `Hydra Override`_ patterns.
 
@@ -32,11 +61,7 @@ def apply_overrides(path: "StrPath", overrides: List[str]) -> None:
     modify_data = MODIFIERS[suffix]
     with modify_data(path) as original_data:
         try:
-            parser = OverridesParser.create()
-            parsed = parser.parse_overrides(overrides=overrides)
-            ConfigLoaderImpl.validate_sweep_overrides_legal(
-                parsed, run_mode=RunMode.RUN, from_shell=True
-            )
+            parsed = to_hydra_overrides(overrides)
 
             new_data = OmegaConf.create(
                 to_omegaconf(original_data),
@@ -51,3 +76,34 @@ def apply_overrides(path: "StrPath", overrides: List[str]) -> None:
 
         merge_dicts(original_data, new_data)
         remove_missing_keys(original_data, new_data)
+
+
+def to_hydra_overrides(path_overrides):
+    parser = OverridesParser.create()
+    return parser.parse_overrides(overrides=path_overrides)
+
+
+def get_hydra_sweeps(path_overrides):
+    merged_overrides = []
+    for path, overrides in path_overrides.items():
+        merged_overrides.extend(
+            [f"{path}___{override}" for override in overrides]
+        )
+
+    hydra_overrides = to_hydra_overrides(merged_overrides)
+    for hydra_override in hydra_overrides:
+        if hydra_override.value_type == ValueType.GLOB_CHOICE_SWEEP:
+            raise InvalidArgumentError(
+                f"Glob override '{hydra_override.input_line}' "
+                "is not supported."
+            )
+
+    splits = BasicSweeper.split_arguments(hydra_overrides, None)[0]
+    sweeps = []
+    for split in splits:
+        sweep_overrides = defaultdict(list)
+        for merged_override in split:
+            path, override = merged_override.split("___")
+            sweep_overrides[path].append(override)
+        sweeps.append(dict(sweep_overrides))
+    return sweeps

pyproject.toml

@@ -30,11 +30,11 @@ line_length = 79
 
 [tool.pytest.ini_options]
 log_level = "debug"
-addopts = "-ra"
+addopts = "-ra -p no:hydra_pytest"
 markers = [
     "needs_internet: Might need network access for the tests",
-"vscode: Tests verifying contract between DVC and VSCode plugin",
-"studio: Tests verifying contract between DVC and Studio"
+    "vscode: Tests verifying contract between DVC and VSCode plugin",
+    "studio: Tests verifying contract between DVC and Studio",
 ]
 xfail_strict = true
 

scripts/build-requirements.txt

@@ -1,3 +1,3 @@
 # NOTE: using strict version due to
 # https://github.com/iterative/dvc/issues/7949
-PyInstaller==5.2
+PyInstaller==5.0

scripts/build.py

@@ -1,5 +1,5 @@
 import argparse
-import importlib
+import importlib.util
 import pathlib
 import textwrap
 from subprocess import STDOUT, check_call

scripts/pyinstaller/build.py

@@ -1,6 +1,6 @@
 import os
 import pathlib
-from subprocess import STDOUT, check_call
+from subprocess import STDOUT, check_call, check_output
 
 path = pathlib.Path(__file__).parent.absolute()
 hooks = path / "hooks"
@@ -21,10 +21,27 @@ check_call(
     stderr=STDOUT,
 )
 
-check_call(
+out = check_output(
     [
         path / "dist" / "dvc" / "dvc",
         "doctor",
     ],
     stderr=STDOUT,
-)
+).decode()
+
+remotes = [
+    "s3",
+    "oss",
+    "gdrive",
+    "gs",
+    "hdfs",
+    "http",
+    "webhdfs",
+    "azure",
+    "ssh",
+    "webdav",
+]
+
+print(out)
+for remote in remotes:
+    assert f"\t{remote}" in out, f"Missing support for {remote}"

scripts/pyinstaller/hooks/hook-dvc.py

@@ -16,5 +16,15 @@ datas += copy_metadata("webdav4")
 datas += copy_metadata("aiohttp")
 datas += copy_metadata("aiohttp_retry")
 
-# https://github.com/pypa/setuptools/issues/1963
-hiddenimports = ["pkg_resources.py2_warn"]
+hiddenimports = [
+    "dvc_azure",
+    "dvc_gdrive",
+    "dvc_gs",
+    "dvc_hdfs",
+    "dvc_oss",
+    "dvc_s3",
+    "dvc_webdav",
+    "dvc_webhdfs",
+    # https://github.com/pypa/setuptools/issues/1963
+    "pkg_resources.py2_warn",
+]

scripts/pyinstaller/hooks/hook-fsspec.py

@@ -0,0 +1 @@
+hiddenimports = ["fsspec.implementations.memory"]

scripts/pyinstaller/hooks/hook-googleapiclient.model.py

@@ -1,8 +0,0 @@
-"""Temporary fix for https://github.com/iterative/dvc/issues/5618."""
-from PyInstaller.utils.hooks import (  # pylint:disable=import-error
-    collect_data_files,
-    copy_metadata,
-)
-
-datas = collect_data_files("googleapiclient.discovery")
-datas += copy_metadata("google_api_python_client")

setup.cfg

@@ -62,14 +62,12 @@ install_requires =
     rich>=10.13.0
     pyparsing>=2.4.7
     typing-extensions>=3.7.4
-    fsspec[http]>=2021.10.1
-    aiohttp-retry>=2.4.5
-    scmrepo==0.0.25
-    dvc-render==0.0.9
+    scmrepo==0.1.1
+    dvc-render==0.0.10
     dvc-task==0.1.2
     dvclive>=0.10.0
-    dvc-data==0.1.17
-    dvc-http==0.0.2
+    dvc-data==0.8.0
+    dvc-http==2.19.1
     hydra-core>=1.1.0
 
