checkout: Expose --allow-missing arg. (#9919)

Implements fsspec's fsid https://github.com/fsspec/filesystem_spec/pull/1122

Required for #9904

.dvc/.gitignore

@@ -0,0 +1,3 @@
+/config.local
+/tmp
+/cache

.flake8

@@ -1,19 +1,20 @@
 [flake8]
-ignore=
+max_line_length = 88
+extend-ignore=
     # Whitespace before ':'
     E203
-    # Too many leading '#' for block comment
-    E266
     # Line break occurred before a binary operator
     W503
     # unindexed parameters in the str.format, see:
     # https://pypi.org/project/flake8-string-format/
     P1
-    # use `!r` instead of surrounding by quotes
-    B028
-max_line_length = 88
 max-complexity = 10
-select = B,C,E,F,W,T4,B902,T,P,TYP
 show_source = true
 count = true
 min_python_version = 3.8.0
+per-file-ignores =
+    dvc/testing/benchmarks/conftest.py:F401
+    dvc/testing/conftest.py:F401
+    dvc/testing/plugin.py:F401
+    tests/conftest.py:F401
+    tests/func/experiments/executor/test_ssh.py:F401

.github/workflows/build.yaml

@@ -0,0 +1,47 @@
+name: Build package
+on:
+  push:
+    branches:
+    - main
+  release:
+    types: [released, prereleased]
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  id-token: write
+
+jobs:
+  pip:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v4
+      with:
+        fetch-depth: 0
+
+    - uses: actions/setup-python@v4
+      with:
+        python-version: 3.8
+
+    - run: python -m pip install --upgrade pip setuptools_scm build twine
+    - name: Force version for Test PyPI upload
+      if: ${{ !startsWith(github.ref, 'refs/tags') }}
+      run: echo version=$(python -m setuptools_scm | awk -F+ '{print $1}' | tail -1) >> $GITHUB_ENV
+
+    - run: |
+        echo 'PKG = "pip"'>dvc/_build.py
+        python -m build
+      env:
+        SETUPTOOLS_SCM_PRETEND_VERSION: ${{ env.version }}
+    - run: twine check --strict dist/*
+
+    - name: Publish to PyPI
+      if: github.event_name == 'release'
+      uses: pypa/gh-action-pypi-publish@release/v1
+
+    - name: Publish to Test PyPI
+      if: ${{ github.event_name == 'release' || (github.event_name == 'push' && github.ref == 'refs/heads/main') }}
+      uses: pypa/gh-action-pypi-publish@release/v1
+      with:
+        repository-url: https://test.pypi.org/legacy/
+        skip-existing: true

.github/workflows/codeql.yml

@@ -28,7 +28,7 @@ jobs:
 
     steps:
     - name: Checkout repository
-      uses: actions/checkout@v3
+      uses: actions/checkout@v4
 
     - name: Initialize CodeQL
       uses: github/codeql-action/init@v2

.github/workflows/packages.yaml

@@ -1,51 +0,0 @@
-name: Build packages
-on:
-  push:
-    branches:
-    - main
-  release:
-    types: [released, prereleased]
-  workflow_dispatch:
-
-permissions:  # added using https://github.com/step-security/secure-workflows
-  contents: read
-
-jobs:
-  pip:
-    runs-on: ubuntu-20.04
-    permissions:
-      id-token: write
-    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
-
-    - name: Install
-      run: pip install --upgrade pip wheel
-
-    - name: Force version for Test PyPI uploads
-      if: ${{ !startsWith(github.ref, 'refs/tags') }}
-      run: |
-        pip install setuptools_scm
-        echo version=$(python -m setuptools_scm | awk -F+ '{print $1}' | tail -1) >> $GITHUB_ENV
-
-    - name: Build packages
-      run: ./scripts/build_package.sh
-      env:
-        SETUPTOOLS_SCM_PRETEND_VERSION: ${{ env.version }}
-
-    - name: Publish packages to PyPI
-      if: github.event_name == 'release'
-      uses: pypa/gh-action-pypi-publish@release/v1
-
-    - name: Publish to Test PyPI
-      if: ${{ github.event_name == 'release' || (github.event_name == 'push' && github.ref == 'refs/heads/main') }}
-      uses: pypa/gh-action-pypi-publish@release/v1
-      with:
-        repository_url: https://test.pypi.org/legacy/
-        skip_existing: true

.github/workflows/plugin_tests.yaml

@@ -23,11 +23,11 @@ jobs:
         plugin: ["dvc-s3"]
 
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         path: dvc
 
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         repository: iterative/${{ matrix.plugin }}
         ref: main

.github/workflows/tests.yaml

@@ -26,7 +26,7 @@ jobs:
         pyv: ["3.8", "3.9", "3.10", "3.11"]
 
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         fetch-depth: 0
     - name: Set up Python
@@ -66,7 +66,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - name: Slack Notification
-      uses: rtCamp/[email protected].0
+      uses: rtCamp/[email protected].1
       env:
         SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
         SLACK_COLOR: ${{ job.status }}

.pre-commit-config.yaml

@@ -23,42 +23,29 @@ repos:
       - id: mixed-line-ending
       - id: sort-simple-yaml
       - id: trailing-whitespace
-  - repo: https://github.com/lovesegfault/beautysh
-    rev: v6.2.1
-    hooks:
-      - id: beautysh
-        args: ["-i", "2"]
-  - repo: https://github.com/asottile/pyupgrade
-    rev: v3.3.1
-    hooks:
-      - id: pyupgrade
-        args:
-          - --py38-plus
   - repo: https://github.com/psf/black
-    rev: 23.3.0
+    rev: 23.7.0
     hooks:
       - id: black
   - repo: https://github.com/PyCQA/isort
     rev: 5.12.0
     hooks:
       - id: isort
-  - repo: https://github.com/charliermarsh/ruff-pre-commit
-    rev: 'v0.0.261'
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: 'v0.0.287'
     hooks:
       - id: ruff
+        args: [--fix, --exit-non-zero-on-fix]
   - repo: https://github.com/codespell-project/codespell
-    rev: v2.2.4
+    rev: v2.2.5
     hooks:
       - id: codespell
         additional_dependencies: ["tomli"]
   - repo: https://github.com/PyCQA/flake8
-    rev: 6.0.0
+    rev: 6.1.0
     hooks:
       - id: flake8
         additional_dependencies:
-          - flake8-bugbear
-          - flake8-comprehensions
-          - flake8-debugger
           - flake8-string-format
           - flake8-typing-imports
   - repo: https://github.com/PyCQA/bandit

README.rst

@@ -280,8 +280,8 @@ Barrak, A., Eghan, E.E. and Adams, B. `On the Co-evolution of ML Pipelines and S
    :target: https://pypi.org/project/dvc
    :alt: PyPI Downloads
 
-.. |Packages| image:: https://img.shields.io/github/v/release/iterative/dvc?label=deb|pkg|rpm|exe&logo=GitHub
-   :target: https://github.com/iterative/dvc/releases/latest
+.. |Packages| image:: https://img.shields.io/badge/deb|pkg|rpm|exe-blue
+   :target: https://dvc.org/doc/install
    :alt: deb|pkg|rpm|exe
 
 .. |DOI| image:: https://img.shields.io/badge/DOI-10.5281/zenodo.3677553-blue.svg

dvc/__pyinstaller/__init__.py

@@ -0,0 +1,5 @@
+import os
+
+
+def get_hook_dirs():
+    return [os.path.dirname(__file__)]

dvc/__pyinstaller/hook-asyncssh.py

@@ -1 +1,2 @@
+# ruff: noqa: N999
 hiddenimports = ["win32timezone"]

dvc/__pyinstaller/hook-celery.py

@@ -1,3 +1,4 @@
+# ruff: noqa: N999
 # pylint: disable=import-error
 from PyInstaller.utils.hooks import collect_submodules, is_module_or_submodule
 

dvc/__pyinstaller/hook-dvc.py

@@ -1,3 +1,4 @@
+# ruff: noqa: N999
 from PyInstaller.utils.hooks import copy_metadata  # pylint:disable=import-error
 
 # needed for `dvc doctor` to show dep versions
@@ -21,6 +22,7 @@ hiddenimports = [
     "dvc_hdfs",
     "dvc_oss",
     "dvc_s3",
+    "dvc_ssh",
     "dvc_webdav",
     "dvc_webhdfs",
     # https://github.com/pypa/setuptools/issues/1963

dvc/__pyinstaller/hook-dvc.utils.flatten.py

@@ -1,3 +1,4 @@
+# ruff: noqa: N999
 from PyInstaller.utils.hooks import copy_metadata  # pylint:disable=import-error
 
 datas = copy_metadata("flatten-dict")

dvc/__pyinstaller/hook-dvc_task.py

@@ -1,3 +1,4 @@
+# ruff: noqa: N999
 # pylint: disable=import-error
 from PyInstaller.utils.hooks import collect_submodules
 

dvc/__pyinstaller/hook-fsspec.py

@@ -1 +1,2 @@
+# ruff: noqa: N999
 hiddenimports = ["fsspec.implementations.memory"]

dvc/__pyinstaller/hook-pydrive2.py

@@ -1,3 +1,4 @@
+# ruff: noqa: N999
 from PyInstaller.utils.hooks import copy_metadata  # pylint:disable=import-error
 
 datas = copy_metadata("pydrive2")

dvc/_debug.py

@@ -198,33 +198,84 @@ def debugtools(args: Optional["Namespace"] = None, **kwargs):
 def add_debugging_flags(parser):
     from argparse import SUPPRESS
 
-    parser.add_argument("--cprofile", action="store_true", default=False, help=SUPPRESS)
-    parser.add_argument("--yappi", action="store_true", default=False, help=SUPPRESS)
+    # For detailed info see:
+    # https://github.com/iterative/dvc/wiki/Debugging,-Profiling-and-Benchmarking-DVC
+    args, _ = parser.parse_known_args()
+    verbose = args.verbose
+
+    def debug_help(msg):
+        if verbose:
+            return msg
+        return SUPPRESS
+
+    parser = parser.add_argument_group("debug options")
+
+    parser.add_argument(
+        "--cprofile",
+        action="store_true",
+        default=False,
+        help=debug_help("Generate cprofile data for tools like snakeviz / tuna"),
+    )
+    parser.add_argument(
+        "--cprofile-dump", help=debug_help("Location to dump cprofile file")
+    )
+    parser.add_argument(
+        "--yappi",
+        action="store_true",
+        default=False,
+        help=debug_help(
+            "Generate a callgrind file for use with tools like "
+            "kcachegrind / qcachegrind"
+        ),
+    )
     parser.add_argument(
         "--yappi-separate-threads",
         action="store_true",
         default=False,
-        help=SUPPRESS,
+        help=debug_help("Generate one callgrind file per thread"),
     )
     parser.add_argument(
-        "--viztracer", action="store_true", default=False, help=SUPPRESS
-    )
-    parser.add_argument("--viztracer-depth", type=int, help=SUPPRESS)
-    parser.add_argument(
-        "--viztracer-async", action="store_true", default=False, help=SUPPRESS
-    )
-    parser.add_argument("--cprofile-dump", help=SUPPRESS)
-    parser.add_argument("--pdb", action="store_true", default=False, help=SUPPRESS)
-    parser.add_argument(
-        "--instrument", action="store_true", default=False, help=SUPPRESS
+        "--viztracer",
+        action="store_true",
+        default=False,
+        help=debug_help("Generate a viztracer file for use with vizviewer"),
     )
     parser.add_argument(
-        "--instrument-open", action="store_true", default=False, help=SUPPRESS
+        "--viztracer-depth",
+        type=int,
+        help=debug_help("Set viztracer maximum stack depth"),
+    )
+    parser.add_argument(
+        "--viztracer-async",
+        action="store_true",
+        default=False,
+        help=debug_help("Treat async tasks as threads"),
+    )
+    parser.add_argument(
+        "--pdb",
+        action="store_true",
+        default=False,
+        help=debug_help("Drop into the pdb/ipdb debugger on any exception"),
+    )
+    parser.add_argument(
+        "--instrument",
+        action="store_true",
+        default=False,
+        help=debug_help("Use pyinstrument CLI profiler"),
+    )
+    parser.add_argument(
+        "--instrument-open",
+        action="store_true",
+        default=False,
+        help=debug_help("Use pyinstrument web profiler"),
     )
     parser.add_argument(
         "--show-stack",
         "--ss",
         action="store_true",
         default=False,
-        help=SUPPRESS,
+        help=debug_help(
+            r"Use Ctrl+T on macOS or Ctrl+\ on Linux to print the stack "
+            "frame currently executing. Unavailable on Windows."
+        ),
     )

dvc/api/data.py

@@ -36,21 +36,22 @@ def get_url(path, repo=None, rev=None, remote=None):
     NOTE: This function does not check for the actual existence of the file or
     directory in the remote storage.
     """
-    with Repo.open(repo, rev=rev, subrepos=True, uninitialized=True) as _repo:
-        with _wrap_exceptions(_repo, path):
-            fs_path = _repo.dvcfs.from_os_path(path)
+    from dvc.config import NoRemoteError
+    from dvc_data.index import StorageKeyError
 
-            with reraise(FileNotFoundError, PathMissingError(path, repo)):
-                info = _repo.dvcfs.info(fs_path)
-
-            dvc_info = info.get("dvc_info")
-            if not dvc_info:
-                raise OutputNotFoundError(path, repo)
-
-            dvc_repo = info["repo"]  # pylint: disable=unsubscriptable-object
-            md5 = dvc_info["md5"]
-
-            return dvc_repo.cloud.get_url_for(remote, checksum=md5)
+    repo_kwargs: Dict[str, Any] = {}
+    if remote:
+        repo_kwargs["config"] = {"core": {"remote": remote}}
+    with Repo.open(
+        repo, rev=rev, subrepos=True, uninitialized=True, **repo_kwargs
+    ) as _repo:
+        index, entry = _repo.get_data_index_entry(path)
+        with reraise(
+            (StorageKeyError, ValueError),
+            NoRemoteError(f"no remote specified in {_repo}"),
+        ):
+            remote_fs, remote_path = index.storage_map.get_remote(entry)
+            return remote_fs.unstrip_protocol(remote_path)
 
 
 class _OpenContextManager(GCM):
@@ -66,7 +67,7 @@ class _OpenContextManager(GCM):
         raise AttributeError("dvc.api.open() should be used in a with statement.")
 
 
-def open(  # noqa, pylint: disable=redefined-builtin
+def open(  # noqa: A001, pylint: disable=redefined-builtin
     path: str,
     repo: Optional[str] = None,
     rev: Optional[str] = None,
@@ -74,6 +75,7 @@ def open(  # noqa, pylint: disable=redefined-builtin
     mode: str = "r",
     encoding: Optional[str] = None,
     config: Optional[Dict[str, Any]] = None,
+    remote_config: Optional[Dict[str, Any]] = None,
 ):
     """
     Opens a file tracked in a DVC project.
@@ -117,6 +119,9 @@ def open(  # noqa, pylint: disable=redefined-builtin
             Mirrors the namesake parameter in builtin `open()`_.
         config(dict, optional): config to be passed to the DVC repository.
             Defaults to None.
+        remote_config(dict, optional): remote config to be passed to the DVC
+            repository.
+            Defaults to None.
 
     Returns:
         _OpenContextManager: A context manager that generatse a corresponding
@@ -213,22 +218,27 @@ def open(  # noqa, pylint: disable=redefined-builtin
         "mode": mode,
         "encoding": encoding,
         "config": config,
+        "remote_config": remote_config,
     }
     return _OpenContextManager(_open, args, kwargs)
 
 
-def _open(path, repo=None, rev=None, remote=None, mode="r", encoding=None, config=None):
-    if remote:
-        if config is not None:
-            raise ValueError(
-                "can't specify both `remote` and `config` at the same time"
-            )
-        config = {"core": {"remote": remote}}
-
+def _open(
+    path,
+    repo=None,
+    rev=None,
+    remote=None,
+    mode="r",
+    encoding=None,
+    config=None,
+    remote_config=None,
+):
     repo_kwargs: Dict[str, Any] = {
         "subrepos": True,
         "uninitialized": True,
+        "remote": remote,
         "config": config,
+        "remote_config": remote_config,
     }
 
     with Repo.open(repo, rev=rev, **repo_kwargs) as _repo:
@@ -264,7 +274,16 @@ def _open(path, repo=None, rev=None, remote=None, mode="r", encoding=None, confi
                 raise DvcIsADirectoryError(f"'{path}' is a directory") from exc
 
 
-def read(path, repo=None, rev=None, remote=None, mode="r", encoding=None, config=None):
+def read(
+    path,
+    repo=None,
+    rev=None,
+    remote=None,
+    mode="r",
+    encoding=None,
+    config=None,
+    remote_config=None,
+):
     """
     Returns the contents of a tracked file (by DVC or Git). For Git repos, HEAD
     is used unless a rev argument is supplied. The default remote is tried
@@ -278,5 +297,6 @@ def read(path, repo=None, rev=None, remote=None, mode="r", encoding=None, config
         mode=mode,
         encoding=encoding,
         config=config,
+        remote_config=remote_config,
     ) as fd:
         return fd.read()

dvc/cachemgr.py

@@ -24,7 +24,7 @@ def _get_odb(
     if not settings:
         return None
 
-    cls, config, fs_path = get_cloud_fs(repo, **settings)
+    cls, config, fs_path = get_cloud_fs(repo.config, **settings)
     fs = fs or cls(**config)
     if prefix:
         fs_path = fs.path.join(fs_path, *prefix)
@@ -42,11 +42,8 @@ class CacheManager:
         self.config = config = repo.config["cache"]
         self._odb = {}
 
-        default = None
-        if repo and repo.local_dvc_dir:
-            default = os.path.join(repo.local_dvc_dir, self.CACHE_DIR)
-
         local = config.get("local")
+        default = self.default_local_cache_dir
 
         if local:
             settings = {"name": local}
@@ -103,6 +100,13 @@ class CacheManager:
         """
         return self.legacy.path
 
+    @property
+    def default_local_cache_dir(self) -> Optional[str]:
+        repo = self._repo
+        if repo and repo.local_dvc_dir:
+            return os.path.join(repo.local_dvc_dir, self.CACHE_DIR)
+        return None
+
 
 def migrate_2_to_3(repo: "Repo", dry: bool = False):
     """Migrate legacy 2.x objects to 3.x cache.

dvc/cli/__init__.py

@@ -192,7 +192,7 @@ def main(argv=None):  # noqa: C901, PLR0912, PLR0915
 
             from dvc import PKG, __version__
 
-            pyv = " ".join([python_implementation(), python_version()])
+            pyv = f"{python_implementation()} {python_version()}"
             pkg = f" ({PKG})" if PKG else ""
             logger.debug("v%s%s, %s on %s", __version__, pkg, pyv, platform())
             logger.debug("command: %s", " ".join(argv or sys.argv))
@@ -230,7 +230,7 @@ def main(argv=None):  # noqa: C901, PLR0912, PLR0915
         logger.exception("")
     except DvcParserError:
         ret = 254
-    except Exception as exc:  # noqa, pylint: disable=broad-except
+    except Exception as exc:  # noqa: BLE001, pylint: disable=broad-except
         # pylint: disable=no-member
         ret = _log_exceptions(exc) or 255
 

dvc/cli/parser.py

@@ -31,7 +31,6 @@ from dvc.commands import (
     install,
     ls,
     ls_url,
-    machine,
     metrics,
     move,
     params,
@@ -89,7 +88,6 @@ COMMANDS = [
     stage,
     experiments,
     check_ignore,
-    machine,
     data,
 ]
 
@@ -137,7 +135,6 @@ def get_parent_parser():
     from dvc._debug import add_debugging_flags
 
     parent_parser = argparse.ArgumentParser(add_help=False)
-    add_debugging_flags(parent_parser)
     log_level_group = parent_parser.add_mutually_exclusive_group()
     log_level_group.add_argument(
         "-q", "--quiet", action="count", default=0, help="Be quiet."
@@ -145,6 +142,7 @@ def get_parent_parser():
     log_level_group.add_argument(
         "-v", "--verbose", action="count", default=0, help="Be verbose."
     )
+    add_debugging_flags(parent_parser)
 
     return parent_parser
 

dvc/cli/utils.py

@@ -1,3 +1,27 @@
+import argparse
+
+
+class DictAction(argparse.Action):
+    def __call__(self, parser, args, values, option_string=None):  # noqa: ARG002
+        d = getattr(args, self.dest) or {}
+
+        if isinstance(values, list):
+            kvs = values
+        else:
+            kvs = [values]
+
+        for kv in kvs:
+            key, value = kv.split("=", 1)
+            if not value:
+                raise argparse.ArgumentError(
+                    self,
+                    f'Could not parse argument "{values}" as k1=v1 k2=v2 ... format',
+                )
+            d[key] = value
+
+        setattr(args, self.dest, d)
+
+
 def fix_subparsers(subparsers):
     """Workaround for bug in Python 3. See more info at:
     https://bugs.python.org/issue16308
@@ -16,9 +40,7 @@ def append_doc_link(help_message, path):
     if not path:
         return help_message
     doc_base = "https://man.dvc.org/"
-    return "{message}\nDocumentation: {link}".format(
-        message=help_message, link=format_link(doc_base + path)
-    )
+    return f"{help_message}\nDocumentation: {format_link(doc_base + path)}"
 
 
 def hide_subparsers_from_help(subparsers):

dvc/commands/checkout.py

@@ -37,6 +37,7 @@ class CmdCheckout(CmdBase):
                 force=self.args.force,
                 relink=self.args.relink,
                 recursive=self.args.recursive,
+                allow_missing=self.args.allow_missing,
             )
         except CheckoutError as _exc:
             exc = _exc
@@ -101,6 +102,12 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Recreate links or copies from cache to workspace.",
     )
+    checkout_parser.add_argument(
+        "--allow-missing",
+        action="store_true",
+        default=False,
+        help="Ignore errors if some of the files or directories are missing.",
+    )
     checkout_parser.add_argument(
         "targets",
         nargs="*",

dvc/commands/commit.py

@@ -47,7 +47,10 @@ def add_parser(subparsers, parent_parser):
         "--force",
         action="store_true",
         default=False,
-        help="Commit even if hash value for dependencies/outputs changed.",
+        help=(
+            "Commit data even if hash values for dependencies or "
+            "outputs did not change."
+        ),
     )
     commit_parser.add_argument(
         "-d",

dvc/commands/completion.py

@@ -9,6 +9,9 @@ from dvc.ui import ui
 logger = logging.getLogger(__name__)
 
 
+SUPPORTED_SHELLS = ["bash", "zsh"]
+
+
 class CmdCompletion(CmdBaseNoRepo):
     def run(self):
         import shtab
@@ -35,6 +38,6 @@ def add_parser(subparsers, parent_parser):
         "--shell",
         help="Shell syntax for completions.",
         default="bash",
-        choices=["bash", "zsh"],
+        choices=SUPPORTED_SHELLS,
     )
     completion_parser.set_defaults(func=CmdCompletion)

dvc/commands/data.py

@@ -90,9 +90,7 @@ class CmdDataStatus(CmdBase):
                     ui.write(f"  ({hint})")
 
             if isinstance(stage_status, dict):
-                items = [
-                    ": ".join([state, file]) for file, state in stage_status.items()
-                ]
+                items = [f"{state}: {file}" for file, state in stage_status.items()]
             else:
                 items = stage_status
 

dvc/commands/data_sync.py

@@ -91,6 +91,8 @@ class CmdDataFetch(CmdDataBase):
                 with_deps=self.args.with_deps,
                 recursive=self.args.recursive,
                 run_cache=self.args.run_cache,
+                max_size=self.args.max_size,
+                types=self.args.types,
             )
             self.log_summary({"fetched": processed_files_count})
         except DvcException:
@@ -322,6 +324,22 @@ def add_parser(subparsers, _parent_parser):
         default=False,
         help="Fetch run history for all stages.",
     )
+    fetch_parser.add_argument(
+        "--max-size",
+        type=int,
+        help="Fetch data files/directories that are each below specified size (bytes).",
+    )
+    fetch_parser.add_argument(
+        "--type",
+        dest="types",
+        action="append",
+        default=[],
+        help=(
+            "Only fetch data files/directories that are of a particular "
+            "type (metrics, plots)."
+        ),
+        choices=["metrics", "plots"],
+    )
     fetch_parser.set_defaults(func=CmdDataFetch)
 
     # Status

dvc/commands/destroy.py

@@ -21,7 +21,7 @@ class CmdDestroy(CmdBase):
             )
 
             if not self.args.force and not ui.confirm(statement):
-                raise DvcException(
+                raise DvcException(  # noqa: TRY301
                     "cannot destroy without a confirmation from the user."
                     " Use `-f` to force."
                 )

dvc/commands/experiments/ls.py

@@ -23,15 +23,22 @@ class CmdExperimentsList(CmdBase):
             git_remote=git_remote,
         )
 
