Delete release-drafter.yml

fix: #8450

1. Add a new function check_rwlock to check and autocorrect rwlock info.
2. Add related tests for it.

.github/release-drafter.yml

@@ -1,49 +0,0 @@
-# Config for https://github.com/apps/release-drafter
-name-template: "$RESOLVED_VERSION 🦉 "
-tag-template: "$RESOLVED_VERSION"
-categories:
-  - title: "🚀 New Features and Enhancements "
-    labels:
-      - "feature"
-      - "enhancement"
-      - "ui"
-  - title: "💣 Breaking Changes "
-    labels:
-      - "breaking-change"
-  - title: "🏇 Optimizations "
-    labels:
-      - "optimize"
-      - "performance"
-  - title: "🐛 Bug Fixes "
-    labels:
-      - "bugfix"
-      - "bug"
-  - title: "🔨 Maintenance "
-    labels:
-      - "maintenance"
-      - "refactoring"
-      - "chore"
-      - "build"
-      - "ci"
-change-template: "- $TITLE (#$NUMBER) @$AUTHOR"
-version-resolver:
-  major:
-    labels:
-      - "major"
-  minor:
-    labels:
-      - "minor"
-  patch:
-    labels:
-      - "patch"
-  default: patch
-exclude-labels:
-  - "skip-changelog"
-template: |
-  > Refer to https://dvc.org/doc/install for installation instructions.
-
-  ## Changes
-
-  $CHANGES
-
-  Thanks again to $CONTRIBUTORS for the contributions! 🎉

.github/workflows/tests.yaml

@@ -66,8 +66,6 @@ jobs:
       run: |
         pip install --upgrade pip setuptools wheel
         pip install -e ".[dev]"
-      env:
-        AIOHTTP_NO_EXTENSIONS: ${{ matrix.pyv == '3.11-dev' && '1' }}
     - name: run tests
       timeout-minutes: 40
       run: pytest -nauto -ra --durations 100 --cov=dvc --cov-report=xml --cov-report=term -k "${{ matrix.pytest-filter }}"

.pre-commit-config.yaml

@@ -6,7 +6,7 @@ repos:
       - id: black
         language_version: python3
     repo: https://github.com/ambv/black
-    rev: 22.6.0
+    rev: 22.10.0
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v4.3.0
     hooks:
@@ -16,9 +16,9 @@ repos:
       - id: codespell
         args: 
           - --ignore-words-list
-          - ba,datas,fo,uptodate,cachable
+          - ba,datas,fo,uptodate,cachable,falsy
     repo: https://github.com/codespell-project/codespell
-    rev: v2.1.0
+    rev: v2.2.2
   - hooks:
       - id: isort
         language_version: python3

README.rst

@@ -82,7 +82,7 @@ To share and back up the *data cache*, DVC supports multiple remote storage plat
 
 |Flowchart|
 
-`DVC pipelines <https://dvc.org/doc/start/data-management/pipelines>`_ (computational graphs) connect code and data together.
+`DVC pipelines <https://dvc.org/doc/start/data-management/data-pipelines>`_ (computational graphs) connect code and data together.
 They specify all steps required to produce a model: input dependencies including code, data, commands to run; and output information to be saved.
 
 Last but not least, `DVC Experiment Versioning <https://dvc.org/doc/start/experiments>`_ lets you prepare and run a large number of experiments.

dvc/_debug.py

@@ -1,16 +1,14 @@
 from contextlib import ExitStack, contextmanager
 from datetime import datetime
-from typing import TYPE_CHECKING, Callable, Union
+from typing import TYPE_CHECKING, Callable, Optional, Union
 
 if TYPE_CHECKING:
     from argparse import Namespace
+    from types import FrameType
 
 
 @contextmanager
-def viztracer_profile(
-    path: Union[Callable[[], str], str],
-    max_stack_depth: int = -1,
-):
+def viztracer_profile(path: Union[Callable[[], str], str], depth: int = -1):
     try:
         import viztracer  # pylint: disable=import-error
     except ImportError:
@@ -18,7 +16,7 @@ def viztracer_profile(
         yield
         return
 
-    tracer = viztracer.VizTracer(max_stack_depth=max_stack_depth)
+    tracer = viztracer.VizTracer(max_stack_depth=depth)
 
     tracer.start()
     yield
@@ -109,6 +107,38 @@ def debug():
         raise  # prevent from jumping ahead
 
 
+def _sigshow(_, frame: Optional["FrameType"]) -> None:
+    from shutil import get_terminal_size
+    from traceback import format_stack
+
+    from dvc.ui import ui
+
+    lines = "\u2015" * get_terminal_size().columns
+    ui.error_write(lines, "\n", *format_stack(frame), lines, sep="")
+
+
+@contextmanager
+def show_stack():
+    r"""Show stack trace on SIGQUIT (Ctrl-\) or SIGINFO (Ctrl-T on macOS)."""
+    import signal
+
+    signal.signal(signal.SIGQUIT, _sigshow)
+    try:
+        signal.signal(signal.SIGINFO, _sigshow)  # only available on macOS
+    except AttributeError:
+        pass
+    yield
+
+
+def _get_path_func(tool: str, ext: str):
+    fmt = f"{tool}.dvc-{{now:%Y%m%d}}_{{now:%H%M%S}}.{ext}"
+
+    def func(now: Optional["datetime"] = None) -> str:
+        return fmt.format(now=now or datetime.now())
+
+    return func
+
+
 @contextmanager
 def debugtools(args: "Namespace" = None, **kwargs):
     kw = vars(args) if args else {}
@@ -121,19 +151,16 @@ def debugtools(args: "Namespace" = None, **kwargs):
             stack.enter_context(profile(kw.get("cprofile_dump")))
         if kw.get("instrument") or kw.get("instrument_open"):
             stack.enter_context(instrument(kw.get("instrument_open", False)))
+        if kw.get("show_stack", False):
+            stack.enter_context(show_stack())
         if kw.get("yappi"):
-            output = "callgrind.dvc-{0:%Y%m%d}_{0:%H%M%S}.out"
-            stack.enter_context(
-                yappi_profile(path=lambda: output.format(datetime.now()))
-            )
+            path_func = _get_path_func("callgrind", "out")
+            stack.enter_context(yappi_profile(path=path_func))
         if kw.get("viztracer") or kw.get("viztracer_depth"):
-            output = "viztracer.dvc-{0:%Y%m%d}_{0:%H%M%S}.json"
-            stack.enter_context(
-                viztracer_profile(
-                    path=lambda: output.format(datetime.now()),
-                    max_stack_depth=kw.get("viztracer_depth") or -1,
-                )
-            )
+            path_func = _get_path_func("viztracer", "json")
+            depth = kw.get("viztracer_depth") or -1
+            prof = viztracer_profile(path=path_func, depth=depth)
+            stack.enter_context(prof)
         yield
 
 
@@ -160,3 +187,10 @@ def add_debugging_flags(parser):
     parser.add_argument(
         "--instrument-open", action="store_true", default=False, help=SUPPRESS
     )
+    parser.add_argument(
+        "--show-stack",
+        "--ss",
+        action="store_true",
+        default=False,
+        help=SUPPRESS,
+    )

dvc/cli/parser.py

@@ -29,6 +29,7 @@ from dvc.commands import (
     init,
     install,
     ls,
+    ls_url,
     machine,
     metrics,
     move,
@@ -75,6 +76,7 @@ COMMANDS = [
     install,
     root,
     ls,
+    ls_url,
     freeze,
     dag,
     daemon,

dvc/commands/checkout.py

@@ -110,7 +110,7 @@ def add_parser(subparsers, parent_parser):
         nargs="*",
         help=(
             "Limit command scope to these tracked files/directories, "
-            ".dvc files, or stage names."
+            ".dvc files and stage names."
         ),
     ).complete = completion.DVC_FILE
     checkout_parser.set_defaults(func=CmdCheckout)

dvc/commands/commit.py

@@ -70,7 +70,9 @@ def add_parser(subparsers, parent_parser):
     commit_parser.add_argument(
         "targets",
         nargs="*",
-        help="stages or .dvc files to commit. Optional. "
-        "(Finds all DVC files in the workspace by default.)",
+        help=(
+            "Limit command scope to these tracked files/directories, "
+            ".dvc files and stage names."
+        ),
     ).complete = completion.DVCFILES_AND_STAGE
     commit_parser.set_defaults(func=CmdCommit)

dvc/commands/dag.py

@@ -204,7 +204,7 @@ def add_parser(subparsers, parent_parser):
     dag_parser.add_argument(
         "target",
         nargs="?",
-        help="Stage or output to show pipeline for (optional). "
+        help="Stage name or output to show pipeline for. "
         "Finds all stages in the workspace by default.",
     )
     dag_parser.set_defaults(func=CmdDAG)

dvc/commands/data_sync.py

@@ -120,7 +120,7 @@ def shared_parent_parser():
         nargs="*",
         help=(
             "Limit command scope to these tracked files/directories, "
-            ".dvc files, or stage names."
+            ".dvc files and stage names."
         ),
     ).complete = completion.DVC_FILE
 
@@ -195,7 +195,7 @@ def add_parser(subparsers, _parent_parser):
         "--glob",
         action="store_true",
         default=False,
-        help="Pull cache for targets matching shell-style wildcards.",
+        help=argparse.SUPPRESS,
     )
     pull_parser.set_defaults(func=CmdDataPull)
 

dvc/commands/experiments/ls.py

@@ -16,13 +16,17 @@ class CmdExperimentsList(CmdBase):
             num=self.args.num,
             git_remote=self.args.git_remote,
         )
+        tags = self.repo.scm.describe(exps)
+        remained = {baseline for baseline, tag in tags.items() if tag is None}
+        base = "refs/heads/"
+        ref_heads = self.repo.scm.describe(remained, base=base)
+
         for baseline in exps:
-            tag = self.repo.scm.describe(baseline)
-            if not tag:
-                branch = self.repo.scm.describe(baseline, base="refs/heads")
-                if branch:
-                    tag = branch.split("/")[-1]
-            name = tag if tag else baseline[:7]
+            name = (
+                tags[baseline]
+                or ref_heads[baseline][len(base) :]
+                or baseline[:7]
+            )
             if not name_only:
                 print(f"{name}:")
             for exp_name in exps[baseline]:

dvc/commands/experiments/show.py

@@ -95,11 +95,8 @@ def _collect_rows(
 
         exp = results.get("data", {})
 
-        if exp.get("running"):
-            state = "Running"
-        elif exp.get("queued"):
-            state = "Queued"
-        else:
+        state = exp.get("status")
+        if state == "Success":
             state = fill_value
 
         is_baseline = rev == "baseline"
@@ -476,6 +473,8 @@ class CmdExperimentsShow(CmdBase):
                 all_branches=self.args.all_branches,
                 all_tags=self.args.all_tags,
                 all_commits=self.args.all_commits,
+                hide_queued=self.args.hide_queued,
+                hide_failed=self.args.hide_failed,
                 revs=self.args.rev,
                 num=self.args.num,
                 sha_only=self.args.sha,
@@ -594,6 +593,18 @@ def add_parser(experiments_subparsers, parent_parser):
         default=False,
         help="Always show git commit SHAs instead of branch/tag names.",
     )
+    experiments_show_parser.add_argument(
+        "--hide-failed",
+        action="store_true",
+        default=False,
+        help="Hide failed experiments in the table.",
+    )
+    experiments_show_parser.add_argument(
+        "--hide-queued",
+        action="store_true",
+        default=False,
+        help="Hide queued experiments in the table.",
+    )
     experiments_show_parser.add_argument(
         "--json",
         "--show-json",

dvc/commands/git_hook.py

@@ -74,7 +74,7 @@ class CmdMergeDriver(CmdHookBase):
             our = Dvcfile(dvc, self.args.our, verify=False)
             their = Dvcfile(dvc, self.args.their, verify=False)
 
-            our.merge(ancestor, their)
+            our.merge(ancestor, their, allowed=["add", "remove", "change"])
 
             return 0
         finally:

dvc/commands/imp_url.py

@@ -25,6 +25,7 @@ class CmdImportUrl(CmdBase):
                 labels=self.args.labels,
                 meta=self.args.meta,
                 jobs=self.args.jobs,
+                version_aware=self.args.version_aware,
             )
         except DvcException:
             logger.exception(
@@ -108,5 +109,14 @@ def add_parser(subparsers, parent_parser):
         ),
         metavar="<number>",
     )
+    import_parser.add_argument(
+        "--version-aware",
+        action="store_true",
+        default=False,
+        help=(
+            "Import using cloud versioning. Implied if the URL contains a "
+            "version ID."
+        ),
+    )
     _add_annotating_args(import_parser)
     import_parser.set_defaults(func=CmdImportUrl)

dvc/commands/ls_url.py

@@ -0,0 +1,43 @@
+import argparse
+import logging
+
+from dvc.cli.command import CmdBaseNoRepo
+from dvc.cli.utils import append_doc_link
+from dvc.ui import ui
+
+from .ls import _prettify
+
+logger = logging.getLogger(__name__)
+
+
+class CmdListUrl(CmdBaseNoRepo):
+    def run(self):
+        from dvc.repo import Repo
+
+        entries = Repo.ls_url(self.args.url, recursive=self.args.recursive)
+        if entries:
+            entries = _prettify(entries, with_color=True)
+            ui.write("\n".join(entries))
+        return 0
+
+
+def add_parser(subparsers, parent_parser):
+    LS_HELP = "List directory contents from URL."
+    lsurl_parser = subparsers.add_parser(
+        "list-url",
+        aliases=["ls-url"],
+        parents=[parent_parser],
+        description=append_doc_link(LS_HELP, "list-url"),
+        help=LS_HELP,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    lsurl_parser.add_argument(
+        "url", help="See `dvc import-url -h` for full list of supported URLs."
+    )
+    lsurl_parser.add_argument(
+        "-R",
+        "--recursive",
+        action="store_true",
+        help="Recursively list files.",
+    )
+    lsurl_parser.set_defaults(func=CmdListUrl)

dvc/commands/plots.py

@@ -25,7 +25,7 @@ def _show_json(renderers, split=False):
 
 
 def _adjust_vega_renderers(renderers):
-    from dvc.render import VERSION_FIELD
+    from dvc.render import REVISION_FIELD, VERSION_FIELD
     from dvc_render import VegaRenderer
 
     for r in renderers:
@@ -42,6 +42,7 @@ def _adjust_vega_renderers(renderers):
                         dp["rev"] = "::".join(vi.values())
             else:
                 for dp in r.datapoints:
+                    dp[REVISION_FIELD] = dp[VERSION_FIELD]["revision"]
                     dp.pop(VERSION_FIELD, {})
 
 
@@ -66,18 +67,6 @@ def _data_versions_count(renderer):
     return len(set(x))
 
 
-def _filter_unhandled_renderers(renderers):
-    # filtering out renderers currently unhandled by vscode extension
-    from dvc_render import VegaRenderer
-
-    def _is_json_viable(r):
-        return not (
-            isinstance(r, VegaRenderer) and _data_versions_count(r) > 1
-        )
-
-    return list(filter(_is_json_viable, renderers))
-
-
 class CmdPlots(CmdBase):
     def _func(self, *args, **kwargs):
         raise NotImplementedError
@@ -156,12 +145,10 @@ class CmdPlots(CmdBase):
                 templates_dir=self.repo.plots.templates_dir,
             )
             if self.args.json:
-                renderers = _filter_unhandled_renderers(renderers)
                 _show_json(renderers, self.args.split)
                 return 0
 
             _adjust_vega_renderers(renderers)
-
             if self.args.show_vega:
                 renderer = first(filter(lambda r: r.TYPE == "vega", renderers))
                 if renderer:

dvc/commands/repro.py

@@ -55,7 +55,16 @@ def add_arguments(repro_parser):
     repro_parser.add_argument(
         "targets",
         nargs="*",
-        help="Stages to reproduce. 'dvc.yaml' by default.",
+        help=(
+            """\
+Stages to reproduce. 'dvc.yaml' by default.
+The targets can be path to a dvc.yaml file or `.dvc` file,
+or a stage name from dvc.yaml file from
+current working directory. To run a stage from dvc.yaml
+from other directories, the target must be a path followed by colon `:`
+and then the stage name name.
+"""
+        ),
     ).complete = completion.DVCFILES_AND_STAGE
     repro_parser.add_argument(
         "-f",

dvc/commands/update.py

@@ -60,7 +60,7 @@ def add_parser(subparsers, parent_parser):
         "--no-download",
         action="store_true",
         default=False,
-        help="Update .dvc file hash value(s)"
+        help="Update .dvc file git revision/hash value(s)"
         " but do not download the file(s).",
     )
     update_parser.add_argument(

dvc/config.py

@@ -176,7 +176,10 @@ class Config(dict):
 
         if fs.exists(filename):
             with fs.open(filename) as fobj:
-                conf_obj = ConfigObj(fobj)
+                try:
+                    conf_obj = ConfigObj(fobj)
+                except UnicodeDecodeError as exc:
+                    raise ConfigError(str(exc)) from exc
         else:
             conf_obj = ConfigObj()
         return _parse_named(_lower_keys(conf_obj.dict()))

dvc/config_schema.py

@@ -180,6 +180,8 @@ SCHEMA = {
                     "user": str,
                     "password": str,
                     "ask_password": Bool,
+                    "passphrase": str,
+                    "ask_passphrase": Bool,
                     "keyfile": str,
                     "timeout": Coerce(int),
                     "gss_auth": Bool,

dvc/data_cloud.py

@@ -11,6 +11,7 @@ if TYPE_CHECKING:
     from dvc_data.hashfile.db import HashFileDB
     from dvc_data.hashfile.hash_info import HashInfo
     from dvc_data.hashfile.status import CompareStatusResult
+    from dvc_data.hashfile.transfer import TransferResult
 
 logger = logging.getLogger(__name__)
 
@@ -106,14 +107,10 @@ class DataCloud:
         dest_odb: "HashFileDB",
         objs: Iterable["HashInfo"],
         **kwargs,
-    ):
-        from dvc.exceptions import FileTransferError
-        from dvc_data.hashfile.transfer import TransferError, transfer
+    ) -> "TransferResult":
+        from dvc_data.hashfile.transfer import transfer
 
-        try:
-            return transfer(src_odb, dest_odb, objs, **kwargs)
-        except TransferError as exc:
-            raise FileTransferError(exc.fails) from exc
+        return transfer(src_odb, dest_odb, objs, **kwargs)
 
     def push(
         self,
@@ -121,7 +118,7 @@ class DataCloud:
         jobs: Optional[int] = None,
         remote: Optional[str] = None,
         odb: Optional["HashFileDB"] = None,
-    ):
+    ) -> "TransferResult":
         """Push data items in a cloud-agnostic way.
 
         Args:
@@ -148,7 +145,7 @@ class DataCloud:
         jobs: Optional[int] = None,
         remote: Optional[str] = None,
         odb: Optional["HashFileDB"] = None,
-    ):
+    ) -> "TransferResult":
         """Pull data items in a cloud-agnostic way.
 
         Args:

dvc/dependency/base.py

@@ -1,6 +1,7 @@
-from typing import Type
+from typing import Dict, Type
 
 from dvc.exceptions import DvcException
+from dvc.fs import download as fs_download
 from dvc.output import Output
 
 
@@ -30,5 +31,34 @@ class Dependency(Output):
     )  # type: Type[DvcException]
     IsStageFileError = DependencyIsStageFileError  # type: Type[DvcException]
 
+    def workspace_status(self) -> Dict[str, str]:
+        if self.fs.version_aware:
+            old_fs_path = self.fs_path
+            try:
+                self.fs_path = self.fs.path.version_path(self.fs_path, None)
+                if self.changed_meta():
+                    return {str(self): "update available"}
+            finally:
+                self.fs_path = old_fs_path
+        return super().workspace_status()
+
     def update(self, rev=None):
-        pass
+        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)
+
+    def save(self):
+        super().save()
+        if self.fs.version_aware:
+            self.fs_path = self.fs.path.version_path(
+                self.fs_path, self.meta.version_id
+            )

dvc/dependency/repo.py

@@ -6,6 +6,7 @@ from typing import TYPE_CHECKING, Dict, Optional, Set, Tuple
 from voluptuous import Required
 
 from dvc.prompt import confirm
+from dvc_data.hashfile.meta import Meta
 
 from .base import Dependency
 
@@ -32,6 +33,7 @@ class RepoDependency(Dependency):
     def __init__(self, def_repo, stage, *args, **kwargs):
         self.def_repo = def_repo
         self._objs: Dict[str, "HashFile"] = {}
+        self._meta: Dict[str, Meta] = {}
         super().__init__(stage, *args, **kwargs)
 
     def _parse_path(self, fs, fs_path):
@@ -94,12 +96,12 @@ class RepoDependency(Dependency):
     def get_used_objs(
         self, **kwargs
     ) -> Dict[Optional["ObjectDB"], Set["HashInfo"]]:
-        used, _ = self._get_used_and_obj(**kwargs)
+        used, _, _ = self._get_used_and_obj(**kwargs)
         return used
 
     def _get_used_and_obj(
         self, obj_only=False, **kwargs
-    ) -> Tuple[Dict[Optional["ObjectDB"], Set["HashInfo"]], "HashFile"]:
+    ) -> Tuple[Dict[Optional["ObjectDB"], Set["HashInfo"]], Meta, "HashFile"]:
         from dvc.config import NoRemoteError
         from dvc.exceptions import NoOutputOrStageError, PathMissingError
         from dvc.utils import as_posix
@@ -131,7 +133,7 @@ class RepoDependency(Dependency):
                     pass
 
             try:
-                object_store, _, obj = build(
+                object_store, meta, obj = build(
                     local_odb,
                     as_posix(self.def_path),
                     repo.dvcfs,
@@ -145,10 +147,12 @@ class RepoDependency(Dependency):
             object_store.read_only = True
 
             self._objs[rev] = obj
+            self._meta[rev] = meta
+
             used_obj_ids[object_store].add(obj.hash_info)
             if isinstance(obj, Tree):
                 used_obj_ids[object_store].update(oid for _, _, oid in obj)
-            return used_obj_ids, obj
+            return used_obj_ids, meta, obj
 
     def _check_circular_import(self, odb, obj_ids):
         from dvc.exceptions import CircularImportError
@@ -190,11 +194,21 @@ class RepoDependency(Dependency):
         rev = self._get_rev(locked=locked)
         if rev in self._objs:
             return self._objs[rev]
-        _, obj = self._get_used_and_obj(
+        _, _, obj = self._get_used_and_obj(
             obj_only=True, filter_info=filter_info, **kwargs
         )
         return obj
 
+    def get_meta(self, filter_info=None, **kwargs):
+        locked = kwargs.get("locked", True)
+        rev = self._get_rev(locked=locked)
+        if rev in self._meta:
+            return self._meta[rev]
+        _, meta, _ = self._get_used_and_obj(
+            obj_only=True, filter_info=filter_info, **kwargs
+        )
+        return meta
+
     def _make_repo(self, locked=True, **kwargs):
         from dvc.external_repo import external_repo
 

dvc/dvcfile.py

@@ -172,7 +172,7 @@ class FileMixin:
     def dump(self, stage, **kwargs):
         raise NotImplementedError
 
-    def merge(self, ancestor, other):
+    def merge(self, ancestor, other, allowed=None):
         raise NotImplementedError
 
 
@@ -204,12 +204,12 @@ class SingleStageFile(FileMixin):
     def remove_stage(self, stage):  # pylint: disable=unused-argument
         self.remove()
 
-    def merge(self, ancestor, other):
+    def merge(self, ancestor, other, allowed=None):
         assert isinstance(ancestor, SingleStageFile)
         assert isinstance(other, SingleStageFile)
 
         stage = self.stage
-        stage.merge(ancestor.stage, other.stage)
+        stage.merge(ancestor.stage, other.stage, allowed=allowed)
         self.dump(stage)
 
 
@@ -237,10 +237,10 @@ class PipelineFile(FileMixin):
             self._dump_pipeline_file(stage)
 
         if update_lock:
-            self._dump_lockfile(stage)
+            self._dump_lockfile(stage, **kwargs)
 