 [options.extras_require]
@@ -82,46 +80,34 @@ all =
     %(s3)s
     %(ssh)s
     %(webdav)s
+    %(webhdfs)s
 dev =
     %(all)s
     %(tests)s
     dvc-data[cli]
 # Extra dependencies for remote integrations
-azure =
-    adlfs>=2021.10.0
-    azure-identity>=1.4.0
-    knack
-gdrive = pydrive2[fsspec]>=1.10.1
-gs = gcsfs>=2021.11.1
+azure = dvc-azure==2.20.2
+gdrive = dvc-gdrive==2.19.0
+gs = dvc-gs==2.19.1
 hdfs =
-    # due to https://github.com/python-poetry/poetry/issues/4683, we have to
-    # explicitly specify pyarrow here
-    pyarrow>=1
-    fsspec[arrow]
-oss = ossfs>=2021.8.0
-s3 =
-    s3fs[boto3]>=2022.02.0
-    aiobotocore[boto3]>2
-ssh =
-    # due to https://github.com/python-poetry/poetry/issues/4683, we have to
-    # explicitly specify bcrypt here
-    bcrypt
-    sshfs[bcrypt]>=2022.6.0
+    dvc-hdfs==2.19.0; python_version < '3.11'
+oss = dvc-oss==2.19.0
+s3 = dvc-s3==2.20.0
+ssh = dvc-ssh==2.19.0
 # gssapi should not be included in all_remotes, because it doesn't have wheels
 # for Linux, so it will fail to compile if user doesn't have all the
 # requirements, including kerberos itself.
-ssh_gssapi = sshfs[gssapi]>=2022.6.0
-webdav = webdav4>=0.9.3
-# not to break `dvc[webhdfs]`
-webhdfs =
+ssh_gssapi = dvc-ssh[gssapi]==2.19.0
+webdav = dvc-webdav==2.19.0
+webhdfs = dvc-webhdfs==2.19.0
 # requests-kerberos requires krb5 & gssapi, which does not provide wheels Linux
-webhdfs_kerberos = requests-kerberos==0.14.0
+webhdfs_kerberos = dvc-webhdfs[kerberos]==2.19.0
 terraform = tpi[ssh]>=2.1.0
 tests =
     %(terraform)s
-    dvc_ssh==0.0.1a0
+    dvc-ssh==2.19.0
     # Test requirements
-    pytest==7.1.2
+    pytest==7.1.3
     pytest-cov==3.0.0
     pytest-xdist==2.5.0
     pytest-mock==3.8.2
@@ -130,11 +116,10 @@ tests =
     # https://github.com/docker/docker-py/issues/2902
     pytest-docker==0.11.0; python_version < '3.10' or sys_platform != 'win32'
     flaky==3.7.0
-    mock==4.0.3
     pytest-timeout==2.1.0
     filelock==3.8.0
     # pylint requirements
-    pylint==2.14.5
+    pylint==2.15.2
     # we use this to suppress pytest-related false positives in our tests.
     pylint-pytest==1.1.2
     # we use this to suppress some messages in tests, eg: foo/bar naming,
@@ -160,6 +145,8 @@ exclude =
 [options.entry_points]
 console_scripts =
     dvc = dvc.cli:main
+fsspec.specs =
+    dvc = dvc.api:DvcFileSystem
 
 [flake8]
 ignore=

tests/conftest.py

@@ -250,3 +250,15 @@ def run_copy_metrics(tmp_dir, run_copy):
         return stage
 
     return run
+
+
+@pytest.fixture(autouse=True)
+def mock_hydra_conf(mocker):
+    if sys.version_info < (3, 11):
+        return
+
+    # `hydra.conf` fails to import in 3.11, it raises ValueError due to changes
+    # in dataclasses. See https://github.com/python/cpython/pull/29867.
+    # NOTE: using sentinel here so that any imports from `hydra.conf`
+    # return a mock.
+    sys.modules["hydra.conf"] = mocker.sentinel

tests/func/api/test_params.py

@@ -1,3 +1,4 @@
+import os
 from textwrap import dedent
 
 import pytest
@@ -67,6 +68,9 @@ def test_params_show_targets(params_repo):
         "bar": 2,
         "foobar": 3,
     }
+    assert api.params_show("params.yaml", stages="stage-1") == {
+        "foo": 5,
+    }
 
 
 def test_params_show_deps(params_repo):
@@ -92,6 +96,27 @@ def test_params_show_stages(params_repo):
         api.params_show(stages="stage-0")
 
 
+def test_params_show_stage_addressing(tmp_dir, dvc):
+    for subdir in {"subdir1", "subdir2"}:
+        subdir = tmp_dir / subdir
+        subdir.mkdir()
+        with subdir.chdir():
+            subdir.gen("params.yaml", "foo: 1")
+
+            dvc.run(name="stage-0", cmd="echo stage-0", params=["foo"])
+
+    for s in {"subdir1", "subdir2"}:
+        dvcyaml = os.path.join(s, "dvc.yaml")
+        assert api.params_show(stages=f"{dvcyaml}:stage-0") == {"foo": 1}
+
+    with subdir.chdir():
+        nested = subdir / "nested"
+        nested.mkdir()
+        with nested.chdir():
+            dvcyaml = os.path.join("..", "dvc.yaml")
+            assert api.params_show(stages=f"{dvcyaml}:stage-0") == {"foo": 1}
+
+
 def test_params_show_revs(params_repo):
     assert api.params_show(rev="HEAD~1") == {
         "params.yaml:foo": 1,
@@ -168,3 +193,15 @@ def test_params_show_stage_without_params(tmp_dir, dvc):
 
     with pytest.raises(DvcException, match="No params found"):
         api.params_show(deps=True)
+
+
+def test_params_show_untracked_target(params_repo, tmp_dir):
+    tmp_dir.gen("params_foo.yaml", "foo: 1")
+
+    assert api.params_show("params_foo.yaml") == {"foo": 1}
+
+    with pytest.raises(DvcException, match="No params found"):
+        api.params_show("params_foo.yaml", stages="stage-0")
+
+    with pytest.raises(DvcException, match="No params found"):
+        api.params_show("params_foo.yaml", deps=True)

tests/func/data/db/test_index.py

@@ -3,7 +3,7 @@ import pytest
 from dvc.exceptions import DownloadError, UploadError
 from dvc.fs import LocalFileSystem
 from dvc.utils.fs import remove
-from dvc_data.db import get_index
+from dvc_data.hashfile.db import get_index
 
 
 @pytest.fixture

tests/func/experiments/conftest.py

@@ -3,7 +3,9 @@ from textwrap import dedent
 import pytest
 
 from tests.func.test_repro_multistage import COPY_SCRIPT
-from tests.unit.repo.experiments.conftest import test_queue  # noqa
+from tests.unit.repo.experiments.conftest import (  # noqa pylint disable=unused-argument
+    test_queue,
+)
 
 DEFAULT_ITERATIONS = 2
 CHECKPOINT_SCRIPT_FORMAT = dedent(

tests/func/experiments/executor/test_ssh.py

@@ -4,9 +4,9 @@ from functools import partial
 from urllib.parse import urlparse
 
 import pytest
+from dvc_ssh import SSHFileSystem
 from dvc_ssh.tests.cloud import TEST_SSH_KEY_PATH, TEST_SSH_USER
 
-from dvc.fs import SSHFileSystem
 from dvc.repo.experiments.executor.base import ExecutorInfo, ExecutorResult
 from dvc.repo.experiments.executor.ssh import SSHExecutor
 from dvc.repo.experiments.refs import EXEC_HEAD, EXEC_MERGE

tests/func/experiments/test_diff.py

@@ -1,6 +1,17 @@
 from funcy import first
 
 
+def test_diff_no_commits(tmp_dir):
+    from scmrepo.git import Git
+
+    from dvc.repo import Repo
+
+    git = Git.init(tmp_dir.fs_path)
+    assert git.no_commits
+
+    assert Repo.init().experiments.diff() == {}
+
+
 def test_diff_empty(tmp_dir, scm, dvc, exp_stage):
     assert dvc.experiments.diff() == {"params": {}, "metrics": {}}
 

tests/func/experiments/test_experiments.py

@@ -104,20 +104,6 @@ def test_failed_exp_workspace(
     )
 