+        from dvc.repo.experiments.utils import describe
+        from dvc.scm import Git
+
+        if name_only or sha_only:
+            names = {}
+        else:
+            assert isinstance(self.repo.scm, Git)
+            names = describe(
+                self.repo.scm,
+                (baseline for baseline in exps),
+                logger=logger,
+            )
+
         for baseline in exps:
             if not (name_only or sha_only):
-                tag_base = "refs/tags/"
-                branch_base = "refs/heads/"
-                name = baseline[:7]
-                if baseline.startswith(tag_base):
-                    name = baseline[len(tag_base) :]
-                elif baseline.startswith(branch_base):
-                    name = baseline[len(branch_base) :]
+                name = names.get(baseline) or baseline[:7]
                 ui.write(f"{name}:")
             for exp_name, rev in exps[baseline]:
                 if name_only:
@@ -52,6 +59,7 @@ def add_parser(experiments_subparsers, parent_parser):
     EXPERIMENTS_LIST_HELP = "List local and remote experiments."
     experiments_list_parser = experiments_subparsers.add_parser(
         "list",
+        aliases=["ls"],
         parents=[parent_parser],
         description=append_doc_link(EXPERIMENTS_LIST_HELP, "exp/list"),
         help=EXPERIMENTS_LIST_HELP,

dvc/commands/experiments/remove.py

@@ -50,6 +50,7 @@ def add_parser(experiments_subparsers, parent_parser):
     EXPERIMENTS_REMOVE_HELP = "Remove experiments."
     experiments_remove_parser = experiments_subparsers.add_parser(
         "remove",
+        aliases=["rm"],
         parents=[parent_parser],
         description=append_doc_link(EXPERIMENTS_REMOVE_HELP, "exp/remove"),
         help=EXPERIMENTS_REMOVE_HELP,

dvc/commands/experiments/run.py

@@ -17,7 +17,6 @@ class CmdExperimentsRun(CmdRepro):
             jobs=self.args.jobs,
             params=self.args.set_param,
             tmp_dir=self.args.tmp_dir,
-            machine=self.args.machine,
             copy_paths=self.args.copy_paths,
             message=self.args.message,
             **self._common_kwargs,
@@ -90,15 +89,6 @@ def _add_run_common(parser):
             "your workspace."
         ),
     )
-    parser.add_argument(
-        "--machine",
-        default=None,
-        help=argparse.SUPPRESS,
-        # help=(
-        #     "Run this experiment on the specified 'dvc machine' instance."
-        # )
-        # metavar="<name>",
-    )
     parser.add_argument(
         "-C",
         "--copy-paths",

dvc/commands/experiments/save.py

@@ -27,10 +27,6 @@ class CmdExperimentsSave(CmdBase):
         else:
             name = self.repo.experiments.get_exact_name([ref])[ref]
             ui.write(f"Experiment has been saved as: {name}")
-            ui.write(
-                "\nTo promote an experiment to a Git branch run:\n\n"
-                "\tdvc exp branch <exp> <branch>\n"
-            )
 
         return 0
 

dvc/commands/get.py

@@ -3,7 +3,7 @@ import logging
 
 from dvc.cli import completion
 from dvc.cli.command import CmdBaseNoRepo
-from dvc.cli.utils import append_doc_link
+from dvc.cli.utils import DictAction, append_doc_link
 from dvc.exceptions import DvcException
 
 logger = logging.getLogger(__name__)
@@ -37,6 +37,9 @@ class CmdGet(CmdBaseNoRepo):
                 rev=self.args.rev,
                 jobs=self.args.jobs,
                 force=self.args.force,
+                config=self.args.config,
+                remote=self.args.remote,
+                remote_config=self.args.remote_config,
             )
             return 0
         except CloneError:
@@ -102,4 +105,27 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Override local file or folder if exists.",
     )
+    get_parser.add_argument(
+        "--config",
+        type=str,
+        help=(
+            "Path to a config file that will be merged with the config "
+            "in the target repository."
+        ),
+    )
+    get_parser.add_argument(
+        "--remote",
+        type=str,
+        help="Remote name to set as a default in the target repository.",
+    )
+    get_parser.add_argument(
+        "--remote-config",
+        type=str,
+        nargs="*",
+        action=DictAction,
+        help=(
+            "Remote config options to merge with a remote's config (default or one "
+            "specified by '--remote') in the target repository."
+        ),
+    )
     get_parser.set_defaults(func=CmdGet)

dvc/commands/get_url.py

@@ -3,7 +3,7 @@ import logging
 
 from dvc.cli import completion
 from dvc.cli.command import CmdBaseNoRepo
-from dvc.cli.utils import append_doc_link
+from dvc.cli.utils import DictAction, append_doc_link
 from dvc.exceptions import DvcException
 
 logger = logging.getLogger(__name__)
@@ -19,6 +19,7 @@ class CmdGetUrl(CmdBaseNoRepo):
                 out=self.args.out,
                 jobs=self.args.jobs,
                 force=self.args.force,
+                fs_config=self.args.fs_config,
             )
             return 0
         except DvcException:
@@ -58,4 +59,11 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Override local file or folder if exists.",
     )
+    get_parser.add_argument(
+        "--fs-config",
+        type=str,
+        nargs="*",
+        action=DictAction,
+        help="Config options for the target url.",
+    )
     get_parser.set_defaults(func=CmdGetUrl)

dvc/commands/imp.py

@@ -3,7 +3,7 @@ import logging
 
 from dvc.cli import completion
 from dvc.cli.command import CmdBase
-from dvc.cli.utils import append_doc_link
+from dvc.cli.utils import DictAction, append_doc_link
 from dvc.exceptions import DvcException
 
 logger = logging.getLogger(__name__)
@@ -22,6 +22,9 @@ class CmdImport(CmdBase):
                 no_exec=self.args.no_exec,
                 no_download=self.args.no_download,
                 jobs=self.args.jobs,
+                config=self.args.config,
+                remote=self.args.remote,
+                remote_config=self.args.remote_config,
             )
         except CloneError:
             logger.exception("failed to import '%s'", self.args.path)
@@ -94,4 +97,27 @@ def add_parser(subparsers, parent_parser):
         ),
         metavar="<number>",
     )
+    import_parser.add_argument(
+        "--config",
+        type=str,
+        help=(
+            "Path to a config file that will be merged with the config "
+            "in the target repository."
+        ),
+    )
+    import_parser.add_argument(
+        "--remote",
+        type=str,
+        help="Remote name to set as a default in the target repository.",
+    )
+    import_parser.add_argument(
+        "--remote-config",
+        type=str,
+        nargs="*",
+        action=DictAction,
+        help=(
+            "Remote config options to merge with a remote's config (default or one "
+            "specified by '--remote') in the target repository."
+        ),
+    )
     import_parser.set_defaults(func=CmdImport)

dvc/commands/imp_url.py

@@ -3,7 +3,7 @@ import logging
 
 from dvc.cli import completion
 from dvc.cli.command import CmdBase
-from dvc.cli.utils import append_doc_link
+from dvc.cli.utils import DictAction, append_doc_link
 from dvc.exceptions import DvcException
 
 logger = logging.getLogger(__name__)
@@ -22,6 +22,7 @@ class CmdImportUrl(CmdBase):
                 jobs=self.args.jobs,
                 force=self.args.force,
                 version_aware=self.args.version_aware,
+                fs_config=self.args.fs_config,
             )
         except DvcException:
             logger.exception(
@@ -114,4 +115,11 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Import using cloud versioning. Implied if the URL contains a version ID.",
     )
+    import_parser.add_argument(
+        "--fs-config",
+        type=str,
+        nargs="*",
+        action=DictAction,
+        help="Config options for the target url.",
+    )
     import_parser.set_defaults(func=CmdImportUrl)

dvc/commands/ls/__init__.py

@@ -3,7 +3,7 @@ import logging
 
 from dvc.cli import completion
 from dvc.cli.command import CmdBaseNoRepo
-from dvc.cli.utils import append_doc_link
+from dvc.cli.utils import DictAction, append_doc_link
 from dvc.commands.ls.ls_colors import LsColors
 from dvc.exceptions import DvcException
 from dvc.ui import ui
@@ -11,7 +11,18 @@ from dvc.ui import ui
 logger = logging.getLogger(__name__)
 
 
-def _prettify(entries, with_color=False):
+def _format_entry(entry, fmt):
+    from dvc.utils.humanize import naturalsize
+
+    size = entry.get("size")
+    if size is None:
+        size = ""
+    else:
+        size = naturalsize(size)
+    return size, fmt(entry)
+
+
+def show_entries(entries, with_color=False, with_size=False):
     if with_color:
         ls_colors = LsColors()
         fmt = ls_colors.format
@@ -20,7 +31,12 @@ def _prettify(entries, with_color=False):
         def fmt(entry):
             return entry["path"]
 
-    return [fmt(entry) for entry in entries]
+    if with_size:
+        ui.table([_format_entry(entry, fmt) for entry in entries])
+        return
+
+    # NOTE: this is faster than ui.table for very large number of entries
+    ui.write("\n".join(fmt(entry) for entry in entries))
 
 
 class CmdList(CmdBaseNoRepo):
@@ -34,12 +50,14 @@ class CmdList(CmdBaseNoRepo):
                 rev=self.args.rev,
                 recursive=self.args.recursive,
                 dvc_only=self.args.dvc_only,
+                config=self.args.config,
+                remote=self.args.remote,
+                remote_config=self.args.remote_config,
             )
             if self.args.json:
                 ui.write_json(entries)
             elif entries:
-                entries = _prettify(entries, with_color=True)
-                ui.write("\n".join(entries))
+                show_entries(entries, with_color=True, with_size=self.args.size)
             return 0
         except DvcException:
             logger.exception("failed to list '%s'", self.args.url)
@@ -80,6 +98,34 @@ def add_parser(subparsers, parent_parser):
         help="Git revision (e.g. SHA, branch, tag)",
         metavar="<commit>",
     )
+    list_parser.add_argument(
+        "--config",
+        type=str,
+        help=(
+            "Path to a config file that will be merged with the config "
+            "in the target repository."
+        ),
+    )
+    list_parser.add_argument(
+        "--remote",
+        type=str,
+        help="Remote name to set as a default in the target repository.",
+    )
+    list_parser.add_argument(
+        "--remote-config",
+        type=str,
+        nargs="*",
+        action=DictAction,
+        help=(
+            "Remote config options to merge with a remote's config (default or one "
+            "specified by '--remote') in the target repository."
+        ),
+    )
+    list_parser.add_argument(
+        "--size",
+        action="store_true",
+        help="Show sizes.",
+    )
     list_parser.add_argument(
         "path",
         nargs="?",

dvc/commands/ls_url.py

@@ -2,10 +2,9 @@ import argparse
 import logging
 
 from dvc.cli.command import CmdBaseNoRepo
-from dvc.cli.utils import append_doc_link
-from dvc.ui import ui
+from dvc.cli.utils import DictAction, append_doc_link
 
-from .ls import _prettify
+from .ls import show_entries
 
 logger = logging.getLogger(__name__)
 
@@ -14,10 +13,13 @@ class CmdListUrl(CmdBaseNoRepo):
     def run(self):
         from dvc.repo import Repo
 
-        entries = Repo.ls_url(self.args.url, recursive=self.args.recursive)
+        entries = Repo.ls_url(
+            self.args.url,
+            recursive=self.args.recursive,
+            fs_config=self.args.fs_config,
+        )
         if entries:
-            entries = _prettify(entries, with_color=True)
-            ui.write("\n".join(entries))
+            show_entries(entries, with_color=True, with_size=self.args.size)
         return 0
 
 
@@ -40,4 +42,16 @@ def add_parser(subparsers, parent_parser):
         action="store_true",
         help="Recursively list files.",
     )
+    lsurl_parser.add_argument(
+        "--size",
+        action="store_true",
+        help="Show sizes.",
+    )
+    lsurl_parser.add_argument(
+        "--fs-config",
+        type=str,
+        nargs="*",
+        action=DictAction,
+        help="Config options for the target url.",
+    )
     lsurl_parser.set_defaults(func=CmdListUrl)

dvc/commands/machine.py

@@ -1,490 +0,0 @@
-import argparse
-from typing import Dict, List
-
-from dvc.cli.command import CmdBase
-from dvc.cli.utils import append_doc_link, fix_subparsers
-from dvc.commands.config import CmdConfig
-from dvc.compare import TabularData
-from dvc.config import ConfigError
-from dvc.exceptions import DvcException
-from dvc.ui import ui
-from dvc.utils import format_link
-
-
-class MachineDisabledError(ConfigError):
-    def __init__(self):
-        super().__init__("Machine feature is disabled")
-
-
-class CmdMachineConfig(CmdConfig):
-    def __init__(self, args):
-        super().__init__(args)
-        if not self.config["feature"].get("machine", False):
-            raise MachineDisabledError
-
-        if getattr(self.args, "name", None):
-            self.args.name = self.args.name.lower()
-
-    def _check_exists(self, conf):
-        if self.args.name not in conf["machine"]:
-            raise ConfigError(f"machine '{self.args.name}' doesn't exist.")
-
-
-class CmdMachineAdd(CmdMachineConfig):
-    def run(self):
-        from dvc.machine import validate_name
-
-        validate_name(self.args.name)
-
-        if self.args.default:
-            ui.write(f"Setting '{self.args.name}' as a default machine.")
-
-        with self.config.edit(self.args.level) as conf:
-            if self.args.name in conf["machine"] and not self.args.force:
-                raise ConfigError(
-                    "machine '{}' already exists. Use `-f|--force` to "
-                    "overwrite it.".format(self.args.name)
-                )
-
-            conf["machine"][self.args.name] = {"cloud": self.args.cloud}
-            if self.args.default:
-                conf["core"]["machine"] = self.args.name
-
-        return 0
-
-
-class CmdMachineRemove(CmdMachineConfig):
-    def run(self):
-        with self.config.edit(self.args.level) as conf:
-            self._check_exists(conf)
-            del conf["machine"][self.args.name]
-
-        up_to_level = self.args.level or "repo"
-        # Remove core.machine refs to this machine in any shadowing configs
-        for level in reversed(self.config.LEVELS):
-            with self.config.edit(level) as conf:
-                if conf["core"].get("machine") == self.args.name:
-                    del conf["core"]["machine"]
-
-            if level == up_to_level:
-                break
-
-        return 0
-
-
-class CmdMachineList(CmdMachineConfig):
-    TABLE_COLUMNS = [
-        "name",
-        "cloud",
-        "region",
-        "image",
-        "spot",
-        "spot_price",
-        "instance_hdd_size",
-        "instance_type",
-        "ssh_private",
-        "startup_script",
-    ]
-
-    PRIVATE_COLUMNS = ["ssh_private", "startup_script"]
-
-    def _hide_private(self, conf):
-        for machine in conf:
-            for column in self.PRIVATE_COLUMNS:
-                if column in conf[machine]:
-                    conf[machine][column] = "***"
-
-    def _show_origin(self):
-        levels = [self.args.level] if self.args.level else self.config.LEVELS
-        for level in levels:
-            conf = self.config.read(level)["machine"]
-            if self.args.name:
-                conf = conf.get(self.args.name, {})
-            self._hide_private(conf)
-            prefix = self._config_file_prefix(True, self.config, level)
-            configs = list(self._format_config(conf, prefix))
-            if configs:
-                ui.write("\n".join(configs))
-
-    def _show_table(self):
-        td = TabularData(self.TABLE_COLUMNS, fill_value="-")
-        conf = self.config.read()["machine"]
-        if self.args.name:
-            conf = {self.args.name: conf.get(self.args.name, {})}
-        self._hide_private(conf)
-        for machine, machine_config in conf.items():
-            machine_config["name"] = machine
-            td.row_from_dict(machine_config)
-        td.dropna("cols", "all")
-        td.render()
-
-    def run(self):
-        if self.args.show_origin:
-            self._show_origin()
-        else:
-            self._show_table()
-        return 0
-
-
-class CmdMachineModify(CmdMachineConfig):
-    def run(self):
-        from dvc.config import merge
-
-        with self.config.edit(self.args.level) as conf:
-            merged = self.config.load_config_to_level(self.args.level)
-            merge(merged, conf)
-            self._check_exists(merged)
-
-            if self.args.name not in conf["machine"]:
-                conf["machine"][self.args.name] = {}
-            section = conf["machine"][self.args.name]
-            if self.args.unset:
-                section.pop(self.args.option, None)
-            else:
-                section[self.args.option] = self.args.value
-        return 0
-
-
-class CmdMachineRename(CmdBase):
-    def _check_exists(self, conf):
-        if self.args.name not in conf["machine"]:
-            raise ConfigError(f"machine '{self.args.name}' doesn't exist.")
-
-    def _rename_default(self, conf):
-        if conf["core"].get("machine") == self.args.name:
-            conf["core"]["machine"] = self.args.new
-
-    def _check_before_rename(self):
-        from dvc.machine import validate_name
-
-        validate_name(self.args.new)
-
-        all_config = self.config.load_config_to_level(None)
-        if self.args.new in all_config.get("machine", {}):
-            raise ConfigError(
-                f"Rename failed. Machine '{self.args.new}' already exists."
-            )
-        ui.write(f"Rename machine '{self.args.name}' to '{self.args.new}'.")
-
-    def run(self):
-        self._check_before_rename()
-
-        with self.config.edit(self.args.level) as conf:
-            self._check_exists(conf)
-            conf["machine"][self.args.new] = conf["machine"][self.args.name]
-            try:
-                assert self.repo.machine
-                self.repo.machine.rename(self.args.name, self.args.new)
-            except DvcException as error:
-                del conf["machine"][self.args.new]
-                raise ConfigError("terraform rename failed") from error
-            del conf["machine"][self.args.name]
-            self._rename_default(conf)
-
-        up_to_level = self.args.level or "repo"
-        for level in reversed(self.config.LEVELS):
-            if level == up_to_level:
-                break
-            with self.config.edit(level) as level_conf:
-                self._rename_default(level_conf)
-
-        return 0
-
-
-class CmdMachineDefault(CmdMachineConfig):
-    def run(self):
-        if self.args.name is None and not self.args.unset:
-            conf = self.config.read(self.args.level)
-            try:
-                ui.write(conf["core"]["machine"])
-            except KeyError:
-                ui.write("No default machine set")
-                return 1
-        else:
-            with self.config.edit(self.args.level) as conf:
-                if self.args.unset:
-                    conf["core"].pop("machine", None)
-                else:
-                    merged_conf = self.config.load_config_to_level(self.args.level)
-                    if (
-                        self.args.name in conf["machine"]
-                        or self.args.name in merged_conf["machine"]
-                    ):
-                        conf["core"]["machine"] = self.args.name
-                    else:
-                        raise ConfigError(
-                            "default machine must be present in machine list."
-                        )
-        return 0
-
-
-class CmdMachineCreate(CmdBase):
-    def run(self):
-        if self.repo.machine is None:
-            raise MachineDisabledError
-
-        self.repo.machine.create(self.args.name)
-        return 0
-
-
-class CmdMachineStatus(CmdBase):
-    INSTANCE_FIELD = ["name", "instance", "status"]
-    SHOWN_FIELD = [
-        "cloud",
-        "instance_ip",
-        "instance_type",
-        "instance_hdd_size",
-        "instance_gpu",
-    ]
-    FILL_VALUE = "-"
-
-    def _add_row(
-        self,
-        name: str,
-        all_status: List[Dict],
-        td: TabularData,
-    ):
-        if not all_status:
-            row = [
-                name,
-                self.FILL_VALUE,
-                "offline",
-            ]  # back to `None` after #7167
-            td.append(row)
-        for i, status in enumerate(all_status, start=1):
-            row = [name, f"num_{i}", "running" if status else "offline"]
-            for field in self.SHOWN_FIELD:
-                value = str(status.get(field, ""))
-                row.append(value)
-            td.append(row)
-
-    def run(self):
-        if self.repo.machine is None:
-            raise MachineDisabledError
-
-        td = TabularData(
-            self.INSTANCE_FIELD + self.SHOWN_FIELD, fill_value=self.FILL_VALUE
-        )
-
-        if self.args.name:
-            all_status = list(self.repo.machine.status(self.args.name))
-            self._add_row(self.args.name, all_status, td)
-        else:
-            name_set = set()
-            for level in self.repo.config.LEVELS:
-                conf = self.repo.config.read(level)["machine"]
-                name_set.update(conf.keys())
-            name_list = list(name_set)
-            for name in sorted(name_list):
-                all_status = list(self.repo.machine.status(name))
-                self._add_row(name, all_status, td)
-
-        td.dropna("cols", "all")
-        td.render()
-        return 0
-
-
-class CmdMachineDestroy(CmdBase):
-    def run(self):
-        if self.repo.machine is None:
-            raise MachineDisabledError
-
-        self.repo.machine.destroy(self.args.name)
-        return 0
-
-
-class CmdMachineSsh(CmdBase):
-    def run(self):
-        if self.repo.machine is None:
-            raise MachineDisabledError
-
-        self.repo.machine.run_shell(self.args.name)
-        return 0
-
-
-def add_parser(subparsers, parent_parser):  # noqa: PLR0915
-    from dvc.commands.config import parent_config_parser
-
-    machine_HELP = "Set up and manage cloud machines."
-    machine_parser = subparsers.add_parser(
-        "machine",
-        parents=[parent_parser],
-        description=append_doc_link(machine_HELP, "machine"),
-        # NOTE: suppress help during development to hide command
-        # help=machine_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-
-    machine_subparsers = machine_parser.add_subparsers(
-        dest="cmd",
-        help="Use `dvc machine CMD --help` for command-specific help.",
-    )
-
-    fix_subparsers(machine_subparsers)
-
-    machine_ADD_HELP = "Add a new data machine."
-    machine_add_parser = machine_subparsers.add_parser(
-        "add",
-        parents=[parent_config_parser, parent_parser],
-        description=append_doc_link(machine_ADD_HELP, "machine/add"),
-        help=machine_ADD_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_add_parser.add_argument("name", help="Name of the machine")
-    machine_add_parser.add_argument(
-        "cloud",
-        help="Machine cloud. See full list of supported clouds at {}".format(
-            format_link(
-                "https://github.com/iterative/terraform-provider-iterative#machine"
-            )
-        ),
-    )
-    machine_add_parser.add_argument(
-        "-d",
-        "--default",
-        action="store_true",
-        default=False,
-        help="Set as default machine.",
-    )
-    machine_add_parser.add_argument(
-        "-f",
-        "--force",
-        action="store_true",
-        default=False,
-        help="Force overwriting existing configs",
-    )
-    machine_add_parser.set_defaults(func=CmdMachineAdd)
-
-    machine_DEFAULT_HELP = "Set/unset the default machine."
-    machine_default_parser = machine_subparsers.add_parser(
-        "default",
-        parents=[parent_config_parser, parent_parser],
-        description=append_doc_link(machine_DEFAULT_HELP, "machine/default"),
-        help=machine_DEFAULT_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_default_parser.add_argument("name", nargs="?", help="Name of the machine")
-    machine_default_parser.add_argument(
-        "-u",
-        "--unset",
-        action="store_true",
-        default=False,
-        help="Unset default machine.",
-    )
-    machine_default_parser.set_defaults(func=CmdMachineDefault)
-
-    machine_LIST_HELP = "List the configuration of one/all machines."
-    machine_list_parser = machine_subparsers.add_parser(
-        "list",
-        parents=[parent_config_parser, parent_parser],
-        description=append_doc_link(machine_LIST_HELP, "machine/list"),
-        help=machine_LIST_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_list_parser.add_argument(
-        "--show-origin",
-        default=False,
-        action="store_true",
-        help="Show the source file containing each config value.",
-    )
-    machine_list_parser.add_argument(
-        "name",
-        nargs="?",
-        type=str,
-        help="name of machine to specify",
-    )
-    machine_list_parser.set_defaults(func=CmdMachineList)
-    machine_MODIFY_HELP = "Modify the configuration of an machine."
-    machine_modify_parser = machine_subparsers.add_parser(
-        "modify",
-        parents=[parent_config_parser, parent_parser],
-        description=append_doc_link(machine_MODIFY_HELP, "machine/modify"),
-        help=machine_MODIFY_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_modify_parser.add_argument("name", help="Name of the machine")
-    machine_modify_parser.add_argument("option", help="Name of the option to modify.")
-    machine_modify_parser.add_argument(
-        "value", nargs="?", help="(optional) Value of the option."
-    )
-    machine_modify_parser.add_argument(
-        "-u",
-        "--unset",
-        default=False,
-        action="store_true",
-        help="Unset option.",
-    )
-    machine_modify_parser.set_defaults(func=CmdMachineModify)
-
-    machine_RENAME_HELP = "Rename a machine "
-    machine_rename_parser = machine_subparsers.add_parser(
-        "rename",
-        parents=[parent_config_parser, parent_parser],
-        description=append_doc_link(machine_RENAME_HELP, "remote/rename"),
-        help=machine_RENAME_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_rename_parser.add_argument("name", help="Machine to be renamed")
-    machine_rename_parser.add_argument("new", help="New name of the machine")
-    machine_rename_parser.set_defaults(func=CmdMachineRename)
-
-    machine_REMOVE_HELP = "Remove an machine."
-    machine_remove_parser = machine_subparsers.add_parser(
-        "remove",
-        parents=[parent_config_parser, parent_parser],
-        description=append_doc_link(machine_REMOVE_HELP, "machine/remove"),
-        help=machine_REMOVE_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_remove_parser.add_argument("name", help="Name of the machine to remove.")
-    machine_remove_parser.set_defaults(func=CmdMachineRemove)
-
-    machine_CREATE_HELP = "Create and start a machine instance."
-    machine_create_parser = machine_subparsers.add_parser(
-        "create",
-        parents=[parent_parser],
-        description=append_doc_link(machine_CREATE_HELP, "machine/create"),
-        help=machine_CREATE_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_create_parser.add_argument("name", help="Name of the machine to create.")
-    machine_create_parser.set_defaults(func=CmdMachineCreate)
-
-    machine_STATUS_HELP = "List the status of running instances for one/all machines."
-    machine_status_parser = machine_subparsers.add_parser(
-        "status",
-        parents=[parent_parser],
-        description=append_doc_link(machine_STATUS_HELP, "machine/status"),
-        help=machine_STATUS_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_status_parser.add_argument(
-        "name", nargs="?", help="(optional) Name of the machine."
-    )
-    machine_status_parser.set_defaults(func=CmdMachineStatus)
-
-    machine_DESTROY_HELP = "Destroy an machine instance."
-    machine_destroy_parser = machine_subparsers.add_parser(
-        "destroy",
-        parents=[parent_parser],
-        description=append_doc_link(machine_DESTROY_HELP, "machine/destroy"),
-        help=machine_DESTROY_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_destroy_parser.add_argument(
-        "name", help="Name of the machine instance to destroy."
-    )
-    machine_destroy_parser.set_defaults(func=CmdMachineDestroy)
-
-    machine_SSH_HELP = "Connect to a machine via SSH."
-    machine_ssh_parser = machine_subparsers.add_parser(
-        "ssh",
-        parents=[parent_parser],
-        description=append_doc_link(machine_SSH_HELP, "machine/ssh"),
-        help=machine_SSH_HELP,
-        formatter_class=argparse.RawDescriptionHelpFormatter,
-    )
-    machine_ssh_parser.add_argument(
-        "name", help="Name of the machine instance to connect to."
-    )
-    machine_ssh_parser.set_defaults(func=CmdMachineSsh)