-    def _dump_lockfile(self, stage):
-        self._lockfile.dump(stage)
+    def _dump_lockfile(self, stage, **kwargs):
+        self._lockfile.dump(stage, **kwargs)
 
     @staticmethod
     def _check_if_parametrized(stage, action: str = "dump") -> None:
@@ -310,7 +310,7 @@ class PipelineFile(FileMixin):
         else:
             super().remove()
 
-    def merge(self, ancestor, other):
+    def merge(self, ancestor, other, allowed=None):
         raise NotImplementedError
 
 
@@ -366,7 +366,7 @@ class Lockfile(FileMixin):
         return {SCHEMA_KWD: version}
 
     def dump(self, stage, **kwargs):
-        stage_data = serialize.to_lockfile(stage)
+        stage_data = serialize.to_lockfile(stage, **kwargs)
 
         with modify_yaml(self.path, fs=self.repo.fs) as data:
             version = LOCKFILE_VERSION.from_dict(data)
@@ -411,7 +411,7 @@ class Lockfile(FileMixin):
         else:
             self.remove()
 
-    def merge(self, ancestor, other):
+    def merge(self, ancestor, other, allowed=None):
         raise NotImplementedError
 
 

dvc/exceptions.py

@@ -278,6 +278,11 @@ class PathMissingError(DvcException):
         self.dvc_only = dvc_only
 
 
+class URLMissingError(DvcException):
+    def __init__(self, url):
+        super().__init__(f"The path '{url}' does not exist")
+
+
 class RemoteCacheRequiredError(DvcException):
     def __init__(self, scheme, fs_path):
 

dvc/fs/__init__.py

@@ -2,6 +2,9 @@ from urllib.parse import urlparse
 
 from dvc_http import HTTPFileSystem, HTTPSFileSystem  # noqa: F401
 
+from dvc.config import ConfigError as RepoConfigError
+from dvc.config_schema import SCHEMA, Invalid
+
 # pylint: disable=unused-import
 from dvc_objects.fs import utils  # noqa: F401
 from dvc_objects.fs import (  # noqa: F401
@@ -23,6 +26,7 @@ from dvc_objects.fs.errors import (  # noqa: F401
 )
 from dvc_objects.fs.path import Path  # noqa: F401
 
+from .callbacks import Callback
 from .data import DataFileSystem  # noqa: F401
 from .dvc import DVCFileSystem  # noqa: F401
 from .git import GitFileSystem  # noqa: F401
@@ -44,7 +48,23 @@ known_implementations.update(
 # pylint: enable=unused-import
 
 
-def get_fs_config(repo, config, **kwargs):
+def download(fs, fs_path, to, jobs=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)
+
+
+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 = fs_cls(**fs_config)
+    return fs, fs_path
+
+
+def get_fs_config(config, **kwargs):
     name = kwargs.get("name")
     if name:
         try:
@@ -55,10 +75,10 @@ def get_fs_config(repo, config, **kwargs):
             raise RemoteNotFoundError(f"remote '{name}' doesn't exist")
     else:
         remote_conf = kwargs
-    return _resolve_remote_refs(repo, config, remote_conf)
+    return _resolve_remote_refs(config, remote_conf)
 
 
-def _resolve_remote_refs(repo, config, remote_conf):
+def _resolve_remote_refs(config, remote_conf):
     # Support for cross referenced remotes.
     # This will merge the settings, shadowing base ref with remote_conf.
     # For example, having:
@@ -84,21 +104,22 @@ def _resolve_remote_refs(repo, config, remote_conf):
     if parsed.scheme != "remote":
         return remote_conf
 
-    base = get_fs_config(repo, config, name=parsed.netloc)
-    cls, _, _ = get_cloud_fs(repo, **base)
+    base = get_fs_config(config, name=parsed.netloc)
+    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):
-    from dvc.config import ConfigError as RepoConfigError
-    from dvc.config_schema import SCHEMA, Invalid
-
     repo_config = repo.config if repo else {}
+    return _get_cloud_fs(repo_config, **kwargs)
+
+
+def _get_cloud_fs(repo_config, **kwargs):
     core_config = repo_config.get("core", {})
 
-    remote_conf = get_fs_config(repo, repo_config, **kwargs)
+    remote_conf = get_fs_config(repo_config, **kwargs)
     try:
         remote_conf = SCHEMA["remote"][str](remote_conf)
     except Invalid as exc:

dvc/fs/callbacks.py

@@ -1,6 +1,6 @@
 # pylint: disable=unused-import
 from contextlib import ExitStack
-from typing import TYPE_CHECKING, Optional
+from typing import TYPE_CHECKING, Any, BinaryIO, Dict, Optional, Union
 
 from funcy import cached_property
 
@@ -24,6 +24,7 @@ class RichCallback(Callback):
         bytes: bool = False,  # pylint: disable=redefined-builtin
         unit: str = None,
         disable: bool = False,
+        transient: bool = True,
     ) -> None:
         self._progress = progress
         self.disable = disable
@@ -35,6 +36,7 @@ class RichCallback(Callback):
             "visible": False,
             "progress_type": None if bytes else "summary",
         }
+        self._transient = transient
         self._stack = ExitStack()
         super().__init__(size=size, value=value)
 
@@ -47,7 +49,7 @@ class RichCallback(Callback):
             return self._progress
 
         progress = RichTransferProgress(
-            transient=True,
+            transient=self._transient,
             disable=self.disable,
             console=ui.error_console,
         )
@@ -61,7 +63,8 @@ class RichCallback(Callback):
         return self
 
     def close(self):
-        self.progress.clear_task(self.task)
+        if self._transient:
+            self.progress.clear_task(self.task)
         self._stack.close()
 
     def call(self, hook_name=None, **kwargs):
@@ -72,8 +75,17 @@ class RichCallback(Callback):
             visible=not self.disable,
         )
 
-    def branch(self, path_1, path_2, kwargs, child: Optional[Callback] = None):
+    def branch(
+        self,
+        path_1: Union[str, BinaryIO],
+        path_2: str,
+        kwargs: Dict[str, Any],
+        child: "Callback" = None,
+    ):
         child = child or RichCallback(
-            progress=self.progress, desc=path_1, bytes=True
+            progress=self.progress,
+            desc=path_1 if isinstance(path_1, str) else path_2,
+            bytes=True,
+            transient=self._transient,
         )
         return super().branch(path_1, path_2, kwargs, child=child)

dvc/fs/dvc.py

@@ -10,9 +10,7 @@ from typing import TYPE_CHECKING, Any, Callable, Optional, Tuple, Type, Union
 from fsspec.spec import AbstractFileSystem
 from funcy import cached_property, wrap_prop, wrap_with
 
-from dvc.utils.fs import makedirs
 from dvc_objects.fs.base import FileSystem
-from dvc_objects.fs.callbacks import DEFAULT_CALLBACK
 from dvc_objects.fs.path import Path
 
 from .data import DataFileSystem
@@ -66,15 +64,7 @@ def _get_dvc_path(dvc_fs, subkey):
 
 
 class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
-    """DVC + git-tracked files fs.
-
-    Args:
-        repo: DVC or git repo.
-        subrepos: traverse to subrepos (by default, it ignores subrepos)
-        repo_factory: A function to initialize subrepo with, default is Repo.
-        kwargs: Additional keyword arguments passed to the `DataFileSystem()`.
-    """
-
+    cachable = False
     root_marker = "/"
 
     def __init__(
@@ -86,6 +76,41 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         repo_factory: Optional[RepoFactory] = None,
         **repo_kwargs: Any,
     ) -> None:
+        """DVC + git-tracked files fs.
+
+        Args:
+            path (str, optional): URL or path to a DVC/Git repository.
+                Defaults to a DVC repository in the current working directory.
+                Both HTTP and SSH protocols are supported for remote Git repos
+                (e.g. [user@]server:project.git).
+            rev (str, optional): Any Git revision such as a branch or tag name,
+                a commit hash or a dvc experiment name.
+                Defaults to the default branch in case of remote repositories.
+                In case of a local repository, if rev is unspecified, it will
+                default to the working directory.
+                If the repo is not a Git repo, this option is ignored.
+            repo (:obj:`Repo`, optional): `Repo` instance.
+            subrepos (bool): traverse to subrepos.
+                By default, it ignores subrepos.
+            repo_factory (callable): A function to initialize subrepo with.
+                The default is `Repo`.
+
+        Examples:
+            - Opening a filesystem from repo in current working directory
+
+            >>> fs = DVCFileSystem()
+
+            - Opening a filesystem from local repository
+
+            >>> fs = DVCFileSystem("path/to/local/repository")
+
+            - Opening a remote repository
+
+            >>> fs = DVCFileSystem(
+            ...    "https://github.com/iterative/example-get-started",
+            ...    rev="main",
+            ... )
+        """
         from pygtrie import Trie
 
         super().__init__()
@@ -238,7 +263,8 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
         dvc_path = _get_dvc_path(dvc_fs, subkey)
         return dvc_fs.open(dvc_path, mode=mode)
 
-    def isdvc(self, path, **kwargs):
+    def isdvc(self, path, **kwargs) -> bool:
+        """Is this entry dvc-tracked?"""
         key = self._get_key_from_relative(path)
         _, dvc_fs, subkey = self._get_subrepo_info(key)
         dvc_path = _get_dvc_path(dvc_fs, subkey)
@@ -295,27 +321,6 @@ class _DVCFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
 
         return infos
 
-    def get_file(  # pylint: disable=arguments-differ
-        self, rpath, lpath, callback=DEFAULT_CALLBACK, **kwargs
-    ):
-        key = self._get_key_from_relative(rpath)
-        fs_path = self._from_key(key)
-        fs = self.repo.fs
-
-        if self.isdir(rpath):
-            makedirs(lpath, exist_ok=True)
-            return None
-
-        try:
-            fs.get_file(fs_path, lpath, callback=callback, **kwargs)
-            return
-        except FileNotFoundError:
-            _, dvc_fs, subkey = self._get_subrepo_info(key)
-            if not dvc_fs:
-                raise
-        dvc_path = _get_dvc_path(dvc_fs, subkey)
-        dvc_fs.get_file(dvc_path, lpath, callback=callback, **kwargs)
-
     def info(self, path, **kwargs):
         key = self._get_key_from_relative(path)
         ignore_subrepos = kwargs.get("ignore_subrepos", True)

dvc/ignore.py

@@ -300,12 +300,25 @@ class DvcIgnoreFilter:
         return [fs_dict[name] for name in chain(dirs, nondirs)]
 
     def walk(self, fs: FileSystem, path: AnyFSPath, **kwargs):
+        detail = kwargs.get("detail", False)
         ignore_subrepos = kwargs.pop("ignore_subrepos", True)
         if fs.protocol == Schemes.LOCAL:
             for root, dirs, files in fs.walk(path, **kwargs):
-                dirs[:], files[:] = self(
-                    root, dirs, files, ignore_subrepos=ignore_subrepos
-                )
+                if detail:
+                    all_dnames = set(dirs.keys())
+                    all_fnames = set(files.keys())
+                    dnames, fnames = self(
+                        root,
+                        all_dnames,
+                        all_fnames,
+                        ignore_subrepos=ignore_subrepos,
+                    )
+                    list(map(dirs.pop, all_dnames - set(dnames)))
+                    list(map(files.pop, all_fnames - set(fnames)))
+                else:
+                    dirs[:], files[:] = self(
+                        root, dirs, files, ignore_subrepos=ignore_subrepos
+                    )
                 yield root, dirs, files
         else:
             yield from fs.walk(path, **kwargs)

dvc/info.py

@@ -76,7 +76,7 @@ def _get_caches(cache):
 
 def _get_remotes(config):
     schemes = (
-        get_fs_cls(get_fs_config(None, config, name=remote)).protocol
+        get_fs_cls(get_fs_config(config, name=remote)).protocol
         for remote in config["remote"]
     )
 

dvc/logger.py

@@ -213,8 +213,6 @@ def set_loggers_level(level: int = logging.INFO) -> None:
 
 
 def setup(level: int = logging.INFO) -> None:
-    colorama.init()
-
     if level >= logging.DEBUG:
         # Unclosed session errors for asyncio/aiohttp are only available
         # on the tracing mode for extensive debug purposes. They are really
@@ -297,3 +295,8 @@ def setup(level: int = logging.INFO) -> None:
             "disable_existing_loggers": False,
         }
     )
+
+    # NOTE: colorama init needs to go after logging config to avoid potential
+    # conflicts with libs that modify logging (like tensorflow/abseil)
+    # https://github.com/iterative/dvc/issues/8387
+    colorama.init()

dvc/output.py

@@ -26,10 +26,12 @@ from dvc_data.hashfile.hash_info import HashInfo
 from dvc_data.hashfile.istextfile import istextfile
 from dvc_data.hashfile.meta import Meta
 from dvc_data.hashfile.transfer import transfer as otransfer
+from dvc_data.index import DataIndexEntry
 from dvc_objects.errors import ObjectFormatError
 
 from .annotations import ANNOTATION_FIELDS, ANNOTATION_SCHEMA, Annotation
 from .fs import LocalFileSystem, RemoteMissingDepsError, Schemes, get_cloud_fs
+from .fs.callbacks import DEFAULT_CALLBACK
 from .utils import relpath
 from .utils.fs import path_isin
 
@@ -38,6 +40,8 @@ if TYPE_CHECKING:
     from dvc_data.index import DataIndexKey
     from dvc_objects.db import ObjectDB
 
+    from .fs.callbacks import Callback
+
 logger = logging.getLogger(__name__)
 
 
@@ -511,6 +515,23 @@ class Output:
             key = self.fs.path.parts(no_drive)[1:]
         return workspace, key
 
+    def get_entry(self) -> "DataIndexEntry":
+        from dvc.config import NoRemoteError
+
+        try:
+            remote = self.repo.cloud.get_remote_odb(self.remote)
+        except NoRemoteError:
+            remote = None
+
+        return DataIndexEntry(
+            meta=self.meta,
+            obj=self.obj,
+            hash_info=self.hash_info,
+            odb=self.odb,
+            cache=self.odb,
+            remote=remote,
+        )
+
     def changed_checksum(self):
         return self.hash_info != self.get_hash()
 
@@ -530,7 +551,7 @@ class Output:
 
     def changed_meta(self) -> bool:
         if self.fs.version_aware and self.meta.version_id:
-            return self.meta.version_id == self.get_meta().version_id
+            return self.meta.version_id != self.get_meta().version_id
         return False
 
     def workspace_status(self):
@@ -765,11 +786,8 @@ class Output:
         if self.remote:
             ret[self.PARAM_REMOTE] = self.remote
 
-        if (
-            self.use_cache
-            and self.is_in_repo
-            and self.hash_info.isdir
-            and (kwargs.get("with_files") or self.files is not None)
+        if self.hash_info.isdir and (
+            kwargs.get("with_files") or self.files is not None
         ):
             if self.obj:
                 obj = self.obj
@@ -800,15 +818,6 @@ class Output:
                 f"binary file '{self.fs_path}' cannot be used as metrics."
             )
 
-    def download(self, to, jobs=None):
-        from dvc.fs.callbacks import Callback
-
-        with Callback.as_tqdm_callback(
-            desc=f"Downloading {self.fs.path.name(self.fs_path)}",
-            unit="files",
-        ) as cb:
-            self.fs.get(self.fs_path, to.fs_path, batch_size=jobs, callback=cb)
-
     def get_obj(
         self, filter_info: Optional[str] = None, **kwargs
     ) -> Optional["HashFile"]:
@@ -828,6 +837,7 @@ class Output:
         else:
             return None
 
+        assert obj
         fs_path = self.fs.path
         if filter_info and filter_info != self.fs_path:
             prefix = fs_path.relparts(filter_info, self.fs_path)
@@ -838,7 +848,7 @@ class Output:
     def checkout(
         self,
         force=False,
-        progress_callback=None,
+        progress_callback: "Callback" = DEFAULT_CALLBACK,
         relink=False,
         filter_info=None,
         allow_missing=False,
@@ -846,9 +856,9 @@ class Output:
         **kwargs,
     ):
         if not self.use_cache:
-            if progress_callback:
-                progress_callback(
-                    self.fs_path, self.get_files_number(filter_info)
+            if progress_callback != DEFAULT_CALLBACK:
+                progress_callback.relative_update(
+                    self.get_files_number(filter_info)
                 )
             return None
 
@@ -1127,7 +1137,7 @@ class Output:
                 "unable to auto-merge outputs that are not directories"
             )
 
-    def merge(self, ancestor, other):
+    def merge(self, ancestor, other, allowed=None):
         from dvc_data.hashfile.tree import MergeError as TreeMergeError
         from dvc_data.hashfile.tree import du, merge
 
@@ -1144,7 +1154,11 @@ class Output:
 
         try:
             merged = merge(
-                self.odb, ancestor_info, self.hash_info, other.hash_info
+                self.odb,
+                ancestor_info,
+                self.hash_info,
+                other.hash_info,
+                allowed=allowed,
             )
         except TreeMergeError as exc:
             raise MergeError(str(exc)) from exc

dvc/render/convert.py

@@ -2,7 +2,13 @@ import json
 from collections import defaultdict
 from typing import Dict, List, Union
 
-from dvc.render import REVISION_FIELD, REVISIONS_KEY, SRC_FIELD, TYPE_KEY
+from dvc.render import (
+    REVISION_FIELD,
+    REVISIONS_KEY,
+    SRC_FIELD,
+    TYPE_KEY,
+    VERSION_FIELD,
+)
 from dvc.render.converter.image import ImageConverter
 from dvc.render.converter.vega import VegaConverter
 
@@ -23,7 +29,7 @@ def _get_converter(
 def _group_by_rev(datapoints):
     grouped = defaultdict(list)
     for datapoint in datapoints:
-        rev = datapoint.pop(REVISION_FIELD)
+        rev = datapoint.get(VERSION_FIELD, {}).get("revision")
         grouped[rev].append(datapoint)
     return dict(grouped)
 

dvc/render/converter/vega.py

@@ -127,13 +127,20 @@ class VegaConverter(Converter):
                 ),
             )
         )
+        self.steps.append(
+            (
+                "infer_y",
+                partial(
+                    self._infer_y,
+                ),
+            )
+        )
 
         self.steps.append(
             (
                 "generate_y",
                 partial(
                     self._generate_y_values,
-                    y_values=self.plot_properties.get("y", None),
                 ),
             )
         )
@@ -156,7 +163,7 @@ class VegaConverter(Converter):
             } - {None}
             self.inferred_properties["fields"] = fields
 
-    def _infer_y(self, datapoints: List[Dict]):
+    def _infer_y(self, datapoints: List[Dict], **kwargs):
         if "y" not in self.plot_properties:
             data_fields = list(first(datapoints))
             skip = (
@@ -178,6 +185,7 @@ class VegaConverter(Converter):
                     )
             else:
                 self.inferred_properties["y"] = inferred_y
+        return datapoints
 
     def convert(
         self,
@@ -197,7 +205,6 @@ class VegaConverter(Converter):
         processed = deepcopy(data)
 
         for step_name, step in self.steps:
-
             if step_name not in skip:
                 processed = step(  # type: ignore
                     processed,
@@ -205,25 +212,28 @@ class VegaConverter(Converter):
                     filename=filename,
                 )
 
-        self._infer_y(processed)  # type: ignore
         return processed, {**self.plot_properties, **self.inferred_properties}
 
     def _generate_y_values(  # noqa: C901
         self,
         datapoints: List[Dict],
-        y_values: Optional[Union[str, List, Dict]],
         revision: str,
         filename: str,
         **kwargs,
     ) -> List[Dict]:
 
+        y_values = self.plot_properties.get(
+            "y", None
+        ) or self.inferred_properties.get("y", None)
+
+        assert y_values is not None
+
         result = []
         properties_update = {}
 
         def _add_version_info(datapoint, version_info):
             tmp = datapoint.copy()
             tmp[VERSION_FIELD] = version_info
-            tmp[REVISION_FIELD] = version_info["revision"]
             return tmp
 
         def _version_info(revision, filename=None, field=None):
@@ -250,12 +260,6 @@ class VegaConverter(Converter):
 
             return tmp
 
-        if not y_values:
-            for dp in datapoints:
-                result.append(
-                    _add_version_info(dp, _version_info(revision, filename))
-                )
-
         if isinstance(y_values, str):
             for datapoint in datapoints:
                 result.append(
@@ -290,13 +294,15 @@ class VegaConverter(Converter):
             ):
                 # if we use the same field from all files,
                 # we dont have to generate it
+                y_field = all_fields.pop()
                 for datapoint in datapoints:
                     result.append(
                         _add_version_info(
-                            datapoint, _version_info(revision, filename)
+                            datapoint,
+                            _version_info(revision, filename, y_field),
                         )
                     )
-                properties_update.update({"y": all_fields.pop()})
+                properties_update.update({"y": y_field})
             else:
                 for def_filename, val in y_values.items():
                     if isinstance(val, str):

dvc/repo/__init__.py

@@ -3,7 +3,7 @@ import os
 from collections import defaultdict
 from contextlib import contextmanager
 from functools import wraps
-from typing import TYPE_CHECKING, Callable, List, Optional
+from typing import TYPE_CHECKING, Callable, Optional
 
 from funcy import cached_property
 
@@ -18,7 +18,6 @@ if TYPE_CHECKING:
     from dvc.fs import FileSystem
     from dvc.repo.scm_context import SCMContext
     from dvc.scm import Base
-    from dvc.stage import Stage
 
 logger = logging.getLogger(__name__)
 
@@ -68,6 +67,7 @@ class Repo:
     from dvc.repo.imp_url import imp_url  # type: ignore[misc]
     from dvc.repo.install import install  # type: ignore[misc]
     from dvc.repo.ls import ls as _ls  # type: ignore[misc]
+    from dvc.repo.ls_url import ls_url as _ls_url  # type: ignore[misc]
     from dvc.repo.move import move  # type: ignore[misc]
     from dvc.repo.pull import pull  # type: ignore[misc]
     from dvc.repo.push import push  # type: ignore[misc]
@@ -80,6 +80,7 @@ class Repo:
     from .data import status as data_status  # type: ignore[misc]
 
     ls = staticmethod(_ls)
+    ls_url = staticmethod(_ls_url)
     get = staticmethod(_get)
     get_url = staticmethod(_get_url)
 
@@ -445,44 +446,6 @@ class Repo:
 
         return used
 
-    def partial_imports(
-        self,
-        targets=None,
-        all_branches=False,
-        all_tags=False,
-        all_commits=False,
-        all_experiments=False,
-        commit_date: Optional[str] = None,
-        recursive=False,
-        revs=None,
-        num=1,
-    ) -> List["Stage"]:
-        """Get the stages related to the given target and collect dependencies
-        which are missing outputs.
-
-        This is useful to retrieve files which have been imported to the repo
-        using --no-download.
-
-        Returns:
-            A list of partially imported stages
-        """
-        from itertools import chain
-
-        partial_imports = chain.from_iterable(
-            self.index.partial_imports(targets, recursive=recursive)
-            for _ in self.brancher(
-                revs=revs,
-                all_branches=all_branches,
-                all_tags=all_tags,
-                all_commits=all_commits,
-                all_experiments=all_experiments,
-                commit_date=commit_date,
-                num=num,
-            )
-        )
-
-        return list(partial_imports)
-
     @property
     def stages(self):  # obsolete, only for backward-compatibility
         return self.index.stages

dvc/repo/brancher.py

@@ -77,6 +77,7 @@ def brancher(  # noqa: E302
         from dvc.fs import GitFileSystem
 
         for sha, names in found_revs.items():
+            self.__dict__.pop("index", None)
             self.fs = GitFileSystem(scm=scm, rev=sha)
             self.root_dir = self.fs.path.join("/", *repo_root_parts)
 

dvc/repo/checkout.py

@@ -7,7 +7,6 @@ from dvc.exceptions import (
     CheckoutErrorSuggestGit,
     NoOutputOrStageError,
 )
-from dvc.progress import Tqdm
 from dvc.utils import relpath
 
 from . import locked
@@ -80,6 +79,7 @@ def checkout(
     allow_missing=False,
     **kwargs,
 ):
+    from dvc.fs.callbacks import Callback
 
     stats = {"added": [], "deleted": [], "modified": [], "failed": []}
     if not targets:
@@ -91,13 +91,16 @@ def checkout(
 
     pairs = _collect_pairs(self, targets, with_deps, recursive)
     total = get_all_files_numbers(pairs)
-    with Tqdm(
-        total=total, unit="file", desc="Checkout", disable=total == 0
-    ) as pbar:
+    with Callback.as_tqdm_callback(
+        unit="file",
+        desc="Checkout",
+        disable=total == 0,
+    ) as cb:
+        cb.set_size(total)
         for stage, filter_info in pairs:
             result = stage.checkout(
                 force=force,
-                progress_callback=pbar.update_msg,
+                progress_callback=cb,
                 relink=relink,
                 filter_info=filter_info,
                 allow_missing=allow_missing,

dvc/repo/experiments/__init__.py

@@ -2,7 +2,7 @@ import logging
 import os
 import re
 import time
-from typing import Any, Dict, Iterable, Optional
+from typing import Dict, Iterable, Optional
 
 from funcy import cached_property, first
 
@@ -16,7 +16,7 @@ from .exceptions import (
     InvalidExpRefError,
     MultipleBranchError,
 )
-from .executor.base import BaseExecutor, ExecutorInfo
+from .executor.base import BaseExecutor
 from .queue.base import BaseStashQueue, QueueEntry
 from .queue.celery import LocalCeleryQueue
 from .queue.tempdir import TempDirQueue
@@ -25,7 +25,6 @@ from .refs import (
     CELERY_FAILED_STASH,
     CELERY_STASH,
     EXEC_APPLY,
-    EXEC_BRANCH,
     EXEC_CHECKPOINT,
     EXEC_NAMESPACE,
     EXPS_NAMESPACE,
@@ -211,8 +210,9 @@ class Experiments:
 
     def _log_reproduced(self, revs: Iterable[str], tmp_dir: bool = False):
         names = []
+        rev_names = self.get_exact_name(revs)
         for rev in revs:
-            name = self.get_exact_name(rev)
+            name = rev_names[rev]
             names.append(name if name else rev[:7])
         ui.write("\nRan experiment(s): {}".format(", ".join(names)))
         if tmp_dir:
@@ -411,86 +411,41 @@ class Experiments:
             raise MultipleBranchError(rev, ref_infos)
         return str(ref_infos[0])
 
-    def get_exact_name(self, rev: str):
+    def get_exact_name(self, revs: Iterable[str]) -> Dict[str, Optional[str]]:
         """Returns preferred name for the specified revision.
 