 
-@pytest.mark.parametrize(
-    "changes, expected",
-    [
-        [["foo=baz"], "foo: baz\ngoo:\n  bag: 3.0\nlorem: false"],
-        [["params.yaml:foo=baz"], "foo: baz\ngoo:\n  bag: 3.0\nlorem: false"],
-    ],
-)
-def test_modify_params(params_repo, dvc, changes, expected):
-    dvc.experiments.run(params=changes)
-    # pylint: disable=unspecified-encoding
-    with open("params.yaml", mode="r") as fobj:
-        assert fobj.read().strip() == expected
-
-
 def test_apply(tmp_dir, scm, dvc, exp_stage):
     from dvc.exceptions import InvalidArgumentError
     from dvc.repo.experiments.exceptions import ApplyConflictError

tests/func/experiments/test_queue.py

@@ -42,6 +42,7 @@ def failed_tasks(tmp_dir, dvc, scm, test_queue, failed_exp_stage):
     return name_list
 
 
+@pytest.mark.xfail(strict=False, reason="pytest-celery flaky")
 @pytest.mark.parametrize("follow", [True, False])
 def test_celery_logs(
     tmp_dir,
@@ -64,7 +65,7 @@ def test_celery_logs(
     assert "failed to reproduce 'failed-copy-file'" in captured.out
 
 
-@pytest.mark.flaky(max_runs=3, min_passes=1)
+@pytest.mark.xfail(strict=False, reason="pytest-celery flaky")
 def test_queue_remove_done(dvc, failed_tasks, success_tasks):
     assert len(dvc.experiments.celery_queue.failed_stash) == 3
     status = to_dict(dvc.experiments.celery_queue.status())

tests/func/experiments/test_set_params.py

@@ -0,0 +1,143 @@
+import sys
+
+import pytest
+
+from dvc.exceptions import InvalidArgumentError
+
+from ..utils.test_hydra import hydra_setup
+
+
+@pytest.mark.parametrize(
+    "changes, expected",
+    [
+        [["foo=baz"], "foo: baz\ngoo:\n  bag: 3.0\nlorem: false"],
+        [["params.yaml:foo=baz"], "foo: baz\ngoo:\n  bag: 3.0\nlorem: false"],
+    ],
+)
+def test_modify_params(params_repo, dvc, changes, expected):
+    dvc.experiments.run(params=changes)
+    # pylint: disable=unspecified-encoding
+    with open("params.yaml", mode="r") as fobj:
+        assert fobj.read().strip() == expected
+
+
+@pytest.mark.parametrize(
+    "hydra_enabled",
+    [
+        pytest.param(
+            True,
+            marks=pytest.mark.skipif(
+                sys.version_info >= (3, 11), reason="unsupported on 3.11"
+            ),
+        ),
+        False,
+    ],
+)
+@pytest.mark.parametrize(
+    "config_dir,config_name",
+    [
+        (None, None),
+        (None, "bar"),
+        ("conf", "bar"),
+    ],
+)
+def test_hydra_compose_and_dump(
+    tmp_dir, params_repo, dvc, hydra_enabled, config_dir, config_name
+):
+    hydra_setup(
+        tmp_dir,
+        config_dir=config_dir or "conf",
+        config_name=config_name or "config",
+    )
+
+    dvc.experiments.run()
+    assert (tmp_dir / "params.yaml").parse() == {
+        "foo": [{"bar": 1}, {"baz": 2}],
+        "goo": {"bag": 3.0},
+        "lorem": False,
+    }
+
+    with dvc.config.edit() as conf:
+        if hydra_enabled:
+            conf["hydra"]["enabled"] = True
+        if config_dir is not None:
+            conf["hydra"]["config_dir"] = config_dir
+        if config_name is not None:
+            conf["hydra"]["config_name"] = config_name
+
+    dvc.experiments.run()
+
+    if hydra_enabled:
+        assert (tmp_dir / "params.yaml").parse() == {
+            "db": {"driver": "mysql", "user": "omry", "pass": "secret"},
+        }
+
+        dvc.experiments.run(params=["db=postgresql"])
+        assert (tmp_dir / "params.yaml").parse() == {
+            "db": {
+                "driver": "postgresql",
+                "user": "foo",
+                "pass": "bar",
+                "timeout": 10,
+            }
+        }
+    else:
+        assert (tmp_dir / "params.yaml").parse() == {
+            "foo": [{"bar": 1}, {"baz": 2}],
+            "goo": {"bag": 3.0},
+            "lorem": False,
+        }
+
+
+@pytest.mark.parametrize(
+    "hydra_enabled,overrides,expected",
+    [
+        (
+            pytest.param(
+                True,
+                marks=pytest.mark.xfail(
+                    sys.version_info >= (3, 11), reason="unsupported on 3.11"
+                ),
+            ),
+            ["db=mysql,postgresql"],
+            [
+                {"params.yaml": ["db=mysql"]},
+                {"params.yaml": ["db=postgresql"]},
+            ],
+        ),
+        (
+            False,
+            ["foo=bar,baz"],
+            [{"params.yaml": ["foo=bar"]}, {"params.yaml": ["foo=baz"]}],
+        ),
+    ],
+)
+def test_hydra_sweep(
+    tmp_dir, params_repo, dvc, mocker, hydra_enabled, overrides, expected
+):
+    patched = mocker.patch.object(dvc.experiments, "queue_one")
+
+    if hydra_enabled:
+        hydra_setup(
+            tmp_dir,
+            config_dir="conf",
+            config_name="config",
+        )
+        with dvc.config.edit() as conf:
+            conf["hydra"]["enabled"] = True
+
+    dvc.experiments.run(params=overrides, queue=True)
+
+    assert patched.call_count == len(expected)
+    for e in expected:
+        patched.assert_any_call(
+            mocker.ANY,
+            params=e,
+            reset=True,
+            targets=None,
+        )
+
+
+def test_hydra_sweep_requires_queue(params_repo, dvc):
+    with pytest.raises(InvalidArgumentError):
+        dvc.experiments.run(params=["db=mysql,postgresql"])

tests/func/experiments/test_show.py

@@ -35,6 +35,18 @@ def make_executor_info(**kwargs):
     return ExecutorInfo(**kwargs)
 