dvc/commands/move.py

@@ -30,6 +30,7 @@ def add_parser(subparsers, parent_parser):
 
     move_parser = subparsers.add_parser(
         "move",
+        aliases=["mv"],
         parents=[parent_parser],
         description=append_doc_link(MOVE_DESCRIPTION, "move"),
         help=MOVE_HELP,

dvc/commands/remote.py

@@ -30,8 +30,8 @@ class CmdRemoteAdd(CmdRemote):
         with self.config.edit(self.args.level) as conf:
             if self.args.name in conf["remote"] and not self.args.force:
                 raise ConfigError(
-                    "remote '{}' already exists. Use `-f|--force` to "
-                    "overwrite it.".format(self.args.name)
+                    f"remote '{self.args.name}' already exists. Use `-f|--force` to "
+                    "overwrite it."
                 )
 
             conf["remote"][self.args.name] = {"url": self.args.url}
@@ -127,9 +127,7 @@ class CmdRemoteRename(CmdRemote):
         all_config = self.config.load_config_to_level(None)
         if self.args.new in all_config.get("remote", {}):
             raise ConfigError(
-                "Rename failed. Remote name '{}' already exists.".format(
-                    {self.args.new}
-                )
+                f"Rename failed. Remote name {self.args.new!r} already exists."
             )
 
         with self.config.edit(self.args.level) as conf:

dvc/commands/remove.py

@@ -26,6 +26,7 @@ def add_parser(subparsers, parent_parser):
     )
     remove_parser = subparsers.add_parser(
         "remove",
+        aliases=["rm"],
         parents=[parent_parser],
         description=append_doc_link(REMOVE_HELP, "remove"),
         help=REMOVE_HELP,

dvc/commands/repro.py

@@ -33,6 +33,7 @@ class CmdRepro(CmdBase):
             "force_downstream": self.args.force_downstream,
             "pull": self.args.pull,
             "allow_missing": self.args.allow_missing,
+            "on_error": self.args.on_error,
         }
 
     @property
@@ -137,6 +138,26 @@ and then the stage name name.
             "Only print the commands that would be executed without actually executing."
         ),
     )
+    repro_parser.add_argument(
+        "-k",
+        "--keep-going",
+        action="store_const",
+        default="fail",
+        const="keep-going",
+        dest="on_error",
+        help=(
+            "Continue executing, skipping stages having dependencies "
+            "on the failed stages"
+        ),
+    )
+    repro_parser.add_argument(
+        "--ignore-errors",
+        action="store_const",
+        default="fail",
+        const="ignore",
+        dest="on_error",
+        help="Ignore errors from stages.",
+    )
 
 
 def add_parser(subparsers, parent_parser):

dvc/commands/stage.py

@@ -195,12 +195,6 @@ def _add_common_args(parser):
         help="Declare output file or directory (do not put into DVC cache).",
         metavar="<filename>",
     ).complete = completion.FILE
-    parser.add_argument(
-        "--external",
-        action="store_true",
-        default=False,
-        help="Allow outputs that are outside of the DVC repository.",
-    )
     parser.add_argument(
         "--outs-persist",
         action="append",

dvc/compare.py

@@ -57,7 +57,7 @@ class TabularData(MutableSequence[Sequence["CellT"]]):
         return self._columns[name]
 
     def items(self) -> ItemsView[str, Column]:
-        projection = {k: self.column(k) for k in self.keys()}  # noqa: SIM118
+        projection = {k: self.column(k) for k in self.keys()}
         return projection.items()
 
     def keys(self) -> List[str]:
@@ -171,8 +171,7 @@ class TabularData(MutableSequence[Sequence["CellT"]]):
                 self.add_column(key)
 
         row: List["CellT"] = [
-            with_value(d.get(key), self._fill_value)
-            for key in self.keys()  # noqa: SIM118
+            with_value(d.get(key), self._fill_value) for key in self.keys()
         ]
         self.append(row)
 
@@ -227,7 +226,7 @@ class TabularData(MutableSequence[Sequence["CellT"]]):
             to_drop -= match_line
 
         if axis == "rows":
-            for name in self.keys():  # noqa: SIM118
+            for name in self.keys():
                 self._columns[name] = Column(
                     [x for n, x in enumerate(self._columns[name]) if n not in to_drop]
                 )
@@ -275,7 +274,7 @@ class TabularData(MutableSequence[Sequence["CellT"]]):
                 else:
                     unique_rows.append(tuple_row)
 
-            for name in self.keys():  # noqa: SIM118
+            for name in self.keys():
                 self._columns[name] = Column(
                     [
                         x

dvc/config.py

@@ -90,20 +90,30 @@ class Config(dict):
     def __init__(
         self,
         dvc_dir: Optional["StrPath"] = None,
+        local_dvc_dir: Optional["StrPath"] = None,
         validate: bool = True,
         fs: Optional["FileSystem"] = None,
         config: Optional["DictStrAny"] = None,
+        remote: Optional[str] = None,
+        remote_config: Optional["DictStrAny"] = None,
     ):  # pylint: disable=super-init-not-called
         from dvc.fs import LocalFileSystem
 
+        dvc_dir = os.fspath(dvc_dir) if dvc_dir else None
         self.dvc_dir = dvc_dir
         self.wfs = LocalFileSystem()
         self.fs = fs or self.wfs
 
         if dvc_dir:
-            self.dvc_dir = self.fs.path.abspath(self.fs.path.realpath(dvc_dir))
+            self.dvc_dir = self.fs.path.abspath(dvc_dir)
 
-        self.load(validate=validate, config=config)
+        self.local_dvc_dir = local_dvc_dir
+        if not fs and not local_dvc_dir:
+            self.local_dvc_dir = dvc_dir
+
+        self.load(
+            validate=validate, config=config, remote=remote, remote_config=remote_config
+        )
 
     @classmethod
     def from_cwd(cls, fs: Optional["FileSystem"] = None, **kwargs):
@@ -136,7 +146,9 @@ class Config(dict):
 
         if self.dvc_dir is not None:
             files["repo"] = self.fs.path.join(self.dvc_dir, self.CONFIG)
-            files["local"] = self.fs.path.join(self.dvc_dir, self.CONFIG_LOCAL)
+
+        if self.local_dvc_dir is not None:
+            files["local"] = self.wfs.path.join(self.local_dvc_dir, self.CONFIG_LOCAL)
 
         return files
 
@@ -154,7 +166,16 @@ class Config(dict):
         with open(config_file, "w+", encoding="utf-8"):
             return Config(dvc_dir)
 
-    def load(self, validate: bool = True, config: Optional["DictStrAny"] = None):
+    def merge(self, config):
+        merge(self, config)
+
+    def load(
+        self,
+        validate: bool = True,
+        config: Optional["DictStrAny"] = None,
+        remote: Optional[str] = None,
+        remote_config: Optional["DictStrAny"] = None,
+    ):
         """Loads config from all the config files.
 
         Raises:
@@ -169,6 +190,17 @@ class Config(dict):
             conf = self.validate(conf)
 
         self.clear()
+
+        if remote:
+            conf["core"]["remote"] = remote
+
+        if remote_config:
+            remote = remote or conf["core"].get("remote")
+            if not remote:
+                raise ValueError("Missing remote name")
+
+            merge(conf, {"remote": {remote: remote_config}})
+
         self.update(conf)
 
     def _get_fs(self, level):
@@ -176,23 +208,32 @@ class Config(dict):
         # the repo.
         return self.fs if level == "repo" else self.wfs
 
-    def _load_config(self, level):
+    @staticmethod
+    def load_file(path, fs=None) -> dict:
         from configobj import ConfigObj, ConfigObjError
 
+        from dvc.fs import localfs
+
+        fs = fs or localfs
+
+        with fs.open(path) as fobj:
+            try:
+                conf_obj = ConfigObj(fobj)
+            except UnicodeDecodeError as exc:
+                raise ConfigError(str(exc)) from exc
+            except ConfigObjError as exc:
+                raise ConfigError(str(exc)) from exc
+
+        return _parse_named(_lower_keys(conf_obj.dict()))
+
+    def _load_config(self, level):
         filename = self.files[level]
         fs = self._get_fs(level)
 
-        if fs.exists(filename):
-            with fs.open(filename) as fobj:
-                try:
-                    conf_obj = ConfigObj(fobj)
-                except UnicodeDecodeError as exc:
-                    raise ConfigError(str(exc)) from exc
-                except ConfigObjError as exc:
-                    raise ConfigError(str(exc)) from exc
-        else:
-            conf_obj = ConfigObj()
-        return _parse_named(_lower_keys(conf_obj.dict()))
+        try:
+            return self.load_file(filename, fs=fs)
+        except FileNotFoundError:
+            return {}
 
     def _save_config(self, level, conf_dict):
         from configobj import ConfigObj
@@ -220,21 +261,30 @@ class Config(dict):
         return conf
 
     @staticmethod
-    def _load_paths(conf, filename):
-        abs_conf_dir = os.path.abspath(os.path.dirname(filename))
+    def _resolve(conf_dir, path):
+        from .config_schema import ExpPath, RelPath
 
-        def resolve(path):
-            from .config_schema import RelPath
+        if re.match(r"\w+://", path):
+            return path
 
-            if os.path.isabs(path) or re.match(r"\w+://", path):
-                return path
+        if os.path.isabs(path):
+            return path
 
-            # on windows convert slashes to backslashes
-            # to have path compatible with abs_conf_dir
-            if os.path.sep == "\\" and "/" in path:
-                path = path.replace("/", "\\")
+        # on windows convert slashes to backslashes
+        # to have path compatible with abs_conf_dir
+        if os.path.sep == "\\" and "/" in path:
+            path = path.replace("/", "\\")
 
-            return RelPath(os.path.join(abs_conf_dir, path))
+        expanded = os.path.expanduser(path)
+        if os.path.isabs(expanded):
+            return ExpPath(expanded, path)
+
+        return RelPath(os.path.abspath(os.path.join(conf_dir, path)))
+
+    @classmethod
+    def _load_paths(cls, conf, filename):
+        conf_dir = os.path.abspath(os.path.dirname(filename))
+        resolve = partial(cls._resolve, conf_dir)
 
         return Config._map_dirs(conf, resolve)
 
@@ -243,11 +293,17 @@ class Config(dict):
         from dvc.fs import localfs
         from dvc.utils import relpath
 
-        from .config_schema import RelPath
+        from .config_schema import ExpPath, RelPath
 
         if re.match(r"\w+://", path):
             return path
 
+        if isinstance(path, ExpPath):
+            return path.def_path
+
+        if os.path.expanduser(path) != path:
+            return localfs.path.as_posix(path)
+
         if isinstance(path, RelPath) or not os.path.isabs(path):
             path = relpath(path, conf_dir)
             return localfs.path.as_posix(path)

dvc/config_schema.py

@@ -8,6 +8,7 @@ from voluptuous import (
     All,
     Any,
     Coerce,
+    Exclusive,
     Invalid,
     Lower,
     Optional,
@@ -74,8 +75,19 @@ def ByUrl(mapping):  # noqa: N802
     return validate
 
 
+class ExpPath(str):
+    __slots__ = ("def_path",)
+
+    def_path: str
+
+    def __new__(cls, string, def_path):
+        ret = super().__new__(cls, string)
+        ret.def_path = def_path
+        return ret
+
+
 class RelPath(str):
-    pass
+    __slots__ = ()
 
 
 class FeatureSchema(Schema):
@@ -324,7 +336,8 @@ SCHEMA = {
     },
     "hydra": {
         Optional("enabled", default=False): Bool,
-        "config_dir": str,
+        Exclusive("config_dir", "config_source"): str,
+        Exclusive("config_module", "config_source"): str,
         "config_name": str,
     },
     "studio": {

dvc/daemon.py

@@ -28,7 +28,9 @@ def _popen(cmd, **kwargs) -> Popen:
             os.path.abspath(os.path.dirname(__file__)), "__main__.py"
         )
         prefix += [main_entrypoint]
-    return Popen(prefix + cmd, close_fds=True, shell=False, **kwargs)  # nosec B603
+    return Popen(
+        prefix + cmd, close_fds=True, shell=False, **kwargs  # nosec B603  # noqa: S603
+    )
 
 
 def _spawn_windows(cmd, env):

dvc/dagascii.py

@@ -209,6 +209,30 @@ def draw(vertices, edges):
     Args:
         vertices (list): list of graph vertices.
         edges (list): list of graph edges.
+
+    Returns:
+        str: ASCII representation
+
+    Example:
+        >>> from dvc.dagascii import draw
+        >>> vertices = [1, 2, 3, 4]
+        >>> edges = [(1, 2), (2, 3), (2, 4), (1, 4)]
+        >>> print(draw(vertices, edges))
+        +---+     +---+
+        | 3 |     | 4 |
+        +---+    *+---+
+          *    **   *
+          *  **     *
+          * *       *
+        +---+       *
+        | 2 |      *
+        +---+     *
+             *    *
+              *  *
+               **
+             +---+
+             | 1 |
+             +---+
     """
     # pylint: disable=too-many-locals
     # NOTE: coordinates might me negative, so we need to shift

dvc/data_cloud.py

@@ -4,6 +4,7 @@ import logging
 from typing import TYPE_CHECKING, Iterable, Optional, Set, Tuple
 
 from dvc.config import NoRemoteError, RemoteConfigError
+from dvc.fs.callbacks import Callback
 from dvc.utils.objects import cached_property
 from dvc_data.hashfile.db import get_index
 from dvc_data.hashfile.transfer import TransferResult
@@ -90,7 +91,7 @@ class DataCloud:
         if name:
             from dvc.fs import get_cloud_fs
 
-            cls, config, fs_path = get_cloud_fs(self.repo, name=name)
+            cls, config, fs_path = get_cloud_fs(self.repo.config, name=name)
 
             if config.get("worktree"):
                 version_aware = config.get("version_aware")
@@ -209,15 +210,20 @@ class DataCloud:
             cache = self.repo.cache.legacy
         else:
             cache = self.repo.cache.local
-        return self.transfer(
-            cache,
-            odb,
-            objs,
-            jobs=jobs,
-            dest_index=get_index(odb),
-            cache_odb=cache,
-            validate_status=self._log_missing,
-        )
+        with Callback.as_tqdm_callback(
+            desc=f"Pushing to {odb.fs.unstrip_protocol(odb.path)}",
+            unit="file",
+        ) as cb:
+            return self.transfer(
+                cache,
+                odb,
+                objs,
+                jobs=jobs,
+                dest_index=get_index(odb),
+                cache_odb=cache,
+                validate_status=self._log_missing,
+                callback=cb,
+            )
 
     def pull(
         self,
@@ -263,16 +269,21 @@ class DataCloud:
             cache = self.repo.cache.legacy
         else:
             cache = self.repo.cache.local
-        return self.transfer(
-            odb,
-            cache,
-            objs,
-            jobs=jobs,
-            src_index=get_index(odb),
-            cache_odb=cache,
-            verify=odb.verify,
-            validate_status=self._log_missing,
-        )
+        with Callback.as_tqdm_callback(
+            desc=f"Fetching from {odb.fs.unstrip_protocol(odb.path)}",
+            unit="file",
+        ) as cb:
+            return self.transfer(
+                odb,
+                cache,
+                objs,
+                jobs=jobs,
+                src_index=get_index(odb),
+                cache_odb=cache,
+                verify=odb.verify,
+                validate_status=self._log_missing,
+                callback=cb,
+            )
 
     def status(
         self,

dvc/dependency/__init__.py

@@ -15,6 +15,7 @@ SCHEMA: Mapping[str, Any] = {
     **ARTIFACT_SCHEMA,
     **RepoDependency.REPO_SCHEMA,
     Output.PARAM_FILES: [DIR_FILES_SCHEMA],
+    Output.PARAM_FS_CONFIG: dict,
 }
 
 
@@ -36,7 +37,10 @@ def loadd_from(stage, d_list):
         p = d.pop(Output.PARAM_PATH, None)
         files = d.pop(Output.PARAM_FILES, None)
         hash_name = d.pop(Output.PARAM_HASH, None)
-        ret.append(_get(stage, p, d, files=files, hash_name=hash_name))
+        fs_config = d.pop(Output.PARAM_FS_CONFIG, None)
+        ret.append(
+            _get(stage, p, d, files=files, hash_name=hash_name, fs_config=fs_config)
+        )
     return ret
 
 

dvc/dependency/base.py

@@ -44,13 +44,10 @@ class Dependency(Output):
         if self.fs.version_aware:
             self.fs_path = self.fs.path.version_path(self.fs_path, rev)
             self.meta = self.get_meta()
-            self.def_path = self.fs.path.version_path(
-                self.def_path, self.meta.version_id
-            )
             self.fs_path = self.fs.path.version_path(self.fs_path, self.meta.version_id)
 
     def download(self, to, jobs=None):
-        fs_download(self.fs, self.fs_path, to, jobs=jobs)
+        fs_download(self.fs, self.fs_path, to.fs_path, jobs=jobs)
 
     def save(self):
         super().save()

dvc/dependency/repo.py

@@ -1,6 +1,7 @@
+from copy import deepcopy
 from typing import TYPE_CHECKING, Dict, Optional, Union
 
-from voluptuous import Required
+from voluptuous import Any, Required
 
 from dvc.utils import as_posix
 
@@ -16,16 +17,20 @@ class RepoDependency(Dependency):
     PARAM_URL = "url"
     PARAM_REV = "rev"
     PARAM_REV_LOCK = "rev_lock"
+    PARAM_CONFIG = "config"
+    PARAM_REMOTE = "remote"
 
     REPO_SCHEMA = {
         PARAM_REPO: {
             Required(PARAM_URL): str,
             PARAM_REV: str,
             PARAM_REV_LOCK: str,
+            PARAM_CONFIG: Any(str, dict),
+            PARAM_REMOTE: Any(str, dict),
         }
     }
 
-    def __init__(self, def_repo: Dict[str, str], stage: "Stage", *args, **kwargs):
+    def __init__(self, def_repo: Dict[str, Any], stage: "Stage", *args, **kwargs):
         self.def_repo = def_repo
         super().__init__(stage, *args, **kwargs)
 
@@ -60,7 +65,28 @@ class RepoDependency(Dependency):
             self.def_repo[self.PARAM_REV_LOCK] = rev
 
     def dumpd(self, **kwargs) -> Dict[str, Union[str, Dict[str, str]]]:
-        return {self.PARAM_PATH: self.def_path, self.PARAM_REPO: self.def_repo}
+        repo = {self.PARAM_URL: self.def_repo[self.PARAM_URL]}
+
+        rev = self.def_repo.get(self.PARAM_REV)
+        if rev:
+            repo[self.PARAM_REV] = self.def_repo[self.PARAM_REV]
+
+        rev_lock = self.def_repo.get(self.PARAM_REV_LOCK)
+        if rev_lock:
+            repo[self.PARAM_REV_LOCK] = rev_lock
+
+        config = self.def_repo.get(self.PARAM_CONFIG)
+        if config:
+            repo[self.PARAM_CONFIG] = config
+
+        remote = self.def_repo.get(self.PARAM_REMOTE)
+        if remote:
+            repo[self.PARAM_REMOTE] = remote
+
+        return {
+            self.PARAM_PATH: self.def_path,
+            self.PARAM_REPO: repo,
+        }
 
     def update(self, rev: Optional[str] = None):
         if rev:
@@ -77,9 +103,24 @@ class RepoDependency(Dependency):
     def _make_fs(
         self, rev: Optional[str] = None, locked: bool = True
     ) -> "DVCFileSystem":
+        from dvc.config import Config
         from dvc.fs import DVCFileSystem
 
-        config = {"cache": self.repo.config["cache"]}
+        rem = self.def_repo.get("remote")
+        if isinstance(rem, dict):
+            remote = None
+            remote_config = rem
+        else:
+            remote = rem
+            remote_config = None
+
+        conf = self.def_repo.get("config", {})
+        if isinstance(conf, dict):
+            config = deepcopy(conf)
+        else:
+            config = Config.load_file(conf)
+
+        config["cache"] = self.repo.config["cache"]
         config["cache"]["dir"] = self.repo.cache.local_cache_dir
 
         return DVCFileSystem(
@@ -87,6 +128,8 @@ class RepoDependency(Dependency):
             rev=rev or self._get_rev(locked=locked),
             subrepos=True,
             config=config,
+            remote=remote,
+            remote_config=remote_config,
         )
 
     def _get_rev(self, locked: bool = True):

dvc/dvcfile.py

@@ -97,9 +97,7 @@ class FileMixin:
         self.verify = verify
 
     def __repr__(self):
-        return "{}: {}".format(
-            self.__class__.__name__, relpath(self.path, self.repo.root_dir)
-        )
+        return f"{self.__class__.__name__}: {relpath(self.path, self.repo.root_dir)}"
 
     def __hash__(self):
         return hash(self.path)

dvc/exceptions.py

@@ -1,9 +1,12 @@
 """Exceptions raised by the dvc."""
 import errno
-from typing import Dict, List
+from typing import TYPE_CHECKING, Dict, List, Set
 
 from dvc.utils import format_link
 
+if TYPE_CHECKING:
+    from dvc.stage import Stage
+
 
 class DvcException(Exception):
     """Base class for all dvc exceptions."""
@@ -17,6 +20,10 @@ class DvcException(Exception):
 class InvalidArgumentError(ValueError, DvcException):
     """Thrown if arguments are invalid."""
 