         Prefers tags, branches (heads), experiments in that orer.
         """
+        result: Dict[str, Optional[str]] = {}
         exclude = f"{EXEC_NAMESPACE}/*"
-        ref = self.scm.describe(rev, base=EXPS_NAMESPACE, exclude=exclude)
-        if ref:
-            try:
-                name = ExpRefInfo.from_ref(ref).name
-                if name:
-                    return name
-            except InvalidExpRefError:
-                pass
-        if rev in self.stash_revs:
-            return self.stash_revs[rev].name
-        return None
-
-    def get_running_exps(self, fetch_refs: bool = True) -> Dict[str, Any]:
-        """Return info for running experiments."""
-        result = {}
-        infofile = self.workspace_queue.get_infofile_path("workspace")
-        result.update(
-            self._fetch_running_exp("workspace", infofile, fetch_refs)
+        ref_dict = self.scm.describe(
+            revs, base=EXPS_NAMESPACE, exclude=exclude
         )
-        for queue in (self.tempdir_queue, self.celery_queue):
-            for entry in queue.iter_active():
-                infofile = queue.get_infofile_path(entry.stash_rev)
-                result.update(
-                    self._fetch_running_exp(
-                        entry.stash_rev, infofile, fetch_refs
-                    )
-                )
+        for rev in revs:
+            name: Optional[str] = None
+            ref = ref_dict[rev]
+            if ref:
+                try:
+                    name = ExpRefInfo.from_ref(ref).name
+                except InvalidExpRefError:
+                    pass
+            if not name:
+                if rev in self.stash_revs:
+                    name = self.stash_revs[rev].name
+                elif rev in self.celery_queue.failed_stash.stash_revs:
+                    name = self.celery_queue.failed_stash.stash_revs[rev].name
+            result[rev] = name
         return result
 
-    def _fetch_running_exp(
-        self, rev: str, infofile: str, fetch_refs: bool
-    ) -> Dict[str, Any]:
-        from dvc.scm import InvalidRemoteSCMRepo
-        from dvc.utils.serialize import load_json
-
-        from .executor.local import TempDirExecutor
-
-        result: Dict[str, Any] = {}
-        try:
-            info = ExecutorInfo.from_dict(load_json(infofile))
-        except OSError:
-            return result
-        if info.result is None:
-            if rev == "workspace":
-                # If we are appending to a checkpoint branch in a workspace
-                # run, show the latest checkpoint as running.
-                last_rev = self.scm.get_ref(EXEC_BRANCH)
-                if last_rev:
-                    result[last_rev] = info.asdict()
-                else:
-                    result[rev] = info.asdict()
-            else:
-                result[rev] = info.asdict()
-                if info.git_url and fetch_refs:
-
-                    def on_diverged(_ref: str, _checkpoint: bool):
-                        return False
-
-                    executor = TempDirExecutor.from_info(info)
-                    try:
-                        for ref in executor.fetch_exps(
-                            self.scm,
-                            on_diverged=on_diverged,
-                        ):
-                            logger.debug(
-                                "Updated running experiment '%s'.", ref
-                            )
-                            last_rev = self.scm.get_ref(ref)
-                            result[rev]["last"] = last_rev
-                            if last_rev:
-                                result[last_rev] = info.asdict()
-                    except InvalidRemoteSCMRepo:
-                        # ignore stale info files
-                        del result[rev]
+    def get_running_exps(self, fetch_refs: bool = True) -> Dict[str, Dict]:
+        """Return info for running experiments."""
+        result = {}
+        for queue in (
+            self.workspace_queue,
+            self.tempdir_queue,
+            self.celery_queue,
+        ):
+            result.update(queue.get_running_exps(fetch_refs))
         return result
 
     def apply(self, *args, **kwargs):

dvc/repo/experiments/executor/base.py

@@ -4,6 +4,7 @@ import pickle
 from abc import ABC, abstractmethod
 from contextlib import contextmanager
 from dataclasses import asdict, dataclass
+from enum import IntEnum
 from functools import partial
 from typing import (
     TYPE_CHECKING,
@@ -19,6 +20,8 @@ from typing import (
     Union,
 )
 
+from scmrepo.exceptions import SCMError
+
 from dvc.env import DVC_EXP_AUTO_PUSH, DVC_EXP_GIT_REMOTE
 from dvc.exceptions import DvcException
 from dvc.stage.serialize import to_lockfile
@@ -35,6 +38,8 @@ from ..refs import (
     EXEC_BASELINE,
     EXEC_BRANCH,
     EXEC_CHECKPOINT,
+    EXEC_HEAD,
+    EXEC_MERGE,
     EXEC_NAMESPACE,
     EXPS_NAMESPACE,
     EXPS_STASH,
@@ -47,10 +52,9 @@ if TYPE_CHECKING:
     from scmrepo.git import Git
 
     from dvc.repo import Repo
+    from dvc.repo.experiments.stash import ExpStashEntry
     from dvc.stage import PipelineStage
 
-    from ..base import ExpStashEntry
-
 logger = logging.getLogger(__name__)
 
 
@@ -64,6 +68,16 @@ class ExecutorResult(NamedTuple):
     force: bool
 
 
+class TaskStatus(IntEnum):
+    PENDING = 0
+    PREPARING = 1
+    RUNNING = 2
+    SUCCESS = 3
+    FAILED = 4
+    CANCELED = 5
+    FINISHED = 6
+
+
 @dataclass
 class ExecutorInfo:
     git_url: str
@@ -73,13 +87,16 @@ class ExecutorInfo:
     dvc_dir: str
     name: Optional[str] = None
     wdir: Optional[str] = None
-    collected: bool = False
     result_hash: Optional[str] = None
     result_ref: Optional[str] = None
     result_force: bool = False
+    status: TaskStatus = TaskStatus.PENDING
 
     @classmethod
     def from_dict(cls, d):
+        if "collected" in d:
+            if d.pop("collected"):
+                d["status"] = TaskStatus.FINISHED
         return cls(**d)
 
     def asdict(self):
@@ -136,6 +153,7 @@ class BaseExecutor(ABC):
         root_dir: str,
         dvc_dir: str,
         baseline_rev: str,
+        status: TaskStatus,
         wdir: Optional[str] = None,
         name: Optional[str] = None,
         location: Optional[str] = None,
@@ -149,9 +167,17 @@ class BaseExecutor(ABC):
         self.baseline_rev = baseline_rev
         self.location: str = location or self.DEFAULT_LOCATION
         self.result = result
+        self.status = status
 
     @abstractmethod
-    def init_git(self, scm: "Git", branch: Optional[str] = None):
+    def init_git(
+        self,
+        scm: "Git",
+        stash_rev: str,
+        entry: "ExpStashEntry",
+        infofile: Optional[str],
+        branch: Optional[str] = None,
+    ):
         """Init git repo and populate it using exp refs from the specified
         SCM instance.
         """
@@ -191,6 +217,7 @@ class BaseExecutor(ABC):
             dvc_dir=self.dvc_dir,
             name=self.name,
             wdir=self.wdir,
+            status=self.status,
             **result_dict,
         )
 
@@ -212,6 +239,7 @@ class BaseExecutor(ABC):
             root_dir=info.root_dir,
             dvc_dir=info.dvc_dir,
             baseline_rev=info.baseline_rev,
+            status=info.status,
             name=info.name,
             wdir=info.wdir,
             result=result,
@@ -222,7 +250,6 @@ class BaseExecutor(ABC):
     def from_stash_entry(
         cls: Type[_T],
         repo: "Repo",
-        stash_rev: str,
         entry: "ExpStashEntry",
         **kwargs,
     ) -> _T:
@@ -232,7 +259,6 @@ class BaseExecutor(ABC):
     def _from_stash_entry(
         cls: Type[_T],
         repo: "Repo",
-        stash_rev: str,
         entry: "ExpStashEntry",
         root_dir: str,
         **kwargs,
@@ -241,12 +267,11 @@ class BaseExecutor(ABC):
             root_dir=root_dir,
             dvc_dir=relpath(repo.dvc_dir, repo.scm.root_dir),
             baseline_rev=entry.baseline_rev,
+            status=TaskStatus.PREPARING,
             name=entry.name,
             wdir=relpath(os.getcwd(), repo.scm.root_dir),
             **kwargs,
         )
-        executor.init_git(repo.scm, branch=entry.branch)
-        executor.init_cache(repo, stash_rev)
         return executor
 
     @staticmethod
@@ -259,8 +284,12 @@ class BaseExecutor(ABC):
                 exp_data.update(to_lockfile(stage))
         return dict_sha256(exp_data)
 
-    def cleanup(self):
-        pass
+    def cleanup(self, infofile: str):
+        if infofile is not None:
+            info = ExecutorInfo.load_json(infofile)
+            if info.status < TaskStatus.FAILED:
+                info.status = TaskStatus.FINISHED
+            info.dump_json(infofile)
 
     # TODO: come up with better way to stash repro arguments
     @staticmethod
@@ -324,30 +353,35 @@ class BaseExecutor(ABC):
                 logger.debug("Replacing existing experiment '%s'", orig_ref)
                 return True
 
+            if on_diverged:
+                return on_diverged(orig_ref, has_checkpoint)
+
             self._raise_ref_conflict(
                 dest_scm, orig_ref, new_rev, has_checkpoint
             )
-            if on_diverged:
-                on_diverged(orig_ref, has_checkpoint)
             logger.debug("Reproduced existing experiment '%s'", orig_ref)
             return False
 
         # fetch experiments
-        dest_scm.fetch_refspecs(
-            self.git_url,
-            [f"{ref}:{ref}" for ref in refs],
-            on_diverged=on_diverged_ref,
-            force=force,
-            **kwargs,
-        )
-        # update last run checkpoint (if it exists)
-        if has_checkpoint:
+        try:
             dest_scm.fetch_refspecs(
                 self.git_url,
-                [f"{EXEC_CHECKPOINT}:{EXEC_CHECKPOINT}"],
-                force=True,
+                [f"{ref}:{ref}" for ref in refs],
+                on_diverged=on_diverged_ref,
+                force=force,
                 **kwargs,
             )
+            # update last run checkpoint (if it exists)
+            if has_checkpoint:
+                dest_scm.fetch_refspecs(
+                    self.git_url,
+                    [f"{EXEC_CHECKPOINT}:{EXEC_CHECKPOINT}"],
+                    force=True,
+                    **kwargs,
+                )
+        except SCMError:
+            pass
+
         return refs
 
     @classmethod
@@ -411,11 +445,9 @@ class BaseExecutor(ABC):
         exp_ref: Optional["ExpRefInfo"] = None
         repro_force: bool = False
 
-        if infofile is not None:
-            info.dump_json(infofile)
-
         with cls._repro_dvc(
             info,
+            infofile,
             log_errors=log_errors,
             **kwargs,
         ) as dvc:
@@ -491,9 +523,6 @@ class BaseExecutor(ABC):
             info.result_ref = ref
             info.result_force = repro_force
 
-        if infofile is not None:
-            info.dump_json(infofile)
-
         # ideally we would return stages here like a normal repro() call, but
         # stages is not currently picklable and cannot be returned across
         # multiprocessing calls
@@ -550,6 +579,7 @@ class BaseExecutor(ABC):
     def _repro_dvc(
         cls,
         info: "ExecutorInfo",
+        infofile: str = None,
         log_errors: bool = True,
         **kwargs,
     ):
@@ -557,6 +587,9 @@ class BaseExecutor(ABC):
         from dvc.stage.monitor import CheckpointKilledError
 
         dvc = Repo(os.path.join(info.root_dir, info.dvc_dir))
+        info.status = TaskStatus.RUNNING
+        if infofile is not None:
+            info.dump_json(infofile)
         if cls.QUIET:
             dvc.scm_context.quiet = cls.QUIET
         old_cwd = os.getcwd()
@@ -568,15 +601,28 @@ class BaseExecutor(ABC):
         try:
             logger.debug("Running repro in '%s'", os.getcwd())
             yield dvc
+            info.status = TaskStatus.SUCCESS
+            if infofile is not None:
+                info.dump_json(infofile)
+
         except CheckpointKilledError:
+            info.status = TaskStatus.FAILED
+            if infofile is not None:
+                info.dump_json(infofile)
             raise
         except DvcException:
             if log_errors:
                 logger.exception("")
+            info.status = TaskStatus.FAILED
+            if infofile is not None:
+                info.dump_json(infofile)
             raise
         except Exception:
             if log_errors:
                 logger.exception("unexpected error")
+            info.status = TaskStatus.FAILED
+            if infofile is not None:
+                info.dump_json(infofile)
             raise
         finally:
             dvc.close()
@@ -714,3 +760,23 @@ class BaseExecutor(ABC):
         disable_other_loggers()
         if level is not None:
             dvc_logger.setLevel(level)
+
+    @contextmanager
+    def set_exec_refs(
+        self, scm: "Git", stash_rev: str, entry: "ExpStashEntry"
+    ):
+        try:
+            # Executor will be initialized with an empty git repo that
+            # we populate by pushing:
+            #   EXEC_HEAD - the base commit for this experiment
+            #   EXEC_MERGE - the unmerged changes (from our stash)
+            #       to be reproduced
+            #   EXEC_BASELINE - the baseline commit for this experiment
+            scm.set_ref(EXEC_HEAD, entry.head_rev)
+            scm.set_ref(EXEC_MERGE, stash_rev)
+            scm.set_ref(EXEC_BASELINE, entry.baseline_rev)
+            yield
+        finally:
+            for ref in (EXEC_HEAD, EXEC_MERGE, EXEC_BASELINE):
+                if scm.get_ref(ref):
+                    scm.remove_ref(ref)

dvc/repo/experiments/executor/local.py

@@ -19,7 +19,7 @@ from ..refs import (
     EXEC_MERGE,
     EXEC_NAMESPACE,
 )
-from .base import EXEC_TMP_DIR, BaseExecutor
+from .base import EXEC_TMP_DIR, BaseExecutor, TaskStatus
 
 if TYPE_CHECKING:
     from scmrepo.git import Git
@@ -46,8 +46,8 @@ class BaseLocalExecutor(BaseExecutor):
     def scm(self):
         return SCM(self.root_dir)
 
-    def cleanup(self):
-        super().cleanup()
+    def cleanup(self, infofile: str):
+        super().cleanup(infofile)
         self.scm.close()
         del self.scm
 
@@ -67,15 +67,28 @@ class TempDirExecutor(BaseLocalExecutor):
     QUIET = True
     DEFAULT_LOCATION = "tempdir"
 
-    def init_git(self, scm: "Git", branch: Optional[str] = None):
+    def init_git(
+        self,
+        scm: "Git",
+        stash_rev: str,
+        entry: "ExpStashEntry",
+        infofile: Optional[str],
+        branch: Optional[str] = None,
+    ):
         from dulwich.repo import Repo as DulwichRepo
 
         from ..utils import push_refspec
 
         DulwichRepo.init(os.fspath(self.root_dir))
 
-        refspec = f"{EXEC_NAMESPACE}/"
-        push_refspec(scm, self.git_url, refspec, refspec)
+        self.status = TaskStatus.PREPARING
+        if infofile:
+            self.info.dump_json(infofile)
+
+        with self.set_exec_refs(scm, stash_rev, entry):
+            refspec = f"{EXEC_NAMESPACE}/"
+            push_refspec(scm, self.git_url, refspec, refspec)
+
         if branch:
             push_refspec(scm, self.git_url, branch, branch)
             self.scm.set_ref(EXEC_BRANCH, branch, symbolic=True)
@@ -84,17 +97,12 @@ class TempDirExecutor(BaseLocalExecutor):
 
         if self.scm.get_ref(EXEC_CHECKPOINT):
             self.scm.remove_ref(EXEC_CHECKPOINT)
-
         # checkout EXEC_HEAD and apply EXEC_MERGE on top of it without
         # committing
         head = EXEC_BRANCH if branch else EXEC_HEAD
         self.scm.checkout(head, detach=True)
         merge_rev = self.scm.get_ref(EXEC_MERGE)
 
-        for ref in (EXEC_HEAD, EXEC_MERGE, EXEC_BASELINE):
-            if scm.get_ref(ref):
-                scm.remove_ref(ref)
-
         try:
             self.scm.merge(merge_rev, squash=True, commit=False)
         except _SCMError as exc:
@@ -114,8 +122,8 @@ class TempDirExecutor(BaseLocalExecutor):
         """Initialize DVC cache."""
         self._config(repo.odb.repo.path)
 
-    def cleanup(self):
-        super().cleanup()
+    def cleanup(self, infofile: str):
+        super().cleanup(infofile)
         logger.debug("Removing tmpdir '%s'", self.root_dir)
         remove(self.root_dir)
 
@@ -123,7 +131,6 @@ class TempDirExecutor(BaseLocalExecutor):
     def from_stash_entry(
         cls,
         repo: "Repo",
-        stash_rev: str,
         entry: "ExpStashEntry",
         wdir: Optional[str] = None,
         **kwargs,
@@ -132,9 +139,7 @@ class TempDirExecutor(BaseLocalExecutor):
         makedirs(parent_dir, exist_ok=True)
         tmp_dir = mkdtemp(dir=parent_dir)
         try:
-            executor = cls._from_stash_entry(
-                repo, stash_rev, entry, tmp_dir, **kwargs
-            )
+            executor = cls._from_stash_entry(repo, entry, tmp_dir, **kwargs)
             logger.debug("Init temp dir executor in '%s'", tmp_dir)
             return executor
         except Exception:
@@ -152,18 +157,31 @@ class WorkspaceExecutor(BaseLocalExecutor):
     def from_stash_entry(
         cls,
         repo: "Repo",
-        stash_rev: str,
         entry: "ExpStashEntry",
         **kwargs,
     ):
         root_dir = repo.scm.root_dir
-        executor = cls._from_stash_entry(
-            repo, stash_rev, entry, root_dir, **kwargs
+        executor: "WorkspaceExecutor" = cls._from_stash_entry(
+            repo, entry, root_dir, **kwargs
         )
         logger.debug("Init workspace executor in '%s'", root_dir)
         return executor
 
-    def init_git(self, scm: "Git", branch: Optional[str] = None):
+    def init_git(
+        self,
+        scm: "Git",
+        stash_rev: str,
+        entry: "ExpStashEntry",
+        infofile: Optional[str],
+        branch: Optional[str] = None,
+    ):
+        self.status = TaskStatus.PREPARING
+        if infofile:
+            self.info.dump_json(infofile)
+
+        scm.set_ref(EXEC_HEAD, entry.head_rev)
+        scm.set_ref(EXEC_MERGE, stash_rev)
+        scm.set_ref(EXEC_BASELINE, entry.baseline_rev)
         self._detach_stack.enter_context(
             self.scm.detach_head(
                 self.scm.get_ref(EXEC_HEAD),
@@ -184,7 +202,9 @@ class WorkspaceExecutor(BaseLocalExecutor):
     def init_cache(self, repo: "Repo", rev: str, run_cache: bool = True):
         pass
 
-    def cleanup(self):
+    def cleanup(self, infofile: str):
+        super().cleanup(infofile)
+        remove(os.path.dirname(infofile))
         with self._detach_stack:
             self.scm.remove_ref(EXEC_BASELINE)
             self.scm.remove_ref(EXEC_MERGE)
@@ -193,4 +213,3 @@ class WorkspaceExecutor(BaseLocalExecutor):
             checkpoint = self.scm.get_ref(EXEC_CHECKPOINT)
             if checkpoint and checkpoint != self._orig_checkpoint:
                 self.scm.set_ref(EXEC_APPLY, checkpoint)
-        super().cleanup()