 
+@pytest.mark.vscode
+def test_show_no_commits(tmp_dir):
+    from scmrepo.git import Git
+
+    from dvc.repo import Repo
+
+    git = Git.init(tmp_dir.fs_path)
+    assert git.no_commits
+
+    assert Repo.init().experiments.show() == {}
+
+
 @pytest.mark.vscode
 def test_show_simple(tmp_dir, scm, dvc, exp_stage):
     assert dvc.experiments.show()["workspace"] == {

tests/func/metrics/test_show.py

@@ -21,6 +21,32 @@ def test_show_simple(tmp_dir, dvc, run_copy_metrics):
     }
 
 
+def test_show_simple_from_subdir(tmp_dir, dvc, run_copy_metrics):
+    subdir = tmp_dir / "subdir"
+    subdir.mkdir()
+    tmp_dir.gen("metrics_t.yaml", "1.1")
+    run_copy_metrics(
+        "metrics_t.yaml",
+        "subdir/metrics.yaml",
+        metrics=["subdir/metrics.yaml"],
+    )
+
+    expected_path = os.path.join("subdir", "metrics.yaml")
+    assert dvc.metrics.show() == {"": {"data": {expected_path: {"data": 1.1}}}}
+
+    expected_path = os.path.join("..", "subdir", "metrics.yaml")
+    with subdir.chdir():
+        assert dvc.metrics.show() == {
+            "": {"data": {expected_path: {"data": 1.1}}}
+        }
+    subdir2 = tmp_dir / "subdir2"
+    subdir2.mkdir()
+    with subdir2.chdir():
+        assert dvc.metrics.show() == {
+            "": {"data": {expected_path: {"data": 1.1}}}
+        }
+
+
 def test_show(tmp_dir, dvc, run_copy_metrics):
     tmp_dir.gen("metrics_t.yaml", "foo: 1.1")
     run_copy_metrics(

tests/func/plots/test_diff.py

@@ -1,3 +1,5 @@
+import pytest
+
 from tests.utils.plots import get_plot
 
 
@@ -46,3 +48,15 @@ def test_diff_dirty(tmp_dir, scm, dvc, run_copy_metrics):
     assert get_plot(
         diff_result, "workspace", "definitions", file="", endkey="data"
     ) == {"metric.json": props}
+
+
+@pytest.mark.vscode
+def test_no_commits(tmp_dir):
+    from scmrepo.git import Git
+
+    from dvc.repo import Repo
+
+    git = Git.init(tmp_dir.fs_path)
+    assert git.no_commits
+
+    assert Repo.init().plots.diff() == {}

tests/func/test_add.py

@@ -12,6 +12,7 @@ import pytest
 
 import dvc as dvc_module
 import dvc_data
+from dvc.annotations import Annotation
 from dvc.cli import main
 from dvc.dvcfile import DVC_FILE_SUFFIX
 from dvc.exceptions import (
@@ -103,7 +104,7 @@ def test_add_unsupported_file(dvc):
 
 
 def test_add_directory(tmp_dir, dvc):
-    from dvc_data import load
+    from dvc_data.hashfile import load
 
     (stage,) = tmp_dir.dvc_gen({"dir": {"file": "file"}})
 
@@ -457,7 +458,7 @@ class TestAddCommit(TestDvc):
 
 def test_should_collect_dir_cache_only_once(mocker, tmp_dir, dvc):
     tmp_dir.gen({"data/data": "foo"})
-    counter = mocker.spy(dvc_data.build, "_build_tree")
+    counter = mocker.spy(dvc_data.hashfile.build, "_build_tree")
     ret = main(["add", "data"])
     assert ret == 0
     assert counter.mock.call_count == 3
@@ -881,7 +882,7 @@ def test_add_with_cache_link_error(tmp_dir, dvc, mocker, capsys):
     tmp_dir.gen("foo", "foo")
 
     mocker.patch(
-        "dvc_data.checkout.test_links",
+        "dvc_data.hashfile.checkout.test_links",
         return_value=[],
     )
     dvc.add("foo")
@@ -903,6 +904,10 @@ def test_add_preserve_fields(tmp_dir, dvc):
         outs:
         - path: foo # out comment
           desc: out desc
+          type: mytype
+          labels:
+          - label1
+          - label2
           remote: testremote
         meta: some metadata
     """
@@ -916,6 +921,10 @@ def test_add_preserve_fields(tmp_dir, dvc):
         outs:
         - path: foo # out comment
           desc: out desc
+          type: mytype
+          labels:
+          - label1
+          - label2
           remote: testremote
           md5: acbd18db4cc2f85cedef654fccc4a4d8
           size: 3
@@ -1161,3 +1170,23 @@ def test_add_ignore_duplicated_targets(tmp_dir, dvc, capsys):
     _, err = capsys.readouterr()
     assert len(stages) == 3
     assert "ignoring duplicated targets: foo, bar" in err
+
+
+def test_add_with_annotations(M, tmp_dir, dvc):
+    tmp_dir.gen("foo", "foo")
+
+    annot = {
+        "desc": "foo desc",
+        "labels": ["l1", "l2"],
+        "type": "t1",
+        "meta": {"key": "value"},
+    }
+    (stage,) = dvc.add("foo", **annot)
+    assert stage.outs[0].annot == Annotation(**annot)
+    assert (tmp_dir / "foo.dvc").parse() == M.dict(outs=[M.dict(**annot)])
+
+    # try to selectively update/overwrite some annotations
+    annot = {**annot, "type": "t2"}
+    (stage,) = dvc.add("foo", type="t2")
+    assert stage.outs[0].annot == Annotation(**annot)
+    assert (tmp_dir / "foo.dvc").parse() == M.dict(outs=[M.dict(**annot)])

tests/func/test_checkout.py

@@ -90,7 +90,7 @@ class TestCheckoutCorruptedCacheFile(TestRepro):
 
 class TestCheckoutCorruptedCacheDir(TestDvc):
     def test(self):
-        from dvc_data import load
+        from dvc_data.hashfile import load
 
         # NOTE: using 'copy' so that cache and link don't have same inode
         ret = main(["config", "cache.type", "copy"])

tests/func/test_commit.py

@@ -47,6 +47,13 @@ def test_commit_preserve_fields(tmp_dir, dvc):
         outs:
         - path: foo # out comment
           desc: out desc
+          type: mytype
+          labels:
+          - label1
+          - label2
+          meta:
+            key1: value1
+            key2: value2
           remote: testremote
         meta: some metadata
     """
@@ -61,6 +68,13 @@ def test_commit_preserve_fields(tmp_dir, dvc):
         outs:
         - path: foo # out comment
           desc: out desc
+          type: mytype
+          labels:
+          - label1
+          - label2
+          meta:
+            key1: value1
+            key2: value2
           remote: testremote
           md5: acbd18db4cc2f85cedef654fccc4a4d8
           size: 3

tests/func/test_data_cloud.py

@@ -13,8 +13,8 @@ from dvc.testing.test_remote import (  # noqa, pylint: disable=unused-import
     TestRemote,
 )
 from dvc.utils.fs import remove
-from dvc_data.db.local import LocalHashFileDB
 from dvc_data.hashfile.db import HashFileDB
+from dvc_data.hashfile.db.local import LocalHashFileDB
 
 
 def test_cloud_cli(tmp_dir, dvc, remote, mocker):
@@ -273,6 +273,20 @@ def test_pull_external_dvc_imports(tmp_dir, dvc, scm, erepo_dir):
     assert (tmp_dir / "new_dir" / "bar").read_text() == "bar"
 