+    def __init__(self, msg, *args):
+        self.msg = msg
+        super().__init__(msg, *args)
+
 
 class OutputDuplicationError(DvcException):
     """Thrown if a file/directory is specified as an output in more than one
@@ -27,24 +34,25 @@ class OutputDuplicationError(DvcException):
         stages (list): list of paths to stages.
     """
 
-    def __init__(self, output, stages):
+    def __init__(self, output: str, stages: Set["Stage"]):
         from funcy import first
 
         assert isinstance(output, str)
         assert all(hasattr(stage, "relpath") for stage in stages)
-        msg = ""
-        stage_names = [s.addressing for s in stages]
-        stages_str = " ".join(stage_names)
         if len(stages) == 1:
-            stage_name = first(stages)
-            msg = f"output '{output}' is already specified in {stage_name}."
+            stage = first(stages)
+            msg = (
+                f"output '{output}' is already specified in {stage}."
+                f"\nUse `dvc remove {stage.addressing}` to stop tracking the "
+                "overlapping output."
+            )
         else:
-            msg = "output '{}' is already specified in stages:\n{}".format(
-                output, "\n".join(f"\t- {s}" for s in stage_names)
+            stage_names = "\n".join(["\t- " + s.addressing for s in stages])
+            msg = (
+                f"output '{output}' is specified in:\n{stage_names}"
+                "\nUse `dvc remove` with any of the above targets to stop tracking the "
+                "overlapping output."
             )
-        msg += (
-            f"\nUse `dvc remove {stages_str}` to stop tracking the overlapping output."
-        )
         super().__init__(msg)
         self.stages = stages
         self.output = output
@@ -63,12 +71,14 @@ class OutputNotFoundError(DvcException):
         self.output = output
         self.repo = repo
         super().__init__(
-            "Unable to find DVC file with output '{path}'".format(
-                path=relpath(self.output)
-            )
+            f"Unable to find DVC file with output {relpath(self.output)!r}"
         )
 
 
+class StageNotFoundError(DvcException):
+    pass
+
+
 class StagePathAsOutputError(DvcException):
     """Thrown if directory that stage is going to be saved in is specified as
     an output of another stage.
@@ -80,11 +90,7 @@ class StagePathAsOutputError(DvcException):
 
     def __init__(self, stage, output):
         assert isinstance(output, str)
-        super().__init__(
-            "{stage} is within an output '{output}' of another stage".format(
-                stage=stage, output=output
-            )
-        )
+        super().__init__(f"{stage} is within an output {output!r} of another stage")
 
 
 class CircularDependencyError(DvcException):
@@ -156,9 +162,7 @@ class CyclicGraphError(DvcException):
 class ConfirmRemoveError(DvcException):
     def __init__(self, path):
         super().__init__(
-            "unable to remove '{}' without a confirmation. Use `-f` to force.".format(
-                path
-            )
+            f"unable to remove {path!r} without a confirmation. Use `-f` to force."
         )
 
 
@@ -167,9 +171,7 @@ class InitError(DvcException):
 
 
 class ReproductionError(DvcException):
-    def __init__(self, name):
-        self.name = name
-        super().__init__(f"failed to reproduce '{name}'")
+    pass
 
 
 class BadMetricError(DvcException):
@@ -198,7 +200,7 @@ class ETagMismatchError(DvcException):
     def __init__(self, etag, cached_etag):
         super().__init__(
             "ETag mismatch detected when copying file to cache! "
-            "(expected: '{}', actual: '{}')".format(etag, cached_etag)
+            f"(expected: '{etag}', actual: '{cached_etag}')"
         )
 
 
@@ -299,7 +301,7 @@ class URLMissingError(DvcException):
         super().__init__(f"The path '{url}' does not exist")
 
 
-class IsADirectoryError(DvcException):  # noqa,pylint:disable=redefined-builtin
+class IsADirectoryError(DvcException):  # noqa: A001,pylint:disable=redefined-builtin
     """Raised when a file operation is requested on a directory."""
 
 

dvc/fs/__init__.py

@@ -1,3 +1,4 @@
+from typing import Optional
 from urllib.parse import urlparse
 
 from dvc_http import HTTPFileSystem, HTTPSFileSystem  # noqa: F401
@@ -18,7 +19,7 @@ from dvc_objects.fs import (  # noqa: F401
     system,
     utils,
 )
-from dvc_objects.fs.base import AnyFSPath, FileSystem  # noqa: F401
+from dvc_objects.fs.base import AnyFSPath, FileSystem  # noqa: F401, TCH001
 from dvc_objects.fs.errors import (  # noqa: F401
     AuthError,
     ConfigError,
@@ -48,18 +49,45 @@ known_implementations.update(
 # pylint: enable=unused-import
 
 
-def download(fs, fs_path, to, jobs=None):
+def download(fs: "FileSystem", fs_path: str, to: str, jobs: Optional[int] = None):
     with Callback.as_tqdm_callback(
         desc=f"Downloading {fs.path.name(fs_path)}",
         unit="files",
     ) as cb:
-        fs.get(fs_path, to.fs_path, batch_size=jobs, callback=cb)
+        # NOTE: We use dvc-objects generic.copy over fs.get since it makes file
+        # download atomic and avoids fsspec glob/regex path expansion.
+        if fs.isdir(fs_path):
+            from_infos = [
+                path
+                for path in fs.find(fs_path)
+                if not path.endswith(fs.path.flavour.sep)
+            ]
+            if not from_infos:
+                return localfs.makedirs(to, exist_ok=True)
+            to_infos = [
+                localfs.path.join(to, *fs.path.relparts(info, fs_path))
+                for info in from_infos
+            ]
+        else:
+            from_infos = [fs_path]
+            to_infos = [to]
+
+        cb.set_size(len(from_infos))
+        jobs = jobs or fs.jobs
+        generic.copy(
+            fs,
+            from_infos,
+            localfs,
+            to_infos,
+            callback=cb,
+            batch_size=jobs,
+        )
 
 
 def parse_external_url(url, config=None):
     remote_config = dict(config) if config else {}
     remote_config["url"] = url
-    fs_cls, fs_config, fs_path = get_cloud_fs(None, **remote_config)
+    fs_cls, fs_config, fs_path = get_cloud_fs({}, **remote_config)
     fs = fs_cls(**fs_config)
     return fs, fs_path
 
@@ -105,18 +133,14 @@ def _resolve_remote_refs(config, remote_conf):
         return remote_conf
 
     base = get_fs_config(config, name=parsed.netloc)
-    cls, _, _ = _get_cloud_fs(config, **base)
+    cls, _, _ = get_cloud_fs(config, **base)
     relpath = parsed.path.lstrip("/").replace("/", cls.sep)
     url = cls.sep.join((base["url"], relpath))
     return {**base, **remote_conf, "url": url}
 
 
-def get_cloud_fs(repo, **kwargs):
-    repo_config = repo.config if repo else {}
-    return _get_cloud_fs(repo_config, **kwargs)
-
-
-def _get_cloud_fs(repo_config, **kwargs):
+def get_cloud_fs(repo_config, **kwargs):
+    repo_config = repo_config or {}
     core_config = repo_config.get("core", {})
 
     remote_conf = get_fs_config(repo_config, **kwargs)

dvc/fs/dvc.py

@@ -5,7 +5,7 @@ import ntpath
 import os
 import posixpath
 import threading
-from contextlib import suppress
+from contextlib import ExitStack, suppress
 from typing import TYPE_CHECKING, Any, Callable, Dict, Optional, Tuple, Type, Union
 
 from fsspec.spec import AbstractFileSystem
@@ -54,6 +54,8 @@ def _merge_info(repo, key, fs_info, dvc_info):
         ret["size"] = dvc_info["size"]
         if not fs_info and "md5" in dvc_info:
             ret["md5"] = dvc_info["md5"]
+        if not fs_info and "md5-dos2unix" in dvc_info:
+            ret["md5-dos2unix"] = dvc_info["md5-dos2unix"]
 
     if fs_info:
         ret["type"] = fs_info["type"]
@@ -121,11 +123,13 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         from pygtrie import Trie
 
         super().__init__()
+        self._repo_stack = ExitStack()
         if repo is None:
             repo = self._make_repo(url=url, rev=rev, subrepos=subrepos, **repo_kwargs)
             assert repo is not None
             # pylint: disable=protected-access
             repo_factory = repo._fs_conf["repo_factory"]
+            self._repo_stack.enter_context(repo)
 
         if not repo_factory:
             from dvc.repo import Repo
@@ -135,7 +139,7 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
             self.repo_factory = repo_factory
 
         def _getcwd():
-            relparts = ()
+            relparts: Tuple[str, ...] = ()
             assert repo is not None
             if repo.fs.path.isin(repo.fs.path.getcwd(), repo.root_dir):
                 relparts = repo.fs.path.relparts(repo.fs.path.getcwd(), repo.root_dir)
@@ -158,7 +162,17 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         if hasattr(repo, "dvc_dir"):
             self._datafss[key] = DataFileSystem(index=repo.index.data["repo"])
 
+    @functools.cached_property
+    def fsid(self) -> str:
+        from fsspec.utils import tokenize
+
+        return "dvcfs_" + tokenize(
+            self.repo.url or self.repo.root_dir,
+            self.repo.get_rev() if self.repo.scm else None,
+        )
+
     def _get_key(self, path: "StrPath") -> Key:
+        path = os.fspath(path)
         parts = self.repo.fs.path.relparts(path, self.repo.root_dir)
         if parts == (os.curdir,):
             return ()
@@ -214,6 +228,7 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
                     scm=self.repo.scm,
                     repo_factory=self.repo_factory,
                 )
+                self._repo_stack.enter_context(repo)
                 self._datafss[key] = DataFileSystem(index=repo.index.data["repo"])
             self._subrepos_trie[key] = repo
 
@@ -256,15 +271,12 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         try:
             return self.repo.fs.open(fs_path, mode=mode)
         except FileNotFoundError:
-            repo, dvc_fs, subkey = self._get_subrepo_info(key)
+            _, dvc_fs, subkey = self._get_subrepo_info(key)
             if not dvc_fs:
                 raise
 
         dvc_path = _get_dvc_path(dvc_fs, subkey)
-        kw = {}
-        if kwargs.get("cache_remote_stream", False):
-            kw["cache_odb"] = repo.cache.local
-        return dvc_fs.open(dvc_path, mode=mode, **kw)
+        return dvc_fs.open(dvc_path, mode=mode, cache=kwargs.get("cache", False))
 
     def isdvc(self, path, **kwargs) -> bool:
         """Is this entry dvc-tracked?"""
@@ -283,7 +295,7 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         dvc_infos = {}
         if dvc_fs:
             dvc_path = _get_dvc_path(dvc_fs, subkey)
-            with suppress(FileNotFoundError):
+            with suppress(FileNotFoundError, NotADirectoryError):
                 for info in dvc_fs.ls(dvc_path, detail=True):
                     dvc_infos[dvc_fs.path.name(info["name"])] = info
             dvc_exists = bool(dvc_infos) or dvc_fs.exists(dvc_path)
@@ -347,7 +359,7 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
                 dvc_info = dvc_fs.fs.index.info(subkey)
                 dvc_path = _get_dvc_path(dvc_fs, subkey)
                 dvc_info["name"] = dvc_path
-            except FileNotFoundError:
+            except KeyError:
                 pass
 
         fs_info = None
@@ -394,6 +406,9 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         dvc_path = _get_dvc_path(dvc_fs, subkey)
         return dvc_fs.get_file(dvc_path, lpath, **kwargs)
 
+    def close(self):
+        self._repo_stack.close()
+
 
 class DVCFileSystem(FileSystem):
     protocol = "local"
@@ -404,9 +419,13 @@ class DVCFileSystem(FileSystem):
 
     @functools.cached_property
     # pylint: disable-next=invalid-overridden-method
-    def fs(self) -> "DVCFileSystem":
+    def fs(self) -> "_DVCFileSystem":
         return _DVCFileSystem(**self.fs_args)
 
+    @property
+    def fsid(self) -> str:
+        return self.fs.fsid
+
     def isdvc(self, path, **kwargs) -> bool:
         return self.fs.isdvc(path, **kwargs)
 
@@ -427,3 +446,7 @@ class DVCFileSystem(FileSystem):
             path = os.path.relpath(path, self.repo.root_dir)
 
         return as_posix(path)
+
+    def close(self):
+        if "fs" in self.__dict__:
+            self.fs.close()

dvc/ignore.py

@@ -148,7 +148,9 @@ class DvcIgnorePatterns(DvcIgnore):
         return bool(self.pattern_list)
 
 
-CheckIgnoreResult = namedtuple("CheckIgnoreResult", ["file", "match", "patterns"])
+CheckIgnoreResult = namedtuple(  # noqa: PYI024
+    "CheckIgnoreResult", ["file", "match", "patterns"]
+)
 
 
 def _no_match(path):

dvc/logger.py

@@ -56,7 +56,7 @@ def add_logging_level(level_name, level_num, method_name=None):
 
 class LoggingException(Exception):
     def __init__(self, record):
-        msg = f"failed to log {str(record)}"
+        msg = f"failed to log {record!s}"
         super().__init__(msg)
 
 
@@ -160,7 +160,7 @@ class LoggerHandler(logging.StreamHandler):
             self.flush()
         except (BrokenPipeError, RecursionError):
             raise
-        except Exception:  # noqa, pylint: disable=broad-except
+        except Exception:  # noqa: BLE001, pylint: disable=broad-except
             self.handleError(record)
 
 

dvc/machine/__init__.py

@@ -1,227 +0,0 @@
-import logging
-import os
-from typing import (
-    TYPE_CHECKING,
-    Any,
-    Dict,
-    Iterable,
-    Iterator,
-    Mapping,
-    Optional,
-    Tuple,
-    Type,
-)
-
-from dvc.exceptions import DvcException
-
-if TYPE_CHECKING:
-    from dvc.repo import Repo
-    from dvc.types import StrPath
-
-    from .backend.base import BaseMachineBackend
-
-    BackendCls = Type[BaseMachineBackend]
-
-logger = logging.getLogger(__name__)
-
-
-RESERVED_NAMES = {"local", "localhost"}
-
-DEFAULT_STARTUP_SCRIPT = """#!/bin/bash
-sudo apt-get update
-sudo apt-get install --yes python3 python3-pip
-python3 -m pip install --upgrade pip
-
-pushd /etc/apt/sources.list.d
-sudo wget https://dvc.org/deb/dvc.list
-sudo apt-get update
-sudo apt-get install --yes dvc
-popd
-
-sudo echo "OK" > /var/log/dvc-machine-init.log
-"""
-
-
-def validate_name(name: str):
-    from dvc.exceptions import InvalidArgumentError
-
-    name = name.lower()
-    if name in RESERVED_NAMES:
-        raise InvalidArgumentError(
-            f"Machine name '{name}' is reserved for internal DVC use."
-        )
-
-
-class MachineBackends(Mapping):
-    try:
-        from .backend.terraform import TerraformBackend
-    except ImportError:
-        TerraformBackend = None  # type: ignore[assignment, misc]
-
-    DEFAULT: Dict[str, Optional["BackendCls"]] = {
-        "terraform": TerraformBackend,
-    }
-
-    def __getitem__(self, key: str) -> "BaseMachineBackend":
-        """Lazily initialize backends and cache it afterwards"""
-        initialized = self.initialized.get(key)
-        if not initialized:
-            backend = self.backends[key]
-            initialized = backend(os.path.join(self.tmp_dir, key), **self.kwargs)
-            self.initialized[key] = initialized
-        return initialized
-
-    def __init__(
-        self,
-        selected: Optional[Iterable[str]],
-        tmp_dir: "StrPath",
-        **kwargs,
-    ) -> None:
-        selected = selected or list(self.DEFAULT)
-        self.backends: Dict[str, "BackendCls"] = {}
-        for key in selected:
-            cls = self.DEFAULT.get(key)
-            if cls is None:
-                raise DvcException(
-                    f"'dvc machine' backend '{key}' is missing required "
-                    "dependencies. Install them with:\n"
-                    f"\tpip install dvc[{key}]"
-                )
-            self.backends[key] = cls
-
-        self.initialized: Dict[str, "BaseMachineBackend"] = {}
-
-        self.tmp_dir = tmp_dir
-        self.kwargs = kwargs
-
-    def __iter__(self):
-        return iter(self.backends)
-
-    def __len__(self) -> int:
-        return len(self.backends)
-
-    def close_initialized(self) -> None:
-        for backend in self.initialized.values():
-            backend.close()
-
-
-class MachineManager:
-    """Class that manages dvc cloud machines.
-
-    Args:
-        repo (dvc.repo.Repo): repo instance that belongs to the repo that
-            we are working on.
-
-    Raises:
-        config.ConfigError: thrown when config has invalid format.
-    """
-
-    CLOUD_BACKENDS = {
-        "aws": "terraform",
-        "azure": "terraform",
-    }
-
-    def __init__(
-        self, repo: "Repo", backends: Optional[Iterable[str]] = None, **kwargs
-    ):
-        self.repo = repo
-        assert self.repo.tmp_dir
-        tmp_dir = os.path.join(self.repo.tmp_dir, "machine")
-        self.backends = MachineBackends(backends, tmp_dir=tmp_dir, **kwargs)
-
-    def get_config_and_backend(
-        self,
-        name: Optional[str] = None,
-    ) -> Tuple[dict, "BaseMachineBackend"]:
-        from dvc.config import NoMachineError
-
-        if not name:
-            name = self.repo.config["core"].get("machine")
-
-        if name:
-            config = self._get_config(name=name)
-            backend = self._get_backend(config["cloud"])
-            return config, backend
-
-        if bool(self.repo.config["machine"]):
-            error_msg = (
-                "no machine specified. Setup default machine with\n"
-                "    dvc machine default <name>\n"
-            )
-        else:
-            error_msg = (
-                "no machine specified. Create a default machine with\n"
-                "    dvc machine add -d <name> <cloud>"
-            )
-
-        raise NoMachineError(error_msg)
-
-    def _get_config(self, **kwargs):
-        config = self.repo.config
-        name = kwargs.get("name")
-        if name:
-            try:
-                conf = config["machine"][name.lower()]
-                conf["name"] = name
-            except KeyError:
-                from dvc.config import MachineNotFoundError
-
-                raise MachineNotFoundError(  # noqa: B904
-                    f"machine '{name}' doesn't exist"
-                )
-        else:
-            conf = kwargs
-        return conf
-
-    def _get_backend(self, cloud: str) -> "BaseMachineBackend":
-        from dvc.config import NoMachineError
-
-        try:
-            backend = self.CLOUD_BACKENDS[cloud]
-            return self.backends[backend]
-        except KeyError:
-            raise NoMachineError(  # noqa: B904
-                f"Machine platform '{cloud}' unsupported"
-            )
-
-    def create(self, name: Optional[str]):
-        """Create and start the specified machine instance."""
-        config, backend = self.get_config_and_backend(name)
-        if "startup_script" in config:
-            with open(config["startup_script"], encoding="utf-8") as fobj:
-                startup_script = fobj.read()
-        else:
-            startup_script = DEFAULT_STARTUP_SCRIPT
-        config["startup_script"] = startup_script
-        config.pop("setup_script", None)
-        return backend.create(**config)
-
-    def destroy(self, name: Optional[str]):
-        """Destroy the specified machine instance."""
-        config, backend = self.get_config_and_backend(name)
-        return backend.destroy(**config)
-
-    def get_sshfs(self, name: Optional[str]):
-        config, backend = self.get_config_and_backend(name)
-        return backend.get_sshfs(**config)
-
-    def run_shell(self, name: Optional[str]):
-        config, backend = self.get_config_and_backend(name)
-        return backend.run_shell(**config)
-
-    def status(self, name: str) -> Iterator[Dict[Any, Any]]:
-        config, backend = self.get_config_and_backend(name)
-        return backend.instances(**config)
-
-    def rename(self, name: str, new: str):
-        """move configuration to a new location if the machine is running."""
-        config, backend = self.get_config_and_backend(name)
-        return backend.rename(new=new, **config)
-
-    def get_executor_kwargs(self, name: Optional[str]):
-        config, backend = self.get_config_and_backend(name)
-        return backend.get_executor_kwargs(**config)
-
-    def get_setup_script(self, name: Optional[str]) -> Optional[str]:
-        config, _backend = self.get_config_and_backend(name)
-        return config.get("setup_script")

dvc/machine/backend/base.py

@@ -1,50 +0,0 @@
-from abc import ABC, abstractmethod
-from contextlib import contextmanager
-from typing import TYPE_CHECKING, Iterator, Optional
-
-if TYPE_CHECKING:
-    from dvc_ssh import SSHFileSystem
-
-    from dvc.types import StrPath
-
-
-class BaseMachineBackend(ABC):
-    def __init__(self, tmp_dir: "StrPath", **kwargs):
-        self.tmp_dir = tmp_dir
-
-    @abstractmethod
-    def create(self, name: Optional[str] = None, **config):
-        """Create and start an instance of the specified machine."""
-
-    @abstractmethod
-    def destroy(self, name: Optional[str] = None, **config):
-        """Stop and destroy all instances of the specified machine."""
-
-    @abstractmethod
-    def instances(self, name: Optional[str] = None, **config) -> Iterator[dict]:
-        """Iterate over status of all instances of the specified machine."""
-
-    def close(self):  # noqa: B027
-        pass
-
-    @abstractmethod
-    def run_shell(self, name: Optional[str] = None, **config):
-        """Spawn an interactive SSH shell for the specified machine."""
-
-    @abstractmethod
-    def get_executor_kwargs(self, name: str, **config) -> dict:
-        """Return SSHExecutor kwargs which can be used for DVC
-        experiment/pipeline execution on the specified machine.
-        """
-
-    @abstractmethod
-    @contextmanager
-    def get_sshfs(
-        self, name: Optional[str] = None, **config
-    ) -> Iterator["SSHFileSystem"]:
-        """Return an sshfs instance for the default directory on the
-        specified machine."""
-
-    @abstractmethod
-    def rename(self, name: str, new: str, **config):
-        """Rename a machine instance."""

dvc/machine/backend/terraform.py

@@ -1,99 +0,0 @@
-import os
-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 .base import BaseMachineBackend
-
-if TYPE_CHECKING:
-    from dvc.types import StrPath
-
-
-@contextmanager
-def _sshfs(resource: dict):
-    from tpi import TerraformProviderIterative
-
-    with TerraformProviderIterative.pemfile(resource) as pem:
-        fs = SSHFileSystem(
-            host=resource["instance_ip"],
-            user="ubuntu",
-            keyfile=pem,
-        )
-        yield fs
-
-
-class TerraformBackend(BaseMachineBackend):
-    def __init__(self, tmp_dir: "StrPath", **kwargs):
-        super().__init__(tmp_dir, **kwargs)
-        os.makedirs(tmp_dir, exist_ok=True)
-
-    @contextmanager
-    def make_tpi(self, name: str):
-        from tpi import TPIError
-        from tpi.terraform import TerraformBackend as TPIBackend
-
-        try:
-            working_dir = os.path.join(self.tmp_dir, name)
-            os.makedirs(working_dir, exist_ok=True)
-            yield TPIBackend(working_dir=working_dir)
-        except TPIError as exc:
-            raise DvcException("TPI operation failed") from exc
-
-    def _tpi_func(self, fname, name: Optional[str] = None, **config):
-        from tpi import TPIError
-
-        assert name
-        with self.make_tpi(name) as tpi:
-            func = getattr(tpi, fname)
-            try:
-                return func(name=name, **config)
-            except TPIError as exc:
-                raise DvcException(f"TPI {fname} failed") from exc
-
-    create = partialmethod(_tpi_func, "create")  # type: ignore[assignment]
-    destroy = partialmethod(_tpi_func, "destroy")  # type: ignore[assignment]
-    instances = partialmethod(_tpi_func, "instances")  # type: ignore[assignment]
-    run_shell = partialmethod(_tpi_func, "run_shell")  # type: ignore[assignment]
-
-    def get_executor_kwargs(self, name: str, **config) -> dict:
-        with self.make_tpi(name) as tpi:
-            resource = tpi.default_resource(name)
-        return {
-            "host": resource["instance_ip"],
-            "port": DEFAULT_PORT,
-            "username": "ubuntu",
-            "fs_factory": partial(_sshfs, dict(resource)),
-        }
-
-    @contextmanager
-    def get_sshfs(  # pylint: disable=unused-argument
-        self, name: Optional[str] = None, **config
-    ) -> Iterator["SSHFileSystem"]:
-        assert name
-        with self.make_tpi(name) as tpi:
-            resource = tpi.default_resource(name)
-        with _sshfs(resource) as fs:
-            yield fs
-
-    def rename(self, name: str, new: str, **config):
-        """rename a dvc machine instance to a new name"""
-        import shutil
-
-        mtype = "iterative_machine"
-
-        new_dir = os.path.join(self.tmp_dir, new)
-        old_dir = os.path.join(self.tmp_dir, name)
-        if os.path.exists(new_dir):
-            raise DvcException(f"rename failed: path {new_dir} already exists")
-
-        if not os.path.exists(old_dir):
-            return
-
-        with self.make_tpi(name) as tpi:
-            tpi.state_mv(f"{mtype}.{name}", f"{mtype}.{new}", **config)
-
-        shutil.move(old_dir, new_dir)

dvc/output.py

@@ -20,6 +20,7 @@ from dvc.exceptions import (
     DvcException,
     MergeError,
 )
+from dvc.utils import format_link
 from dvc.utils.objects import cached_property
 from dvc_data.hashfile import check as ocheck
 from dvc_data.hashfile import load as oload
@@ -96,6 +97,7 @@ def loadd_from(stage, d_list):
         files = d.pop(Output.PARAM_FILES, None)
         push = d.pop(Output.PARAM_PUSH, True)
         hash_name = d.pop(Output.PARAM_HASH, None)
+        fs_config = d.pop(Output.PARAM_FS_CONFIG, None)
         ret.append(
             _get(
                 stage,
@@ -110,6 +112,7 @@ def loadd_from(stage, d_list):
                 files=files,
                 push=push,
                 hash_name=hash_name,
+                fs_config=fs_config,
             )
         )
     return ret
@@ -312,6 +315,7 @@ class Output:
     PARAM_PUSH = "push"
     PARAM_CLOUD = "cloud"
     PARAM_HASH = "hash"
+    PARAM_FS_CONFIG = "fs_config"
 
     DoesNotExistError: Type[DvcException] = OutputDoesNotExistError
     IsNotFileOrDirError: Type[DvcException] = OutputIsNotFileOrDirError
@@ -350,10 +354,15 @@ class Output:
         if meta.version_id or files:
             fs_kwargs["version_aware"] = True
 