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

@@ -1,269 +0,0 @@
-import logging
-import os
-from abc import ABC
-from collections import defaultdict, deque
-from collections.abc import Mapping
-from typing import TYPE_CHECKING, Deque, Dict, Generator, Optional, Tuple, Type
-
-from dvc.proc.manager import ProcessManager
-
-from ...exceptions import CheckpointExistsError, ExperimentExistsError
-from ...refs import EXEC_BASELINE, EXEC_HEAD, EXEC_MERGE, ExpRefInfo
-from ...stash import ExpStashEntry
-from ..base import EXEC_PID_DIR, BaseExecutor
-from ..local import TempDirExecutor, WorkspaceExecutor
-
-if TYPE_CHECKING:
-    from scmrepo.git import Git
-
-    from dvc.repo import Repo
-
-logger = logging.getLogger(__name__)
-
-
-class BaseExecutorManager(ABC, Mapping):  # noqa: B024
-    """Manages executors for a collection of experiments to be run."""
-
-    EXECUTOR_CLS: Type = BaseExecutor
-
-    def __init__(
-        self,
-        scm: "Git",
-        wdir: str,
-        **kwargs,
-    ):
-        from dvc.utils.fs import makedirs
-
-        self.scm = scm
-        makedirs(wdir, exist_ok=True)
-        self.wdir = wdir
-        self.proc = ProcessManager(self.pid_dir)
-        self._attached: Dict[str, "BaseExecutor"] = {}
-        self._detached: Dict[str, "BaseExecutor"] = dict(self._load_infos())
-        self._queue: Deque[Tuple[str, "BaseExecutor"]] = deque()
-
-    def __getitem__(self, key: str) -> "BaseExecutor":
-        try:
-            return self._attached[key]
-        except KeyError:
-            pass
-        return self._detached[key]
-
-    def __iter__(self):
-        yield from self._attached
-        yield from self._detached
-
-    def __len__(self):
-        return len(self._attached) + len(self._detached)
-
-    @property
-    def pid_dir(self) -> str:
-        return os.path.join(self.wdir, EXEC_PID_DIR)
-
-    def enqueue(self, rev: str, executor: "BaseExecutor"):
-        assert rev not in self
-        self._queue.append((rev, executor))
-
-    def _load_infos(self) -> Generator[Tuple[str, "BaseExecutor"], None, None]:
-        import json
-        from urllib.parse import urlparse
-
-        from ..base import ExecutorInfo
-        from ..ssh import SSHExecutor
-
-        def make_executor(info: "ExecutorInfo"):
-            if info.git_url:
-                scheme = urlparse(info.git_url).scheme
-                if scheme == "file":
-                    cls: Type = TempDirExecutor
-                elif scheme == "ssh":
-                    cls = SSHExecutor
-                else:
-                    raise NotImplementedError
-            else:
-                cls = WorkspaceExecutor
-            return cls.from_info(info)
-
-        for name in self.proc:
-            infofile = self.get_infofile_path(name)
-            try:
-                with open(infofile, encoding="utf-8") as fobj:
-                    info = ExecutorInfo.from_dict(json.load(fobj))
-                yield name, make_executor(info)
-            except OSError:
-                continue
-
-    def get_infofile_path(self, name: str) -> str:
-        return os.path.join(
-            self.pid_dir,
-            name,
-            f"{name}{BaseExecutor.INFOFILE_EXT}",
-        )
-
-    @classmethod
-    def from_stash_entries(
-        cls,
-        scm: "Git",
-        wdir: str,
-        repo: "Repo",
-        to_run: Dict[str, ExpStashEntry],
-        **kwargs,
-    ):
-        manager = cls(scm, wdir)
-        manager._enqueue_stash_entries(scm, repo, to_run, **kwargs)
-        return manager
-
-    def _enqueue_stash_entries(
-        self,
-        scm: "Git",
-        repo: "Repo",
-        to_run: Dict[str, ExpStashEntry],
-        **kwargs,
-    ):
-        try:
-            for stash_rev, entry in to_run.items():
-                scm.set_ref(EXEC_HEAD, entry.head_rev)
-                scm.set_ref(EXEC_MERGE, stash_rev)
-                scm.set_ref(EXEC_BASELINE, entry.baseline_rev)
-
-                # Executor will be initialized with an empty git repo that
-                # we populate by pushing:
-                #   EXEC_HEAD - the base commit for this experiment
-                #   EXEC_MERGE - the unmerged changes (from our stash)
-                #       to be reproduced
-                #   EXEC_BASELINE - the baseline commit for this experiment
-                executor = self.EXECUTOR_CLS.from_stash_entry(
-                    repo,
-                    stash_rev,
-                    entry,
-                    **kwargs,
-                )
-                self.enqueue(stash_rev, executor)
-        finally:
-            for ref in (EXEC_HEAD, EXEC_MERGE, EXEC_BASELINE):
-                scm.remove_ref(ref)
-
-    def exec_queue(
-        self, repo: "Repo", jobs: Optional[int] = 1, detach: bool = False
-    ):
-        """Run dvc repro for queued executors in parallel."""
-        if detach:
-            raise NotImplementedError
-            # TODO use ProcessManager.spawn() to support detached runs
-        return self._exec_attached(repo, jobs=jobs)
-
-    def _exec_attached(self, repo: "Repo", jobs: Optional[int] = 1):
-        import signal
-        from concurrent.futures import (
-            CancelledError,
-            ProcessPoolExecutor,
-            wait,
-        )
-        from multiprocessing import Manager
-
-        from dvc.stage.monitor import CheckpointKilledError
-
-        result: Dict[str, Dict[str, str]] = defaultdict(dict)
-
-        manager = Manager()
-        pid_q = manager.Queue()
-
-        with ProcessPoolExecutor(max_workers=jobs) as workers:
-            futures = {}
-            while self._queue:
-                rev, executor = self._queue.popleft()
-                infofile = self.get_infofile_path(rev)
-                future = workers.submit(
-                    executor.reproduce,
-                    info=executor.info,
-                    rev=rev,
-                    queue=pid_q,
-                    infofile=infofile,
-                    log_level=logger.getEffectiveLevel(),
-                )
-                futures[future] = (rev, executor)
-                self._attached[rev] = executor
-
-            try:
-                wait(futures)
-            except KeyboardInterrupt:
-                # forward SIGINT to any running executor processes and
-                # cancel any remaining futures
-                workers.shutdown(wait=False)
-                pids = {}
-                for future, (rev, _) in futures.items():
-                    if future.running():
-                        # if future has already been started by the scheduler
-                        # we still have to wait until it tells us its PID
-                        while rev not in pids:
-                            rev, pid = pid_q.get()
-                            pids[rev] = pid
-                        os.kill(pids[rev], signal.SIGINT)
-                    elif not future.done():
-                        future.cancel()
-
-            for future, (rev, executor) in futures.items():
-                rev, executor = futures[future]
-
-                try:
-                    exc = future.exception()
-                    if exc is None:
-                        exec_result = future.result()
-                        result[rev].update(
-                            self._collect_executor(repo, executor, exec_result)
-                        )
-                    elif not isinstance(exc, CheckpointKilledError):
-                        logger.error(
-                            "Failed to reproduce experiment '%s'", rev[:7]
-                        )
-                except CancelledError:
-                    logger.error(
-                        "Cancelled before attempting to reproduce experiment "
-                        "'%s'",
-                        rev[:7],
-                    )
-                finally:
-                    self.cleanup_executor(rev, executor)
-
-        return result
-
-    def _collect_executor(self, repo, executor, exec_result) -> Dict[str, str]:
-        # NOTE: GitPython Repo instances cannot be re-used
-        # after process has received SIGINT or SIGTERM, so we
-        # need this hack to re-instantiate git instances after
-        # checkpoint runs. See:
-        # https://github.com/gitpython-developers/GitPython/issues/427
-        # del self.repo.scm
-
-        results = {}
-
-        def on_diverged(ref: str, checkpoint: bool):
-            ref_info = ExpRefInfo.from_ref(ref)
-            if checkpoint:
-                raise CheckpointExistsError(ref_info.name)
-            raise ExperimentExistsError(ref_info.name)
-
-        for ref in executor.fetch_exps(
-            self.scm,
-            force=exec_result.force,
-            on_diverged=on_diverged,
-        ):
-            exp_rev = self.scm.get_ref(ref)
-            if exp_rev:
-                logger.debug("Collected experiment '%s'.", exp_rev[:7])
-                results[exp_rev] = exec_result.exp_hash
-
-        if exec_result.ref_info is not None:
-            executor.collect_cache(repo, exec_result.ref_info)
-
-        return results
-
-    def cleanup_executor(self, rev: str, executor: "BaseExecutor"):
-        from dvc.utils.fs import remove
-
-        executor.cleanup()
-        try:
-            self.proc.remove(rev)
-        except KeyError:
-            pass
-        remove(os.path.join(self.pid_dir, rev))

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

@@ -1,107 +0,0 @@
-import logging
-from collections import defaultdict
-from typing import TYPE_CHECKING, Dict, Optional
-
-from ...refs import EXEC_BASELINE, EXEC_BRANCH, EXEC_HEAD, EXEC_MERGE
-from ...stash import ExpStashEntry
-from ..local import TempDirExecutor, WorkspaceExecutor
-from .base import BaseExecutorManager
-
-if TYPE_CHECKING:
-    from scmrepo.git import Git
-
-    from dvc.repo import Repo
-
-logger = logging.getLogger(__name__)
-
-
-class TempDirExecutorManager(BaseExecutorManager):
-    EXECUTOR_CLS = TempDirExecutor
-
-
-class WorkspaceExecutorManager(BaseExecutorManager):
-    EXECUTOR_CLS = WorkspaceExecutor
-
-    @classmethod
-    def from_stash_entries(
-        cls,
-        scm: "Git",
-        wdir: str,
-        repo: "Repo",
-        to_run: Dict[str, ExpStashEntry],
-        **kwargs,
-    ):
-        manager = cls(scm, wdir)
-        try:
-            assert len(to_run) == 1
-            for stash_rev, entry in to_run.items():
-                scm.set_ref(EXEC_HEAD, entry.head_rev)
-                scm.set_ref(EXEC_MERGE, stash_rev)
-                scm.set_ref(EXEC_BASELINE, entry.baseline_rev)
-
-                executor = cls.EXECUTOR_CLS.from_stash_entry(
-                    repo,
-                    stash_rev,
-                    entry,
-                    **kwargs,
-                )
-                manager.enqueue(stash_rev, executor)
-        finally:
-            for ref in (EXEC_MERGE,):
-                scm.remove_ref(ref)
-        return manager
-
-    def _collect_executor(self, repo, executor, exec_result) -> Dict[str, str]:
-        results = {}
-        exp_rev = self.scm.get_ref(EXEC_BRANCH)
-        if exp_rev:
-            logger.debug("Collected experiment '%s'.", exp_rev[:7])
-            results[exp_rev] = exec_result.exp_hash
-        return results
-
-    def exec_queue(
-        self, repo: "Repo", jobs: Optional[int] = 1, detach: bool = False
-    ):
-        """Run a single WorkspaceExecutor.
-
-        Workspace execution is done within the main DVC process
-        (rather than in multiprocessing context)
-        """
-        from dvc.exceptions import DvcException
-        from dvc.stage.monitor import CheckpointKilledError
-
-        assert len(self._queue) == 1
-        assert not detach
-        result: Dict[str, Dict[str, str]] = defaultdict(dict)
-        rev, executor = self._queue.popleft()
-
-        exec_name = "workspace"
-        infofile = self.get_infofile_path(exec_name)
-        try:
-            exec_result = executor.reproduce(
-                info=executor.info,
-                rev=rev,
-                infofile=infofile,
-                log_level=logger.getEffectiveLevel(),
-                log_errors=not isinstance(executor, WorkspaceExecutor),
-            )
-            if not exec_result.exp_hash:
-                raise DvcException(
-                    f"Failed to reproduce experiment '{rev[:7]}'"
-                )
-            if exec_result.ref_info:
-                result[rev].update(
-                    self._collect_executor(repo, executor, exec_result)
-                )
-        except CheckpointKilledError:
-            # Checkpoint errors have already been logged
-            return {}
-        except DvcException:
-            raise
-        except Exception as exc:
-            raise DvcException(
-                f"Failed to reproduce experiment '{rev[:7]}'"
-            ) from exc
-        finally:
-            self.cleanup_executor(exec_name, executor)
-        return result

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

@@ -1,109 +0,0 @@
-import logging
-import posixpath
-from collections import defaultdict
-from typing import TYPE_CHECKING, Callable, Dict, Generator, Optional, Tuple
-
-from ...stash import ExpStashEntry
-from ..base import BaseExecutor
-from ..ssh import SSHExecutor, _sshfs
-from .base import BaseExecutorManager
-
-if TYPE_CHECKING:
-    from scmrepo.git import Git
-
-    from dvc.repo import Repo
-
-logger = logging.getLogger(__name__)
-
-
-class SSHExecutorManager(BaseExecutorManager):
-    EXECUTOR_CLS = SSHExecutor
-
-    def __init__(
-        self,
-        scm: "Git",
-        wdir: str,
-        host: Optional[str] = None,
-        port: Optional[int] = None,
-        username: Optional[str] = None,
-        fs_factory: Optional[Callable] = None,
-        **kwargs,
-    ):
-        assert host
-        super().__init__(scm, wdir, **kwargs)
-        self.host = host
-        self.port = port
-        self.username = username
-        self._fs_factory = fs_factory
-
-    def _load_infos(self) -> Generator[Tuple[str, "BaseExecutor"], None, None]:
-        # TODO: load existing infos using sshfs
-        yield from []
-
-    @classmethod
-    def from_stash_entries(
-        cls,
-        scm: "Git",
-        wdir: str,
-        repo: "Repo",
-        to_run: Dict[str, ExpStashEntry],
-        **kwargs,
-    ):
-        machine_name: Optional[str] = kwargs.get("machine_name", None)
-        manager = cls(
-            scm, wdir, **repo.machine.get_executor_kwargs(machine_name)
-        )
-        manager._enqueue_stash_entries(scm, repo, to_run, **kwargs)
-        return manager
-
-    def sshfs(self):
-        return _sshfs(self._fs_factory, host=self.host, port=self.port)
-
-    def get_infofile_path(self, name: str) -> str:
-        return f"{name}{BaseExecutor.INFOFILE_EXT}"
-
-    def _exec_attached(self, repo: "Repo", jobs: Optional[int] = 1):
-        from dvc.exceptions import DvcException
-        from dvc.stage.monitor import CheckpointKilledError
-
-        assert len(self._queue) == 1
-        result: Dict[str, Dict[str, str]] = defaultdict(dict)
-        rev, executor = self._queue.popleft()
-        info = executor.info
-        infofile = posixpath.join(
-            info.root_dir,
-            info.dvc_dir,
-            "tmp",
-            self.get_infofile_path(rev),
-        )
-        try:
-            exec_result = executor.reproduce(
-                info=executor.info,
-                rev=rev,
-                infofile=infofile,
-                log_level=logger.getEffectiveLevel(),
-                fs_factory=self._fs_factory,
-            )
-            if not exec_result.exp_hash:
-                raise DvcException(
-                    f"Failed to reproduce experiment '{rev[:7]}'"
-                )
-            if exec_result.ref_info:
-                result[rev].update(
-                    self._collect_executor(repo, executor, exec_result)
-                )
-        except CheckpointKilledError:
-            # Checkpoint errors have already been logged
-            return {}
-        except DvcException:
-            raise
-        except Exception as exc:
-            raise DvcException(
-                f"Failed to reproduce experiment '{rev[:7]}'"
-            ) from exc
-        finally:
-            self.cleanup_executor(rev, executor)
-        return result
-
-    def cleanup_executor(self, rev: str, executor: "BaseExecutor"):
-        executor.cleanup()

dvc/repo/experiments/executor/ssh.py

@@ -15,7 +15,7 @@ from ..refs import (
     EXEC_MERGE,
     EXEC_NAMESPACE,
 )
-from .base import BaseExecutor, ExecutorInfo, ExecutorResult
+from .base import BaseExecutor, ExecutorInfo, ExecutorResult, TaskStatus
 
 if TYPE_CHECKING:
     from queue import Queue
@@ -23,8 +23,8 @@ if TYPE_CHECKING:
     from scmrepo.git import Git
 
     from dvc.repo import Repo
-
-    from ..base import ExpRefInfo, ExpStashEntry
+    from dvc.repo.experiments.refs import ExpRefInfo
+    from dvc.repo.experiments.stash import ExpStashEntry
 
 logger = logging.getLogger(__name__)
 
@@ -75,14 +75,12 @@ class SSHExecutor(BaseExecutor):
     def from_stash_entry(
         cls,
         repo: "Repo",
-        stash_rev: str,
         entry: "ExpStashEntry",
         **kwargs,
     ):
         machine_name: Optional[str] = kwargs.pop("machine_name", None)
         executor = cls._from_stash_entry(
             repo,
-            stash_rev,
             entry,
             cls.gen_dirname(entry.name),
             location=machine_name,
@@ -119,9 +117,20 @@ class SSHExecutor(BaseExecutor):
         }
         return kwargs
 
-    def init_git(self, scm: "Git", branch: Optional[str] = None):
+    def init_git(
+        self,
+        scm: "Git",
+        stash_rev: str,
+        entry: "ExpStashEntry",
+        infofile: Optional[str],
+        branch: Optional[str] = None,
+    ):
         from ..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 .")
@@ -137,8 +146,11 @@ class SSHExecutor(BaseExecutor):
             # TODO: support multiple client key retries in git backends
             # (see https://github.com/iterative/dvc/issues/6508)
             kwargs = self._git_client_args(fs)
-            refspec = f"{EXEC_NAMESPACE}/"
-            push_refspec(scm, self.git_url, refspec, refspec, **kwargs)
+
+            with self.set_exec_refs(scm, stash_rev, entry):
+                refspec = f"{EXEC_NAMESPACE}/"
+                push_refspec(scm, self.git_url, refspec, refspec, **kwargs)
+
             if branch:
                 push_refspec(scm, self.git_url, branch, branch, **kwargs)
                 self._ssh_cmd(fs, f"git symbolic-ref {EXEC_BRANCH} {branch}")

dvc/repo/experiments/queue/base.py

@@ -18,11 +18,12 @@ from typing import (
     Union,
 )
 
-from funcy import cached_property
+from funcy import cached_property, retry
 
 from dvc.dependency import ParamsDependency
 from dvc.env import DVCLIVE_RESUME
 from dvc.exceptions import DvcException
+from dvc.lock import LockError
 from dvc.ui import ui
 
 from ..exceptions import CheckpointExistsError, ExperimentExistsError
@@ -33,7 +34,7 @@ from ..executor.base import (
     ExecutorResult,
 )
 from ..executor.local import WorkspaceExecutor
-from ..refs import EXEC_BASELINE, EXEC_HEAD, EXEC_MERGE, ExpRefInfo
+from ..refs import ExpRefInfo
 from ..stash import ExpStash, ExpStashEntry
 from ..utils import exp_refs_by_rev, scm_locked
 
@@ -560,8 +561,9 @@ class BaseStashQueue(ABC):
         self.scm.add(list(params.keys()))
 
     @staticmethod
+    @retry(180, errors=LockError, timeout=1)
     @scm_locked
-    def setup_executor(
+    def init_executor(
         exp: "Experiments",
         queue_entry: QueueEntry,
         executor_cls: Type[BaseExecutor] = WorkspaceExecutor,
@@ -576,21 +578,23 @@ class BaseStashQueue(ABC):
         )
         if stash_entry.stash_index is not None:
             stash.drop(stash_entry.stash_index)
-
-        scm.set_ref(EXEC_HEAD, stash_entry.head_rev)
-        scm.set_ref(EXEC_MERGE, stash_rev)
-        scm.set_ref(EXEC_BASELINE, stash_entry.baseline_rev)
-
-        # Executor will be initialized with an empty git repo that
-        # we populate by pushing:
-        #   EXEC_HEAD - the base commit for this experiment
-        #   EXEC_MERGE - the unmerged changes (from our stash)
-        #       to be reproduced
-        #   EXEC_BASELINE - the baseline commit for this experiment
-        return executor_cls.from_stash_entry(
-            exp.repo, stash_rev, stash_entry, **kwargs
+        executor = executor_cls.from_stash_entry(
+            exp.repo, stash_entry, **kwargs
         )
 
+        infofile = exp.celery_queue.get_infofile_path(stash_rev)
+        executor.init_git(
+            exp.repo.scm,
+            stash_rev,
+            stash_entry,
+            infofile,
+            branch=stash_entry.branch,
+        )
+
+        executor.init_cache(exp.repo, stash_rev)
+
+        return executor
+
     def get_infofile_path(self, name: str) -> str:
         return os.path.join(
             self.pid_dir,
@@ -599,8 +603,9 @@ class BaseStashQueue(ABC):
         )
 
     @staticmethod
+    @retry(180, errors=LockError, timeout=1)
     @scm_locked
-    def collect_executor(
+    def collect_git(
         exp: "Experiments",
         executor: BaseExecutor,
         exec_result: ExecutorResult,
@@ -624,6 +629,17 @@ class BaseStashQueue(ABC):
                 logger.debug("Collected experiment '%s'.", exp_rev[:7])
                 results[exp_rev] = exec_result.exp_hash
 
+        return results
+
+    @classmethod
+    def collect_executor(
+        cls,
+        exp: "Experiments",
+        executor: BaseExecutor,
+        exec_result: ExecutorResult,
+    ) -> Dict[str, str]:
+        results = cls.collect_git(exp, executor, exec_result)
+
         if exec_result.ref_info is not None:
             executor.collect_cache(exp.repo, exec_result.ref_info)
 
@@ -690,3 +706,11 @@ class BaseStashQueue(ABC):
                 entry.stash_rev,
                 message=f"commit: {msg}",
             )
+
+    @abstractmethod
+    def get_running_exps(self, fetch_refs: bool = True) -> Dict[str, Dict]:
+        """Get the execution info of the currently running experiments
+
+        Args:
+            fetch_ref (bool): fetch completed checkpoints or not.
+        """

dvc/repo/experiments/queue/celery.py

@@ -23,9 +23,15 @@ from dvc.exceptions import DvcException
 from dvc.ui import ui
 
 from ..exceptions import UnresolvedQueueExpNamesError
-from ..executor.base import EXEC_TMP_DIR, ExecutorInfo, ExecutorResult
+from ..executor.base import (
+    EXEC_TMP_DIR,
+    ExecutorInfo,
+    ExecutorResult,
+    TaskStatus,
+)
 from .base import BaseStashQueue, QueueDoneResult, QueueEntry, QueueGetResult
 from .tasks import run_exp
+from .utils import fetch_running_exp_from_temp_dir
 
 if TYPE_CHECKING:
     from dvc_task.app import FSApp
@@ -248,7 +254,7 @@ class LocalCeleryQueue(BaseStashQueue):
 
         def _load_collected(rev: str) -> Optional[ExecutorResult]:
             executor_info = _load_info(rev)
-            if executor_info.collected:
+            if executor_info.status > TaskStatus.SUCCESS:
                 return executor_info.result
             raise FileNotFoundError
 
@@ -368,3 +374,18 @@ class LocalCeleryQueue(BaseStashQueue):
         from .remove import celery_remove
 
         return celery_remove(self, *args, **kwargs)
+
+    def get_running_exps(self, fetch_refs: bool = True) -> Dict[str, Dict]:
+        """Get the execution info of the currently running experiments
+
+        Args:
+            fetch_ref (bool): fetch completed checkpoints or not.
+        """
+        result: Dict[str, Dict] = {}
+        for entry in self.iter_active():
+            result.update(
+                fetch_running_exp_from_temp_dir(
+                    self, entry.stash_rev, fetch_refs
+                )
+            )
+        return result

dvc/repo/experiments/queue/tasks.py

@@ -3,8 +3,6 @@ from typing import Any, Dict
 from celery import shared_task
 from celery.utils.log import get_task_logger
 
-from dvc.utils.fs import remove
-
 from ..executor.base import ExecutorInfo
 from ..executor.local import TempDirExecutor
 from .base import BaseStashQueue, QueueEntry
@@ -13,7 +11,7 @@ logger = get_task_logger(__name__)
 
 
 @shared_task
-def setup_exp(entry_dict: Dict[str, Any]) -> str:
+def setup_exp(entry_dict: Dict[str, Any]) -> TempDirExecutor:
     """Setup an experiment.
 
     Arguments:
@@ -28,7 +26,7 @@ def setup_exp(entry_dict: Dict[str, Any]) -> str:
     repo = Repo(entry.dvc_root)
     # TODO: split executor.init_cache into separate subtask - we can release
     # exp.scm_lock before DVC push
-    executor = BaseStashQueue.setup_executor(
+    executor = BaseStashQueue.init_executor(
         repo.experiments,
         entry,
         TempDirExecutor,
@@ -36,7 +34,7 @@ def setup_exp(entry_dict: Dict[str, Any]) -> str:
     )
     infofile = repo.experiments.celery_queue.get_infofile_path(entry.stash_rev)
     executor.info.dump_json(infofile)
-    return executor.root_dir
+    return executor
 
 
 @shared_task
@@ -79,25 +77,14 @@ def collect_exp(
 
 
 @shared_task
-def cleanup_exp(  # pylint: disable=unused-argument
-    tmp_dir: str, entry_dict: Dict[str, Any]
-) -> None:
+def cleanup_exp(executor: TempDirExecutor, infofile: str) -> None:
     """Cleanup after an experiment.
 