 
+def test_pull_partial_import(tmp_dir, dvc, local_workspace):
+    local_workspace.gen("file", "file content")
+    dst = tmp_dir / "file"
+    stage = dvc.imp_url(
+        "remote://workspace/file", os.fspath(dst), no_download=True
+    )
+
+    result = dvc.pull("file")
+    assert result["fetched"] == 1
+    assert dst.exists()
+
+    assert stage.outs[0].get_hash().value == "d10b4c3ff123b26dc068d43a8bef2d23"
+
+
 def test_pull_external_dvc_imports_mixed(
     tmp_dir, dvc, scm, erepo_dir, local_remote
 ):

tests/func/test_data_ls.py

@@ -0,0 +1,31 @@
+from operator import itemgetter
+
+from dvc.repo.data import ls
+
+
+def test_data_ls(tmp_dir, dvc):
+    assert not list(ls(dvc))
+
+    tmp_dir.dvc_gen("bar", "bar")
+    tmp_dir.gen("foo", "foo")
+    dvc.add(
+        "foo",
+        meta={"key": "value"},
+        labels=["l1", "l2"],
+        type="t1",
+        desc="foo",
+    )
+
+    foo_entry = {
+        "path": "foo",
+        "desc": "foo",
+        "type": "t1",
+        "labels": ["l1", "l2"],
+        "meta": {"key": "value"},
+    }
+    assert sorted(ls(dvc), key=itemgetter("path")) == [
+        {"path": "bar"},
+        foo_entry,
+    ]
+    assert list(ls(dvc, targets=["foo"])) == [foo_entry]
+    assert list(ls(dvc, targets=["foo"], recursive=True)) == [foo_entry]

tests/func/test_data_status.py

@@ -1,6 +1,10 @@
+from os import fspath
 from os.path import join
 
+import pytest
+
 from dvc.repo import Repo
+from dvc.repo.data import _transform_git_paths_to_dvc, posixpath_to_os_path
 from dvc.testing.tmp_dir import make_subrepo
 from dvc.utils.fs import remove
 
@@ -14,24 +18,55 @@ EMPTY_STATUS = {
 }
 
 
-def test_file(M, tmp_dir, scm, dvc):
+@pytest.mark.parametrize("path", [None, ("sub", "repo")])
+def test_git_to_dvc_path_wdir_transformation(tmp_dir, scm, path):
+    struct = {"dir": {"foo": "foo", "bar": "bar"}, "file": "file", "dir2": {}}
+    tmp_dir.gen(struct)
+
+    subdir = tmp_dir.joinpath(*path) if path else tmp_dir
+    make_subrepo(subdir, scm)
+    dvc = subdir.dvc
+
+    with subdir.chdir():
+        subdir.gen(struct)
+        _, _, untracked = scm.status(untracked_files="all")
+        # make order independent of the platforms for easier test assertions
+        untracked = sorted(map(posixpath_to_os_path, untracked), reverse=True)
+        assert _transform_git_paths_to_dvc(dvc, untracked) == [
+            "file",
+            join("dir", "foo"),
+            join("dir", "bar"),
+        ]
+        with (subdir / "dir").chdir():
+            assert _transform_git_paths_to_dvc(dvc, untracked) == [
+                join("..", "file"),
+                "foo",
+                "bar",
+            ]
+        with (subdir / "dir2").chdir():
+            assert _transform_git_paths_to_dvc(dvc, untracked) == [
+                join("..", "file"),
+                join("..", "dir", "foo"),
+                join("..", "dir", "bar"),
+            ]
+
+
+def test_file(M, tmp_dir, dvc, scm):
     tmp_dir.dvc_gen("foo", "foo", commit="add foo")
     tmp_dir.dvc_gen("foo", "foobar")
     remove(tmp_dir / "foo")
 
     expected = {
+        **EMPTY_STATUS,
         "committed": {"modified": ["foo"]},
         "uncommitted": {"deleted": ["foo"]},
-        "git": M.instance_of(dict),
-        "not_in_cache": [],
-        "unchanged": [],
-        "untracked": [],
+        "git": M.dict(),
     }
     assert dvc.data_status() == expected
     assert dvc.data_status(granular=True) == expected
 
 
-def test_directory(M, tmp_dir, scm, dvc):
+def test_directory(M, tmp_dir, dvc, scm):
     tmp_dir.dvc_gen({"dir": {"foo": "foo"}}, commit="add dir")
     tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar", "foobar": "foobar"}})
     remove(tmp_dir / "dir")
@@ -39,12 +74,10 @@ def test_directory(M, tmp_dir, scm, dvc):
     tmp_dir.gen("untracked", "untracked")
 
     assert dvc.data_status() == {
+        **EMPTY_STATUS,
         "committed": {"modified": [join("dir", "")]},
         "uncommitted": {"modified": [join("dir", "")]},
-        "git": M.instance_of(dict),
-        "not_in_cache": [],
-        "unchanged": [],
-        "untracked": [],
+        "git": M.dict(),
     }
 
     assert dvc.data_status(granular=True, untracked_files="all") == {
@@ -52,20 +85,52 @@ def test_directory(M, tmp_dir, scm, dvc):
             "added": M.unordered(
                 join("dir", "bar"),
                 join("dir", "foobar"),
-            )
+            ),
+            "modified": [join("dir", "")],
         },
         "uncommitted": {
             "added": [join("dir", "baz")],
-            "modified": [join("dir", "bar")],
+            "modified": M.unordered(join("dir", ""), join("dir", "bar")),
             "deleted": [join("dir", "foobar")],
         },
-        "git": M.instance_of(dict),
+        "git": M.dict(),
         "not_in_cache": [],
         "unchanged": [join("dir", "foo")],
         "untracked": ["untracked"],
     }
 
 
+def test_tracked_directory_deep(M, tmp_dir, dvc, scm):
+    """Test for a directory not in cwd, but nested inside other directories."""
+    (tmp_dir / "sub").gen({"dir": {"foo": "foo"}})
+    dvc.add(fspath(tmp_dir / "sub" / "dir"))
+    scm.add_commit(["sub/dir.dvc", "sub/.gitignore"], message="add sub/dir")
+
+    (tmp_dir / "sub" / "dir").gen("bar", "bar")
+    dvc.commit(None, force=True)
+    (tmp_dir / "sub" / "dir").gen("foobar", "foobar")
+
+    assert dvc.data_status() == {
+        **EMPTY_STATUS,
+        "committed": {"modified": [join("sub", "dir", "")]},
+        "uncommitted": {"modified": [join("sub", "dir", "")]},
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True, untracked_files="all") == {
+        **EMPTY_STATUS,
+        "committed": {
+            "added": [join("sub", "dir", "bar")],
+            "modified": [join("sub", "dir", "")],
+        },
+        "uncommitted": {
+            "added": [join("sub", "dir", "foobar")],
+            "modified": [join("sub", "dir", "")],
+        },
+        "git": M.dict(),
+        "unchanged": [join("sub", "dir", "foo")],
+    }
+
+
 def test_new_empty_git_repo(M, tmp_dir, scm):
     dvc = Repo.init()
     assert dvc.data_status() == {
@@ -81,52 +146,19 @@ def test_noscm_repo(dvc):
     assert dvc.data_status() == EMPTY_STATUS
 