+        self.def_fs_config = fs_config
         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)
+        fs_cls, fs_config, fs_path = get_cloud_fs(
+            self.repo.config if self.repo else {},
+            url=path,
+            **fs_kwargs,
+        )
         self.fs = fs_cls(**fs_config)
 
         if (
@@ -470,9 +479,7 @@ class Output:
         return fs.path.abspath(fs.path.normpath(fs_path))
 
     def __repr__(self):
-        return "{class_name}: '{def_path}'".format(
-            class_name=type(self).__name__, def_path=self.def_path
-        )
+        return f"{type(self).__name__}: {self.def_path!r}"
 
     def __str__(self):
         if self.fs.protocol != "local":
@@ -485,6 +492,9 @@ class Output:
         ):
             return str(self.def_path)
 
+        if not self.fs.path.isin(self.fs_path, self.repo.root_dir):
+            return self.fs_path
+
         cur_dir = self.fs.path.getcwd()
         if self.fs.path.isin(cur_dir, self.repo.root_dir):
             return self.fs.path.relpath(self.fs_path, cur_dir)
@@ -547,12 +557,24 @@ class Output:
         _, hash_info = self._get_hash_meta()
         return hash_info
 
+    def _build(
+        self, *args, no_progress_bar=False, **kwargs
+    ) -> Tuple["HashFileDB", "Meta", "HashFile"]:
+        from dvc.ui import ui
+
+        with ui.progress(
+            unit="file",
+            desc=f"Collecting files and computing hashes in {self}",
+            disable=no_progress_bar,
+        ) as pb:
+            return build(*args, callback=pb.as_callback(), **kwargs)
+
     def _get_hash_meta(self):
         if self.use_cache:
             odb = self.cache
         else:
             odb = self.local_cache
-        _, meta, obj = build(
+        _, meta, obj = self._build(
             odb,
             self.fs_path,
             self.fs,
@@ -678,6 +700,14 @@ class Output:
     # pylint: enable=no-member
 
     def save(self) -> None:
+        if self.use_cache and not self.is_in_repo:
+            raise DvcException(
+                f"Saving cached external output {self!s} is not supported "
+                "since DVC 3.0. See "
+                f"{format_link('https://dvc.org/doc/user-guide/upgrade')} "
+                "for more info."
+            )
+
         if not self.exists:
             raise self.DoesNotExistError(self)
 
@@ -694,7 +724,7 @@ class Output:
 
         self._update_legacy_hash_name()
         if self.use_cache:
-            _, self.meta, self.obj = build(
+            _, self.meta, self.obj = self._build(
                 self.cache,
                 self.fs_path,
                 self.fs,
@@ -702,7 +732,7 @@ class Output:
                 ignore=self.dvcignore,
             )
         else:
-            _, self.meta, self.obj = build(
+            _, self.meta, self.obj = self._build(
                 self.local_cache,
                 self.fs_path,
                 self.fs,
@@ -754,20 +784,25 @@ class Output:
             if granular:
                 obj = self._commit_granular_dir(filter_info, hardlink)
             else:
-                staging, _, obj = build(
+                staging, _, obj = self._build(
                     self.cache,
                     filter_info or self.fs_path,
                     self.fs,
                     self.hash_name,
                     ignore=self.dvcignore,
                 )
-                otransfer(
-                    staging,
-                    self.cache,
-                    {obj.hash_info},
-                    shallow=False,
-                    hardlink=hardlink,
-                )
+                with Callback.as_tqdm_callback(
+                    desc=f"Committing {self} to cache",
+                    unit="file",
+                ) as cb:
+                    otransfer(
+                        staging,
+                        self.cache,
+                        {obj.hash_info},
+                        shallow=False,
+                        hardlink=hardlink,
+                        callback=cb,
+                    )
             if relink:
                 rel = self.fs.path.relpath(filter_info or self.fs_path)
                 with CheckoutCallback(desc=f"Checking out {rel}", unit="files") as cb:
@@ -785,7 +820,7 @@ class Output:
 
     def _commit_granular_dir(self, filter_info, hardlink) -> Optional["HashFile"]:
         prefix = self.fs.path.parts(self.fs.path.relpath(filter_info, self.fs_path))
-        staging, _, obj = build(
+        staging, _, obj = self._build(
             self.cache,
             self.fs_path,
             self.fs,
@@ -796,13 +831,18 @@ class Output:
         save_obj = obj.filter(prefix)
         assert isinstance(save_obj, Tree)
         checkout_obj = save_obj.get_obj(self.cache, prefix)
-        otransfer(
-            staging,
-            self.cache,
-            {save_obj.hash_info} | {oid for _, _, oid in save_obj},
-            shallow=True,
-            hardlink=hardlink,
-        )
+        with Callback.as_tqdm_callback(
+            desc=f"Committing {self} to cache",
+            unit="file",
+        ) as cb:
+            otransfer(
+                staging,
+                self.cache,
+                {save_obj.hash_info} | {oid for _, _, oid in save_obj},
+                shallow=True,
+                hardlink=hardlink,
+                callback=cb,
+            )
         return checkout_obj
 
     def dumpd(self, **kwargs):  # noqa: C901, PLR0912
@@ -836,6 +876,9 @@ class Output:
 
         ret[self.PARAM_PATH] = path
 
+        if self.def_fs_config:
+            ret[self.PARAM_FS_CONFIG] = self.def_fs_config
+
         if not self.IS_DEPENDENCY:
             ret.update(self.annot.to_dict())
             if not self.use_cache:
@@ -998,7 +1041,9 @@ class Output:
         if odb is None:
             odb = self.cache
 
-        cls, config, from_info = get_cloud_fs(self.repo, url=source)
+        cls, config, from_info = get_cloud_fs(
+            self.repo.config if self.repo else {}, url=source
+        )
         from_fs = cls(**config)
 
         # When running import-url --to-remote / add --to-remote/-o ... we
@@ -1012,7 +1057,7 @@ class Output:
 
         upload = not (update and from_fs.isdir(from_info))
         jobs = jobs or min((from_fs.jobs, odb.fs.jobs))
-        staging, self.meta, obj = build(
+        staging, self.meta, obj = self._build(
             odb,
             from_info,
             from_fs,
@@ -1020,14 +1065,19 @@ class Output:
             upload=upload,
             no_progress_bar=no_progress_bar,
         )
-        otransfer(
-            staging,
-            odb,
-            {obj.hash_info},
-            jobs=jobs,
-            hardlink=False,
-            shallow=False,
-        )
+        with Callback.as_tqdm_callback(
+            desc=f"Transferring to {odb.fs.unstrip_protocol(odb.path)}",
+            unit="file",
+        ) as cb:
+            otransfer(
+                staging,
+                odb,
+                {obj.hash_info},
+                jobs=jobs,
+                hardlink=False,
+                shallow=False,
+                callback=cb,
+            )
 
         self.hash_info = obj.hash_info
         self.files = None
@@ -1047,7 +1097,7 @@ class Output:
         return len(obj) if obj else 0
 
     def unprotect(self):
-        if self.exists:
+        if self.exists and self.use_cache:
             self.cache.unprotect(self.fs_path)
 
     def get_dir_cache(self, **kwargs) -> Optional["Tree"]:
@@ -1099,7 +1149,7 @@ class Output:
             if not force and not prompt.confirm(msg.format(self.fs_path)):
                 raise CollectCacheError(  # noqa: B904
                     "unable to fully collect used cache"
-                    " without cache for directory '{}'".format(self)
+                    f" without cache for directory '{self}'"
                 )
             return None
 
@@ -1128,19 +1178,15 @@ class Output:
 
         if not self.hash_info:
             msg = (
-                "Output '{}'({}) is missing version info. "
+                f"Output '{self}'({self.stage}) is missing version info. "
                 "Cache for it will not be collected. "
-                "Use `dvc repro` to get your pipeline up to date.".format(
-                    self, self.stage
-                )
+                "Use `dvc repro` to get your pipeline up to date."
             )
             if self.exists:
                 msg += (
                     "\n"
-                    "You can also use `dvc commit {stage.addressing}` "
-                    "to associate existing '{out}' with {stage}.".format(
-                        out=self, stage=self.stage
-                    )
+                    f"You can also use `dvc commit {self.stage.addressing}` "
+                    f"to associate existing '{self}' with {self.stage}."
                 )
             logger.warning(msg)
             return {}
@@ -1270,7 +1316,7 @@ class Output:
             raise FileNotFoundError(  # noqa: B904
                 errno.ENOENT,
                 os.strerror(errno.ENOENT),
-                self.fs.path.relpath(path),
+                path,
             )
 
         new = tree.from_trie(trie)
@@ -1348,7 +1394,7 @@ class Output:
         new: "HashFile"
         try:
             assert self.hash_name
-            staging, meta, obj = build(
+            staging, meta, obj = self._build(
                 cache,
                 path,
                 self.fs,
@@ -1390,11 +1436,23 @@ class Output:
 
         assert staging
         assert obj.hash_info
-        otransfer(staging, self.cache, {obj.hash_info}, hardlink=relink, shallow=False)
+        with Callback.as_tqdm_callback(
+            desc=f"Adding {self} to cache",
+            unit="file",
+        ) as cb:
+            otransfer(
+                staging,
+                self.cache,
+                {obj.hash_info},
+                hardlink=relink,
+                shallow=False,
+                callback=cb,
+            )
 
         if relink:
-            rel = self.fs.path.relpath(path)
-            with CheckoutCallback(desc=f"Checking out {rel}", unit="files") as callback:
+            with CheckoutCallback(
+                desc=f"Checking out {path}", unit="files"
+            ) as callback:
                 self._checkout(
                     path,
                     self.fs,
@@ -1486,4 +1544,5 @@ SCHEMA = {
     Output.PARAM_REMOTE: str,
     Output.PARAM_PUSH: bool,
     Output.PARAM_FILES: [DIR_FILES_SCHEMA],
+    Output.PARAM_FS_CONFIG: dict,
 }

dvc/parsing/__init__.py

@@ -2,6 +2,7 @@ import logging
 import os
 from collections.abc import Mapping, Sequence
 from copy import deepcopy
+from itertools import product
 from typing import (
     TYPE_CHECKING,
     Any,
@@ -51,6 +52,7 @@ VARS_KWD = "vars"
 WDIR_KWD = "wdir"
 PARAMS_KWD = "params"
 FOREACH_KWD = "foreach"
+MATRIX_KWD = "matrix"
 DO_KWD = "do"
 
 DEFAULT_PARAMS_FILE = "params.yaml"
@@ -105,7 +107,7 @@ def is_map_or_seq(data: Any) -> bool:
     return not isinstance(data, str) and _is_map_or_seq(data)
 
 
-def split_foreach_name(name: str) -> Tuple[str, Optional[str]]:
+def split_group_name(name: str) -> Tuple[str, Optional[str]]:
     group, *keys = name.rsplit(JOIN, maxsplit=1)
     return group, first(keys)
 
@@ -120,13 +122,15 @@ def check_interpolations(data: "DictStrAny", where: str, path: str):
     return recurse(func)(data)
 
 
-Definition = Union["ForeachDefinition", "EntryDefinition"]
+Definition = Union["ForeachDefinition", "EntryDefinition", "MatrixDefinition"]
 
 
 def make_definition(
     resolver: "DataResolver", name: str, definition: "DictStrAny", **kwargs
 ) -> Definition:
     args = resolver, resolver.context, name, definition
+    if MATRIX_KWD in definition:
+        return MatrixDefinition(*args, **kwargs)
     if FOREACH_KWD in definition:
         return ForeachDefinition(*args, **kwargs)
     return EntryDefinition(*args, **kwargs)
@@ -135,6 +139,7 @@ def make_definition(
 class DataResolver:
     def __init__(self, repo: "Repo", wdir: str, d: dict):
         self.fs = fs = repo.fs
+        self.parsing_config = repo.config.get("parsing", {})
 
         if os.path.isabs(wdir):
             wdir = fs.path.relpath(wdir)
@@ -158,7 +163,8 @@ class DataResolver:
         self.tracked_vars: Dict[str, Mapping] = {}
 
         stages_data = d.get(STAGES_KWD, {})
-        # we wrap the definitions into ForeachDefinition and EntryDefinition,
+        # we wrap the definitions into:
+        # ForeachDefinition, MatrixDefinition, and EntryDefinition
         # that helps us to optimize, cache and selectively load each one of
         # them as we need, and simplify all of this DSL/parsing logic.
         self.definitions: Dict[str, Definition] = {
@@ -167,12 +173,13 @@ class DataResolver:
         }
 
     def resolve_one(self, name: str):
-        group, key = split_foreach_name(name)
+        group, key = split_group_name(name)
 
         if not self._has_group_and_key(group, key):
             raise EntryNotFound(f"Could not find '{name}'")
 
-        # all of the checks for `key` not being None for `ForeachDefinition`
+        # all of the checks for `key` not being None for
+        # `ForeachDefinition`/`MatrixDefinition`
         # and/or `group` not existing in the `interim`, etc. should be
         # handled by the `self.has_key()` above.
         definition = self.definitions[group]
@@ -189,7 +196,7 @@ class DataResolver:
         return {STAGES_KWD: data}
 
     def has_key(self, key: str):
-        return self._has_group_and_key(*split_foreach_name(key))
+        return self._has_group_and_key(*split_group_name(key))
 
     def _has_group_and_key(self, group: str, key: Optional[str] = None):
         try:
@@ -197,16 +204,14 @@ class DataResolver:
         except KeyError:
             return False
 
-        if key:
-            return isinstance(definition, ForeachDefinition) and definition.has_member(
-                key
-            )
-        return not isinstance(definition, ForeachDefinition)
+        if not isinstance(definition, (ForeachDefinition, MatrixDefinition)):
+            return key is None
+        return key is not None and definition.has_member(key)
 
     @collecting
     def get_keys(self):
         for name, definition in self.definitions.items():
-            if isinstance(definition, ForeachDefinition):
+            if isinstance(definition, (ForeachDefinition, MatrixDefinition)):
                 yield from definition.get_generated_names()
                 continue
             yield name
@@ -255,8 +260,9 @@ class EntryDefinition:
         name = self.name
         if not skip_checks:
             # we can check for syntax errors as we go for interpolated entries,
-            # but for foreach-generated ones, once is enough, which it does
-            # that itself. See `ForeachDefinition.do_definition`.
+            # but for foreach and matrix generated ones, once is enough, which it does
+            # that itself. See `ForeachDefinition.template`
+            # and `MatrixDefinition.template`.
             check_syntax_errors(self.definition, name, self.relpath)
 
         # we need to pop vars from generated/evaluated data
@@ -301,7 +307,10 @@ class EntryDefinition:
     ) -> "DictStrAny":
         try:
             return context.resolve(
-                value, skip_interpolation_checks=skip_checks, key=key
+                value,
+                skip_interpolation_checks=skip_checks,
+                key=key,
+                config=self.resolver.parsing_config,
             )
         except (ParseError, KeyNotInContext) as exc:
             format_and_raise(exc, f"'{self.where}.{self.name}.{key}'", self.relpath)
@@ -327,17 +336,18 @@ class ForeachDefinition:
         self.name = name
 
         assert DO_KWD in definition
+        assert MATRIX_KWD not in definition
         self.foreach_data = definition[FOREACH_KWD]
-        self._do_definition = definition[DO_KWD]
+        self._template = definition[DO_KWD]
 
         self.pair = IterationPair()
         self.where = where
 
     @cached_property
-    def do_definition(self):
+    def template(self):
         # optimization: check for syntax errors only once for `foreach` stages
-        check_syntax_errors(self._do_definition, self.name, self.relpath)
-        return self._do_definition
+        check_syntax_errors(self._template, self.name, self.relpath)
+        return self._template
 
     @cached_property
     def resolved_iterable(self):
@@ -440,16 +450,111 @@ class ForeachDefinition:
             # i.e. quadratic complexity).
             generated = self._generate_name(key)
             entry = EntryDefinition(
-                self.resolver, self.context, generated, self.do_definition
+                self.resolver, self.context, generated, self.template
             )
             try:
                 # optimization: skip checking for syntax errors on each foreach
-                # generated stages. We do it once when accessing do_definition.
+                # generated stages. We do it once when accessing template.
                 return entry.resolve_stage(skip_checks=True)
             except ContextError as exc:
                 format_and_raise(exc, f"stage '{generated}'", self.relpath)
 
-            # let mypy know that this state is unreachable as format_and_raise
-            # does not return at all (it's not able to understand it for some
-            # reason)
-            raise AssertionError("unreachable")
+
+class MatrixDefinition:
+    def __init__(
+        self,
+        resolver: DataResolver,
+        context: Context,
+        name: str,
+        definition: "DictStrAny",
+        where: str = STAGES_KWD,
+    ):
+        self.resolver = resolver
+        self.relpath = self.resolver.relpath
+        self.context = context
+        self.name = name
+
+        assert MATRIX_KWD in definition
+        assert DO_KWD not in definition
+        assert FOREACH_KWD not in definition
+
+        self._template = definition.copy()
+        self.matrix_data = self._template.pop(MATRIX_KWD)
+
+        self.pair = IterationPair()
+        self.where = where
+
+    @cached_property
+    def template(self) -> "DictStrAny":
+        # optimization: check for syntax errors only once for `matrix` stages
+        check_syntax_errors(self._template, self.name, self.relpath)
+        return self._template
+
+    @cached_property
+    def resolved_iterable(self) -> Dict[str, List]:
+        return self._resolve_matrix_data()
+
+    def _resolve_matrix_data(self) -> Dict[str, List]:
+        try:
+            iterable = self.context.resolve(self.matrix_data, unwrap=False)
+        except (ContextError, ParseError) as exc:
+            format_and_raise(exc, f"'{self.where}.{self.name}.matrix'", self.relpath)
+        return iterable
+
+    @cached_property
+    def normalized_iterable(self) -> Dict[str, "DictStrAny"]:
+        """Convert sequence to Mapping with keys normalized."""
+        iterable = self.resolved_iterable
+        assert isinstance(iterable, Mapping)
+
+        ret: Dict[str, "DictStrAny"] = {}
+        matrix = {key: enumerate(v) for key, v in iterable.items()}
+        for combination in product(*matrix.values()):
+            d: "DictStrAny" = {}
+            fragments: List[str] = []
+            for k, (i, v) in zip(matrix.keys(), combination):
+                d[k] = v
+                fragments.append(f"{k}{i}" if is_map_or_seq(v) else to_str(v))
+
+            key = "-".join(fragments)
+            ret[key] = d
+        return ret
+
+    def has_member(self, key: str) -> bool:
+        return key in self.normalized_iterable
+
+    def get_generated_names(self) -> List[str]:
+        return list(map(self._generate_name, self.normalized_iterable))
+
+    def _generate_name(self, key: str) -> str:
+        return f"{self.name}{JOIN}{key}"
+
+    def resolve_all(self) -> "DictStrAny":
+        return join(map(self.resolve_one, self.normalized_iterable))
+
+    def resolve_one(self, key: str) -> "DictStrAny":
+        return self._each_iter(key)
+
+    def _each_iter(self, key: str) -> "DictStrAny":
+        err_message = f"Could not find '{key}' in matrix group '{self.name}'"
+        with reraise(KeyError, EntryNotFound(err_message)):
+            value = self.normalized_iterable[key]
+
+        temp_dict = {self.pair.value: value}
+        with self.context.set_temporarily(temp_dict, reserve=True):
+            # optimization: item and key can be removed on __exit__() as they
+            # are top-level values, and are not merged recursively.
+            # This helps us avoid cloning context, which is slower
+            # (increasing the size of the context might increase
+            # the no. of items to be generated which means more cloning,
+            # i.e. quadratic complexity).
+            generated = self._generate_name(key)
+            entry = EntryDefinition(
+                self.resolver, self.context, generated, self.template
+            )
+            try:
+                # optimization: skip checking for syntax errors on each matrix
+                # generated stages. We do it once when accessing template.
+                return entry.resolve_stage(skip_checks=True)
+            except ContextError as exc:
+                format_and_raise(exc, f"stage '{generated}'", self.relpath)

dvc/parsing/context.py

@@ -495,7 +495,12 @@ class Context(CtxDict):
                 self.data.pop(key, None)
 
     def resolve(
-        self, src, unwrap=True, skip_interpolation_checks=False, key=None
+        self,
+        src,
+        unwrap=True,
+        skip_interpolation_checks=False,
+        key=None,
+        config=None,
     ) -> Any:
         """Recursively resolves interpolation and returns resolved data.
 
@@ -511,10 +516,15 @@ class Context(CtxDict):
         {'lst': [1, 2, 3]}
         """
         func = recurse(self.resolve_str)
-        return func(src, unwrap, skip_interpolation_checks, key)
+        return func(src, unwrap, skip_interpolation_checks, key, config)
 
     def resolve_str(
-        self, src: str, unwrap=True, skip_interpolation_checks=False, key=None
+        self,
+        src: str,
+        unwrap=True,
+        skip_interpolation_checks=False,
+        key=None,
+        config=None,
     ) -> str:
         """Resolves interpolated string to it's original value,
         or in case of multiple interpolations, a combined string.
@@ -535,7 +545,12 @@ class Context(CtxDict):
             return value
         # but not "${num} days"
         return str_interpolate(
-            src, matches, self, skip_checks=skip_interpolation_checks, key=key
+            src,
+            matches,
+            self,
+            skip_checks=skip_interpolation_checks,
+            key=key,
+            config=config,
         )
 
 

dvc/parsing/interpolate.py

@@ -82,20 +82,18 @@ def escape_str(value):
 
 
 @singledispatch
-def to_str(obj) -> str:
+def to_str(obj, config=None) -> str:  # noqa: ARG001, pylint: disable=unused-argument
     return str(obj)
 
 
 @to_str.register(bool)
-def _(obj: bool):
+def _(obj: bool, config=None):  # noqa: ARG001, pylint: disable=unused-argument
     return "true" if obj else "false"
 
 
 @to_str.register(dict)
-def _(obj: dict):  # noqa: C901
-    from dvc.config import Config
-
-    config = Config.from_cwd().get("parsing", {})
+def _(obj: dict, config=None):  # noqa: C901
+    config = config or {}
 
     result = ""
     for k, v in flatten(obj).items():
@@ -211,6 +209,7 @@ def str_interpolate(
     context: "Context",
     skip_checks: bool = False,
     key=None,
+    config=None,
 ):
     index, buf = 0, ""
     for match in matches:
@@ -218,7 +217,7 @@ def str_interpolate(
         expr = get_expression(match, skip_checks=skip_checks)
         value = context.select(expr, unwrap=True)
         validate_value(value, key)
-        buf += template[index:start] + to_str(value)
+        buf += template[index:start] + to_str(value, config=config)
         index = end
     buf += template[index:]
     # regex already backtracks and avoids any `${` starting with

dvc/pathspec_math.py

@@ -8,7 +8,7 @@ from pathspec.util import normalize_file
 
 from dvc.utils import relpath
 
-PatternInfo = namedtuple("PatternInfo", ["patterns", "file_info"])
+PatternInfo = namedtuple("PatternInfo", ["patterns", "file_info"])  # noqa: PYI024
 
 
 def _not_ignore(rule):

dvc/render/match.py

@@ -1,3 +1,4 @@
+import logging
 import os
 from collections import defaultdict
 from typing import TYPE_CHECKING, DefaultDict, Dict, List, NamedTuple, Optional
@@ -17,6 +18,7 @@ if TYPE_CHECKING:
 
 
 dpath.options.ALLOW_EMPTY_STRING_KEYS = True
+logger = logging.getLogger(__name__)
 
 
 def _squash_plots_properties(data: List) -> Dict:
@@ -113,6 +115,7 @@ def match_defs_renderers(  # noqa: C901, PLR0912
             try:
                 dps, rev_props = converter.flat_datapoints(rev)
             except Exception as e:  # noqa: BLE001, pylint: disable=broad-except
+                logger.warning("In %r, %s", rev, str(e).lower())
                 def_errors[rev] = e
                 continue
 

dvc/repo/__init__.py

@@ -15,8 +15,6 @@ from typing import (
 
 from dvc.exceptions import NotDvcRepoError, OutputNotFoundError
 from dvc.ignore import DvcIgnoreFilter
-from dvc.utils import env2bool
-from dvc.utils.fs import path_isin
 from dvc.utils.objects import cached_property
 
 if TYPE_CHECKING:
@@ -24,12 +22,11 @@ if TYPE_CHECKING:
     from dvc.fs.data import DataFileSystem
     from dvc.fs.dvc import DVCFileSystem
     from dvc.lock import LockBase
-    from dvc.machine import MachineManager
     from dvc.scm import Git, NoSCM
     from dvc.stage import Stage
     from dvc.types import DictStrAny
     from dvc_data.hashfile.state import StateBase
-    from dvc_data.index import DataIndex
+    from dvc_data.index import DataIndex, DataIndexEntry
 
     from .experiments import Experiments
     from .index import Index
@@ -93,6 +90,7 @@ class Repo:
     from dvc.repo.status import status  # type: ignore[misc]
     from dvc.repo.update import update  # type: ignore[misc]
 
+    from .cache import check_missing as cache_check_missing  # type: ignore[misc]
     from .data import status as data_status  # type: ignore[misc]
 
     ls = staticmethod(_ls)
@@ -133,7 +131,7 @@ class Repo:
         assert root_dir
         return root_dir, dvc_dir
 