     Arguments:
         tmp_dir: Temp directory to be removed.
         entry_dict: Serialized QueueEntry for this experiment.
     """
-    remove(tmp_dir)
-
-
-def _set_collected(infofile: str):
-    try:
-        executor_info = ExecutorInfo.load_json(infofile)
-        executor_info.collected = True
-        executor_info.dump_json(infofile)
-    except FileNotFoundError:
-        pass
+    executor.cleanup(infofile)
 
 
 @shared_task
@@ -115,11 +102,10 @@ def run_exp(entry_dict: Dict[str, Any]) -> None:
     repo = Repo(entry.dvc_root)
     queue = repo.experiments.celery_queue
     infofile = queue.get_infofile_path(entry.stash_rev)
-    root_dir = setup_exp.s(entry_dict)()
+    executor = setup_exp.s(entry_dict)()
     try:
         cmd = ["dvc", "exp", "exec-run", "--infofile", infofile]
         proc_dict = queue.proc.run_signature(cmd, name=entry.stash_rev)()
         collect_exp.s(proc_dict, entry_dict)()
     finally:
-        cleanup_exp.s(root_dir, entry_dict)()
-        _set_collected(infofile)
+        cleanup_exp.s(executor, infofile)()

dvc/repo/experiments/queue/tempdir.py

@@ -11,9 +11,11 @@ from ..executor.base import (
     BaseExecutor,
     ExecutorInfo,
     ExecutorResult,
+    TaskStatus,
 )
 from ..executor.local import TempDirExecutor
 from .base import BaseStashQueue, QueueEntry, QueueGetResult
+from .utils import fetch_running_exp_from_temp_dir
 from .workspace import WorkspaceQueue
 
 if TYPE_CHECKING:
@@ -60,7 +62,7 @@ class TempDirQueue(WorkspaceQueue):
             stash_entry.name,
             stash_entry.head_rev,
         )
-        executor = self.setup_executor(
+        executor = self.init_executor(
             self.repo.experiments,
             entry,
             TempDirExecutor,
@@ -75,8 +77,9 @@ class TempDirQueue(WorkspaceQueue):
         for stash_rev in self.proc:
             infofile = self.get_infofile_path(stash_rev)
             executor_info = ExecutorInfo.load_json(infofile)
-            if not executor_info.collected and os.path.exists(
-                executor_info.root_dir
+            if (
+                not executor_info.status <= TaskStatus.SUCCESS
+                and os.path.exists(executor_info.root_dir)
             ):
                 yield QueueEntry(
                     self.repo.root_dir,
@@ -96,3 +99,13 @@ class TempDirQueue(WorkspaceQueue):
         exec_result: ExecutorResult,
     ) -> Dict[str, str]:
         return BaseStashQueue.collect_executor(exp, executor, exec_result)
+
+    def get_running_exps(self, fetch_refs: bool = True) -> Dict[str, Dict]:
+        result: Dict[str, Dict] = {}
+        for entry in self.iter_active():
+            result.update(
+                fetch_running_exp_from_temp_dir(
+                    self, entry.stash_rev, fetch_refs
+                )
+            )
+        return result

dvc/repo/experiments/queue/utils.py

@@ -0,0 +1,58 @@
+import logging
+from typing import TYPE_CHECKING, Dict
+
+from ..executor.base import ExecutorInfo, TaskStatus
+
+logger = logging.getLogger(__name__)
+
+
+if TYPE_CHECKING:
+    from .base import BaseStashQueue
+
+
+def fetch_running_exp_from_temp_dir(
+    queue: "BaseStashQueue", rev: str, fetch_refs: bool
+) -> Dict[str, Dict]:
+    """Fetch status of running exps out of current working directory
+
+    Args:
+        queue (BaseStashQueue):
+        rev (str): stash revision of the experiment
+        fetch_refs (bool): fetch running checkpoint results to local or not.
+
+    Returns:
+        Dict[str, Dict]: _description_
+    """
+    from dvc.scm import InvalidRemoteSCMRepo
+    from dvc.utils.serialize import load_json
+
+    from ..executor.local import TempDirExecutor
+
+    result: Dict[str, Dict] = {}
+    infofile = queue.get_infofile_path(rev)
+    try:
+        info = ExecutorInfo.from_dict(load_json(infofile))
+    except OSError:
+        return result
+    if info.status < TaskStatus.FAILED:
+        result[rev] = info.asdict()
+        if info.git_url and fetch_refs and info.status > TaskStatus.PREPARING:
+
+            def on_diverged(_ref: str, _checkpoint: bool):
+                return False
+
+            executor = TempDirExecutor.from_info(info)
+            try:
+                for ref in executor.fetch_exps(
+                    queue.scm,
+                    on_diverged=on_diverged,
+                ):
+                    logger.debug("Updated running experiment '%s'.", ref)
+                    last_rev = queue.scm.get_ref(ref)
+                    result[rev]["last"] = last_rev
+                    if last_rev:
+                        result[last_rev] = info.asdict()
+            except InvalidRemoteSCMRepo:
+                # ignore stale info files
+                del result[rev]
+    return result

dvc/repo/experiments/queue/workspace.py

@@ -1,15 +1,25 @@
+import json
 import logging
-import os
 from collections import defaultdict
-from typing import TYPE_CHECKING, Collection, Dict, Generator, Optional
+from typing import TYPE_CHECKING, Collection, Dict, Generator, List, Optional
 
+import psutil
 from funcy import first
+from voluptuous import Invalid
 
 from dvc.exceptions import DvcException
+from dvc.lock import make_lock
+from dvc.rwlock import RWLOCK_FILE, RWLOCK_LOCK, SCHEMA
+from dvc.utils import relpath
 from dvc.utils.fs import remove
 
 from ..exceptions import ExpQueueEmptyError
-from ..executor.base import BaseExecutor, ExecutorResult
+from ..executor.base import (
+    BaseExecutor,
+    ExecutorInfo,
+    ExecutorResult,
+    TaskStatus,
+)
 from ..executor.local import WorkspaceExecutor
 from ..refs import EXEC_BRANCH
 from .base import BaseStashQueue, QueueDoneResult, QueueEntry, QueueGetResult
@@ -41,7 +51,7 @@ class WorkspaceQueue(BaseStashQueue):
             stash_entry.name,
             stash_entry.head_rev,
         )
-        executor = self.setup_executor(self.repo.experiments, entry)
+        executor = self.init_executor(self.repo.experiments, entry)
         return QueueGetResult(entry, executor)
 
     def iter_queued(self) -> Generator[QueueEntry, None, None]:
@@ -118,9 +128,7 @@ class WorkspaceQueue(BaseStashQueue):
                 f"Failed to reproduce experiment '{rev[:7]}'"
             ) from exc
         finally:
-            if self._EXEC_NAME == exec_name:
-                remove(os.path.join(self.pid_dir, exec_name))
-            executor.cleanup()
+            executor.cleanup(infofile)
         return results
 
     @staticmethod
@@ -135,6 +143,7 @@ class WorkspaceQueue(BaseStashQueue):
             assert exec_result.exp_hash
             logger.debug("Collected experiment '%s'.", exp_rev[:7])
             results[exp_rev] = exec_result.exp_hash
+
         return results
 
     def get_result(self, entry: QueueEntry) -> Optional[ExecutorResult]:
@@ -153,3 +162,89 @@ class WorkspaceQueue(BaseStashQueue):
         follow: bool = False,
     ):
         raise NotImplementedError
+
+    def check_rwlock(
+        self,
+        hardlink: bool = False,
+        autocorrect: bool = False,
+    ) -> bool:
+        """Check and autocorrect the RWLock status for file paths.
+
+        Args:
+            hardlink (bool): use hardlink lock to guard rwlock file when on
+                            edit.
+            autocorrect (bool): autocorrect corrupted rwlock file.
+
+        Return:
+            (bool): if the pid alive.
+        """
+        path = self.repo.fs.path.join(self.repo.tmp_dir, RWLOCK_FILE)
+
+        rwlock_guard = make_lock(
+            self.repo.fs.path.join(self.repo.tmp_dir, RWLOCK_LOCK),
+            tmp_dir=self.repo.tmp_dir,
+            hardlink_lock=hardlink,
+        )
+        with rwlock_guard:
+            try:
+                with self.repo.fs.open(path, encoding="utf-8") as fobj:
+                    lock: Dict[str, List[Dict]] = SCHEMA(json.load(fobj))
+                file_path = first(lock["read"])
+                if not file_path:
+                    return False
+                lock_info = first(lock["read"][file_path])
+                pid = int(lock_info["pid"])
+                if psutil.pid_exists(pid):
+                    return True
+                cmd = lock_info["cmd"]
+                logger.warning(
+                    "Process '%s' with (Pid %s), in RWLock-file '%s'"
+                    " had been killed.",
+                    cmd,
+                    pid,
+                    relpath(path),
+                )
+            except FileNotFoundError:
+                return False
+            except json.JSONDecodeError:
+                logger.warning(
+                    "Unable to read RWLock-file '%s'. JSON structure is"
+                    " corrupted",
+                    relpath(path),
+                )
+            except Invalid:
+                logger.warning("RWLock-file '%s' format error.", relpath(path))
+            if autocorrect:
+                logger.warning(
+                    "Delete corrupted RWLock-file '%s'", relpath(path)
+                )
+                remove(path)
+            return False
+
+    def get_running_exps(self, fetch_refs: bool = True) -> Dict[str, Dict]:
+        from dvc.utils.serialize import load_json
+
+        assert self._EXEC_NAME
+        result: Dict[str, Dict] = {}
+
+        if not self.check_rwlock(autocorrect=True):
+            return result
+
+        infofile = self.get_infofile_path(self._EXEC_NAME)
+
+        try:
+            info = ExecutorInfo.from_dict(load_json(infofile))
+        except OSError:
+            return result
+
+        if info.status < TaskStatus.FAILED:
+            # If we are appending to a checkpoint branch in a workspace
+            # run, show the latest checkpoint as running.
+            if info.status == TaskStatus.SUCCESS:
+                return result
+            last_rev = self.scm.get_ref(EXEC_BRANCH)
+            if last_rev:
+                result[last_rev] = info.asdict()
+            else:
+                result[self._EXEC_NAME] = info.asdict()
+        return result

dvc/repo/experiments/show.py

@@ -1,6 +1,7 @@
 import logging
 from collections import OrderedDict, defaultdict
 from datetime import datetime
+from enum import Enum
 from itertools import chain
 from typing import TYPE_CHECKING, Any, Callable, Dict, List, Optional, Union
 
@@ -17,16 +18,21 @@ if TYPE_CHECKING:
 logger = logging.getLogger(__name__)
 
 
+class ExpStatus(Enum):
+    Success = 0
+    Queued = 1
+    Running = 2
+    Failed = 3
+
+
 @error_handler
 def _collect_experiment_commit(
-    repo,
-    exp_rev,
-    stash=False,
-    sha_only=True,
+    repo: "Repo",
+    exp_rev: str,
+    status: ExpStatus = ExpStatus.Success,
     param_deps=False,
     running=None,
     onerror: Optional[Callable] = None,
-    is_baseline: bool = False,
 ):
     from dvc.dependency import ParamsDependency, RepoDependency
 
@@ -67,37 +73,35 @@ def _collect_experiment_commit(
             if not (out.is_metric or out.is_plot)
         }
 
-        res["queued"] = stash
-        if running is not None and exp_rev in running:
-            res["running"] = True
-            res["executor"] = running[exp_rev].get("location")
+        res["status"] = status.name
+        if status == ExpStatus.Running:
+            res["executor"] = running.get(exp_rev, {}).get("location", None)
         else:
-            res["running"] = False
             res["executor"] = None
-        if not stash:
+
+        if status == ExpStatus.Failed:
+            res["error"] = {
+                "msg": "Experiment run failed.",
+                "type": "",
+            }
+
+        if status not in {ExpStatus.Queued, ExpStatus.Failed}:
             vals = _gather_metrics(
                 repo, targets=None, rev=rev, recursive=False, onerror=onerror
             )
             res["metrics"] = vals
 
-        if not sha_only and rev != "workspace":
-            name: Optional[str] = None
-            if is_baseline:
-                for refspec in ["refs/tags", "refs/heads"]:
-                    name = repo.scm.describe(rev, base=refspec)
-                    if name:
-                        break
-            if not name:
-                name = repo.experiments.get_exact_name(rev)
-            if name:
-                name = name.rsplit("/")[-1]
-                res["name"] = name
-
     return res
 
 
 def _collect_experiment_branch(
-    res, repo, branch, baseline, onerror: Optional[Callable] = None, **kwargs
+    res,
+    repo,
+    branch,
+    baseline,
+    onerror: Optional[Callable] = None,
+    running=None,
+    **kwargs,
 ):
     from dvc.scm import resolve_rev
 
@@ -105,8 +109,14 @@ def _collect_experiment_branch(
     prev = None
     revs = list(repo.scm.branch_revs(exp_rev, baseline))
     for rev in revs:
+        status = ExpStatus.Running if rev in running else ExpStatus.Success
         collected_exp = _collect_experiment_commit(
-            repo, rev, onerror=onerror, **kwargs
+            repo,
+            rev,
+            onerror=onerror,
+            status=status,
+            running=running,
+            **kwargs,
         )
         if len(revs) > 1:
             exp = {"checkpoint_tip": exp_rev}
@@ -129,12 +139,129 @@ def _collect_experiment_branch(
     return res
 
 
+def get_names(repo: "Repo", result: Dict[str, Dict[str, Any]]):
+
+    rev_set = set()
+    baseline_set = set()
+    for baseline in result:
+        for rev in result[baseline]:
+            if rev == "baseline":
+                rev = baseline
+                baseline_set.add(rev)
+            if rev != "workspace":
+                rev_set.add(rev)
+
+    names: Dict[str, Optional[str]] = {}
+    for base in ("refs/tags/", "refs/heads/"):
+        if rev_set:
+            names.update(
+                (rev, ref[len(base) :])
+                for rev, ref in repo.scm.describe(
+                    baseline_set, base=base
+                ).items()
+                if ref is not None
+            )
+            rev_set.difference_update(names.keys())
+
+    exact_name = repo.experiments.get_exact_name(rev_set)
+
+    for baseline, baseline_results in result.items():
+        for rev, rev_result in baseline_results.items():
+            name: Optional[str] = None
+            if rev == "baseline":
+                rev = baseline
+                if rev == "workspace":
+                    continue
+                name = names.get(rev, None)
+            name = name or exact_name[rev]
+            if name:
+                rev_result["data"]["name"] = name
+
+
+# flake8: noqa: C901
+def _collect_active_experiment(
+    repo: "Repo",
+    found_revs: Dict[str, List[str]],
+    running: Dict[str, Any],
+    **kwargs,
+) -> Dict[str, Dict[str, Any]]:
+    result: Dict[str, Dict] = defaultdict(OrderedDict)
+    for entry in chain(
+        repo.experiments.tempdir_queue.iter_active(),
+        repo.experiments.celery_queue.iter_active(),
+    ):
+        stash_rev = entry.stash_rev
+        if entry.baseline_rev in found_revs and (
+            stash_rev not in running or not running[stash_rev].get("last")
+        ):
+            result[entry.baseline_rev][stash_rev] = _collect_experiment_commit(
+                repo,
+                stash_rev,
+                status=ExpStatus.Running,
+                running=running,
+                **kwargs,
+            )
+    return result
+
+
+def _collect_queued_experiment(
+    repo: "Repo",
+    found_revs: Dict[str, List[str]],
+    running: Dict[str, Any],
+    **kwargs,
+) -> Dict[str, Dict[str, Any]]:
+    result: Dict[str, Dict] = defaultdict(OrderedDict)
+    for entry in repo.experiments.celery_queue.iter_queued():
+        stash_rev = entry.stash_rev
+        if entry.baseline_rev in found_revs:
+            result[entry.baseline_rev][stash_rev] = _collect_experiment_commit(
+                repo,
+                stash_rev,
+                status=ExpStatus.Queued,
+                running=running,
+                **kwargs,
+            )
+    return result
+
+
+def _collect_failed_experiment(
+    repo: "Repo",
+    found_revs: Dict[str, List[str]],
+    running: Dict[str, Any],
+    **kwargs,
+) -> Dict[str, Dict[str, Any]]:
+    result: Dict[str, Dict] = defaultdict(OrderedDict)
+    for queue_done_result in repo.experiments.celery_queue.iter_failed():
+        entry = queue_done_result.entry
+        stash_rev = entry.stash_rev
+        if entry.baseline_rev in found_revs:
+            experiment = _collect_experiment_commit(
+                repo,
+                stash_rev,
+                status=ExpStatus.Failed,
+                running=running,
+                **kwargs,
+            )
+            result[entry.baseline_rev][stash_rev] = experiment
+    return result
+
+
+def update_new(
+    to_dict: Dict[str, Dict[str, Any]], from_dict: Dict[str, Dict[str, Any]]
+):
+    for baseline, experiments in from_dict.items():
+        for rev, experiment in experiments.items():
+            to_dict[baseline][rev] = to_dict[baseline].get(rev, experiment)
+
+
 def show(
     repo: "Repo",
     all_branches=False,
     all_tags=False,
     revs: Union[List[str], str, None] = None,
     all_commits=False,
+    hide_queued=False,
+    hide_failed=False,
     sha_only=False,
     num=1,
     param_deps=False,
@@ -160,17 +287,52 @@ def show(
         iter_revs(repo.scm, revs, num, all_branches, all_tags, all_commits)
     )
 
-    running = repo.experiments.get_running_exps(fetch_refs=fetch_running)
+    running: Dict[str, Dict] = repo.experiments.get_running_exps(
+        fetch_refs=fetch_running
+    )
+
+    queued_experiment = (
+        _collect_queued_experiment(
+            repo,
+            found_revs,
+            running,
+            param_deps=param_deps,
+            onerror=onerror,
+        )
+        if not hide_queued
+        else {}
+    )
+
+    active_experiment = _collect_active_experiment(
+        repo,
+        found_revs,
+        running,
+        param_deps=param_deps,
+        onerror=onerror,
+    )
+
+    failed_experiments = (
+        _collect_failed_experiment(
+            repo,
+            found_revs,
+            running,
+            sha_only=sha_only,
+            param_deps=param_deps,
+            onerror=onerror,
+        )
+        if not hide_failed
+        else {}
+    )
 
     for rev in found_revs:
+        status = ExpStatus.Running if rev in running else ExpStatus.Success
         res[rev]["baseline"] = _collect_experiment_commit(
             repo,
             rev,
-            sha_only=sha_only,
+            status=status,
             param_deps=param_deps,
             running=running,
             onerror=onerror,
-            is_baseline=True,
         )
 
         if rev == "workspace":
@@ -191,31 +353,20 @@ def show(
                 repo,
                 exp_ref,
                 rev,
-                sha_only=sha_only,
                 param_deps=param_deps,
                 running=running,
                 onerror=onerror,
             )
 
-        # collect standalone & celery experiments
-        for entry in chain(
-            repo.experiments.tempdir_queue.iter_active(),
-            repo.experiments.celery_queue.iter_active(),
-            repo.experiments.celery_queue.iter_queued(),
-        ):
-            stash_rev = entry.stash_rev
-            if entry.baseline_rev in found_revs:
-                if stash_rev not in running or not running[stash_rev].get(
-                    "last"
-                ):
-                    experiment = _collect_experiment_commit(
-                        repo,
-                        stash_rev,
-                        sha_only=sha_only,
-                        stash=stash_rev not in running,
-                        param_deps=param_deps,
-                        running=running,
-                        onerror=onerror,
-                    )
-                    res[entry.baseline_rev][stash_rev] = experiment
+    if not hide_failed:
+        update_new(res, failed_experiments)
+
+    update_new(res, active_experiment)
+
+    if not hide_queued:
+        update_new(res, queued_experiment)
+
+    if not sha_only:
+        get_names(repo, res)
+
     return res

dvc/repo/fetch.py

@@ -3,12 +3,15 @@ from contextlib import suppress
 from typing import TYPE_CHECKING, Optional
 
 from dvc.config import NoRemoteError
-from dvc.exceptions import DownloadError, FileTransferError
+from dvc.exceptions import DownloadError
 from dvc.fs import Schemes
 
 from . import locked
 
 if TYPE_CHECKING:
+    from dvc.repo import Repo
+    from dvc.types import TargetType
+    from dvc_data.hashfile.transfer import TransferResult
     from dvc_objects.db.base import ObjectDB
 
 logger = logging.getLogger(__name__)
@@ -25,22 +28,6 @@ def _fetch_worktree(repo, remote):
             *key,
         )
     save(index)
-
-    for stage in repo.index.stages:
-        for out in stage.outs:
-            if not out.use_cache:
-                continue
-
-            if not out.is_in_repo:
-                continue
-
-            workspace, key = out.index_key
-            entry = repo.index.data[workspace][key]
-            out.hash_info = entry.hash_info
-            out.meta = entry.meta
-
-        stage.dvcfile.dump(stage)
-
     return len(index)
 
 
@@ -58,7 +45,7 @@ def fetch(
     run_cache=False,
     revs=None,
     odb: Optional["ObjectDB"] = None,
-):
+) -> int:
     """Download data items from a cloud and imported repositories
 
     Returns:
@@ -71,6 +58,9 @@ def fetch(
         config.NoRemoteError: thrown when downloading only local files and no
             remote is configured
     """
+    from dvc.repo.imports import save_imports
+    from dvc_data.hashfile.transfer import TransferResult
+
     if isinstance(targets, str):
         targets = [targets]
 
@@ -79,110 +69,102 @@ def fetch(
         if _remote.worktree:
             return _fetch_worktree(self, _remote)
 
-    used = self.used_objs(
-        targets,
-        all_branches=all_branches,
-        all_tags=all_tags,
-        all_commits=all_commits,
-        with_deps=with_deps,
-        force=True,
-        remote=remote,
-        jobs=jobs,
-        recursive=recursive,
-        revs=revs,
-    )
-
-    downloaded = 0
-    failed = 0
+    failed_count = 0
 
     try:
         if run_cache:
             self.stage_cache.pull(remote)
     except DownloadError as exc:
-        failed += exc.amount
+        failed_count += exc.amount
 
+    no_remote_msg: Optional[str] = None
+    result = TransferResult(set(), set())
+    try:
+        d, f = _fetch(
+            self,
+            targets,
+            all_branches=all_branches,
+            all_tags=all_tags,
+            all_commits=all_commits,
+            with_deps=with_deps,
+            force=True,
+            remote=remote,
+            jobs=jobs,
+            recursive=recursive,
+            revs=revs,
+            odb=odb,
+        )
+        result.transferred.update(d)
+        result.failed.update(f)
+    except NoRemoteError as exc:
+        no_remote_msg = str(exc)
+
+    for rev in self.brancher(
+        revs=revs,
+        all_branches=all_branches,
+        all_tags=all_tags,
+        all_commits=all_commits,
+    ):
+        imported = save_imports(
+            self,
+            targets,
+            unpartial=not rev or rev == "workspace",
+            recursive=recursive,
+        )
+        result.transferred.update(imported)
+        result.failed.difference_update(imported)
+
+    failed_count += len(result.failed)
+
+    if failed_count:
+        if no_remote_msg:
+            logger.error(no_remote_msg)
+        raise DownloadError(failed_count)
+
+    return len(result.transferred)
+
+
+def _fetch(
+    repo: "Repo",
+    targets: "TargetType",
+    remote: Optional[str] = None,
+    jobs: Optional[int] = None,
+    odb: Optional["ObjectDB"] = None,
+    **kwargs,
+) -> "TransferResult":
+    from dvc_data.hashfile.transfer import TransferResult
+
+    result = TransferResult(set(), set())
+    used = repo.used_objs(
+        targets,
+        remote=remote,
+        jobs=jobs,
+        **kwargs,
+    )
     if odb:
         all_ids = set()
         for _odb, obj_ids in used.items():
             all_ids.update(obj_ids)
-        d, f = _fetch(
-            self,
+        d, f = repo.cloud.pull(
             all_ids,
             jobs=jobs,
             remote=remote,
             odb=odb,
         )
-        downloaded += d
-        failed += f
+        result.transferred.update(d)
+        result.failed.update(f)
     else:
         for src_odb, obj_ids in sorted(
             used.items(),
             key=lambda item: item[0] is not None
             and item[0].fs.protocol == Schemes.MEMORY,
         ):
-            d, f = _fetch(
-                self,
+            d, f = repo.cloud.pull(
                 obj_ids,
                 jobs=jobs,
                 remote=remote,
                 odb=src_odb,
             )
-            downloaded += d
-            failed += f
-
-    d, f = _fetch_partial_imports(
-        self,
-        targets,
-        all_branches=all_branches,
-        all_tags=all_tags,
-        all_commits=all_commits,
-        recursive=recursive,
-        revs=revs,
-    )
-    downloaded += d
-    failed += f
-
-    if failed:
-        raise DownloadError(failed)
-
-    return downloaded
-
-
-def _fetch(repo, obj_ids, **kwargs):
-    downloaded = 0
-    failed = 0
-    try:
-        downloaded += repo.cloud.pull(obj_ids, **kwargs)
-    except FileTransferError as exc:
-        failed += exc.amount
-    return downloaded, failed
-
-
-def _fetch_partial_imports(repo, targets, **kwargs):
-    from dvc.stage.exceptions import DataSourceChanged
-
-    downloaded = 0
-    failed = 0
-    for stage in repo.partial_imports(targets, **kwargs):
-        dep = stage.deps[0]
-        out = stage.outs[0]
-        try:
-            if dep.changed_checksum():
-                raise DataSourceChanged(f"{stage} ({dep})")
-
-            logger.info("Importing '%s' -> '%s'", dep, out)
-            out.transfer(
-                dep.def_path, jobs=kwargs.get("jobs"), odb=repo.odb.local
-            )
-        except DataSourceChanged as exc:
-            logger.warning(f"{exc}")
-            failed += 1
-            continue
-        if not any(
-            kwargs.get(kw, None)
-            for kw in ("all_branches", "all_tags", "all_commits", "revs")
-        ):
-            stage.dump()
-
-        downloaded += 1
-    return downloaded, failed
+            result.transferred.update(d)
+            result.failed.update(f)
+    return result

dvc/repo/get_url.py

@@ -1,19 +1,17 @@
 import os
 
+from dvc import output
+from dvc.exceptions import URLMissingError
+from dvc.fs import download, parse_external_url
+from dvc.utils import resolve_output
 