 
-def test_unchanged(M, tmp_dir, scm, dvc):
+def test_unchanged(M, tmp_dir, dvc, scm):
     tmp_dir.dvc_gen({"dir": {"foo": "foo"}}, commit="add dir")
     tmp_dir.dvc_gen("bar", "bar", commit="add foo")
 
     assert dvc.data_status() == {
         **EMPTY_STATUS,
-        "git": M.instance_of(dict),
+        "git": M.dict(),
         "unchanged": M.unordered("bar", join("dir", "")),
     }
     assert dvc.data_status(granular=True) == {
         **EMPTY_STATUS,
-        "git": M.instance_of(dict),
-        "unchanged": M.unordered("bar", join("dir", "foo")),
-    }
-
-
-def test_not_in_cache(M, tmp_dir, scm, dvc):
-    # TODO: investigation required, might return wrong results
-    tmp_dir.dvc_gen({"dir": {"foo": "foo"}}, commit="add dir")
-    tmp_dir.dvc_gen("bar", "bar", commit="add foo")
-    remove(dvc.odb.local.path)
-
-    assert dvc.data_status() == {
-        **EMPTY_STATUS,
-        "not_in_cache": M.unordered("bar", join("dir", "")),
-        "git": M.instance_of(dict),
-        "unchanged": ["bar"],
-        "uncommitted": {"added": [join("dir", "")]},
-    }
-    assert dvc.data_status(granular=True) == {
-        **EMPTY_STATUS,
-        "not_in_cache": M.unordered("bar"),
-        "git": M.instance_of(dict),
-        "unchanged": [],
-        "committed": {"modified": ["bar"]},
-        "uncommitted": {"modified": ["bar"], "added": [join("dir", "foo")]},
-    }
-
-
-def test_withdirs(M, tmp_dir, scm, dvc):
-    tmp_dir.dvc_gen({"dir": {"foo": "foo"}}, commit="add dir")
-    tmp_dir.dvc_gen("bar", "bar", commit="add foo")
-    assert dvc.data_status(granular=True, with_dirs=True) == {
-        **EMPTY_STATUS,
-        "git": M.instance_of(dict),
-        "unchanged": M.unordered("bar", join("dir", "foo"), join("dir", "")),
+        "git": M.dict(),
+        "unchanged": M.unordered("bar", join("dir", ""), join("dir", "foo")),
     }
 
 
@@ -146,7 +178,8 @@ def test_skip_uncached_pipeline_outputs(tmp_dir, dvc, run_copy_metrics):
 
 def test_output_with_newly_added_stage(tmp_dir, dvc):
     dvc.stage.add(deps=["bar"], outs=["foo"], name="copy", cmd="cp foo bar")
-    assert dvc.data_status() == {**EMPTY_STATUS, "not_in_cache": ["foo"]}
+    assert dvc.data_status() == EMPTY_STATUS
+    assert dvc.data_status(granular=True) == EMPTY_STATUS
 
 
 def test_subdir(M, tmp_dir, scm):
@@ -161,7 +194,243 @@ def test_subdir(M, tmp_dir, scm):
         dvc = subrepo.dvc
         assert dvc.data_status(granular=True, untracked_files="all") == {
             **EMPTY_STATUS,
-            "git": M.instance_of(dict),
-            "unchanged": M.unordered("bar", join("dir", "foo")),
+            "git": M.dict(),
+            "unchanged": M.unordered(
+                "bar", join("dir", ""), join("dir", "foo")
+            ),
             "untracked": ["untracked"],
         }
+
+
+def test_untracked_newly_added_files(M, tmp_dir, dvc, scm):
+    tmp_dir.gen({"dir": {"foo": "foo", "bar": "bar"}})
+    tmp_dir.gen("foobar", "foobar")
+
+    expected = {
+        **EMPTY_STATUS,
+        "untracked": M.unordered(
+            join("dir", "foo"), join("dir", "bar"), "foobar"
+        ),
+        "git": M.dict(),
+    }
+    assert dvc.data_status(untracked_files="all") == expected
+    assert dvc.data_status(granular=True, untracked_files="all") == expected
+
+
+def test_missing_cache_workspace_exists(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}})
+    tmp_dir.dvc_gen("foobar", "foobar")
+    remove(dvc.odb.repo.path)
+
+    assert dvc.data_status(untracked_files="all") == {
+        **EMPTY_STATUS,
+        "untracked": M.unordered("foobar.dvc", "dir.dvc", ".gitignore"),
+        "committed": {"added": M.unordered("foobar", join("dir", ""))},
+        "not_in_cache": M.unordered("foobar", join("dir", "")),
+        "git": M.dict(),
+    }
+
+    assert dvc.data_status(granular=True, untracked_files="all") == {
+        **EMPTY_STATUS,
+        "untracked": M.unordered("foobar.dvc", "dir.dvc", ".gitignore"),
+        "committed": {"added": M.unordered("foobar", join("dir", ""))},
+        "uncommitted": {
+            "unknown": M.unordered(join("dir", "foo"), join("dir", "bar"))
+        },
+        "not_in_cache": M.unordered(
+            "foobar",
+            join("dir", ""),
+        ),
+        "git": M.dict(),
+    }
+
+
+def test_missing_cache_missing_workspace(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}})
+    tmp_dir.dvc_gen("foobar", "foobar")
+    for path in [dvc.odb.repo.path, "dir", "foobar"]:
+        remove(path)
+
+    assert dvc.data_status(untracked_files="all") == {
+        **EMPTY_STATUS,
+        "untracked": M.unordered("foobar.dvc", "dir.dvc", ".gitignore"),
+        "uncommitted": {"deleted": M.unordered("foobar", join("dir", ""))},
+        "committed": {"added": M.unordered("foobar", join("dir", ""))},
+        "not_in_cache": M.unordered("foobar", join("dir", "")),
+        "git": M.dict(),
+    }
+
+    assert dvc.data_status(granular=True, untracked_files="all") == {
+        **EMPTY_STATUS,
+        "untracked": M.unordered("foobar.dvc", "dir.dvc", ".gitignore"),
+        "uncommitted": {"deleted": M.unordered("foobar", join("dir", ""))},
+        "committed": {"added": M.unordered("foobar", join("dir", ""))},
+        "not_in_cache": M.unordered("foobar", join("dir", "")),
+        "git": M.dict(),
+    }
+
+
+def test_git_committed_missing_cache_workspace_exists(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}}, commit="add dir")
+    tmp_dir.dvc_gen("foobar", "foobar", commit="add foobar")
+    remove(dvc.odb.local.path)
+
+    assert dvc.data_status(untracked_files="all") == {
+        **EMPTY_STATUS,
+        "not_in_cache": M.unordered("foobar", join("dir", "")),
+        "git": M.dict(),
+        "unchanged": M.unordered("foobar", join("dir", "")),
+    }
+    assert dvc.data_status(granular=True) == {
+        **EMPTY_STATUS,
+        "not_in_cache": M.unordered(
+            "foobar",
+            join("dir", ""),
+        ),
+        "uncommitted": {
+            "unknown": M.unordered(join("dir", "foo"), join("dir", "bar"))
+        },
+        "git": M.dict(),
+        "unchanged": M.unordered("foobar", join("dir", "")),
+    }
+
+
+def test_git_committed_missing_cache_missing_workspace(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}}, commit="add dir")
+    tmp_dir.dvc_gen("foobar", "foobar", commit="add foobar")
+    for path in [dvc.odb.repo.path, "dir", "foobar"]:
+        remove(path)
+
+    assert dvc.data_status(untracked_files="all") == {
+        **EMPTY_STATUS,
+        "uncommitted": {"deleted": M.unordered(join("dir", ""), "foobar")},
+        "not_in_cache": M.unordered(join("dir", ""), "foobar"),
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True, untracked_files="all") == {
+        **EMPTY_STATUS,
+        "committed": {},
+        "uncommitted": {"deleted": M.unordered(join("dir", ""), "foobar")},
+        "not_in_cache": M.unordered(join("dir", ""), "foobar"),
+        "git": M.dict(),
+    }
+
+
+def test_partial_missing_cache(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}})
+
+    # remove "foo" from cache
+    odb = dvc.odb.repo
+    odb.fs.rm(odb.oid_to_path("acbd18db4cc2f85cedef654fccc4a4d8"))
+
+    assert dvc.data_status() == {
+        **EMPTY_STATUS,
+        "committed": {"added": [join("dir", "")]},
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True) == {
+        **EMPTY_STATUS,
+        "committed": {
+            "added": M.unordered(
+                join("dir", ""), join("dir", "foo"), join("dir", "bar")
+            )
+        },
+        "not_in_cache": [join("dir", "foo")],
+        "git": M.dict(),
+    }
+
+
+def test_missing_dir_object_from_head(M, tmp_dir, dvc, scm):
+    (stage,) = tmp_dir.dvc_gen(
+        {"dir": {"foo": "foo", "bar": "bar"}}, commit="add dir"
+    )
+    remove("dir")
+    tmp_dir.dvc_gen({"dir": {"foobar": "foobar"}})
+    odb = dvc.odb.repo
+    odb.fs.rm(odb.oid_to_path(stage.outs[0].hash_info.value))
+
+    assert dvc.data_status() == {
+        **EMPTY_STATUS,
+        "committed": {"modified": [join("dir", "")]},
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True) == {
+        **EMPTY_STATUS,
+        "committed": {
+            "modified": [join("dir", "")],
+            "unknown": [join("dir", "foobar")],
+        },
+        "git": M.dict(),
+    }
+
+
+def test_missing_dir_object_from_index(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}}, commit="add dir")
+    remove("dir")
+    (stage,) = tmp_dir.dvc_gen({"dir": {"foobar": "foobar"}})
+    odb = dvc.odb.repo
+    odb.fs.rm(odb.oid_to_path(stage.outs[0].hash_info.value))
+
+    assert dvc.data_status() == {
+        **EMPTY_STATUS,
+        "committed": {"modified": [join("dir", "")]},
+        "not_in_cache": [join("dir", "")],
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True) == {
+        **EMPTY_STATUS,
+        "committed": {
+            "modified": [join("dir", "")],
+        },
+        "uncommitted": {"unknown": [join("dir", "foobar")]},
+        "not_in_cache": [join("dir", "")],
+        "git": M.dict(),
+    }
+
+
+def test_root_from_dir_to_file(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen({"data": {"foo": "foo", "bar": "bar"}})
+    remove("data")
+    tmp_dir.gen("data", "file")
+
+    assert dvc.data_status() == {
+        **EMPTY_STATUS,
+        "committed": {"added": [join("data", "")]},
+        "uncommitted": {"modified": ["data"]},
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True) == {
+        **EMPTY_STATUS,
+        "committed": {
+            "added": M.unordered(
+                join("data", ""), join("data", "foo"), join("data", "bar")
+            )
+        },
+        "uncommitted": {
+            "deleted": M.unordered(join("data", "foo"), join("data", "bar")),
+            "modified": ["data"],
+        },
+        "git": M.dict(),
+    }
+
+
+def test_root_from_file_to_dir(M, tmp_dir, dvc, scm):
+    tmp_dir.dvc_gen("data", "file")
+    remove("data")
+    tmp_dir.gen({"data": {"foo": "foo", "bar": "bar"}})
+
+    assert dvc.data_status() == {
+        **EMPTY_STATUS,
+        "committed": {"added": ["data"]},
+        "uncommitted": {"modified": [join("data", "")]},
+        "git": M.dict(),
+    }
+    assert dvc.data_status(granular=True) == {
+        **EMPTY_STATUS,
+        "committed": {"added": ["data"]},
+        "uncommitted": {
+            "modified": [join("data", "")],
+            "added": M.unordered(join("data", "foo"), join("data", "bar")),
+        },
+        "git": M.dict(),
+    }