-    def __init__(  # noqa: PLR0915
+    def __init__(  # noqa: PLR0915, PLR0913
         self,
         root_dir: Optional[str] = None,
         fs: Optional["FileSystem"] = None,
@@ -144,9 +142,10 @@ class Repo:
         url: Optional[str] = None,
         repo_factory: Optional[Callable] = None,
         scm: Optional[Union["Git", "NoSCM"]] = None,
+        remote: Optional[str] = None,
+        remote_config: Optional["DictStrAny"] = None,
     ):
         from dvc.cachemgr import CacheManager
-        from dvc.config import Config
         from dvc.data_cloud import DataCloud
         from dvc.fs import GitFileSystem, LocalFileSystem, localfs
         from dvc.lock import LockNoop, make_lock
@@ -163,6 +162,9 @@ class Repo:
         self._fs_conf = {"repo_factory": repo_factory}
         self._fs = fs or localfs
         self._scm = scm
+        self._config = config
+        self._remote = remote
+        self._remote_config = remote_config
         self._data_index = None
 
         if rev and not fs:
@@ -182,7 +184,6 @@ class Repo:
             scm=scm,
         )
 
-        self.config: Config = Config(self.dvc_dir, fs=self.fs, config=config)
         self._uninitialized = uninitialized
 
         # used by DVCFileSystem to determine if it should traverse subrepos
@@ -234,6 +235,19 @@ class Repo:
     def __str__(self):
         return self.url or self.root_dir
 
+    @cached_property
+    def config(self):
+        from dvc.config import Config
+
+        return Config(
+            self.dvc_dir,
+            local_dvc_dir=self.local_dvc_dir,
+            fs=self.fs,
+            config=self._config,
+            remote=self._remote,
+            remote_config=self._remote_config,
+        )
+
     @cached_property
     def local_dvc_dir(self):
         from dvc.fs import GitFileSystem, LocalFileSystem
@@ -285,7 +299,7 @@ class Repo:
             new.check_graph()
 
     @staticmethod
-    def open(url, *args, **kwargs):  # noqa: A003
+    def open(url: Optional[str], *args, **kwargs) -> "Repo":  # noqa: A003
         from .open_repo import open_repo
 
         return open_repo(url, *args, **kwargs)
@@ -335,16 +349,6 @@ class Repo:
 
         return Experiments(self)
 
-    @cached_property
-    def machine(self) -> Optional["MachineManager"]:
-        from dvc.machine import MachineManager
-
-        if self.tmp_dir and (
-            self.config["feature"].get("machine", False) or env2bool("DVC_TEST")
-        ):
-            return MachineManager(self)
-        return None
-
     @property
     def fs(self) -> "FileSystem":
         return self._fs
@@ -375,6 +379,28 @@ class Repo:
         self.data_index.commit()
         self._reset()
 
+    def get_data_index_entry(
+        self,
+        path: str,
+        workspace: str = "repo",
+    ) -> Tuple["DataIndex", "DataIndexEntry"]:
+        if self.subrepos:
+            fs_path = self.dvcfs.from_os_path(path)
+            fs = self.dvcfs.fs
+            # pylint: disable-next=protected-access
+            key = fs._get_key_from_relative(fs_path)
+            # pylint: disable-next=protected-access
+            subrepo, _, key = fs._get_subrepo_info(key)
+            index = subrepo.index.data[workspace]
+        else:
+            index = self.index.data[workspace]
+            key = self.fs.path.relparts(path, self.root_dir)
+
+        try:
+            return index, index[key]
+        except KeyError as exc:
+            raise OutputNotFoundError(path, self) from exc
+
     def __repr__(self):
         return f"{self.__class__.__name__}: '{self.root_dir}'"
 
@@ -384,7 +410,7 @@ class Repo:
 
         fs = fs or localfs
         root = root or os.curdir
-        root_dir = fs.path.realpath(root)
+        root_dir = fs.path.abspath(root)
 
         if not fs.isdir(root_dir):
             raise NotDvcRepoError(f"directory '{root}' does not exist")
@@ -428,8 +454,9 @@ class Repo:
         flist = [self.config.files["local"]]
         if tmp_dir := self.tmp_dir:
             flist.append(tmp_dir)
-        if path_isin(self.cache.legacy.path, self.root_dir):
-            flist.append(self.cache.legacy.path)
+
+        if cache_dir := self.cache.default_local_cache_dir:
+            flist.append(cache_dir)
 
         for file in flist:
             self.scm_context.ignore(file)
@@ -557,6 +584,7 @@ class Repo:
 
         from dvc.dirs import site_cache_dir
         from dvc.fs import GitFileSystem
+        from dvc.version import version_tuple
 
         cache_dir = self.config["core"].get("site_cache_dir") or site_cache_dir()
 
@@ -577,8 +605,21 @@ class Repo:
         finally:
             os.umask(umask)
 
+        # NOTE: Some number to change the generated token if none of the
+        # components were changed (useful to prevent newer dvc versions from
+        # using older broken cache). Please reset this back to 0 if other parts
+        # of the token components are changed.
+        salt = 0
+
         md5 = hashlib.md5(  # noqa: S324  # nosec B324, B303
-            str((root_dir, getpass.getuser())).encode()
+            str(
+                (
+                    root_dir,
+                    getpass.getuser(),
+                    version_tuple[0],
+                    salt,
+                )
+            ).encode()
         )
         repo_token = md5.hexdigest()
         return os.path.join(repos_dir, repo_token)
@@ -586,16 +627,21 @@ class Repo:
     def close(self):
         self.scm.close()
         self.state.close()
+        if "dvcfs" in self.__dict__:
+            self.dvcfs.close()
         if self._data_index is not None:
             self._data_index.close()
 
     def _reset(self):
         self.scm._reset()  # pylint: disable=protected-access
         self.state.close()
+        if "dvcfs" in self.__dict__:
+            self.dvcfs.close()
         self.__dict__.pop("index", None)
         self.__dict__.pop("dvcignore", None)
         self.__dict__.pop("dvcfs", None)
         self.__dict__.pop("datafs", None)
+        self.__dict__.pop("config", None)
 
     def __enter__(self):
         return self

dvc/repo/add.py

@@ -120,7 +120,7 @@ def translate_graph_error(stages: List["Stage"]) -> Iterator[None]:
         )
     except OutputDuplicationError as exc:
         raise OutputDuplicationError(  # noqa: B904
-            exc.output, list(set(exc.stages) - set(stages))
+            exc.output, set(exc.stages) - set(stages)
         )
 
 

dvc/repo/brancher.py

@@ -57,16 +57,17 @@ def brancher(  # noqa: E302
 
     from dvc.fs import LocalFileSystem
 
-    repo_root_parts = ()
+    repo_root_parts: Tuple[str, ...] = ()
     if self.fs.path.isin(self.root_dir, self.scm.root_dir):
         repo_root_parts = self.fs.path.relparts(self.root_dir, self.scm.root_dir)
 
-    cwd_parts = ()
+    cwd_parts: Tuple[str, ...] = ()
     if self.fs.path.isin(self.fs.path.getcwd(), self.scm.root_dir):
         cwd_parts = self.fs.path.relparts(self.fs.path.getcwd(), self.scm.root_dir)
 
     saved_fs = self.fs
     saved_root = self.root_dir
+    saved_dvc_dir = self.dvc_dir
 
     scm = self.scm
 
@@ -101,6 +102,8 @@ def brancher(  # noqa: E302
     finally:
         self.fs = saved_fs
         self.root_dir = saved_root
+        self.dvc_dir = saved_dvc_dir
+        self._reset()  # pylint: disable=protected-access
 
 
 def _switch_fs(
@@ -125,6 +128,8 @@ def _switch_fs(
 
     repo.fs = fs
     repo.root_dir = root_dir
+    repo.dvc_dir = fs.path.join(root_dir, repo.DVC_DIR)
+    repo._reset()  # pylint: disable=protected-access
 
     if cwd_parts:
         cwd = repo.fs.path.join("/", *cwd_parts)
@@ -139,19 +144,22 @@ def switch(repo: "Repo", rev: str) -> Iterator[str]:
     if rev != "workspace":
         rev = resolve_rev(repo.scm, rev)
 
-    repo_root_parts = ()
+    repo_root_parts: Tuple[str, ...] = ()
     if repo.fs.path.isin(repo.root_dir, repo.scm.root_dir):
         repo_root_parts = repo.fs.path.relparts(repo.root_dir, repo.scm.root_dir)
 
-    cwd_parts = ()
+    cwd_parts: Tuple[str, ...] = ()
     if repo.fs.path.isin(repo.fs.path.getcwd(), repo.scm.root_dir):
         cwd_parts = repo.fs.path.relparts(repo.fs.path.getcwd(), repo.scm.root_dir)
 
     saved_fs = repo.fs
     saved_root = repo.root_dir
+    saved_dvc_dir = repo.dvc_dir
     try:
         _switch_fs(repo, rev, repo_root_parts, cwd_parts)
         yield rev
     finally:
         repo.fs = saved_fs
         repo.root_dir = saved_root
+        repo.dvc_dir = saved_dvc_dir
+        repo._reset()  # pylint: disable=protected-access

dvc/repo/cache.py

@@ -0,0 +1,39 @@
+import os
+
+
+def check_missing(repo, rev=None, max_size=None, types=None):
+    from dvc_data.index import StorageKeyError
+
+    with repo.switch(rev or "workspace"):
+        idx = repo.index.targets_view(
+            None,
+            max_size=max_size,
+            types=types,
+        )
+
+        index = idx.data["repo"]
+
+    def onerror(_entry, _exc):
+        pass
+
+    index.onerror = onerror
+
+    ret = []
+    for _, entry in index.iteritems():
+        try:
+            fs, path = index.storage_map.get_cache(entry)
+        except (StorageKeyError, ValueError):
+            continue
+
+        if not fs.exists(path):
+            typ = "directory" if (entry.meta and entry.meta.isdir) else "file"
+            ret.append(
+                (
+                    typ,
+                    entry.hash_info.name,
+                    entry.hash_info.value,
+                    os.path.join(*entry.key),
+                )
+            )
+
+    return ret

dvc/repo/checkout.py

@@ -51,20 +51,16 @@ def _build_out_changes(index, changes):
         out_keys.append(key)
 
     out_changes = {}
-    for objects in changes.values():
-        for change in objects:
-            for out_key in out_keys:
-                if (
-                    len(out_key) > len(change.key)
-                    or change.key[: len(out_key)] != out_key
-                ):
-                    continue
+    for key, change in changes.items():
+        for out_key in out_keys:
+            if len(out_key) > len(key) or key[: len(out_key)] != out_key:
+                continue
 
-                if change.key == out_key:
-                    out_changes[out_key] = change.typ
-                elif not out_changes.get(out_key):
-                    out_changes[out_key] = MODIFY
-                break
+            if key == out_key:
+                out_changes[out_key] = change.typ
+            elif not out_changes.get(out_key):
+                out_changes[out_key] = MODIFY
+            break
 
     return out_changes
 
@@ -85,7 +81,7 @@ def _check_can_delete(
         if cache_fs.exists(cache_path):
             continue
 
-        entry_paths.append(fs.path.join(path, *entry.key))
+        entry_paths.append(fs.path.join(path, *(entry.key or ())))
 
     if not entry_paths:
         return

dvc/repo/experiments/__init__.py

@@ -136,10 +136,6 @@ class Experiments:
         **kwargs,
     ) -> "QueueEntry":
         """Queue a single experiment."""
-        if kwargs.pop("machine", None) is not None:
-            # TODO: decide how to handle queued remote execution
-            raise NotImplementedError
-
         return self.new(
             queue,
             **kwargs,
@@ -214,10 +210,6 @@ class Experiments:
             )
         else:
             ui.write("Experiment results have been applied to your workspace.")
-        ui.write(
-            "\nTo promote an experiment to a Git branch run:\n\n"
-            "\tdvc exp branch <exp> <branch>\n"
-        )
 
     def new(
         self,

dvc/repo/experiments/collect.py

@@ -1,6 +1,7 @@
 import itertools
 import logging
 import os
+from dataclasses import fields
 from datetime import datetime
 from typing import (
     TYPE_CHECKING,
@@ -22,6 +23,7 @@ from dvc.scm import Git, SCMError, iter_revs
 from .exceptions import InvalidExpRefError
 from .refs import EXEC_BRANCH, ExpRefInfo
 from .serialize import ExpRange, ExpState, SerializableError, SerializableExp
+from .utils import describe
 
 if TYPE_CHECKING:
     from dvc.repo import Repo
@@ -152,7 +154,23 @@ def collect_queued(
     """
     if not baseline_revs:
         return {}
-    return repo.experiments.celery_queue.collect_queued_data(baseline_revs, **kwargs)
+    queued_data = {}
+    for rev, ranges in repo.experiments.celery_queue.collect_queued_data(
+        baseline_revs, **kwargs
+    ).items():
+        for exp_range in ranges:
+            for exp_state in exp_range.revs:
+                if exp_state.data:
+                    attrs = [f.name for f in fields(SerializableExp)]
+                    exp_state.data = SerializableExp(
+                        **{
+                            attr: getattr(exp_state.data, attr)
+                            for attr in attrs
+                            if attr != "metrics"
+                        }
+                    )
+        queued_data[rev] = ranges
+    return queued_data
 
 
 def collect_active(
@@ -291,7 +309,9 @@ def collect(
     if sha_only:
         baseline_names: Dict[str, Optional[str]] = {}
     else:
-        baseline_names = _describe(repo.scm, baseline_revs, refs=cached_refs)
+        baseline_names = describe(
+            repo.scm, baseline_revs, refs=cached_refs, logger=logger
+        )
 
     workspace_data = collect_rev(repo, "workspace", **kwargs)
     result: List["ExpState"] = [workspace_data]
@@ -320,55 +340,6 @@ def collect(
     return result
 
 
-def _describe(
-    scm: "Git",
-    revs: Iterable[str],
-    refs: Optional[Iterable[str]] = None,
-) -> Dict[str, Optional[str]]:
-    """Describe revisions using a tag, branch.
-
-    The first matching name will be returned for each rev. Names are preferred in this
-    order:
-        - current branch (if rev matches HEAD and HEAD is a branch)
-        - tags
-        - branches
-
-    Returns:
-        Dict mapping revisions from revs to a name.
-    """
-
-    head_rev = scm.get_rev()
-    head_ref = scm.get_ref("HEAD", follow=False)
-    if head_ref and head_ref.startswith("refs/heads/"):
-        head_branch = head_ref[len("refs/heads/") :]
-    else:
-        head_branch = None
-
-    tags = {}
-    branches = {}
-    ref_it = iter(refs) if refs else scm.iter_refs()
-    for ref in ref_it:
-        is_tag = ref.startswith("refs/tags/")
-        is_branch = ref.startswith("refs/heads/")
-        if not (is_tag or is_branch):
-            continue
-        rev = scm.get_ref(ref)
-        if not rev:
-            logger.debug("unresolved ref %s", ref)
-            continue
-        if is_tag and rev not in tags:
-            tags[rev] = ref[len("refs/tags/") :]
-        if is_branch and rev not in branches:
-            branches[rev] = ref[len("refs/heads/") :]
-    names: Dict[str, Optional[str]] = {}
-    for rev in revs:
-        if rev == head_rev and head_branch:
-            names[rev] = head_branch
-        else:
-            names[rev] = tags.get(rev) or branches.get(rev)
-    return names
-
-
 def _sorted_ranges(exp_ranges: Iterable["ExpRange"]) -> List["ExpRange"]:
     """Return list of ExpRange sorted by (timestamp, rev)."""
 

dvc/repo/experiments/executor/base.py

@@ -44,7 +44,7 @@ if TYPE_CHECKING:
     from dvc.repo import Repo
     from dvc.repo.experiments.stash import ExpStashEntry
     from dvc.scm import Git
-    from dvc.stage import PipelineStage
+    from dvc.stage import PipelineStage, Stage
 
 logger = logging.getLogger(__name__)
 
@@ -129,7 +129,6 @@ class BaseExecutor(ABC):
 
     PACKED_ARGS_FILE = "repro.dat"
     WARN_UNTRACKED = False
-    QUIET = False
     INFOFILE_EXT = ".run"
     DEFAULT_LOCATION: str = "workspace"
 
@@ -459,25 +458,16 @@ class BaseExecutor(ABC):
             should force overwrite any existing duplicates.
         """
         from dvc.repo.checkout import checkout as dvc_checkout
-        from dvc.repo.reproduce import reproduce as dvc_reproduce
-        from dvc.stage import PipelineStage
         from dvc.ui import ui
 
         auto_push = env2bool(DVC_EXP_AUTO_PUSH)
         git_remote = os.getenv(DVC_EXP_GIT_REMOTE, None)
 
-        unchanged = []
-
         if queue is not None:
             queue.put((rev, os.getpid()))
         if log_errors and log_level is not None:
             cls._set_log_level(log_level)
 
-        def filter_pipeline(stages):
-            unchanged.extend(
-                [stage for stage in stages if isinstance(stage, PipelineStage)]
-            )
-
         exp_hash: Optional[str] = None
         exp_ref: Optional["ExpRefInfo"] = None
         repro_force: bool = False
@@ -519,12 +509,8 @@ class BaseExecutor(ABC):
                     recursive=kwargs.get("recursive", False),
                 )
 
-            stages = dvc_reproduce(
-                dvc,
-                *args,
-                on_unchanged=filter_pipeline,
-                **kwargs,
-            )
+            kwargs["repro_fn"] = cls._repro_and_track
+            stages = dvc.reproduce(*args, **kwargs)
             if paths := cls._get_top_level_paths(dvc):
                 logger.debug("Staging top-level files: %s", paths)
                 dvc.scm_context.add(paths)
@@ -549,6 +535,17 @@ class BaseExecutor(ABC):
         # multiprocessing calls
         return ExecutorResult(exp_hash, exp_ref, repro_force)
 
+    @staticmethod
+    def _repro_and_track(stage: "Stage", **kwargs) -> Optional["Stage"]:
+        from dvc.repo.reproduce import _reproduce_stage
+        from dvc.stage.utils import _get_stage_files
+
+        ret = _reproduce_stage(stage, **kwargs)
+        if not kwargs.get("dry") and (paths := _get_stage_files(stage)):
+            logger.debug("Staging stage-related files: %s", paths)
+            stage.repo.scm_context.add(paths)
+        return ret
+
     @classmethod
     def _repro_commit(
         cls,
@@ -604,12 +601,11 @@ class BaseExecutor(ABC):
             info.status = TaskStatus.RUNNING
             if infofile is not None:
                 info.dump_json(infofile)
-            if cls.QUIET:
-                dvc.scm_context.quiet = cls.QUIET
+            dvc.scm_context.quiet = True
             old_cwd = os.getcwd()
 
             for path in copy_paths or []:
-                cls._copy_path(os.path.realpath(path), os.path.join(dvc.root_dir, path))
+                cls._copy_path(os.path.abspath(path), os.path.join(dvc.root_dir, path))
 
             if info.wdir:
                 os.chdir(os.path.join(dvc.scm.root_dir, info.wdir))

dvc/repo/experiments/executor/local.py

@@ -65,7 +65,6 @@ class TempDirExecutor(BaseLocalExecutor):
     # debugging user code), and suppress other DVC hints (like `git add`
     # suggestions) that are not applicable outside of workspace runs
     WARN_UNTRACKED = True
-    QUIET = True
     DEFAULT_LOCATION = "tempdir"
 
     @retry(180, errors=LockError, timeout=1)

dvc/repo/experiments/executor/ssh.py

@@ -1,292 +0,0 @@
-import logging
-import os
-import posixpath
-import sys
-from contextlib import contextmanager
-from typing import TYPE_CHECKING, Callable, Iterable, List, Optional
-
-from dvc_ssh import SSHFileSystem
-from funcy import first
-
-from dvc.repo.experiments.refs import (
-    EXEC_BASELINE,
-    EXEC_BRANCH,
-    EXEC_HEAD,
-    EXEC_MERGE,
-    EXEC_NAMESPACE,
-)
-
-from .base import BaseExecutor, ExecutorInfo, ExecutorResult, TaskStatus
-
-if TYPE_CHECKING:
-    from queue import Queue
-
-    from dvc.repo import Repo
-    from dvc.repo.experiments.refs import ExpRefInfo
-    from dvc.repo.experiments.stash import ExpStashEntry
-    from dvc.scm import Git
-
-logger = logging.getLogger(__name__)
-
-
-@contextmanager
-def _sshfs(fs_factory, **kwargs):
-    if fs_factory:
-        with fs_factory() as fs:
-            yield fs
-        return
-    yield SSHFileSystem(**kwargs)
-
-
-class SSHExecutor(BaseExecutor):
-    """SSH experiment executor."""
-
-    WARN_UNTRACKED = True
-    QUIET = True
-    SETUP_SCRIPT_FILENAME = "exec-setup.sh"
-
-    def __init__(
-        self,
-        *args,
-        host: Optional[str] = None,
-        port: Optional[int] = None,
-        username: Optional[str] = None,
-        fs_factory: Optional[Callable] = None,
-        setup_script: Optional[str] = None,
-        **kwargs,
-    ):
-        assert host
-
-        super().__init__(*args, **kwargs)
-        self.host: str = host
-        self.port = port
-        self.username = username
-        self._fs_factory = fs_factory
-        self._repo_abspath = ""
-        self._setup_script = setup_script
-
-    @classmethod
-    def gen_dirname(cls, name: Optional[str] = None):
-        from shortuuid import uuid
-
-        return "-".join([name or "dvc-exp", "executor", uuid()])
-
-    @classmethod
-    def from_stash_entry(
-        cls,
-        repo: "Repo",
-        entry: "ExpStashEntry",
-        **kwargs,
-    ):
-        machine_name: Optional[str] = kwargs.pop("machine_name", None)
-        assert repo.machine
-        executor = cls._from_stash_entry(
-            repo,
-            entry,
-            cls.gen_dirname(entry.name),
-            location=machine_name,
-            **repo.machine.get_executor_kwargs(machine_name),
-            setup_script=repo.machine.get_setup_script(machine_name),
-        )
-        logger.debug("Init SSH executor for host '%s'", executor.host)
-        return executor
-
-    def sshfs(self):
-        return _sshfs(self._fs_factory, host=self.host, port=self.port)
-
-    @property
-    def git_url(self) -> str:
-        user = f"{self.username}@" if self.username else ""
-        port = f":{self.port}" if self.port is not None else ""
-        path = f"{self.root_dir}" if self.root_dir else ""
-        if path and not posixpath.isabs(path):
-            path = f"/~/{path}"
-        return f"ssh://{user}{self.host}{port}{path}"
-
-    @property
-    def abs_url(self) -> str:
-        assert self._repo_abspath
-        user = f"{self.username}@" if self.username else ""
-        port = f":{self.port}" if self.port is not None else ""
-        return f"ssh://{user}{self.host}{port}{self._repo_abspath}"
-
-    @staticmethod
-    def _git_client_args(fs):
-        return {
-            "password": fs.fs_args.get("password"),
-            "key_filename": first(fs.fs_args.get("client_keys", [])),
-        }
-
-    def init_git(
-        self,
-        repo: "Repo",  # noqa: ARG002
-        scm: "Git",
-        stash_rev: str,
-        entry: "ExpStashEntry",
-        infofile: Optional[str],
-        branch: Optional[str] = None,
-    ):
-        from dvc.repo.experiments.utils import push_refspec
-
-        self.status = TaskStatus.PREPARING
-        if infofile:
-            self.info.dump_json(infofile)
-
-        with self.sshfs() as fs:
-            fs.makedirs(self.root_dir)
-            self._ssh_cmd(fs, "git init .")
-            self._ssh_cmd(fs, "git config user.name dvc-exp")
-            self._ssh_cmd(fs, "git config user.email [email protected]")
-
-            result = self._ssh_cmd(fs, "pwd")
-            path = result.stdout.strip()
-            self._repo_abspath = path
-
-            # TODO: support multiple client key retries in git backends
-            # (see https://github.com/iterative/dvc/issues/6508)
-            kwargs = self._git_client_args(fs)
-
-            ref_dict = {
-                EXEC_HEAD: entry.head_rev,
-                EXEC_MERGE: stash_rev,
-                EXEC_BASELINE: entry.baseline_rev,
-            }
-            with self.set_temp_refs(scm, ref_dict):
-                exec_namespace = f"{EXEC_NAMESPACE}/"
-                refspec = [(exec_namespace, exec_namespace)]
-                push_refspec(scm, self.git_url, refspec, **kwargs)
-
-            if branch:
-                push_refspec(scm, self.git_url, [(branch, branch)], **kwargs)
-                self._ssh_cmd(fs, f"git symbolic-ref {EXEC_BRANCH} {branch}")
-            else:
-                self._ssh_cmd(fs, f"git symbolic-ref -d {EXEC_BRANCH}", check=False)
-
-            # checkout EXEC_HEAD and apply EXEC_MERGE on top of it without
-            # committing
-            head = EXEC_BRANCH if branch else EXEC_HEAD
-            self._ssh_cmd(fs, f"git checkout {head}")
-            merge_rev = scm.get_ref(EXEC_MERGE)
-            self._ssh_cmd(fs, f"git stash apply {merge_rev}")
-
-            if self._setup_script:
-                self._init_setup_script(fs)
-
-    @classmethod
-    def _setup_script_path(cls, dvc_dir: str):
-        return posixpath.join(
-            dvc_dir,
-            "tmp",
-            cls.SETUP_SCRIPT_FILENAME,
-        )
-
-    def _init_setup_script(self, fs: "SSHFileSystem"):
-        assert self._repo_abspath
-        script_path = self._setup_script_path(
-            posixpath.join(self._repo_abspath, self.dvc_dir)
-        )
-        assert self._setup_script
-        fs.put_file(self._setup_script, script_path)
-
-    def _ssh_cmd(self, sshfs, cmd, chdir=None, **kwargs):
-        working_dir = chdir or self.root_dir
-        return sshfs.fs.execute(f"cd {working_dir};{cmd}", **kwargs)
-
-    def init_cache(self, repo: "Repo", rev: str, run_cache: bool = True):
-        from dvc.repo.push import push
-
-        with self.get_odb() as odb:
-            push(
-                repo,
-                revs=[rev],
-                run_cache=run_cache,
-                odb=odb,
-                include_imports=True,
-            )
-
-    def collect_cache(
-        self, repo: "Repo", exp_ref: "ExpRefInfo", run_cache: bool = True
-    ):
-        """Collect DVC cache."""
-        from dvc.repo.experiments.pull import _pull_cache
-
-        _pull_cache(repo, exp_ref, run_cache=run_cache)
-
-    @contextmanager
-    def get_odb(self):
-        from dvc.cachemgr import CacheManager, get_odb
-
-        cache_path = posixpath.join(
-            self._repo_abspath,
-            self.dvc_dir,
-            CacheManager.CACHE_DIR,
-        )
-
-        with self.sshfs() as fs:
-            yield get_odb(fs, cache_path, **fs.config)
-
-    def fetch_exps(self, *args, **kwargs) -> Iterable[str]:
-        with self.sshfs() as fs:
-            kwargs.update(self._git_client_args(fs))
-            return super().fetch_exps(*args, **kwargs)
-
-    @classmethod
-    def reproduce(
-        cls,
-        info: "ExecutorInfo",
-        rev: str,  # noqa: ARG003
-        queue: Optional["Queue"] = None,  # noqa: ARG003
-        infofile: Optional[str] = None,
-        log_errors: bool = True,
-        log_level: Optional[int] = None,
-        copy_paths: Optional[List[str]] = None,  # noqa: ARG003
-        message: Optional[str] = None,  # noqa: ARG003
-        **kwargs,
-    ) -> "ExecutorResult":
-        """Reproduce an experiment on a remote machine over SSH.
-
-        Internally uses 'dvc exp exec-run' over SSH.
-        """
-        import json
-        import time
-        from tempfile import TemporaryFile
-
-        from asyncssh import ProcessError
-
-        fs_factory: Optional[Callable] = kwargs.pop("fs_factory", None)
-        if log_errors and log_level is not None:
-            cls._set_log_level(log_level)
-
-        with _sshfs(fs_factory) as fs:
-            while not fs.exists("/var/log/dvc-machine-init.log"):
-                logger.info("Waiting for dvc-machine startup script to complete...")
-                time.sleep(5)
-            logger.info("Reproducing experiment on '%s'", fs.fs_args.get("host"))
-            with TemporaryFile(mode="w+", encoding="utf-8") as fobj:
-                json.dump(info.asdict(), fobj)
-                fobj.seek(0)
-                fs.put_file(fobj, infofile)
-            cmd = ["source ~/.profile"]
-            script_path = cls._setup_script_path(info.dvc_dir)
-            if fs.exists(posixpath.join(info.root_dir, script_path)):
-                cmd.extend([f"pushd {info.root_dir}", f"source {script_path}", "popd"])
-            exec_cmd = f"dvc exp exec-run --infofile {infofile}"
-            if log_level is not None:
-                if log_level <= logging.TRACE:  # type: ignore[attr-defined]
-                    exec_cmd += " -vv"
-                elif log_level <= logging.DEBUG:
-                    exec_cmd += " -v"
-            cmd.append(exec_cmd)
-            try:
-                sys.stdout.flush()
-                sys.stderr.flush()
-                stdout = os.dup(sys.stdout.fileno())
-                stderr = os.dup(sys.stderr.fileno())
-                fs.fs.execute("; ".join(cmd), stdout=stdout, stderr=stderr)
-                with fs.open(infofile) as fobj:
-                    result_info = ExecutorInfo.from_dict(json.load(fobj))
-                if result_info.result_hash:
-                    return result_info.result
-            except ProcessError:
-                pass
-            return ExecutorResult(None, None, False)