-def get_url(url, out=None, jobs=None):
-    from dvc import dependency, output
-    from dvc.utils import resolve_output
 
+def get_url(url, out=None, *, config=None, jobs=None):
     out = resolve_output(url, out)
-
-    if os.path.exists(url):
-        url = os.path.abspath(url)
-
     out = os.path.abspath(out)
-
-    (dep,) = dependency.loads_from(None, [url])
     (out,) = output.loads_from(None, [out], use_cache=False)
-    if not dep.exists:
-        raise dep.DoesNotExistError(dep)
-    dep.download(out, jobs=jobs)
+
+    src_fs, src_path = parse_external_url(url, config)
+    if not src_fs.exists(src_path):
+        raise URLMissingError(url)
+    download(src_fs, src_path, out, jobs=jobs)

dvc/repo/imp_url.py

@@ -1,4 +1,5 @@
 import os
+from typing import TYPE_CHECKING
 
 from dvc.repo.scm_context import scm_context
 from dvc.utils import relpath, resolve_output, resolve_paths
@@ -7,6 +8,9 @@ from dvc.utils.fs import path_isin
 from ..exceptions import InvalidArgumentError, OutputDuplicationError
 from . import locked
 
+if TYPE_CHECKING:
+    from dvc.dvcfile import DVCFile
+
 
 @locked
 @scm_context
@@ -27,6 +31,7 @@ def imp_url(
     meta=None,
     jobs=None,
     fs_config=None,
+    version_aware: bool = False,
 ):
     from dvc.dvcfile import Dvcfile
     from dvc.stage import Stage, create_stage, restore_fields
@@ -54,6 +59,11 @@ def imp_url(
     ):
         url = relpath(url, wdir)
 
+    if version_aware:
+        if fs_config is None:
+            fs_config = {}
+        fs_config["version_aware"] = True
+
     stage = create_stage(
         Stage,
         self,
@@ -68,7 +78,7 @@ def imp_url(
 
     out_obj = stage.outs[0]
     out_obj.annot.update(desc=desc, type=type, labels=labels, meta=meta)
-    dvcfile = Dvcfile(self, stage.path)
+    dvcfile: "DVCFile" = Dvcfile(self, stage.path)  # type: ignore
     dvcfile.remove()
 
     try:

dvc/repo/imports.py

@@ -0,0 +1,132 @@
+import logging
+from tempfile import TemporaryDirectory
+from typing import TYPE_CHECKING, List, Set, Tuple, Union
+
+if TYPE_CHECKING:
+    from dvc.dependency.base import Dependency
+    from dvc.repo import Repo
+    from dvc.repo.index import Index, IndexView
+    from dvc.stage import Stage
+    from dvc.types import TargetType
+    from dvc_data.hashfile import HashInfo
+
+logger = logging.getLogger(__name__)
+
+
+def unfetched_view(
+    index: "Index", targets: "TargetType", unpartial: bool = False, **kwargs
+) -> Tuple["IndexView", List["Dependency"]]:
+    """Return index view of imports which have not been fetched.
+
+    Returns:
+        Tuple in the form (view, changed_deps) where changed_imports is a list
+        of import dependencies that cannot be fetched due to changed data
+        source.
+    """
+    changed_deps: List["Dependency"] = []
+
+    def need_fetch(stage: "Stage") -> bool:
+        if (
+            not stage.is_import
+            or stage.is_repo_import
+            or (stage.is_partial_import and not unpartial)
+        ):
+            return False
+
+        out = stage.outs[0]
+        if not out.changed_cache():
+            return False
+
+        dep = stage.deps[0]
+        if dep.changed_checksum():
+            changed_deps.append(dep)
+            return False
+
+        return True
+
+    unfetched = index.targets_view(targets, filter_fn=need_fetch, **kwargs)
+    return unfetched, changed_deps
+
+
+def partial_view(
+    index: "Index", targets: "TargetType", **kwargs
+) -> "IndexView":
+    return index.targets_view(
+        targets,
+        filter_fn=lambda s: s.is_partial_import,
+        **kwargs,
+    )
+
+
+def unpartial_imports(index: Union["Index", "IndexView"]) -> int:
+    """Update any outs in the index which are no longer partial imports.
+
+    Returns:
+        Total number of files which were unpartialed.
+    """
+    updated = 0
+    for out in index.outs:
+        # we need to use view[key] here and since the out fields have not been
+        # updated yet (out.get_entry() would return the partial-import state)
+        workspace, key = out.index_key
+        entry = index.data[workspace][key]
+        if out.stage.is_partial_import:
+            if out.stage.is_repo_import:
+                dep = out.stage.deps[0]
+                out.hash_info = dep.get_obj().hash_info
+                out.meta = dep.get_meta()
+            else:
+                out.hash_info = entry.hash_info
+                out.meta = entry.meta
+            out.stage.dump()
+            updated += out.meta.nfiles if out.meta.nfiles is not None else 1
+    return updated
+
+
+def save_imports(
+    repo: "Repo", targets: "TargetType", unpartial: bool = False, **kwargs
+) -> Set["HashInfo"]:
+    """Save (download) imports from their original source location.
+
+    Imports which are already cached will not be downloaded.
+
+    Returns:
+        Objects which were downloaded from source location.
+    """
+    from dvc.stage.exceptions import DataSourceChanged
+    from dvc.utils.fs import makedirs
+    from dvc_data.index import checkout, md5, save
+    from dvc_objects.fs.callbacks import Callback
+
+    downloaded: Set["HashInfo"] = set()
+
+    unfetched, changed = unfetched_view(
+        repo.index, targets, unpartial=unpartial, **kwargs
+    )
+    for dep in changed:
+        logger.warning(str(DataSourceChanged(f"{dep.stage} ({dep})")))
+
+    data_view = unfetched.data["repo"]
+    if len(data_view):
+        cache = repo.odb.local
+        if not cache.fs.exists(cache.path):
+            makedirs(cache.path)
+        with TemporaryDirectory(dir=cache.path) as tmpdir:
+            with Callback.as_tqdm_callback(
+                desc="Downloading imports from source",
+                unit="files",
+            ) as cb:
+                checkout(data_view, tmpdir, cache.fs, callback=cb)
+            md5(data_view)
+            save(data_view, odb=cache, hardlink=True)
+
+        downloaded.update(
+            entry.hash_info
+            for _, entry in data_view.iteritems()
+            if not entry.meta.isdir and entry.hash_info is not None
+        )
+
+    if unpartial:
+        unpartial_imports(partial_view(repo.index, targets, **kwargs))
+
+    return downloaded

dvc/repo/index.py

@@ -1,13 +1,13 @@
 from contextlib import suppress
 from typing import (
     TYPE_CHECKING,
+    Any,
     Callable,
     Dict,
     Iterable,
     Iterator,
     List,
     Optional,
-    Sequence,
     Set,
 )
 
@@ -22,11 +22,11 @@ if TYPE_CHECKING:
     from dvc.dependency import Dependency, ParamsDependency
     from dvc.fs import FileSystem
     from dvc.output import Output
-    from dvc.repo.stage import StageLoad
+    from dvc.repo.stage import StageInfo, StageLoad
     from dvc.stage import Stage
     from dvc.types import StrPath, TargetType
     from dvc_data.hashfile.hash_info import HashInfo
-    from dvc_data.index import DataIndex
+    from dvc_data.index import DataIndex, DataIndexKey, DataIndexView
     from dvc_objects.db import ObjectDB
 
 
@@ -64,6 +64,7 @@ class Index:
         self.stage_collector: "StageLoad" = repo.stage
         if stages is not None:
             self.stages: List["Stage"] = stages
+        self._collected_targets: Dict[int, List["StageInfo"]] = {}
 
     @cached_property
     def stages(self) -> List["Stage"]:  # pylint: disable=method-hidden
@@ -113,6 +114,62 @@ class Index:
 
         return self.filter(is_stage_inside_path)
 
+    @staticmethod
+    def _hash_targets(
+        targets: Iterable[Optional[str]],
+        **kwargs: Any,
+    ) -> int:
+        return hash(
+            (
+                frozenset(targets),
+                kwargs.get("with_deps", False),
+                kwargs.get("recursive", False),
+            )
+        )
+
+    def _collect_targets(
+        self, targets: Optional["TargetType"], **kwargs: Any
+    ) -> List["StageInfo"]:
+        from itertools import chain
+
+        from dvc.repo.stage import StageInfo
+        from dvc.utils.collections import ensure_list
+
+        targets = ensure_list(targets)
+        if not targets:
+            return [StageInfo(stage) for stage in self.stages]
+        targets_hash = self._hash_targets(targets, **kwargs)
+        if targets_hash not in self._collected_targets:
+            self._collected_targets[targets_hash] = list(
+                chain.from_iterable(
+                    self.stage_collector.collect_granular(target, **kwargs)
+                    for target in targets
+                )
+            )
+        return self._collected_targets[targets_hash]
+
+    def targets_view(
+        self,
+        targets: "TargetType",
+        filter_fn: Optional[Callable[["Stage"], bool]] = None,
+        **kwargs: Any,
+    ) -> "IndexView":
+        """Return read-only view of index for the specified targets.
+
+        Args:
+            targets: Targets to collect
+            filter_fn: Optional stage filter to be applied after collecting
+                targets.
+
+        Additional kwargs will be passed into the stage collector.
+        """
+        stage_infos = [
+            stage_info
+            for stage_info in self._collect_targets(targets, **kwargs)
+            if not filter_fn or filter_fn(stage_info.stage)
+        ]
+        return IndexView(self, stage_infos)
+
     @property
     def outs(self) -> Iterator["Output"]:
         for stage in self:
@@ -171,8 +228,7 @@ class Index:
     def data(self) -> "Dict[str, DataIndex]":
         from collections import defaultdict
 
-        from dvc.config import NoRemoteError
-        from dvc_data.index import DataIndex, DataIndexEntry
+        from dvc_data.index import DataIndex
 
         by_workspace: dict = defaultdict(DataIndex)
 
@@ -184,22 +240,18 @@ class Index:
                 continue
 
             workspace, key = out.index_key
-
-            try:
-                remote = self.repo.cloud.get_remote_odb(out.remote)
-            except NoRemoteError:
-                remote = None
-
             data_index = by_workspace[workspace]
 
-            data_index[key] = DataIndexEntry(
-                meta=out.meta,
-                obj=out.get_obj() if out.files else out.obj,
-                hash_info=out.hash_info,
-                odb=out.odb,
-                cache=out.odb,
-                remote=remote,
-            )
+            if out.files:
+                out.obj = out.get_obj()
+
+            entry = out.get_entry()
+            entry.key = key
+            if out.stage.is_import and not out.stage.is_repo_import:
+                entry.fs = out.stage.deps[0].fs
+                entry.path = out.stage.deps[0].fs_path
+                entry.meta = out.stage.deps[0].meta
+            data_index.add(entry)
 
         return dict(by_workspace)
 
@@ -213,22 +265,11 @@ class Index:
         jobs: int = None,
     ) -> "ObjectContainer":
         from collections import defaultdict
-        from itertools import chain
-
-        from dvc.utils.collections import ensure_list
 
         used: "ObjectContainer" = defaultdict(set)
-        collect_targets: Sequence[Optional[str]] = (None,)
-        if targets:
-            collect_targets = ensure_list(targets)
-
-        pairs = chain.from_iterable(
-            self.stage_collector.collect_granular(
-                target, recursive=recursive, with_deps=with_deps
-            )
-            for target in collect_targets
+        pairs = self._collect_targets(
+            targets, recursive=recursive, with_deps=with_deps
         )
-
         for stage, filter_info in pairs:
             for odb, objs in stage.get_used_objs(
                 remote=remote,
@@ -239,28 +280,6 @@ class Index:
                 used[odb].update(objs)
         return used
 
-    def partial_imports(
-        self,
-        targets: "TargetType" = None,
-        recursive: bool = False,
-    ) -> List["Stage"]:
-        from itertools import chain
-
-        from dvc.utils.collections import ensure_list
-
-        collect_targets: Sequence[Optional[str]] = (None,)
-        if targets:
-            collect_targets = ensure_list(targets)
-
-        pairs = chain.from_iterable(
-            self.stage_collector.collect_granular(
-                target, recursive=recursive, with_deps=True
-            )
-            for target in collect_targets
-        )
-
-        return [stage for stage, _ in pairs if stage.is_partial_import]
-
     # Following methods help us treat the collection as a set-like structure
     # and provides faux-immutability.
     # These methods do not preserve stages order.
@@ -326,6 +345,89 @@ class Index:
         return dict_md5(self.dumpd())
 
 
+class IndexView:
+    """Read-only view of Index.data using filtered stages."""
+
+    def __init__(  # pylint: disable=redefined-outer-name
+        self, index: Index, stage_infos: Iterable["StageInfo"]
+    ):
+        self._index = index
+        self._stage_infos = stage_infos
+        # NOTE: stage_infos might have the same stage multiple times but with
+        # different filter_info
+        self._stages = list({stage for stage, _ in stage_infos})
+
+    def __len__(self) -> int:
+        return len(self._stages)
+
+    def __contains__(self, stage: "Stage") -> bool:
+        # as we are keeping stages inside a list, it might be slower.
+        return stage in self._stages
+
+    def __iter__(self) -> Iterator["Stage"]:
+        yield from self._stages
+
+    def __getitem__(self, item: str) -> "Stage":
+        """Get a stage by its addressing attribute."""
+        for stage in self:
+            if stage.addressing == item:
+                return stage
+        raise KeyError(f"{item} - available stages are {self.stages}")
+
+    @property
+    def stages(self):
+        return self._stages
+
+    @property
+    def deps(self) -> Iterator["Dependency"]:
+        for stage in self:
+            yield from stage.deps
+
+    @property
+    def outs(self) -> Iterator["Output"]:
+        outs = set()
+        for stage, filter_info in self._stage_infos:
+            for out in stage.filter_outs(filter_info):
+                outs.add(out)
+        yield from outs
+
+    @cached_property
+    def _data_prefixes(self) -> Dict[str, Set["DataIndexKey"]]:
+        from collections import defaultdict
+
+        prefixes: Dict[str, Set["DataIndexKey"]] = defaultdict(set)
+        for stage, filter_info in self._stage_infos:
+            for out in stage.filter_outs(filter_info):
+                workspace, key = out.index_key
+                if filter_info and out.fs.path.isin(filter_info, out.fs_path):
+                    key = (
+                        *key,
+                        out.fs.path.relparts(filter_info, out.fs_path),
+                    )
+                prefixes[workspace].add(key)
+                prefixes[workspace].update(
+                    key[:i] for i in range(len(key), 0, -1)
+                )
+        return prefixes
+
+    @cached_property
+    def data(self) -> "Dict[str, DataIndexView]":
+        from functools import partial
+
+        from dvc_data.index import view
+
+        def key_filter(workspace: str, key: "DataIndexKey"):
+            try:
+                return key in self._data_prefixes[workspace]
+            except KeyError:
+                return False
+
+        return {
+            workspace: view(index, partial(key_filter, workspace))
+            for workspace, index in self._index.data.items()
+        }
+
+
 if __name__ == "__main__":
     import logging
 

dvc/repo/ls_url.py

@@ -0,0 +1,29 @@
+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)
+    try:
+        info = fs.info(fs_path)
+    except FileNotFoundError as exc:
+        raise URLMissingError(url) from exc
+    if info["type"] != "directory":
+        return [{"path": info["name"], "isdir": False}]
+
+    ret = []
+    for _, dirs, files in fs.walk(fs_path, detail=True):
+        if not recursive:
+            files.update(dirs)
+
+        for info in files.values():
+            ls_info = {
+                "path": fs.path.relpath(info["name"], fs_path),
+                "isdir": info["type"] == "directory",
+            }
+            ret.append(ls_info)
+
+        if not recursive:
+            break
+
+    return ret

dvc/repo/plots/__init__.py

@@ -100,7 +100,7 @@ class Plots:
                             "filename":
                             {
                                 "data_source": callable loading the data,
-                                "props": propreties for the file if it is
+                                "props": properties for the file if it is
                                          plots type output
                             }
                         }
@@ -444,6 +444,10 @@ def _collect_pipeline_files(repo, targets: List[str], props, onerror=None):
         if isinstance(dvcfile, PipelineFile):
             dvcfile_path = _relpath(repo.dvcfs, dvcfile.path)
             dvcfile_defs = dvcfile.load().get("plots", {})
+            if isinstance(dvcfile_defs, list):
+                dvcfile_defs = {
+                    k: v for elem in dvcfile_defs for k, v in elem.items()
+                }
             resolved = _resolve_definitions(
                 repo.dvcfs,
                 targets,

dvc/repo/push.py

@@ -1,6 +1,6 @@
 from typing import TYPE_CHECKING, Optional
 
-from dvc.exceptions import FileTransferError, UploadError
+from dvc.exceptions import UploadError
 
 from ..utils import glob_targets
 from . import locked
@@ -10,14 +10,11 @@ if TYPE_CHECKING:
 
 
 def _push_worktree(repo, remote):
-    from dvc_data.hashfile.tree import tree_from_index
-    from dvc_data.index import build, checkout, collect
+    from dvc_data.index import checkout
+    from dvc_data.index.save import build_tree
 
     index = repo.index.data["repo"]
-    checkout(index, remote.path, remote.fs, force=True)
-    build(index, remote.path, remote.fs)
-    if any(out.isdir() for out in repo.index.outs):
-        collect(index, remote.path, remote.fs)
+    checkout(index, remote.path, remote.fs)
 
     for stage in repo.index.stages:
         for out in stage.outs:
@@ -42,8 +39,7 @@ def _push_worktree(repo, remote):
                         repo.fs.path.relparts(fs_path, out.fs_path)
                     )
                     entry.hash_info = hash_info
-                tree_meta, new_tree = tree_from_index(index, key)
-                new_tree.digest()
+                tree_meta, new_tree = build_tree(index, key)
                 out.obj = new_tree
                 out.hash_info = new_tree.hash_info
                 out.meta = tree_meta
@@ -106,18 +102,18 @@ def push(
             if not include_imports and dest_odb and dest_odb.read_only:
                 continue
             all_ids.update(obj_ids)
-        try:
-            pushed += self.cloud.push(all_ids, jobs, remote=remote, odb=odb)
-        except FileTransferError as exc:
-            raise UploadError(exc.amount)
+        result = self.cloud.push(all_ids, jobs, remote=remote, odb=odb)
+        if result.failed:
+            raise UploadError(len(result.failed))
+        pushed += len(result.transferred)
     else:
         for dest_odb, obj_ids in used.items():
             if dest_odb and dest_odb.read_only:
                 continue
-            try:
-                pushed += self.cloud.push(
-                    obj_ids, jobs, remote=remote, odb=odb or dest_odb
-                )
-            except FileTransferError as exc:
-                raise UploadError(exc.amount)
+            result = self.cloud.push(
+                obj_ids, jobs, remote=remote, odb=odb or dest_odb
+            )
+            if result.failed:
+                raise UploadError(len(result.failed))
+            pushed += len(result.transferred)
     return pushed

dvc/repo/stage.py

@@ -225,10 +225,9 @@ class StageLoad:
         path, name = parse_target(target)
         return self.load_one(path=path, name=name)
 
-    @staticmethod
-    def _get_filepath(path: str = None, name: str = None) -> str:
+    def _get_filepath(self, path: str = None, name: str = None) -> str:
         if path:
-            return path
+            return self.repo.fs.path.realpath(path)
 
         path = PIPELINE_FILE
         logger.debug("Assuming '%s' to be a stage inside '%s'", name, path)

dvc/repo/status.py

@@ -14,7 +14,9 @@ def _joint_status(pairs):
     status_info = {}
 
     for stage, filter_info in pairs:
-        if stage.frozen and not stage.is_repo_import:
+        if stage.frozen and not (
+            stage.is_repo_import or stage.is_versioned_import
+        ):
             logger.warning(
                 "%s is frozen. Its dependencies are"
                 " not going to be shown in the status output.",

dvc/rwlock.py

@@ -1,4 +1,5 @@
 import json
+import logging
 import os
 from collections import defaultdict
 from contextlib import contextmanager
@@ -10,6 +11,9 @@ from .fs import localfs
 from .lock import make_lock
 from .utils import relpath
 
+logger = logging.getLogger(__name__)
+
+
 INFO_SCHEMA = {Required("pid"): int, Required("cmd"): str}
 
 SCHEMA = Schema(
@@ -19,6 +23,9 @@ SCHEMA = Schema(
     }
 )
 
+RWLOCK_FILE = "rwlock"
+RWLOCK_LOCK = "rwlock.lock"
+
 
 class RWLockFileCorruptedError(DvcException):
     def __init__(self, path):
@@ -35,10 +42,10 @@ class RWLockFileFormatError(DvcException):
 
 @contextmanager
 def _edit_rwlock(lock_dir, fs, hardlink):
-    path = fs.path.join(lock_dir, "rwlock")
+    path = fs.path.join(lock_dir, RWLOCK_FILE)
 
     rwlock_guard = make_lock(
-        fs.path.join(lock_dir, "rwlock.lock"),
+        fs.path.join(lock_dir, RWLOCK_LOCK),
         tmp_dir=lock_dir,
         hardlink_lock=hardlink,
     )

dvc/schema.py

@@ -118,7 +118,7 @@ PLOT_DEFINITION = {
     Output.PARAM_PLOT_TITLE: str,
     Output.PARAM_PLOT_TEMPLATE: str,
 }
-PLOTS_SCHEMA = {str: Any(PLOT_DEFINITION, None)}
+SINGLE_PLOT_SCHEMA = {str: Any(PLOT_DEFINITION, None)}
 FOREACH_IN = {
     Required(FOREACH_KWD): Any(dict, list, str),
     Required(DO_KWD): STAGE_DEFINITION,
@@ -127,9 +127,9 @@ SINGLE_PIPELINE_STAGE_SCHEMA = {
     str: either_or(STAGE_DEFINITION, FOREACH_IN, [FOREACH_KWD, DO_KWD])
 }
 MULTI_STAGE_SCHEMA = {
+    PLOTS: Any(SINGLE_PLOT_SCHEMA, [SINGLE_PLOT_SCHEMA]),
     STAGES: SINGLE_PIPELINE_STAGE_SCHEMA,
     VARS_KWD: VARS_SCHEMA,
-    PLOTS: PLOTS_SCHEMA,
 }
 
 COMPILED_SINGLE_STAGE_SCHEMA = Schema(SINGLE_STAGE_SCHEMA)

dvc/stage/__init__.py

@@ -257,6 +257,10 @@ class Stage(params.StageParams):
 
         return isinstance(self.deps[0], RepoDependency)
 
+    @property
+    def is_versioned_import(self):
+        return self.is_import and self.deps[0].fs.version_aware
+
     @property
     def is_checkpoint(self):
         """
@@ -568,6 +572,8 @@ class Stage(params.StageParams):
             if kwargs.get("checkpoint_func", None) or no_download:
                 allow_missing = True
             self.save(allow_missing=allow_missing)
+            if no_download:
+                self.ignore_outs()
             if not no_commit:
                 self.commit(allow_missing=allow_missing)
 
@@ -618,7 +624,9 @@ class Stage(params.StageParams):
     @rwlocked(read=["deps", "outs"])
     def status(self, check_updates=False, filter_info=None):
         ret = []
-        show_import = self.is_repo_import and check_updates
+        show_import = (
+            self.is_repo_import or self.is_versioned_import
+        ) and check_updates
 
         if not self.frozen or show_import:
             self._status_deps(ret)
@@ -681,7 +689,7 @@ class Stage(params.StageParams):
         self, *args, **kwargs
     ) -> Dict[Optional["ObjectDB"], Set["HashInfo"]]:
         """Return set of object IDs used by this stage."""
-        if self.is_partial_import:
+        if self.is_partial_import and not self.is_repo_import:
             return {}
 
         used_objs = defaultdict(set)
@@ -706,7 +714,7 @@ class Stage(params.StageParams):
                 "unable to auto-merge DVC files with deleted outputs"
             )
 
-    def merge(self, ancestor, other):
+    def merge(self, ancestor, other, allowed=None):
         assert other
 
         if not other.outs:
@@ -726,7 +734,7 @@ class Stage(params.StageParams):
         self._check_can_merge(self, ancestor_out)
         self._check_can_merge(other, ancestor_out)
 
-        self.outs[0].merge(ancestor_out, other.outs[0])
+        self.outs[0].merge(ancestor_out, other.outs[0], allowed=allowed)
 
     def dump(self, **kwargs):
         self.dvcfile.dump(self, **kwargs)
@@ -768,5 +776,5 @@ class PipelineStage(Stage):
     def _changed_stage_entry(self):
         return f"'cmd' of {self} has changed."
 