tests/func/test_dvcfile.py

@@ -3,6 +3,7 @@ import textwrap
 
 import pytest
 
+from dvc.annotations import Annotation
 from dvc.dvcfile import (
     PIPELINE_FILE,
     PIPELINE_LOCK,
@@ -335,7 +336,7 @@ def test_dvcfile_dump_preserves_desc(tmp_dir, dvc, run_copy):
     (tmp_dir / dvcfile.path).dump(data)
 
     assert stage.desc == stage_desc
-    stage.outs[0].desc = out_desc
+    stage.outs[0].annot.desc = out_desc
     dvcfile.dump(stage)
     loaded = dvcfile._load()[0]
     assert loaded == data
@@ -382,3 +383,31 @@ def test_dvcfile_try_dumping_parametrized_stage(tmp_dir, dvc, data, name):
         dvcfile.dump(stage)
 
     assert str(exc.value) == f"cannot dump a parametrized stage: '{name}'"
+
+
+def test_dvcfile_load_dump_stage_with_desc_meta(tmp_dir, dvc):
+    data = {
+        "stages": {
+            "stage1": {
+                "cmd": "cp foo bar",
+                "desc": "stage desc",
+                "meta": {"key1": "value1", "key2": "value2"},
+                "deps": ["foo"],
+                "outs": [
+                    {"bar": {"desc": "bar desc", "meta": {"key": "value"}}}
+                ],
+            }
+        }
+    }
+    (tmp_dir / "dvc.yaml").dump(data)
+
+    stage = dvc.stage.load_one(name="stage1")
+    assert stage.meta == {"key1": "value1", "key2": "value2"}
+    assert stage.desc == "stage desc"
+    assert stage.outs[0].annot == Annotation(
+        desc="bar desc", meta={"key": "value"}
+    )
+
+    # sanity check
+    stage.dump()
+    assert (tmp_dir / "dvc.yaml").parse() == data

tests/func/test_external_repo.py

@@ -7,8 +7,8 @@ from dvc.external_repo import CLONES, external_repo
 from dvc.testing.tmp_dir import make_subrepo
 from dvc.utils import relpath
 from dvc.utils.fs import makedirs, remove
-from dvc_data.build import build
-from dvc_data.transfer import transfer
+from dvc_data.hashfile.build import build
+from dvc_data.hashfile.transfer import transfer
 
 
 def test_external_repo(erepo_dir, mocker):

tests/func/test_gc.py

@@ -11,7 +11,7 @@ from dvc.exceptions import CollectCacheError
 from dvc.fs import LocalFileSystem
 from dvc.repo import Repo as DvcRepo
 from dvc.utils.fs import remove
-from dvc_data.db.local import LocalHashFileDB
+from dvc_data.hashfile.db.local import LocalHashFileDB
 from tests.basic_env import TestDir, TestDvcGit
 