dvc/repo/experiments/ls.py

@@ -31,23 +31,15 @@ def ls(
             rev = [rev]
         revs = iter_revs(repo.scm, rev, num)
         rev_set = set(revs.keys())
+
     ref_info_dict = exp_refs_by_baseline(repo.scm, rev_set, git_remote)
-
-    tags = repo.scm.describe(ref_info_dict.keys())
-    remained = {baseline for baseline, tag in tags.items() if tag is None}
-    base = "refs/heads"
-    ref_heads = repo.scm.describe(remained, base=base)
-
     results = defaultdict(list)
     for baseline in ref_info_dict:
-        name = baseline
-        if tags[baseline] or ref_heads[baseline]:
-            name = tags[baseline] or ref_heads[baseline]
         for info in ref_info_dict[baseline]:
             if git_remote:
                 exp_rev = None
             else:
                 exp_rev = repo.scm.get_ref(str(info))
-            results[name].append((info.name, exp_rev))
+            results[baseline].append((info.name, exp_rev))
 
     return results

dvc/repo/experiments/queue/base.py

@@ -21,7 +21,7 @@ from dvc_studio_client.post_live_metrics import get_studio_config
 from funcy import retry
 
 from dvc.dependency import ParamsDependency
-from dvc.env import DVC_EXP_BASELINE_REV, DVC_EXP_NAME
+from dvc.env import DVC_EXP_BASELINE_REV, DVC_EXP_NAME, DVC_ROOT
 from dvc.lock import LockError
 from dvc.repo.experiments.exceptions import ExperimentExistsError
 from dvc.repo.experiments.executor.base import BaseExecutor
@@ -348,6 +348,9 @@ class BaseStashQueue(ABC):
                     if not name:
                         name = get_random_exp_name(self.scm, baseline_rev)
                     run_env[DVC_EXP_NAME] = name
+                    # Override DVC_ROOT env var to point to the parent DVC repo
+                    # root (and not an executor tempdir root)
+                    run_env[DVC_ROOT] = self.repo.root_dir
 
                     # save studio config to read later by dvc and dvclive
                     studio_config = get_studio_config(
@@ -476,13 +479,17 @@ class BaseStashQueue(ABC):
         hydra_output_file = ParamsDependency.DEFAULT_PARAMS_FILE
         for path, overrides in params.items():
             if hydra_enabled and path == hydra_output_file:
-                config_dir = os.path.join(
-                    self.repo.root_dir, hydra_config.get("config_dir", "conf")
-                )
+                if (config_module := hydra_config.get("config_module")) is None:
+                    config_dir = os.path.join(
+                        self.repo.root_dir, hydra_config.get("config_dir", "conf")
+                    )
+                else:
+                    config_dir = None
                 config_name = hydra_config.get("config_name", "config")
                 compose_and_dump(
                     path,
                     config_dir,
+                    config_module,
                     config_name,
                     overrides,
                 )

dvc/repo/experiments/queue/tempdir.py

@@ -115,7 +115,9 @@ class TempDirQueue(WorkspaceQueue):
                 message=message,
             )
             if not exec_result.exp_hash:
-                raise DvcException(f"Failed to reproduce experiment '{rev[:7]}'")
+                raise DvcException(  # noqa: TRY301
+                    f"Failed to reproduce experiment '{rev[:7]}'"
+                )
             if exec_result.ref_info:
                 results[rev].update(
                     self.collect_executor(self.repo.experiments, executor, exec_result)

dvc/repo/experiments/queue/workspace.py

@@ -121,7 +121,9 @@ class WorkspaceQueue(BaseStashQueue):
                 message=kwargs.get("message"),
             )
             if not exec_result.exp_hash:
-                raise DvcException(f"Failed to reproduce experiment '{rev[:7]}'")
+                raise DvcException(  # noqa: TRY301
+                    f"Failed to reproduce experiment '{rev[:7]}'"
+                )
             if exec_result.ref_info:
                 results[rev].update(
                     self.collect_executor(self.repo.experiments, executor, exec_result)

dvc/repo/experiments/show.py

@@ -177,26 +177,31 @@ def _build_rows(
                 )
 
 
-def _sort_column(
+def _sort_column(  # noqa: C901
     sort_by: str,
     metric_names: Mapping[str, Iterable[str]],
     param_names: Mapping[str, Iterable[str]],
 ) -> Tuple[str, str, str]:
-    path, _, sort_name = sort_by.rpartition(":")
+    sep = ":"
+    parts = sort_by.split(sep)
     matches: Set[Tuple[str, str, str]] = set()
 
-    if path:
+    for split_num in range(len(parts)):
+        path = sep.join(parts[:split_num])
+        sort_name = sep.join(parts[split_num:])
+        if not path:  # handles ':metric_name' case
+            sort_by = sort_name
         if path in metric_names and sort_name in metric_names[path]:
             matches.add((path, sort_name, "metrics"))
         if path in param_names and sort_name in param_names[path]:
             matches.add((path, sort_name, "params"))
-    else:
+    if not matches:
         for path in metric_names:
-            if sort_name in metric_names[path]:
-                matches.add((path, sort_name, "metrics"))
+            if sort_by in metric_names[path]:
+                matches.add((path, sort_by, "metrics"))
         for path in param_names:
-            if sort_name in param_names[path]:
-                matches.add((path, sort_name, "params"))
+            if sort_by in param_names[path]:
+                matches.add((path, sort_by, "params"))
 
     if len(matches) == 1:
         return matches.pop()
@@ -238,19 +243,16 @@ def _exp_range_rows(
     is_base: bool = False,
     **kwargs,
 ) -> Iterator[Tuple["CellT", ...]]:
-    for i, exp in enumerate(exp_range.revs):
+    from funcy import first
+
+    if len(exp_range.revs) > 1:
+        logger.debug("Returning tip commit for legacy checkpoint exp")
+    exp = first(exp_range.revs)
+    if exp:
         row: Dict[str, "CellT"] = {k: fill_value for k in all_headers}
         row["Experiment"] = exp.name or ""
         row["rev"] = exp.rev[:7] if Git.is_sha(exp.rev) else exp.rev
-        if len(exp_range.revs) > 1:
-            if i == 0:
-                row["typ"] = "checkpoint_tip"
-            elif i == len(exp_range.revs) - 1:
-                row["typ"] = "checkpoint_base"
-            else:
-                row["typ"] = "checkpoint_commit"
-        else:
-            row["typ"] = "branch_base" if is_base else "branch_commit"
+        row["typ"] = "branch_base" if is_base else "branch_commit"
         row["parent"] = ""
         if exp_range.executor:
             row["State"] = exp_range.executor.state.capitalize()

dvc/repo/fetch.py

@@ -1,12 +1,79 @@
 import logging
+from typing import List, Tuple
 
 from dvc.exceptions import DownloadError
+from dvc_data.index import DataIndex, FileStorage
 
 from . import locked
 
 logger = logging.getLogger(__name__)
 
 
+def _make_index_onerror(onerror, rev):
+    def _onerror(entry, exc):
+        if onerror:
+            return onerror(rev, entry, exc)
+
+    return _onerror
+
+
+def _collect_indexes(  # noqa: PLR0913
+    repo,
+    targets=None,
+    remote=None,
+    all_branches=False,
+    with_deps=False,
+    all_tags=False,
+    recursive=False,
+    all_commits=False,
+    revs=None,
+    max_size=None,
+    types=None,
+    config=None,
+    onerror=None,
+):
+    indexes = {}
+    collection_exc = None
+
+    config = config or {}
+    if remote:
+        core = config.get("core") or {}
+        core["remote"] = remote
+        config["core"] = core
+
+    for rev in repo.brancher(
+        revs=revs,
+        all_branches=all_branches,
+        all_tags=all_tags,
+        all_commits=all_commits,
+    ):
+        try:
+            repo.config.merge(config)
+
+            idx = repo.index.targets_view(
+                targets,
+                with_deps=with_deps,
+                recursive=recursive,
+                max_size=max_size,
+                types=types,
+            )
+
+            data = idx.data["repo"]
+            data.onerror = _make_index_onerror(onerror, rev)
+
+            indexes[idx.data_tree.hash_info.value] = data
+        except Exception as exc:  # pylint: disable=broad-except
+            if onerror:
+                onerror(rev, None, exc)
+            collection_exc = exc
+            logger.exception("failed to collect '%s'", rev or "workspace")
+
+    if not indexes and collection_exc:
+        raise collection_exc
+
+    return indexes
+
+
 @locked
 def fetch(  # noqa: C901, PLR0913
     self,
@@ -20,6 +87,10 @@ def fetch(  # noqa: C901, PLR0913
     all_commits=False,
     run_cache=False,
     revs=None,
+    max_size=None,
+    types=None,
+    config=None,
+    onerror=None,
 ) -> int:
     """Download data items from a cloud and imported repositories
 
@@ -51,39 +122,37 @@ def fetch(  # noqa: C901, PLR0913
     except DownloadError as exc:
         failed_count += exc.amount
 
-    indexes = []
-    index_keys = set()
-    for _ in self.brancher(
-        revs=revs,
+    indexes = _collect_indexes(
+        self,
+        targets=targets,
+        remote=remote,
         all_branches=all_branches,
+        with_deps=with_deps,
         all_tags=all_tags,
+        recursive=recursive,
         all_commits=all_commits,
-    ):
-        saved_remote = self.config["core"].get("remote")
-        try:
-            if remote:
-                self.config["core"]["remote"] = remote
+        revs=revs,
+        max_size=max_size,
+        types=types,
+        config=config,
+        onerror=onerror,
+    )
 
-            idx = self.index.targets_view(
-                targets,
-                with_deps=with_deps,
-                recursive=recursive,
-            )
-            index_keys.add(idx.data_tree.hash_info.value)
-            indexes.append(idx.data["repo"])
-        finally:
-            if remote:
-                self.config["core"]["remote"] = saved_remote
-
-    cache_key = ("fetch", tokenize(sorted(index_keys)))
+    cache_key = ("fetch", tokenize(sorted(indexes.keys())))
 
     with Callback.as_tqdm_callback(
         desc="Collecting",
         unit="entry",
     ) as cb:
         data = collect(
-            indexes, cache_index=self.data_index, cache_key=cache_key, callback=cb
+            indexes.values(),
+            "remote",
+            cache_index=self.data_index,
+            cache_key=cache_key,
+            callback=cb,
         )
+    data, unversioned_count = _log_unversioned(data)
+    failed_count += unversioned_count
 
     with Callback.as_tqdm_callback(
         desc="Fetching",
@@ -96,7 +165,7 @@ def fetch(  # noqa: C901, PLR0913
                 callback=cb,
             )  # pylint: disable=assignment-from-no-return
         finally:
-            for fs_index in data.values():
+            for fs_index in data:
                 fs_index.close()
 
     if fetch_transferred:
@@ -109,3 +178,34 @@ def fetch(  # noqa: C901, PLR0913
         raise DownloadError(failed_count)
 
     return transferred_count
+
+
+def _log_unversioned(data: List["DataIndex"]) -> Tuple[List["DataIndex"], int]:
+    ret: List["DataIndex"] = []
+    unversioned: List[str] = []
+    for fs_index in data:
+        remote = fs_index.storage_map[()].remote
+        if not isinstance(remote, FileStorage) or not remote.fs.version_aware:
+            ret.append(fs_index)
+            continue
+
+        fs = remote.fs
+        index = DataIndex()
+        index.storage_map = fs_index.storage_map
+        for key, entry in fs_index.iteritems():
+            if entry.meta and not entry.meta.isdir and entry.meta.version_id is None:
+                unversioned.append(fs.unstrip_protocol(fs.path.join(remote.path, *key)))
+            else:
+                index[key] = entry
+        fs_index.close()
+        ret.append(index)
+
+    if unversioned:
+        logger.warning(
+            (
+                "Some files are missing cloud version information and will not be "
+                "fetched from the remote:\n%s"
+            ),
+            "\n".join(unversioned),
+        )
+    return ret, len(unversioned)

dvc/repo/get.py

@@ -19,22 +19,39 @@ class GetDVCFileError(DvcException):
         )
 
 
-def get(url, path, out=None, rev=None, jobs=None, force=False):
+def get(
+    url,
+    path,
+    out=None,
+    rev=None,
+    jobs=None,
+    force=False,
+    config=None,
+    remote=None,
+    remote_config=None,
+):
+    from dvc.config import Config
     from dvc.dvcfile import is_valid_filename
-    from dvc.fs.callbacks import Callback
     from dvc.repo import Repo
 
     out = resolve_output(path, out, force=force)
 
     if is_valid_filename(out):
-        raise GetDVCFileError()
+        raise GetDVCFileError
+
+    if config and not isinstance(config, dict):
+        config = Config.load_file(config)
 
     with Repo.open(
         url=url,
         rev=rev,
         subrepos=True,
         uninitialized=True,
+        config=config,
+        remote=remote,
+        remote_config=remote_config,
     ) as repo:
+        from dvc.fs import download
         from dvc.fs.data import DataFileSystem
 
         fs: Union[DataFileSystem, "DVCFileSystem"]
@@ -44,14 +61,9 @@ def get(url, path, out=None, rev=None, jobs=None, force=False):
         else:
             fs = repo.dvcfs
             fs_path = fs.from_os_path(path)
-
-        with Callback.as_tqdm_callback(
-            desc=f"Downloading {fs.path.name(path)}",
-            unit="files",
-        ) as cb:
-            fs.get(
-                fs_path,
-                os.path.abspath(out),
-                batch_size=jobs,
-                callback=cb,
-            )
+        download(
+            fs,
+            fs_path,
+            os.path.abspath(out),
+            jobs=jobs,
+        )

dvc/repo/get_url.py

@@ -6,12 +6,12 @@ from dvc.fs import download, parse_external_url
 from dvc.utils import resolve_output
 
 
-def get_url(url, out=None, *, config=None, jobs=None, force=False):
+def get_url(url, out=None, *, fs_config=None, jobs=None, force=False):
     out = resolve_output(url, out, force=force)
     out = os.path.abspath(out)
     (out,) = output.loads_from(None, [out], use_cache=False)
 
-    src_fs, src_path = parse_external_url(url, config)
+    src_fs, src_path = parse_external_url(url, fs_config)
     if not src_fs.exists(src_path):
         raise URLMissingError(url)
-    download(src_fs, src_path, out, jobs=jobs)
+    download(src_fs, src_path, out.fs_path, jobs=jobs)

dvc/repo/graph.py

@@ -1,4 +1,4 @@
-from typing import TYPE_CHECKING, Iterator, List, Set
+from typing import TYPE_CHECKING, Any, Iterator, List, Optional, Set, TypeVar
 
 from dvc.fs import localfs
 from dvc.utils.fs import path_isin
@@ -8,6 +8,8 @@ if TYPE_CHECKING:
 
     from dvc.stage import Stage
 
+T = TypeVar("T")
+
 
 def check_acyclic(graph: "DiGraph") -> None:
     import networkx as nx
@@ -29,6 +31,9 @@ def check_acyclic(graph: "DiGraph") -> None:
 
 def get_pipeline(pipelines, node):
     found = [i for i in pipelines if i.has_node(node)]
+    if not found:
+        return None
+
     assert len(found) == 1
     return found[0]
 
@@ -39,10 +44,28 @@ def get_pipelines(graph: "DiGraph"):
     return [graph.subgraph(c).copy() for c in nx.weakly_connected_components(graph)]
 
 
+def get_subgraph_of_nodes(
+    graph: "DiGraph", sources: Optional[List[Any]] = None, downstream: bool = False
+) -> "DiGraph":
+    from networkx import dfs_postorder_nodes, reverse_view
+
+    if not sources:
+        return graph
+
+    g = reverse_view(graph) if downstream else graph
+    nodes = []
+    for source in sources:
+        nodes.extend(dfs_postorder_nodes(g, source))
+    return graph.subgraph(nodes)
+
+
 def collect_pipeline(stage: "Stage", graph: "DiGraph") -> Iterator["Stage"]:
     import networkx as nx
 
     pipeline = get_pipeline(get_pipelines(graph), stage)
+    if not pipeline:
+        return iter([])
+
     return nx.dfs_postorder_nodes(pipeline, stage)
 
 

dvc/repo/imp.py

@@ -1,6 +1,44 @@
-def imp(self, url, path, out=None, rev=None, **kwargs):
+def imp(
+    self,
+    url,
+    path,
+    out=None,
+    rev=None,
+    config=None,
+    remote=None,
+    remote_config=None,
+    **kwargs,
+):
     erepo = {"url": url}
     if rev is not None:
         erepo["rev"] = rev
 
+    if remote and remote_config and isinstance(config, str):
+        raise ValueError(
+            "Can't specify config path together with both remote and remote_config"
+        )
+
+    if config is not None:
+        erepo["config"] = config
+
+    if remote is not None and remote_config is not None:
+        conf = erepo.get("config") or {}
+
+        core = conf.get("core") or {}
+        core["remote"] = remote
+
+        remotes = conf.get("remote") or {}
+        remote_conf = remotes.get(remote) or {}
+        remote_conf.update(remote_config)
+        remotes[remote] = remote_conf
+
+        conf["core"] = core
+        conf["remote"] = remotes
+
+        erepo["config"] = conf
+    elif remote is not None:
+        erepo["remote"] = remote
+    elif remote_config is not None:
+        erepo["remote"] = remote_config
+
     return self.imp_url(path, out=out, erepo=erepo, frozen=True, **kwargs)

dvc/repo/index.py

@@ -374,6 +374,50 @@ class Index:
 
         return dict(by_workspace)
 
+    @cached_property
+    def metric_keys(self) -> Dict[str, Set["DataIndexKey"]]:
+        from collections import defaultdict
+
+        from .metrics.show import _collect_top_level_metrics
+
+        by_workspace: Dict[str, Set["DataIndexKey"]] = defaultdict(set)
+
+        by_workspace["repo"] = set()
+
+        for out in self.outs:
+            if not out.metric:
+                continue
+
+            workspace, key = out.index_key
+            by_workspace[workspace].add(key)
+
+        for path in _collect_top_level_metrics(self.repo):
+            key = self.repo.fs.path.relparts(path, self.repo.root_dir)
+            by_workspace["repo"].add(key)
+
+        return dict(by_workspace)
+
+    @cached_property
+    def plot_keys(self) -> Dict[str, Set["DataIndexKey"]]:
+        from collections import defaultdict
+
+        by_workspace: Dict[str, Set["DataIndexKey"]] = defaultdict(set)
+
+        by_workspace["repo"] = set()
+
+        for out in self.outs:
+            if not out.plot:
+                continue
+
+            workspace, key = out.index_key
+            by_workspace[workspace].add(key)
+
+        for path in self._plot_sources:
+            key = self.repo.fs.path.parts(path)
+            by_workspace["repo"].add(key)
+
+        return dict(by_workspace)
+
     @cached_property
     def data_tree(self):
         from dvc_data.hashfile.tree import Tree
@@ -487,11 +531,31 @@ class Index:
                 used[odb].update(objs)
         return used
 
+    def _types_filter(self, types, out):
+        ws, okey = out.index_key
+        for typ in types:
+            if typ == "plots":
+                keys = self.plot_keys
+            elif typ == "metrics":
+                keys = self.metric_keys
+            else:
+                raise ValueError(f"unsupported type {typ}")
+
+            for key in keys.get(ws, []):
+                if (len(key) >= len(okey) and key[: len(okey)] == okey) or (
+                    len(key) < len(okey) and okey[: len(key)] == key
+                ):
+                    return True
+
+        return False
+
     def targets_view(
         self,
         targets: Optional["TargetType"],
         stage_filter: Optional[Callable[["Stage"], bool]] = None,
         outs_filter: Optional[Callable[["Output"], bool]] = None,
+        max_size: Optional[int] = None,
+        types: Optional[List[str]] = None,
         **kwargs: Any,
     ) -> "IndexView":
         """Return read-only view of index for the specified targets.
@@ -514,7 +578,20 @@ class Index:
             for stage_info in self.collect_targets(targets, **kwargs)
             if not stage_filter or stage_filter(stage_info.stage)
         ]
-        return IndexView(self, stage_infos, outs_filter=outs_filter)
+
+        def _outs_filter(out):
+            if max_size and out.meta and out.meta.size and out.meta.size >= max_size:
+                return False
+
+            if types and not self._types_filter(types, out):
+                return False
+
+            if outs_filter:
+                return outs_filter(out)
+
+            return True
+
+        return IndexView(self, stage_infos, outs_filter=_outs_filter)
 
 
 class _DataPrefixes(NamedTuple):

dvc/repo/init.py

@@ -37,7 +37,7 @@ def init(root_dir=os.curdir, no_scm=False, force=False, subdir=False):  # noqa:
             "Cannot initialize repo with `--no-scm` and `--subdir`"
         )
 
-    root_dir = os.path.realpath(root_dir)
+    root_dir = os.path.abspath(root_dir)
     dvc_dir = os.path.join(root_dir, Repo.DVC_DIR)
 
     try:

dvc/repo/ls.py

@@ -13,6 +13,9 @@ def ls(
     rev: Optional[str] = None,
     recursive: Optional[bool] = None,
     dvc_only: bool = False,
+    config: Optional[str] = None,
+    remote: Optional[str] = None,
+    remote_config: Optional[dict] = None,
 ):
     """Methods for getting files and outputs for the repo.
 