-    def merge(self, ancestor, other):
+    def merge(self, ancestor, other, allowed=None):
         raise NotImplementedError

dvc/stage/cache.py

@@ -9,7 +9,6 @@ from funcy import cached_property, first
 from dvc import fs
 from dvc.exceptions import DvcException
 from dvc.utils import dict_sha256, relpath
-from dvc_data.hashfile.transfer import _log_exceptions
 
 if TYPE_CHECKING:
     from dvc_objects.db import ObjectDB
@@ -222,7 +221,7 @@ class StageCache:
 
         from_fs = from_odb.fs
         to_fs = to_odb.fs
-        func = _log_exceptions(fs.generic.copy)
+        func = fs.generic.log_exceptions(fs.generic.copy)
         runs = from_fs.path.join(from_odb.path, "runs")
 
         http_odb = next(

dvc/stage/imports.py

@@ -56,7 +56,10 @@ def sync_import(
         stage.outs[0].checkout()
     else:
         stage.save_deps()
-        if not no_download:
+        if no_download:
+            if stage.is_repo_import:
+                stage.deps[0].update()
+        else:
             stage.deps[0].download(
                 stage.outs[0],
                 jobs=jobs,

dvc/stage/loader.py

@@ -73,6 +73,7 @@ class StageLoader(Mapping):
             item.hash_info = HashInfo(item.hash_name, hash_value)
             if item.hash_info and item.hash_info.isdir:
                 item.meta.isdir = True
+            item.files = get_in(checksums, [key, path, item.PARAM_FILES])
 
     @classmethod
     def load_stage(cls, dvcfile, name, stage_data, lock_data=None):

dvc/stage/serialize.py

@@ -152,7 +152,7 @@ def to_pipeline_file(stage: "PipelineStage"):
     }
 
 
-def to_single_stage_lockfile(stage: "Stage") -> dict:
+def to_single_stage_lockfile(stage: "Stage", **kwargs) -> dict:
     assert stage.cmd
 
     def _dumpd(item):
@@ -164,6 +164,13 @@ def to_single_stage_lockfile(stage: "Stage") -> dict:
             *meta_d.items(),
         ]
 
+        if item.hash_info.isdir and kwargs.get("with_files"):
+            if item.obj:
+                obj = item.obj
+            else:
+                obj = item.get_obj()
+            ret.append((item.PARAM_FILES, obj.as_list(with_meta=True)))
+
         return OrderedDict(ret)
 
     res = OrderedDict([("cmd", stage.cmd)])
@@ -183,9 +190,9 @@ def to_single_stage_lockfile(stage: "Stage") -> dict:
     return res
 
 
-def to_lockfile(stage: "PipelineStage") -> dict:
+def to_lockfile(stage: "PipelineStage", **kwargs) -> dict:
     assert stage.name
-    return {stage.name: to_single_stage_lockfile(stage)}
+    return {stage.name: to_single_stage_lockfile(stage, **kwargs)}
 
 
 def to_single_stage_file(stage: "Stage", **kwargs):

dvc/testing/api_tests.py

@@ -38,7 +38,7 @@ class TestAPI:
     )
     def test_filesystem(
         self,
-        M,  # from `pytest-test-utils`
+        M,
         tmp_dir,
         make_tmp_dir,
         scm,
@@ -47,6 +47,8 @@ class TestAPI:
         fs_kwargs,
         clear_cache,
     ):
+        fs_kwargs = fs_kwargs.copy()
+
         tmp_dir.scm_gen({"scripts": {"script1": "script1"}}, commit="scripts")
         tmp_dir.dvc_gen({"data": {"foo": "foo", "bar": "bar"}}, commit="data")
         dvc.push()

dvc/testing/workspace_tests.py

@@ -2,6 +2,10 @@ import os
 
 import pytest
 
+from dvc.exceptions import URLMissingError
+from dvc.repo import Repo
+from dvc.repo.ls_url import ls_url, parse_external_url
+
 
 class TestImport:
     def test_import(self, tmp_dir, dvc, workspace):
@@ -134,3 +138,103 @@ class TestAdd:
         assert (
             workspace / "cache" / dir_hash_value[:2] / dir_hash_value[2:]
         ).is_file()
+
+
+def match_files(fs, entries, expected):
+    entries_content = {
+        (fs.path.normpath(d["path"]), d["isdir"]) for d in entries
+    }
+    expected_content = {
+        (fs.path.normpath(d["path"]), d["isdir"]) for d in expected
+    }
+    assert entries_content == expected_content
+
+
+class TestLsUrl:
+    @pytest.mark.parametrize("fname", ["foo", "foo.dvc", "dir/foo"])
+    def test_file(self, cloud, fname):
+        cloud.gen({fname: "foo contents"})
+        fs, fs_path = parse_external_url(cloud.url, cloud.config)
+        result = ls_url(str(cloud / fname), config=cloud.config)
+        match_files(
+            fs,
+            result,
+            [{"path": fs.path.join(fs_path, fname), "isdir": False}],
+        )
+
+    def test_dir(self, cloud):
+        cloud.gen(
+            {"dir/foo": "foo contents", "dir/subdir/bar": "bar contents"}
+        )
+        if not (cloud / "dir").is_dir():
+            pytest.skip("Cannot create directories on this cloud")
+        fs, _ = parse_external_url(cloud.url, cloud.config)
+        result = ls_url(str(cloud / "dir"), config=cloud.config)
+        match_files(
+            fs,
+            result,
+            [
+                {"path": "foo", "isdir": False},
+                {"path": "subdir", "isdir": True},
+            ],
+        )
+
+    def test_recursive(self, cloud):
+        cloud.gen(
+            {"dir/foo": "foo contents", "dir/subdir/bar": "bar contents"}
+        )
+        if not (cloud / "dir").is_dir():
+            pytest.skip("Cannot create directories on this cloud")
+        fs, _ = parse_external_url(cloud.url, cloud.config)
+        result = ls_url(
+            str(cloud / "dir"), config=cloud.config, recursive=True
+        )
+        match_files(
+            fs,
+            result,
+            [
+                {"path": "foo", "isdir": False},
+                {"path": "subdir/bar", "isdir": False},
+            ],
+        )
+
+    def test_nonexistent(self, cloud):
+        with pytest.raises(URLMissingError):
+            ls_url(str(cloud / "dir"), config=cloud.config)
+
+
+class TestGetUrl:
+    def test_get_file(self, cloud, tmp_dir):
+        cloud.gen({"foo": "foo contents"})
+
+        Repo.get_url(str(cloud / "foo"), "foo_imported", config=cloud.config)
+
+        assert (tmp_dir / "foo_imported").is_file()
+        assert (tmp_dir / "foo_imported").read_text() == "foo contents"
+
+    def test_get_dir(self, cloud, tmp_dir):
+        cloud.gen({"foo": {"foo": "foo contents"}})
+        if not (cloud / "foo").is_dir():
+            pytest.skip("Cannot create directories on this cloud")
+
+        Repo.get_url(str(cloud / "foo"), "foo_imported", config=cloud.config)
+
+        assert (tmp_dir / "foo_imported").is_dir()
+        assert (tmp_dir / "foo_imported" / "foo").is_file()
+        assert (tmp_dir / "foo_imported" / "foo").read_text() == "foo contents"
+
+    @pytest.mark.parametrize("dname", [".", "dir", "dir/subdir"])
+    def test_get_url_to_dir(self, cloud, tmp_dir, dname):
+        cloud.gen({"src": {"foo": "foo contents"}})
+        if not (cloud / "src").is_dir():
+            pytest.skip("Cannot create directories on this cloud")
+        tmp_dir.gen({"dir": {"subdir": {}}})
+
+        Repo.get_url(str(cloud / "src" / "foo"), dname, config=cloud.config)
+
+        assert (tmp_dir / dname).is_dir()
+        assert (tmp_dir / dname / "foo").read_text() == "foo contents"
+
+    def test_get_url_nonexistent(self, cloud):
+        with pytest.raises(URLMissingError):
+            Repo.get_url(str(cloud / "nonexistent"), config=cloud.config)

dvc/utils/hydra.py

@@ -1,4 +1,3 @@
-from collections import defaultdict
 from pathlib import Path
 from typing import TYPE_CHECKING, List
 
@@ -83,27 +82,21 @@ def to_hydra_overrides(path_overrides):
     return parser.parse_overrides(overrides=path_overrides)
 
 
+def dict_product(dicts):
+    import itertools
+
+    return [dict(zip(dicts, x)) for x in itertools.product(*dicts.values())]
+
+
 def get_hydra_sweeps(path_overrides):
-    merged_overrides = []
+    path_sweeps = {}
     for path, overrides in path_overrides.items():
-        merged_overrides.extend(
-            [f"{path}___{override}" for override in overrides]
-        )
-
-    hydra_overrides = to_hydra_overrides(merged_overrides)
-    for hydra_override in hydra_overrides:
-        if hydra_override.value_type == ValueType.GLOB_CHOICE_SWEEP:
-            raise InvalidArgumentError(
-                f"Glob override '{hydra_override.input_line}' "
-                "is not supported."
-            )
-
-    splits = BasicSweeper.split_arguments(hydra_overrides, None)[0]
-    sweeps = []
-    for split in splits:
-        sweep_overrides = defaultdict(list)
-        for merged_override in split:
-            path, override = merged_override.split("___")
-            sweep_overrides[path].append(override)
-        sweeps.append(dict(sweep_overrides))
-    return sweeps
+        overrides = to_hydra_overrides(overrides)
+        for override in overrides:
+            if override.value_type == ValueType.GLOB_CHOICE_SWEEP:
+                raise InvalidArgumentError(
+                    f"Glob override '{override.input_line}' "
+                    "is not supported."
+                )
+        path_sweeps[path] = BasicSweeper.split_arguments(overrides, None)[0]
+    return dict_product(path_sweeps)

setup.cfg

@@ -22,6 +22,7 @@ classifiers =
     Programming Language :: Python :: 3.8
     Programming Language :: Python :: 3.9
     Programming Language :: Python :: 3.10
+    Programming Language :: Python :: 3.11
 
 [options]
 setup_requires =
@@ -62,12 +63,12 @@ install_requires =
     rich>=10.13.0
     pyparsing>=2.4.7
     typing-extensions>=3.7.4
-    scmrepo==0.1.1
-    dvc-render==0.0.11
-    dvc-task==0.1.2
+    scmrepo==0.1.2
+    dvc-render==0.0.12
+    dvc-task==0.1.4
     dvclive>=0.10.0
-    dvc-data==0.10.0
-    dvc-http==2.19.1
+    dvc-data==0.22.0
+    dvc-http==2.27.2
     hydra-core>=1.1.0
     iterative-telemetry==0.0.5
 
@@ -94,11 +95,11 @@ hdfs =
     dvc-hdfs==2.19.0; python_version < '3.11'
 oss = dvc-oss==2.19.0
 s3 = dvc-s3==2.20.1
-ssh = dvc-ssh==2.19.0
+ssh = dvc-ssh==2.20.0
 # gssapi should not be included in all_remotes, because it doesn't have wheels
 # for Linux, so it will fail to compile if user doesn't have all the
 # requirements, including kerberos itself.
-ssh_gssapi = dvc-ssh[gssapi]==2.19.0
+ssh_gssapi = dvc-ssh[gssapi]==2.20.0
 webdav = dvc-webdav==2.19.0
 webhdfs = dvc-webhdfs==2.19.0
 # requests-kerberos requires krb5 & gssapi, which does not provide wheels Linux
@@ -108,12 +109,12 @@ testing =
     pytest-test-utils==0.0.8
 tests =
     %(terraform)s
-    dvc-ssh==2.19.0
+    dvc-ssh==2.20.0
     # Test requirements
     pytest==7.1.3
-    pytest-cov==3.0.0
+    pytest-cov==4.0.0
     pytest-xdist==2.5.0
-    pytest-mock==3.8.2
+    pytest-mock==3.10.0
     pytest-lazy-fixture==0.6.3
     %(testing)s
     # https://github.com/docker/docker-py/issues/2902
@@ -122,16 +123,14 @@ tests =
     pytest-timeout==2.1.0
     filelock==3.8.0
     # pylint requirements
-    pylint==2.15.2
+    pylint==2.15.4
     # we use this to suppress pytest-related false positives in our tests.
     pylint-pytest==1.1.2
-    # see https://github.com/PyCQA/pylint/issues/7488
-    astroid==2.12.9
     # we use this to suppress some messages in tests, eg: foo/bar naming,
     # and, protected method calls in our tests
     pylint-plugin-utils==0.7
     # type-checking
-    mypy==0.971
+    mypy==0.982
     types-requests>=2.27.15
     types-tabulate>=0.8.6
     types-toml>=0.10.4

tests/func/api/test_data.py

@@ -9,7 +9,9 @@ from dvc.exceptions import (
     OutputNotFoundError,
     PathMissingError,
 )
-from dvc.testing.test_api import TestAPI  # noqa, pylint: disable=unused-import
+from dvc.testing.api_tests import (  # noqa, pylint: disable=unused-import
+    TestAPI,
+)
 from dvc.testing.tmp_dir import make_subrepo
 from dvc.utils.fs import remove
 

tests/func/experiments/conftest.py

@@ -1,122 +1,12 @@
-from textwrap import dedent
-
 import pytest
 
-from tests.func.test_repro_multistage import COPY_SCRIPT
-from tests.unit.repo.experiments.conftest import (  # noqa pylint disable=unused-argument
+from tests.unit.repo.experiments.conftest import (  # noqa, pylint disable=unused-argument
+    checkpoint_stage,
+    exp_stage,
+    failed_exp_stage,
     test_queue,
 )
 
-DEFAULT_ITERATIONS = 2
-CHECKPOINT_SCRIPT_FORMAT = dedent(
-    """\
-    import os
-    import sys
-    import shutil
-
-    from dvc.api import make_checkpoint
-
-    checkpoint_file = {}
-    checkpoint_iterations = int({})
-    if os.path.exists(checkpoint_file):
-        with open(checkpoint_file) as fobj:
-            try:
-                value = int(fobj.read())
-            except ValueError:
-                value = 0
-    else:
-        with open(checkpoint_file, "w"):
-            pass
-        value = 0
-
-    shutil.copyfile({}, {})
-
-    if os.getenv("DVC_CHECKPOINT"):
-        for _ in range(checkpoint_iterations):
-            value += 1
-            with open(checkpoint_file, "w") as fobj:
-                fobj.write(str(value))
-            make_checkpoint()
-"""
-)
-CHECKPOINT_SCRIPT = CHECKPOINT_SCRIPT_FORMAT.format(
-    "sys.argv[1]", "sys.argv[2]", "sys.argv[3]", "sys.argv[4]"
-)
-
-
-@pytest.fixture
-def exp_stage(tmp_dir, scm, dvc):
-    tmp_dir.gen("copy.py", COPY_SCRIPT)
-    tmp_dir.gen("params.yaml", "foo: 1")
-    stage = dvc.run(
-        cmd="python copy.py params.yaml metrics.yaml",
-        metrics_no_cache=["metrics.yaml"],
-        params=["foo"],
-        name="copy-file",
-        deps=["copy.py"],
-    )
-    scm.add(
-        [
-            "dvc.yaml",
-            "dvc.lock",
-            "copy.py",
-            "params.yaml",
-            "metrics.yaml",
-            ".gitignore",
-        ]
-    )
-    scm.commit("init")
-    return stage
-
-
-@pytest.fixture
-def checkpoint_stage(tmp_dir, scm, dvc, mocker):
-    mocker.patch("dvc.stage.run.Monitor.AWAIT", 0.01)
-
-    tmp_dir.gen("checkpoint.py", CHECKPOINT_SCRIPT)
-    tmp_dir.gen("params.yaml", "foo: 1")
-    stage = dvc.run(
-        cmd=(
-            f"python checkpoint.py foo {DEFAULT_ITERATIONS} "
-            "params.yaml metrics.yaml"
-        ),
-        metrics_no_cache=["metrics.yaml"],
-        params=["foo"],
-        checkpoints=["foo"],
-        deps=["checkpoint.py"],
-        no_exec=True,
-        name="checkpoint-file",
-    )
-    scm.add(["dvc.yaml", "checkpoint.py", "params.yaml", ".gitignore"])
-    scm.commit("init")
-    stage.iterations = DEFAULT_ITERATIONS
-    return stage
-
-
-@pytest.fixture
-def failed_exp_stage(tmp_dir, scm, dvc):
-    tmp_dir.gen("copy.py", COPY_SCRIPT)
-    tmp_dir.gen("params.yaml", "foo: 1")
-    stage = dvc.stage.add(
-        cmd="python -c 'import sys; sys.exit(1)'",
-        metrics_no_cache=["failed-metrics.yaml"],
-        params=["foo"],
-        name="failed-copy-file",
-        deps=["copy.py"],
-    )
-    scm.add(
-        [
-            "dvc.yaml",
-            "dvc.lock",
-            "copy.py",
-            "params.yaml",
-            "failed-metrics.yaml",
-            ".gitignore",
-        ]
-    )
-    scm.commit("init")
-    return stage
-
 
 @pytest.fixture
 def http_auth_patch(mocker):

tests/func/experiments/executor/test_ssh.py

@@ -29,7 +29,6 @@ def test_init_from_stash(tmp_dir, scm, dvc, machine_instance, mocker):
     mock_entry.name = ""
     SSHExecutor.from_stash_entry(
         dvc,
-        "abc123",
         mock_entry,
         machine_name="foo",
     )
@@ -39,14 +38,14 @@ def test_init_from_stash(tmp_dir, scm, dvc, machine_instance, mocker):
 
 @pytest.mark.needs_internet
 @pytest.mark.parametrize("cloud", [pytest.lazy_fixture("git_ssh")])
-def test_init_git(tmp_dir, scm, cloud):
+def test_init_git(tmp_dir, scm, cloud, mocker):
     tmp_dir.scm_gen({"foo": "foo", "dir": {"bar": "bar"}}, commit="init")
     baseline_rev = scm.get_rev()
-    scm.set_ref(EXEC_HEAD, baseline_rev)
     tmp_dir.gen("foo", "stashed")
     scm.gitpython.git.stash()
     rev = scm.resolve_rev("stash@{0}")
-    scm.set_ref(EXEC_MERGE, rev)
+
+    mock = mocker.Mock(baseline_rev=baseline_rev, head_rev=baseline_rev)
 
     root_url = cloud / SSHExecutor.gen_dirname()
 
@@ -59,7 +58,8 @@ def test_init_git(tmp_dir, scm, cloud):
         username=TEST_SSH_USER,
         fs_factory=partial(_ssh_factory, cloud),
     )
-    executor.init_git(scm)
+    infofile = str((root_url / "foo.run").path)
+    executor.init_git(scm, rev, mock, infofile=infofile)
     assert root_url.path == executor._repo_abspath
 
     fs = cloud._ssh
@@ -117,7 +117,6 @@ def test_reproduce(tmp_dir, scm, dvc, cloud, exp_stage, mocker):
     SSHExecutor.reproduce(
         info,
         rev,
-        infofile=infofile,
         fs_factory=partial(_ssh_factory, cloud),
     )
     assert mock_execute.called_once()

tests/func/experiments/test_experiments.py

@@ -38,7 +38,7 @@ def test_new_simple(tmp_dir, scm, dvc, exp_stage, mocker, name, workspace):
         assert (tmp_dir / "metrics.yaml").read_text().strip() == "foo: 2"
 
     exp_name = name if name else ref_info.name
-    assert dvc.experiments.get_exact_name(exp) == exp_name
+    assert dvc.experiments.get_exact_name([exp])[exp] == exp_name
     assert resolve_rev(scm, exp_name) == exp
 
 
@@ -481,7 +481,7 @@ def test_subdir(tmp_dir, scm, dvc, workspace):
     with fs.open("dir/metrics.yaml", mode="r", encoding="utf-8") as fobj:
         assert fobj.read().strip() == "foo: 2"
 
-    assert dvc.experiments.get_exact_name(exp) == ref_info.name
+    assert dvc.experiments.get_exact_name([exp])[exp] == ref_info.name
     assert resolve_rev(scm, ref_info.name) == exp
 
 
@@ -525,7 +525,7 @@ def test_subrepo(tmp_dir, scm, workspace):
     with fs.open("dir/repo/metrics.yaml", mode="r", encoding="utf-8") as fobj:
         assert fobj.read().strip() == "foo: 2"
 
-    assert subrepo.dvc.experiments.get_exact_name(exp) == ref_info.name
+    assert subrepo.dvc.experiments.get_exact_name([exp])[exp] == ref_info.name
     assert resolve_rev(scm, ref_info.name) == exp
 
 

tests/func/experiments/test_show.py

@@ -12,14 +12,25 @@ from dvc.repo.experiments.executor.base import (
     EXEC_TMP_DIR,
     BaseExecutor,
     ExecutorInfo,
+    TaskStatus,
 )
 from dvc.repo.experiments.queue.base import QueueEntry
 from dvc.repo.experiments.refs import CELERY_STASH, ExpRefInfo
 from dvc.repo.experiments.utils import exp_refs_by_rev
+from dvc.utils import relpath
 from dvc.utils.fs import makedirs
 from dvc.utils.serialize import YAMLFileCorruptedError
 from tests.func.test_repro_multistage import COPY_SCRIPT
 
+LOCK_CONTENTS = {
+    "read": {
+        "data/MNIST": [{"pid": 54062, "cmd": "dvc exp run"}],
+    },
+    "write": {
+        "data/MNIST": {"pid": 54062, "cmd": "dvc exp run"},
+    },
+}
+
 
 def make_executor_info(**kwargs):
     # set default values for required info fields
@@ -47,6 +58,21 @@ def test_show_no_commits(tmp_dir):
     assert Repo.init().experiments.show() == {}
 