 

tests/func/test_ignore.py

@@ -10,7 +10,7 @@ from dvc.pathspec_math import PatternInfo, merge_patterns
 from dvc.repo import Repo
 from dvc.testing.tmp_dir import TmpDir
 from dvc.types import List
-from dvc_data.build import IgnoreInCollectedDirError
+from dvc_data.hashfile.build import IgnoreInCollectedDirError
 from dvc_data.hashfile.utils import get_mtime_and_size
 
 

tests/func/test_import.py

@@ -6,6 +6,7 @@ import pytest
 from funcy import first
 from scmrepo.git import Git
 
+from dvc.annotations import Annotation
 from dvc.config import NoRemoteError
 from dvc.dvcfile import Dvcfile
 from dvc.exceptions import DownloadError, PathMissingError
@@ -527,7 +528,7 @@ def test_import_with_no_exec(tmp_dir, dvc, erepo_dir):
 
 
 def test_import_with_jobs(mocker, dvc, erepo_dir):
-    import dvc_data.transfer as otransfer
+    import dvc_data.hashfile.transfer as otransfer
 
     with erepo_dir.chdir():
         erepo_dir.dvc_gen(
@@ -627,3 +628,26 @@ def test_parameterized_repo(tmp_dir, dvc, scm, erepo_dir, paths):
         "url": os.fspath(erepo_dir),
         "rev_lock": erepo_dir.scm.get_rev(),
     }
+
+
+def test_import_with_annotations(M, tmp_dir, scm, dvc, erepo_dir):
+    with erepo_dir.chdir():
+        erepo_dir.dvc_gen("foo", "foo content", commit="create foo")
+
+    annot = {
+        "desc": "foo desc",
+        "labels": ["l1", "l2"],
+        "type": "t1",
+        "meta": {"key": "value"},
+    }
+    stage = dvc.imp(os.fspath(erepo_dir), "foo", "foo", no_exec=True, **annot)
+    assert stage.outs[0].annot == Annotation(**annot)
+    assert (tmp_dir / "foo.dvc").parse() == M.dict(outs=[M.dict(**annot)])
+
+    # try to selectively update/overwrite some annotations
+    annot = {**annot, "type": "t2"}
+    stage = dvc.imp(
+        os.fspath(erepo_dir), "foo", "foo", no_exec=True, type="t2"
+    )
+    assert stage.outs[0].annot == Annotation(**annot)
+    assert (tmp_dir / "foo.dvc").parse() == M.dict(outs=[M.dict(**annot)])

tests/func/test_import_url.py

@@ -5,8 +5,9 @@ from uuid import uuid4
 
 import pytest
 
+from dvc.annotations import Annotation
 from dvc.cli import main
-from dvc.dependency.base import DependencyDoesNotExistError
+from dvc.dependency.base import Dependency, DependencyDoesNotExistError
 from dvc.exceptions import InvalidArgumentError
 from dvc.stage import Stage
 from dvc.testing.test_workspace import TestImport as _TestImport
@@ -132,7 +133,7 @@ class TestImport(_TestImport):
         return False
 
 
-def test_import_url_preserve_meta(tmp_dir, dvc):
+def test_import_url_preserve_fields(tmp_dir, dvc):
     text = textwrap.dedent(
         """\
         # top comment
@@ -142,6 +143,12 @@ def test_import_url_preserve_meta(tmp_dir, dvc):
         outs:
         - path: bar # out comment
           desc: out desc
+          type: mytype
+          labels:
+          - label1
+          - label2
+          meta:
+            key: value
         meta: some metadata
     """
     )
@@ -160,6 +167,12 @@ def test_import_url_preserve_meta(tmp_dir, dvc):
         outs:
         - path: bar # out comment
           desc: out desc
+          type: mytype
+          labels:
+          - label1
+          - label2
+          meta:
+            key: value
           md5: acbd18db4cc2f85cedef654fccc4a4d8
           size: 3
         meta: some metadata
@@ -267,3 +280,71 @@ def test_import_url_to_remote_status(tmp_dir, dvc, local_cloud, local_remote):
 
     status = dvc.status()
     assert len(status) == 0
+
+
+def test_import_url_no_download(tmp_dir, dvc, local_workspace):
+    local_workspace.gen("file", "file content")
+    dst = tmp_dir / "file"
+    stage = dvc.imp_url(
+        "remote://workspace/file", os.fspath(dst), no_download=True
+    )
+
+    assert stage.deps[0].hash_info.value == "d10b4c3ff123b26dc068d43a8bef2d23"
+
+    assert not dst.exists()
+
+    out = stage.outs[0]
+    assert not out.hash_info
+    assert out.meta.size is None
+
+    status = dvc.status()
+    assert status["file.dvc"] == [
+        {"changed outs": {"file": "deleted"}},
+    ]
+
+
+def test_imp_url_with_annotations(M, tmp_dir, dvc, local_workspace):
+    local_workspace.gen("foo", "foo")
+    annot = {
+        "desc": "foo desc",
+        "labels": ["l1", "l2"],
+        "type": "t1",
+        "meta": {"key": "value"},
+    }
+    stage = dvc.imp_url(
+        "remote://workspace/foo",
+        os.fspath(tmp_dir / "foo"),
+        no_exec=True,
+        **annot,
+    )
+    assert stage.outs[0].annot == Annotation(**annot)
+    assert (tmp_dir / "foo.dvc").parse() == M.dict(outs=[M.dict(**annot)])
+
+    # try to selectively update/overwrite some annotations
+    annot = {**annot, "type": "t2"}
+    stage = dvc.imp_url(
+        "remote://workspace/foo",
+        os.fspath(tmp_dir / "foo"),
+        no_exec=True,
+        type="t2",
+    )
+    assert stage.outs[0].annot == Annotation(**annot)
+    assert (tmp_dir / "foo.dvc").parse() == M.dict(outs=[M.dict(**annot)])
+
+
+def test_import_url_fs_config(tmp_dir, dvc, workspace, mocker):
+    import dvc.fs as dvc_fs
+
+    workspace.gen("foo", "foo")
+
+    url = "remote://workspace/foo"
+    get_fs_config = mocker.spy(dvc_fs, "get_fs_config")
+    dep_init = mocker.spy(Dependency, "__init__")
+    dvc.imp_url(url, fs_config={"jobs": 42})
+
+    dep_init_kwargs = dep_init.call_args[1]
+    assert dep_init_kwargs.get("fs_config") == {"jobs": 42}
+
+    assert get_fs_config.call_args_list[0][1] == {"url": "foo"}
+    assert get_fs_config.call_args_list[1][1] == {"url": url, "jobs": 42}
+    assert get_fs_config.call_args_list[2][1] == {"name": "workspace"}

tests/func/test_odb.py

@@ -43,7 +43,7 @@ def test_cache(tmp_dir, dvc):
 
 
 def test_cache_load_bad_dir_cache(tmp_dir, dvc):
-    from dvc_data import load
+    from dvc_data.hashfile import load
 
     dir_hash = "123.dir"
     fname = os.fspath(dvc.odb.local.oid_to_path(dir_hash))