@@ -22,6 +25,9 @@ def ls(
         rev (str, optional): SHA commit, branch or tag name
         recursive (bool, optional): recursively walk the repo
         dvc_only (bool, optional): show only DVC-artifacts
+        config (str, optional): path to config file
+        remote (str, optional): remote name to set as a default remote in the repo
+        remote_config (str, dict): remote config to merge with a remote in the repo
 
     Returns:
         list of `entry`
@@ -35,9 +41,24 @@ def ls(
             "isexec": bool,
         }
     """
+    from dvc.config import Config
+
     from . import Repo
 
-    with Repo.open(url, rev=rev, subrepos=True, uninitialized=True) as repo:
+    if config and not isinstance(config, dict):
+        config_dict = Config.load_file(config)
+    else:
+        config_dict = None
+
+    with Repo.open(
+        url,
+        rev=rev,
+        subrepos=True,
+        uninitialized=True,
+        config=config_dict,
+        remote=remote,
+        remote_config=remote_config,
+    ) as repo:
         path = path or ""
 
         ret = _ls(repo, path, recursive, dvc_only)
@@ -88,6 +109,7 @@ def _ls(
             "isout": dvc_info.get("isout", False),
             "isdir": info["type"] == "directory",
             "isexec": info.get("isexec", False),
+            "size": info.get("size"),
         }
 
     return ret

dvc/repo/ls_url.py

@@ -2,8 +2,8 @@ from dvc.exceptions import URLMissingError
 from dvc.fs import parse_external_url
 
 
-def ls_url(url, *, config=None, recursive=False):
-    fs, fs_path = parse_external_url(url, config=config)
+def ls_url(url, *, fs_config=None, recursive=False):
+    fs, fs_path = parse_external_url(url, config=fs_config)
     try:
         info = fs.info(fs_path)
     except FileNotFoundError as exc:
@@ -20,6 +20,7 @@ def ls_url(url, *, config=None, recursive=False):
             ls_info = {
                 "path": fs.path.relpath(info["name"], fs_path),
                 "isdir": info["type"] == "directory",
+                "size": info.get("size"),
             }
             ret.append(ls_info)
 

dvc/repo/open_repo.py

@@ -2,8 +2,7 @@ import logging
 import os
 import tempfile
 import threading
-from contextlib import contextmanager
-from typing import TYPE_CHECKING, Dict, Iterator, Optional, Tuple
+from typing import TYPE_CHECKING, Dict, Optional, Tuple
 
 from funcy import retry, wrap_with
 
@@ -18,16 +17,14 @@ if TYPE_CHECKING:
 logger = logging.getLogger(__name__)
 
 
-@contextmanager
 @map_scm_exception()
 def _external_repo(
     url,
     rev: Optional[str] = None,
-    for_write: bool = False,
     **kwargs,
-) -> Iterator["Repo"]:
+) -> "Repo":
     logger.debug("Creating external repo %s@%s", url, rev)
-    path = _cached_clone(url, rev, for_write=for_write)
+    path = _cached_clone(url, rev)
     # Local HEAD points to the tip of whatever branch we first cloned from
     # (which may not be the default branch), use origin/HEAD here to get
     # the tip of the default branch
@@ -38,11 +35,6 @@ def _external_repo(
     config.update(kwargs.pop("config", None) or {})
 
     main_root = "/"
-    if for_write:
-        # we already checked out needed revision
-        rev = None
-        main_root = path
-
     repo_kwargs = dict(
         root_dir=path,
         url=url,
@@ -52,14 +44,7 @@ def _external_repo(
         **kwargs,
     )
 
-    repo = Repo(**repo_kwargs)
-
-    try:
-        yield repo
-    finally:
-        repo.close()
-        if for_write:
-            _remove(path)
+    return Repo(**repo_kwargs)
 
 
 def open_repo(url, *args, **kwargs):
@@ -67,6 +52,7 @@ def open_repo(url, *args, **kwargs):
         url = os.getcwd()
 
     if os.path.exists(url):
+        url = os.path.abspath(url)
         try:
             config = _get_remote_config(url)
             config.update(kwargs.get("config") or {})
@@ -139,20 +125,19 @@ def _get_remote_config(url):
         repo.close()
 
 
-def _cached_clone(url, rev, for_write=False):
+def _cached_clone(url, rev):
     """Clone an external git repo to a temporary directory.
 
     Returns the path to a local temporary directory with the specified
-    revision checked out. If for_write is set prevents reusing this dir via
-    cache.
+    revision checked out.
     """
     from shutil import copytree
 
     # even if we have already cloned this repo, we may need to
     # fetch/fast-forward to get specified rev
-    clone_path, shallow = _clone_default_branch(url, rev, for_write=for_write)
+    clone_path, shallow = _clone_default_branch(url, rev)
 
-    if not for_write and (url) in CLONES:
+    if url in CLONES:
         return CLONES[url][0]
 
     # Copy to a new dir to keep the clone clean
@@ -160,16 +145,12 @@ def _cached_clone(url, rev, for_write=False):
     logger.debug("erepo: making a copy of %s clone", url)
     copytree(clone_path, repo_path)
 
-    # Check out the specified revision
-    if for_write:
-        _git_checkout(repo_path, rev)
-    else:
-        CLONES[url] = (repo_path, shallow)
+    CLONES[url] = (repo_path, shallow)
     return repo_path
 
 
 @wrap_with(threading.Lock())
-def _clone_default_branch(url, rev, for_write=False):  # noqa: C901, PLR0912
+def _clone_default_branch(url, rev):  # noqa: C901, PLR0912
     """Get or create a clean clone of the url.
 
     The cloned is reactualized with git pull unless rev is a known sha.
@@ -203,7 +184,7 @@ def _clone_default_branch(url, rev, for_write=False):  # noqa: C901, PLR0912
 
             logger.debug("erepo: git clone '%s' to a temporary dir", url)
             clone_path = tempfile.mkdtemp("dvc-clone")
-            if not for_write and rev and not Git.is_sha(rev):
+            if rev and not Git.is_sha(rev):
                 # If rev is a tag or branch name try shallow clone first
 
                 try:
@@ -248,17 +229,6 @@ def _merge_upstream(git: "Git"):
         pass
 
 
-def _git_checkout(repo_path, rev):
-    from dvc.scm import Git
-
-    logger.debug("erepo: git checkout %s@%s", repo_path, rev)
-    git = Git(repo_path)
-    try:
-        git.checkout(rev)
-    finally:
-        git.close()
-
-
 def _remove(path):
     from dvc.utils.fs import remove
 

dvc/repo/plots/__init__.py

@@ -46,8 +46,8 @@ SUPPORTED_IMAGE_EXTENSIONS = ImageRenderer.EXTENSIONS
 class PlotMetricTypeError(DvcException):
     def __init__(self, file):
         super().__init__(
-            "'{}' - file type error\n"
-            "Only JSON, YAML, CSV and TSV formats are supported.".format(file)
+            f"'{file}' - file type error\n"
+            "Only JSON, YAML, CSV and TSV formats are supported."
         )
 
 
@@ -134,7 +134,7 @@ class Plots:
             if "workspace" in revs:
                 # reorder revs to match repo.brancher ordering
                 revs.remove("workspace")
-                revs = ["workspace"] + revs
+                revs = ["workspace", *revs]
         for rev in revs:
             with switch_repo(self.repo, rev) as (repo, _):
                 res: Dict = {}
@@ -212,7 +212,7 @@ class Plots:
             short_rev = "workspace"
             if rev := getattr(self.repo.fs, "rev", None):
                 short_rev = rev[:7]
-            _resolve_data_sources(data, short_rev, cache_remote_stream=True)
+            _resolve_data_sources(data, short_rev, cache=True)
             result.update(data)
 
         errored = errored_revisions(result)
@@ -281,9 +281,7 @@ def _is_plot(out: "Output") -> bool:
     return bool(out.plot)
 
 
-def _resolve_data_sources(
-    plots_data: Dict, rev: str, cache_remote_stream: bool = False
-):
+def _resolve_data_sources(plots_data: Dict, rev: str, cache: bool = False):
     from dvc.progress import Tqdm
 
     values = list(plots_data.values())
@@ -298,7 +296,7 @@ def _resolve_data_sources(
     def resolve(value):
         data_source = value.pop("data_source")
         assert callable(data_source)
-        value.update(data_source(cache_remote_stream=cache_remote_stream))
+        value.update(data_source(cache=cache))
 
     if not to_resolve:
         return
@@ -448,6 +446,7 @@ def _resolve_definitions(
     definitions: "DictStrAny",
     onerror: Optional[Callable[[Any], Any]] = None,
 ):
+    config_path = os.fspath(config_path)
     config_dir = fs.path.dirname(config_path)
     result: Dict[str, Dict] = {}
     for plot_id, plot_props in definitions.items():
@@ -483,7 +482,7 @@ def _collect_pipeline_files(repo, targets: List[str], props, onerror=None):
             if isinstance(elem, str):
                 dvcfile_defs_dict[elem] = None
             else:
-                k, v = list(elem.items())[0]
+                k, v = next(iter(elem.items()))
                 dvcfile_defs_dict[k] = v
 
         resolved = _resolve_definitions(

dvc/repo/pull.py

@@ -1,5 +1,6 @@
 import logging
 
+from dvc.exceptions import CheckoutError
 from dvc.repo import locked
 from dvc.utils import glob_targets
 
@@ -38,13 +39,17 @@ def pull(  # noqa: PLR0913
         recursive=recursive,
         run_cache=run_cache,
     )
-    stats = self.checkout(
-        targets=expanded_targets,
-        with_deps=with_deps,
-        force=force,
-        recursive=recursive,
-        allow_missing=allow_missing,
-    )
+    try:
+        stats = self.checkout(
+            targets=expanded_targets,
+            with_deps=with_deps,
+            force=force,
+            recursive=recursive,
+            allow_missing=allow_missing,
+        )
+    except CheckoutError as exc:
+        exc.stats["fetched"] = processed_files_count
+        raise
 
     stats["fetched"] = processed_files_count
     return stats

dvc/repo/reproduce.py

@@ -1,130 +1,85 @@
 import logging
-from typing import TYPE_CHECKING, Iterator, List
+from typing import (
+    TYPE_CHECKING,
+    Callable,
+    Dict,
+    Iterable,
+    List,
+    NoReturn,
+    Optional,
+    Set,
+    Tuple,
+    TypeVar,
+    Union,
+    cast,
+)
+
+from funcy import ldistinct
 
 from dvc.exceptions import ReproductionError
 from dvc.repo.scm_context import scm_context
 from dvc.stage.cache import RunCacheNotSupported
+from dvc.utils import humanize
+from dvc.utils.collections import ensure_list
 
 from . import locked
 
 if TYPE_CHECKING:
+    from networkx import DiGraph
+
     from dvc.stage import Stage
 
     from . import Repo
 
 logger = logging.getLogger(__name__)
+T = TypeVar("T")
 
 
-def _reproduce_stage(stage: "Stage", **kwargs) -> List["Stage"]:
-    if stage.frozen and not stage.is_import:
-        logger.warning(
-            "%s is frozen. Its dependencies are not going to be reproduced.",
-            stage,
-        )
-
-    stage = stage.reproduce(**kwargs)
-    if not stage:
-        return []
-
-    if not kwargs.get("dry", False):
-        stage.dump(update_pipeline=False)
-        _track_stage(stage)
-
-    return [stage]
+def collect_stages(
+    repo: "Repo",
+    targets: Iterable[str],
+    recursive: bool = False,
+    glob: bool = False,
+) -> List["Stage"]:
+    stages: List["Stage"] = []
+    for target in targets:
+        stages.extend(repo.stage.collect(target, recursive=recursive, glob=glob))
+    return ldistinct(stages)
 
 
-def _get_stage_files(stage: "Stage") -> Iterator[str]:
-    yield stage.dvcfile.relpath
-    for dep in stage.deps:
-        if (
-            not dep.use_scm_ignore
-            and dep.is_in_repo
-            and not stage.repo.dvcfs.isdvc(stage.repo.dvcfs.from_os_path(str(dep)))
-        ):
-            yield dep.fs_path
-    for out in stage.outs:
-        if not out.use_scm_ignore and out.is_in_repo:
-            yield out.fs_path
+def get_subgraph(
+    graph: "DiGraph",
+    nodes: Optional[List] = None,
+    pipeline: bool = False,
+    downstream: bool = False,
+) -> "DiGraph":
+    import networkx as nx
+
+    from .graph import get_pipeline, get_pipelines, get_subgraph_of_nodes
+
+    if not pipeline or not nodes:
+        return get_subgraph_of_nodes(graph, nodes, downstream=downstream)
+
+    pipelines = get_pipelines(graph)
+    used_pipelines = [get_pipeline(pipelines, node) for node in nodes]
+    return nx.compose_all(used_pipelines)
 
 
-def _track_stage(stage: "Stage") -> None:
-    from dvc.utils import relpath
-
-    context = stage.repo.scm_context
-    for path in _get_stage_files(stage):
-        context.track_file(relpath(path))
-    return context.track_changed_files()
+def get_active_graph(graph: "DiGraph") -> "DiGraph":
+    g = cast("DiGraph", graph.copy())
+    for stage in graph:
+        if stage.frozen:
+            # NOTE: disconnect frozen stage from its dependencies
+            g.remove_edges_from(graph.out_edges(stage))
+    return g
 
 
-@locked
-@scm_context
-def reproduce(  # noqa: C901, PLR0912
-    self: "Repo",
-    targets=None,
-    recursive=False,
-    pipeline=False,
-    all_pipelines=False,
-    **kwargs,
-):
-    from .graph import get_pipeline, get_pipelines
-
-    glob = kwargs.pop("glob", False)
-
-    if isinstance(targets, str):
-        targets = [targets]
-
-    if not all_pipelines and not targets:
-        from dvc.dvcfile import PROJECT_FILE
-
-        targets = [PROJECT_FILE]
-
-    interactive = kwargs.get("interactive", False)
-    if not interactive:
-        kwargs["interactive"] = self.config["core"].get("interactive", False)
-
-    stages = set()
-    if pipeline or all_pipelines:
-        pipelines = get_pipelines(self.index.graph)
-        if all_pipelines:
-            used_pipelines = pipelines
-        else:
-            used_pipelines = []
-            for target in targets:
-                stage = self.stage.get_target(target)
-                used_pipelines.append(get_pipeline(pipelines, stage))
-
-        for pline in used_pipelines:
-            for stage in pline:
-                if pline.in_degree(stage) == 0:
-                    stages.add(stage)
-    else:
-        for target in targets:
-            stages.update(
-                self.stage.collect(
-                    target,
-                    recursive=recursive,
-                    glob=glob,
-                )
-            )
-
-    if kwargs.get("pull", False) and kwargs.get("run_cache", True):
-        logger.debug("Pulling run cache")
-        try:
-            self.stage_cache.pull(None)
-        except RunCacheNotSupported as e:
-            logger.warning("Failed to pull run cache: %s", e)
-
-    return _reproduce_stages(self.index.graph, list(stages), **kwargs)
-
-
-def _reproduce_stages(  # noqa: C901
-    graph,
-    stages,
-    downstream=False,
-    single_item=False,
-    on_unchanged=None,
-    **kwargs,
-):
+def plan_repro(
+    graph: "DiGraph",
+    stages: Optional[List["T"]] = None,
+    pipeline: bool = False,
+    downstream: bool = False,
+) -> List["T"]:
     r"""Derive the evaluation of the given node for the given graph.
 
     When you _reproduce a stage_, you want to _evaluate the descendants_
@@ -160,80 +115,146 @@ def _reproduce_stages(  # noqa: C901
 
     The derived evaluation of _downstream_ B would be: [B, D, E]
     """
-    steps = _get_steps(graph, stages, downstream, single_item)
+    import networkx as nx
+
+    sub = get_subgraph(graph, stages, pipeline=pipeline, downstream=downstream)
+    return list(nx.dfs_postorder_nodes(sub))
+
+
+def _reproduce_stage(stage: "Stage", **kwargs) -> Optional["Stage"]:
+    if stage.frozen and not stage.is_import:
+        msg = "%s is frozen. Its dependencies are not going to be reproduced."
+        logger.warning(msg, stage)
+
+    ret = stage.reproduce(**kwargs)
+    if ret and not kwargs.get("dry", False):
+        stage.dump(update_pipeline=False)
+    return ret
+
+
+def _get_upstream_downstream_nodes(
+    graph: Optional["DiGraph"], node: T
+) -> Tuple[List[T], List[T]]:
+    succ = list(graph.successors(node)) if graph else []
+    pre = list(graph.predecessors(node)) if graph else []
+    return succ, pre
+
+
+def _repr(stages: Iterable["Stage"]) -> str:
+    return humanize.join(repr(stage.addressing) for stage in stages)
+
+
+def handle_error(
+    graph: Optional["DiGraph"], on_error: str, exc: Exception, stage: "Stage"
+) -> Set["Stage"]:
+    import networkx as nx
+
+    logger.warning("%s%s", exc, " (ignored)" if on_error == "ignore" else "")
+    if not graph or on_error == "ignore":
+        return set()
+
+    dependents = set(nx.dfs_postorder_nodes(graph.reverse(), stage)) - {stage}
+    if dependents:
+        names = _repr(dependents)
+        msg = "%s %s will be skipped due to this failure"
+        logger.warning(msg, "Stages" if len(dependents) > 1 else "Stage", names)
+    return dependents
+
+
+def _raise_error(exc: Optional[Exception], *stages: "Stage") -> NoReturn:
+    names = _repr(stages)
+    segment = " stages:" if len(stages) > 1 else ""
+    raise ReproductionError(f"failed to reproduce{segment} {names}") from exc
+
+
+def _reproduce(
+    stages: List["Stage"],
+    graph: Optional["DiGraph"] = None,
+    force_downstream: bool = False,
+    on_error: str = "fail",
+    force: bool = False,
+    repro_fn: Callable = _reproduce_stage,
+    **kwargs,
+) -> List["Stage"]:
+    assert on_error in ("fail", "keep-going", "ignore")
 
-    force_downstream = kwargs.pop("force_downstream", False)
     result: List["Stage"] = []
-    unchanged: List["Stage"] = []
-    # `ret` is used to add a cosmetic newline.
-    ret: List["Stage"] = []
+    failed: List["Stage"] = []
+    to_skip: Dict["Stage", "Stage"] = {}
+    ret: Optional["Stage"] = None
+
+    force_state = {node: force for node in stages}
+
+    for stage in stages:
+        if stage in to_skip:
+            continue
 
-    for stage in steps:
         if ret:
-            logger.info("")
+            logger.info("")  # add a newline
+
+        upstream, downstream = _get_upstream_downstream_nodes(graph, stage)
+        force_stage = force_state[stage]
 
         try:
-            kwargs["upstream"] = result + unchanged
-            ret = _reproduce_stage(stage, **kwargs)
+            ret = repro_fn(stage, upstream=upstream, force=force_stage, **kwargs)
+        except Exception as exc:  # noqa: BLE001, pylint: disable=broad-exception-caught
+            failed.append(stage)
+            if on_error == "fail":
+                _raise_error(exc, stage)
 
-            if len(ret) == 0:
-                unchanged.extend([stage])
-            elif force_downstream:
-                # NOTE: we are walking our pipeline from the top to the
-                # bottom. If one stage is changed, it will be reproduced,
-                # which tells us that we should force reproducing all of
-                # the other stages down below, even if their direct
-                # dependencies didn't change.
-                kwargs["force"] = True
+            dependents = handle_error(graph, on_error, exc, stage)
+            to_skip.update({node: stage for node in dependents})
+            continue
 
-            if ret:
-                result.extend(ret)
-        except Exception as exc:  # noqa: BLE001
-            raise ReproductionError(stage.addressing) from exc
+        if force_downstream and (ret or force_stage):
+            force_state.update({node: True for node in downstream})
 
-    if on_unchanged is not None:
-        on_unchanged(unchanged)
+        if ret:
+            result.append(ret)
+
+    if on_error != "ignore" and failed:
+        _raise_error(None, *failed)
     return result
 
 
-def _get_steps(graph, stages, downstream, single_item):
-    import networkx as nx
+@locked
+@scm_context
+def reproduce(
+    self: "Repo",
+    targets: Union[Iterable[str], str, None] = None,
+    recursive: bool = False,
+    pipeline: bool = False,
+    all_pipelines: bool = False,
+    downstream: bool = False,
+    single_item: bool = False,
+    glob: bool = False,
+    on_error: Optional[str] = "fail",
+    **kwargs,
+):
+    from dvc.dvcfile import PROJECT_FILE
 
-    active = graph.copy()
+    if all_pipelines or pipeline:
+        single_item = False
+        downstream = False
+
+    if not kwargs.get("interactive", False):
+        kwargs["interactive"] = self.config["core"].get("interactive", False)
+
+    stages: List["Stage"] = []
+    if not all_pipelines:
+        targets_list = ensure_list(targets or PROJECT_FILE)
+        stages = collect_stages(self, targets_list, recursive=recursive, glob=glob)
+
+    if kwargs.get("pull", False) and kwargs.get("run_cache", True):
+        logger.debug("Pulling run cache")
+        try:
+            self.stage_cache.pull(None)
+        except RunCacheNotSupported as e:
+            logger.warning("Failed to pull run cache: %s", e)
+
+    graph = None
+    steps = stages
     if not single_item:
-        # NOTE: frozen stages don't matter for single_item
-        for stage in graph:
-            if stage.frozen:
-                # NOTE: disconnect frozen stage from its dependencies
-                active.remove_edges_from(graph.out_edges(stage))
-
-    all_pipelines: List["Stage"] = []
-    for stage in stages:
-        if downstream:
-            # NOTE (py3 only):
-            # Python's `deepcopy` defaults to pickle/unpickle the object.
-            # Stages are complex objects (with references to `repo`,
-            # `outs`, and `deps`) that cause struggles when you try
-            # to serialize them. We need to create a copy of the graph
-            # itself, and then reverse it, instead of using
-            # graph.reverse() directly because it calls `deepcopy`
-            # underneath -- unless copy=False is specified.
-            nodes = nx.dfs_postorder_nodes(active.reverse(copy=False), stage)
-            all_pipelines += reversed(list(nodes))
-        else:
-            all_pipelines += nx.dfs_postorder_nodes(active, stage)
-
-    steps = []
-    for stage in all_pipelines:
-        if stage not in steps:
-            # NOTE: order of steps still matters for single_item
-            if single_item and stage not in stages:
-                continue
-
-            steps.append(stage)
-
-    return steps
-
-
-def _repro_callback(experiments_callback, unchanged, stages):
-    experiments_callback(unchanged, stages)
+        graph = get_active_graph(self.index.graph)
+        steps = plan_repro(graph, stages, pipeline=pipeline, downstream=downstream)
+    return _reproduce(steps, graph=graph, on_error=on_error or "fail", **kwargs)

dvc/repo/stage.py

@@ -37,11 +37,19 @@ StageSet = Set["Stage"]
 
 
 def _collect_with_deps(stages: StageList, graph: "DiGraph") -> StageSet:
+    from dvc.exceptions import StageNotFoundError
     from dvc.repo.graph import collect_pipeline
 
     res: StageSet = set()
     for stage in stages:
-        res.update(collect_pipeline(stage, graph=graph))
+        pl = list(collect_pipeline(stage, graph=graph))
+        if not pl:
+            raise StageNotFoundError(
+                f"Stage {stage} is not found in the project. "
+                "Check that there are no symlinks in the parents "
+                "leading up to it within the project."
+            )
+        res.update(pl)
     return res
 
 
@@ -211,7 +219,7 @@ class StageLoad:
         self, path: Optional[str] = None, name: Optional[str] = None
     ) -> str:
         if path:
-            return self.repo.fs.path.realpath(path)
+            return self.repo.fs.path.abspath(path)
 
         path = PROJECT_FILE
         logger.debug("Assuming '%s' to be a stage inside '%s'", name, path)
@@ -235,7 +243,7 @@ class StageLoad:
     ) -> Iterable[str]:
         if not name:
             return stages.keys()
-        if accept_group and stages.is_foreach_generated(name):
+        if accept_group and stages.is_foreach_or_matrix_generated(name):
             return self._get_group_keys(stages, name)
         if glob:
             return fnmatch.filter(stages.keys(), name)