 
+@pytest.mark.vscode
+def test_show_branch_and_tag_name(tmp_dir, scm, dvc, exp_stage):
+    with tmp_dir.branch("new/branch", new=True):
+        tmp_dir.scm_gen("branch", "branch", "commit")
+        branch_rev = scm.get_rev()
+
+    result = dvc.experiments.show(all_branches=True)
+    assert result[branch_rev]["baseline"]["data"]["name"] == "new/branch"
+
+    scm.tag("new/tag")
+    tag_rev = scm.get_rev()
+    result = dvc.experiments.show(all_tags=True)
+    assert result[tag_rev]["baseline"]["data"]["name"] == "new/tag"
+
+
 @pytest.mark.vscode
 def test_show_simple(tmp_dir, scm, dvc, exp_stage):
     assert dvc.experiments.show()["workspace"] == {
@@ -62,8 +88,7 @@ def test_show_simple(tmp_dir, scm, dvc, exp_stage):
                 "metrics": {"metrics.yaml": {"data": {"foo": 1}}},
                 "outs": {},
                 "params": {"params.yaml": {"data": {"foo": 1}}},
-                "queued": False,
-                "running": False,
+                "status": "Success",
                 "executor": None,
                 "timestamp": None,
             }
@@ -96,8 +121,7 @@ def test_show_experiment(tmp_dir, scm, dvc, exp_stage, workspace):
             "metrics": {"metrics.yaml": {"data": {"foo": 1}}},
             "outs": {},
             "params": {"params.yaml": {"data": {"foo": 1}}},
-            "queued": False,
-            "running": False,
+            "status": "Success",
             "executor": None,
             "timestamp": timestamp,
             "name": "master",
@@ -129,7 +153,7 @@ def test_show_queued(tmp_dir, scm, dvc, exp_stage):
     assert len(results) == 2
     exp = results[exp_rev]["data"]
     assert exp["name"] == "test_name"
-    assert exp["queued"]
+    assert exp["status"] == "Queued"
     assert exp["params"]["params.yaml"] == {"data": {"foo": 2}}
 
     # test that only queued experiments for the current baseline are returned
@@ -144,10 +168,68 @@ def test_show_queued(tmp_dir, scm, dvc, exp_stage):
     results = dvc.experiments.show()[new_rev]
     assert len(results) == 2
     exp = results[exp_rev]["data"]
-    assert exp["queued"]
+    assert exp["status"] == "Queued"
     assert exp["params"]["params.yaml"] == {"data": {"foo": 3}}
 
 
+@pytest.mark.vscode
+@pytest.mark.xfail(strict=False, reason="pytest-celery flaky")
+def test_show_failed_experiment(tmp_dir, scm, dvc, failed_exp_stage):
+    baseline_rev = scm.get_rev()
+    timestamp = datetime.fromtimestamp(
+        scm.gitpython.repo.rev_parse(baseline_rev).committed_date
+    )
+
+    dvc.experiments.run(
+        failed_exp_stage.addressing, params=["foo=2"], queue=True
+    )
+    exp_rev = dvc.experiments.scm.resolve_rev(f"{CELERY_STASH}@{{0}}")
+    dvc.experiments.run(run_all=True)
+    experiments = dvc.experiments.show()[baseline_rev]
+
+    expected_baseline = {
+        "data": {
+            "deps": {
+                "copy.py": {
+                    "hash": ANY,
+                    "size": ANY,
+                    "nfiles": None,
+                }
+            },
+            "metrics": {},
+            "outs": {},
+            "params": {"params.yaml": {"data": {"foo": 1}}},
+            "status": "Success",
+            "executor": None,
+            "timestamp": timestamp,
+            "name": "master",
+        }
+    }
+
+    expected_failed = {
+        "data": {
+            "timestamp": ANY,
+            "params": {"params.yaml": {"data": {"foo": 2}}},
+            "deps": {"copy.py": {"hash": None, "size": None, "nfiles": None}},
+            "outs": {},
+            "status": "Failed",
+            "executor": None,
+            "error": {
+                "msg": "Experiment run failed.",
+                "type": "",
+            },
+        }
+    }
+
+    assert len(experiments) == 2
+    for rev, exp in experiments.items():
+        if rev == "baseline":
+            assert exp == expected_baseline
+        else:
+            assert rev == exp_rev
+            assert exp == expected_failed
+
+
 @pytest.mark.vscode
 @pytest.mark.parametrize("workspace", [True, False])
 def test_show_checkpoint(
@@ -174,7 +256,7 @@ def test_show_checkpoint(
 
     for i, rev in enumerate(checkpoints):
         if i == 0:
-            name = dvc.experiments.get_exact_name(rev)
+            name = dvc.experiments.get_exact_name([rev])[rev]
             name = f"{rev[:7]} [{name}]"
             fs = "╓"
         elif i == len(checkpoints) - 1:
@@ -338,9 +420,24 @@ def test_show_sort(tmp_dir, scm, dvc, exp_stage, caplog):
 
 
 @pytest.mark.vscode
-def test_show_running_workspace(tmp_dir, scm, dvc, exp_stage, capsys):
+@pytest.mark.parametrize(
+    "status, pid_exists",
+    [
+        (TaskStatus.RUNNING, True),
+        (TaskStatus.RUNNING, False),
+        (TaskStatus.FAILED, False),
+    ],
+)
+def test_show_running_workspace(
+    tmp_dir, scm, dvc, exp_stage, capsys, caplog, status, pid_exists, mocker
+):
+    from dvc.rwlock import RWLOCK_FILE
+
     pid_dir = os.path.join(dvc.tmp_dir, EXEC_TMP_DIR, EXEC_PID_DIR)
-    info = make_executor_info(location=BaseExecutor.DEFAULT_LOCATION)
+    lock_file = relpath(os.path.join(dvc.tmp_dir, RWLOCK_FILE), str(tmp_dir))
+    info = make_executor_info(
+        location=BaseExecutor.DEFAULT_LOCATION, status=status
+    )
     pidfile = os.path.join(
         pid_dir,
         "workspace",
@@ -348,8 +445,11 @@ def test_show_running_workspace(tmp_dir, scm, dvc, exp_stage, capsys):
     )
     makedirs(os.path.dirname(pidfile), True)
     (tmp_dir / pidfile).dump_json(info.asdict())
+    (tmp_dir / lock_file).dump_json(LOCK_CONTENTS)
 
-    assert dvc.experiments.show()["workspace"] == {
+    mocker.patch("psutil.pid_exists", return_value=pid_exists)
+
+    assert dvc.experiments.show().get("workspace") == {
         "baseline": {
             "data": {
                 "deps": {
@@ -362,19 +462,33 @@ def test_show_running_workspace(tmp_dir, scm, dvc, exp_stage, capsys):
                 "metrics": {"metrics.yaml": {"data": {"foo": 1}}},
                 "params": {"params.yaml": {"data": {"foo": 1}}},
                 "outs": {},
-                "queued": False,
-                "running": True,
-                "executor": info.location,
+                "status": "Running"
+                if status == TaskStatus.RUNNING and pid_exists
+                else "Success",
+                "executor": info.location
+                if status == TaskStatus.RUNNING and pid_exists
+                else None,
                 "timestamp": None,
             }
         }
     }
-
     capsys.readouterr()
     assert main(["exp", "show", "--csv"]) == 0
     cap = capsys.readouterr()
-    assert "Running" in cap.out
-    assert info.location in cap.out
+    if status == TaskStatus.RUNNING:
+        if pid_exists:
+            assert "Running" in cap.out
+            assert info.location in cap.out
+        else:
+            cmd = LOCK_CONTENTS["read"]["data/MNIST"][0]["cmd"]
+            pid = LOCK_CONTENTS["read"]["data/MNIST"][0]["pid"]
+            assert (
+                f"Process '{cmd}' with (Pid {pid}), in RWLock-file "
+                f"'{lock_file}' had been killed."
+            ) in caplog.text
+            assert (
+                f"Delete corrupted RWLock-file '{lock_file}'"
+            ) in caplog.text
 
 
 def test_show_running_tempdir(tmp_dir, scm, dvc, exp_stage, mocker):
@@ -409,20 +523,20 @@ def test_show_running_tempdir(tmp_dir, scm, dvc, exp_stage, mocker):
     makedirs(os.path.dirname(pidfile), True)
     (tmp_dir / pidfile).dump_json(info.asdict())
     mock_fetch = mocker.patch.object(
-        dvc.experiments,
-        "_fetch_running_exp",
+        dvc.experiments.tempdir_queue,
+        "get_running_exps",
         return_value={exp_rev: info.asdict()},
     )
 
     results = dvc.experiments.show()
     mock_fetch.assert_has_calls(
-        [mocker.call(stash_rev, pidfile, True)],
+        [mocker.call(True)],
     )
     exp_data = get_in(results, [baseline_rev, exp_rev, "data"])
-    assert exp_data["running"]
+    assert exp_data["status"] == "Running"
     assert exp_data["executor"] == info.location
 
-    assert not results["workspace"]["baseline"]["data"]["running"]
+    assert results["workspace"]["baseline"]["data"]["status"] == "Success"
 
 
 def test_show_running_celery(tmp_dir, scm, dvc, exp_stage, mocker):
@@ -444,11 +558,10 @@ def test_show_running_celery(tmp_dir, scm, dvc, exp_stage, mocker):
 
     results = dvc.experiments.show()
     exp_data = get_in(results, [baseline_rev, exp_rev, "data"])
-    assert not exp_data["queued"]
-    assert exp_data["running"]
+    assert exp_data["status"] == "Running"
     assert exp_data["executor"] == info.location
 
-    assert not results["workspace"]["baseline"]["data"]["running"]
+    assert results["workspace"]["baseline"]["data"]["status"] == "Success"
 
 
 def test_show_running_checkpoint(tmp_dir, scm, dvc, checkpoint_stage, mocker):
@@ -470,11 +583,17 @@ def test_show_running_checkpoint(tmp_dir, scm, dvc, checkpoint_stage, mocker):
         "iter_active",
         return_value=entries,
     )
+    mocker.patch.object(
+        dvc.experiments.celery_queue,
+        "iter_failed",
+        return_value=[],
+    )
     pidfile = queue.get_infofile_path(entries[0].stash_rev)
     info = make_executor_info(
         git_url="foo.git",
         baseline_rev=baseline_rev,
         location=TempDirExecutor.DEFAULT_LOCATION,
+        status=TaskStatus.RUNNING,
     )
     makedirs(os.path.dirname(pidfile), True)
     (tmp_dir / pidfile).dump_json(info.asdict())
@@ -486,10 +605,10 @@ def test_show_running_checkpoint(tmp_dir, scm, dvc, checkpoint_stage, mocker):
     results = dvc.experiments.show()
 
     checkpoint_res = get_in(results, [baseline_rev, checkpoint_rev, "data"])
-    assert checkpoint_res["running"]
+    assert checkpoint_res["status"] == "Running"
     assert checkpoint_res["executor"] == info.location
 
-    assert not results["workspace"]["baseline"]["data"]["running"]
+    assert results["workspace"]["baseline"]["data"]["status"] == "Success"
 
 
 def test_show_with_broken_repo(tmp_dir, scm, dvc, exp_stage, caplog):

tests/func/test_add.py

@@ -34,7 +34,7 @@ from dvc.stage.exceptions import (
     StageExternalOutputsError,
     StagePathNotFoundError,
 )
-from dvc.testing.test_workspace import TestAdd
+from dvc.testing.workspace_tests import TestAdd
 from dvc.utils import LARGE_DIR_SIZE, relpath
 from dvc.utils.fs import path_isin
 from dvc.utils.serialize import YAMLFileCorruptedError, load_yaml

tests/func/test_data_cloud.py

@@ -9,7 +9,7 @@ import dvc_data
 from dvc.cli import main
 from dvc.external_repo import clean_repos
 from dvc.stage.exceptions import StageNotFound
-from dvc.testing.test_remote import (  # noqa, pylint: disable=unused-import
+from dvc.testing.remote_tests import (  # noqa, pylint: disable=unused-import
     TestRemote,
 )
 from dvc.utils.fs import remove

tests/func/test_dvcfile.py

@@ -15,6 +15,16 @@ from dvc.stage.exceptions import StageFileDoesNotExistError
 from dvc.stage.loader import StageNotFound
 from dvc.utils.strictyaml import YAMLValidationError
 
+STAGE_EXAMPLE = {
+    "stage1": {
+        "cmd": "cp foo bar",
+        "desc": "stage desc",
+        "meta": {"key1": "value1", "key2": "value2"},
+        "deps": ["foo"],
+        "outs": [{"bar": {"desc": "bar desc", "meta": {"key": "value"}}}],
+    }
+}
+
 
 def test_run_load_one_for_multistage(tmp_dir, dvc):
     tmp_dir.gen("foo", "foo")
@@ -386,19 +396,7 @@ def test_dvcfile_try_dumping_parametrized_stage(tmp_dir, dvc, data, name):
 
 
 def test_dvcfile_load_dump_stage_with_desc_meta(tmp_dir, dvc):
-    data = {
-        "stages": {
-            "stage1": {
-                "cmd": "cp foo bar",
-                "desc": "stage desc",
-                "meta": {"key1": "value1", "key2": "value2"},
-                "deps": ["foo"],
-                "outs": [
-                    {"bar": {"desc": "bar desc", "meta": {"key": "value"}}}
-                ],
-            }
-        }
-    }
+    data = {"stages": STAGE_EXAMPLE}
     (tmp_dir / "dvc.yaml").dump(data)
 
     stage = dvc.stage.load_one(name="stage1")
@@ -411,3 +409,34 @@ def test_dvcfile_load_dump_stage_with_desc_meta(tmp_dir, dvc):
     # sanity check
     stage.dump()
     assert (tmp_dir / "dvc.yaml").parse() == data
+
+
+@pytest.mark.parametrize(
+    "data",
+    (
+        {
+            "plots": {
+                "path/to/plot": {"x": "value", "y": "value"},
+                "path/to/another/plot": {"x": "value", "y": "value"},
+            },
+            "stages": STAGE_EXAMPLE,
+        },
+        {
+            "plots": [
+                {"path/to/plot": {"x": "value", "y": "value"}},
+                {"path/to/another/plot": {"x": "value", "y": "value"}},
+            ],
+            "stages": STAGE_EXAMPLE,
+        },
+    ),
+)
+def test_dvcfile_load_with_plots(tmp_dir, dvc, data):
+    (tmp_dir / "dvc.yaml").dump(data)
+    plots = list(dvc.plots.collect())
+    top_level_plots = plots[0]["workspace"]["definitions"]["data"]["dvc.yaml"][
+        "data"
+    ]
+    assert all(
+        name in top_level_plots
+        for name in ("path/to/plot", "path/to/another/plot")
+    )

tests/func/test_fs.py

@@ -1,7 +1,6 @@
 import os
 from operator import itemgetter
 
-from dvc.fs.callbacks import Callback
 from dvc.repo import Repo
 
 
@@ -42,57 +41,3 @@ def test_walk_dont_ignore_subrepos(tmp_dir, scm, dvc):
 
     assert set(get_dirs(next(dvc_fs.walk(path)))) == {".dvc", "subdir", ".git"}
     assert set(get_dirs(next(scm_fs.walk("/")))) == {".dvc", "subdir"}
-
-
-def test_callback_on_dvcfs(tmp_dir, dvc, scm, mocker):
-    tmp_dir.dvc_gen({"dir": {"bar": "bar"}}, commit="dvc")
-    tmp_dir.scm_gen({"dir": {"foo": "foo"}}, commit="git")
-
-    fs = dvc.dvcfs
-
-    callback = Callback()
-    fs.get(
-        "dir",
-        (tmp_dir / "dir2").fs_path,
-        callback=callback,
-    )
-
-    assert (tmp_dir / "dir2").read_text() == {"foo": "foo", "bar": "bar"}
-    assert callback.size == 2
-    assert callback.value == 2
-
-    callback = Callback()
-    branch = mocker.spy(callback, "branch")
-    fs.get(
-        os.path.join("dir", "foo"),
-        (tmp_dir / "foo").fs_path,
-        callback=callback,
-    )
-
-    size = os.path.getsize(tmp_dir / "dir" / "foo")
-    assert (tmp_dir / "foo").read_text() == "foo"
-    assert callback.size == 1
-    assert callback.value == 1
-
-    assert branch.call_count == 1
-    assert branch.spy_return.size == size
-    assert branch.spy_return.value == size
-
-    branch.reset_mock()
-
-    callback = Callback()
-    branch = mocker.spy(callback, "branch")
-    fs.get(
-        os.path.join("dir", "bar"),
-        (tmp_dir / "bar").fs_path,
-        callback=callback,
-    )
-
-    size = os.path.getsize(tmp_dir / "dir" / "bar")
-    assert (tmp_dir / "bar").read_text() == "bar"
-    assert callback.size == 1
-    assert callback.value == 1
-
-    assert branch.call_count == 1
-    assert branch.spy_return.size == size
-    assert branch.spy_return.value == size

tests/func/test_get_url.py

@@ -2,8 +2,9 @@ import os
 
 import pytest
 
-from dvc.dependency.base import DependencyDoesNotExistError
+from dvc.exceptions import URLMissingError
 from dvc.repo import Repo
+from dvc.testing.workspace_tests import TestGetUrl as _TestGetUrl
 
 
 def test_get_file(tmp_dir):
@@ -26,7 +27,7 @@ def test_get_dir(tmp_dir):
 
 
 @pytest.mark.parametrize("dname", [".", "dir", "dir/subdir"])
-def test_get_url_to_dir(tmp_dir, tmp_path_factory, dname):
+def test_get_url_to_dir(tmp_dir, dname):
     tmp_dir.gen({"src": {"foo": "foo contents"}, "dir": {"subdir": {}}})
 
     Repo.get_url(os.path.join("src", "foo"), dname)
@@ -36,5 +37,9 @@ def test_get_url_to_dir(tmp_dir, tmp_path_factory, dname):
 
 
 def test_get_url_nonexistent(tmp_dir):
-    with pytest.raises(DependencyDoesNotExistError):
+    with pytest.raises(URLMissingError):
         Repo.get_url("nonexistent")
+
+
+class TestGetUrl(_TestGetUrl):
+    pass

tests/func/test_ignore.py

@@ -18,10 +18,6 @@ def _to_pattern_info_list(str_list: List):
     return [PatternInfo(a, "") for a in str_list]
 
 
-def walk_files(dvc, *args):
-    yield from dvc.dvcignore.find(*args)
-
-
 @pytest.mark.parametrize("filename", ["ignored", "тест"])
 def test_ignore(tmp_dir, dvc, filename):
     tmp_dir.gen({"dir": {filename: filename, "other": "text2"}})
@@ -29,13 +25,68 @@ def test_ignore(tmp_dir, dvc, filename):
 
     dvc._reset()
 
-    result = walk_files(dvc, dvc.fs, tmp_dir)
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir)
     assert set(result) == {
         (tmp_dir / DvcIgnore.DVCIGNORE_FILE).fs_path,
         (tmp_dir / "dir" / "other").fs_path,
     }
 
 
+def test_walk(tmp_dir, dvc):
+    tmp_dir.gen(
+        {
+            "foo": "foo",
+            "bar": "bar",
+            "dir": {
+                "foo": "foo",
+                "bar": "bar",
+                "baz": "baz",
+                "subdir": {"foo": "foo", "qux": "qux"},
+            },
+        }
+    )
+    tmp_dir.gen(DvcIgnore.DVCIGNORE_FILE, "dir/bar\nfoo")
+
+    dvc._reset()
+
+    result = list(dvc.dvcignore.walk(dvc.fs, tmp_dir))
+    assert result[0][0] == str(tmp_dir)
+    assert result[0][1] == ["dir"]
+    assert set(result[0][2]) == {"bar", ".dvcignore"}
+    assert result[1][0] == str(tmp_dir / "dir")
+    assert result[1][1] == ["subdir"]
+    assert result[1][2] == ["baz"]
+    assert result[2][0] == str(tmp_dir / "dir" / "subdir")
+    assert result[2][1] == []
+    assert result[2][2] == ["qux"]
+
+    result = list(dvc.dvcignore.walk(dvc.fs, tmp_dir, detail=True))
+    assert result == [
+        (
+            str(tmp_dir),
+            {"dir": dvc.fs.info(tmp_dir / "dir")},
+            {
+                "bar": dvc.fs.info(tmp_dir / "bar"),
+                ".dvcignore": dvc.fs.info(tmp_dir / ".dvcignore"),
+            },
+        ),
+        (
+            str(tmp_dir / "dir"),
+            {
+                "subdir": dvc.fs.info(tmp_dir / "dir" / "subdir"),
+            },
+            {
+                "baz": dvc.fs.info(tmp_dir / "dir" / "baz"),
+            },
+        ),
+        (
+            str(tmp_dir / "dir" / "subdir"),
+            {},
+            {"qux": dvc.fs.info(tmp_dir / "dir" / "subdir" / "qux")},
+        ),
+    ]
+
+
 def test_rename_ignored_file(tmp_dir, dvc):
     tmp_dir.gen({"dir": {"ignored": "...", "other": "text"}})
 
@@ -133,7 +184,7 @@ def test_ignore_on_branch(tmp_dir, scm, dvc):
 
     dvc._reset()
 
-    result = walk_files(dvc, dvc.fs, tmp_dir)
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir)
     assert set(result) == {
         (tmp_dir / "foo").fs_path,
         (tmp_dir / "bar").fs_path,
@@ -155,7 +206,7 @@ def test_match_nested(tmp_dir, dvc):
         }
     )
     dvc._reset()
-    result = walk_files(dvc, dvc.fs, tmp_dir)
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir)
     assert set(result) == {
         (tmp_dir / DvcIgnore.DVCIGNORE_FILE).fs_path,
         (tmp_dir / "foo").fs_path,
@@ -167,7 +218,7 @@ def test_ignore_external(tmp_dir, scm, dvc, tmp_path_factory):
     ext_dir = TmpDir(os.fspath(tmp_path_factory.mktemp("external_dir")))
     ext_dir.gen({"y.backup": "y", "tmp": {"file": "ext tmp"}})
 
-    result = walk_files(dvc, dvc.fs, ext_dir)
+    result = dvc.dvcignore.find(dvc.fs, ext_dir)
     assert set(result) == {
         (ext_dir / "y.backup").fs_path,
         (ext_dir / "tmp" / "file").fs_path,
@@ -207,7 +258,7 @@ def test_ignore_blank_line(tmp_dir, dvc):
     tmp_dir.gen({"dir": {"ignored": "text", "other": "text2"}})
     tmp_dir.gen(DvcIgnore.DVCIGNORE_FILE, "foo\n\ndir/ignored")
     dvc._reset()
-    result = walk_files(dvc, dvc.fs, tmp_dir / "dir")
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir / "dir")
     assert set(result) == {(tmp_dir / "dir" / "other").fs_path}
 
 
@@ -242,7 +293,7 @@ def test_ignore_file_in_parent_path(
     tmp_dir.gen(data_struct)
     tmp_dir.gen(DvcIgnore.DVCIGNORE_FILE, "\n".join(pattern_list))
     dvc._reset()
-    result = walk_files(dvc, dvc.fs, tmp_dir / "dir")
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir / "dir")
     assert set(result) == {
         (tmp_dir / relpath).fs_path for relpath in result_set
     }
@@ -265,7 +316,7 @@ def test_ignore_sub_directory(tmp_dir, dvc):
     tmp_dir.gen({"dir": {DvcIgnore.DVCIGNORE_FILE: "doc/fortz"}})
 
     dvc._reset()
-    result = walk_files(dvc, dvc.fs, tmp_dir / "dir")
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir / "dir")
     assert set(result) == {
         (tmp_dir / "dir" / "a" / "doc" / "fortz" / "a").fs_path,
         (tmp_dir / "dir" / DvcIgnore.DVCIGNORE_FILE).fs_path,
@@ -277,7 +328,7 @@ def test_ignore_directory(tmp_dir, dvc):
     tmp_dir.gen({"dir": {"fortz": {}, "a": {"fortz": {}}}})
     tmp_dir.gen({"dir": {DvcIgnore.DVCIGNORE_FILE: "fortz"}})
     dvc._reset()
-    result = walk_files(dvc, dvc.fs, tmp_dir / "dir")
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir / "dir")
     assert set(result) == {
         (tmp_dir / "dir" / DvcIgnore.DVCIGNORE_FILE).fs_path
     }
@@ -288,7 +339,7 @@ def test_multi_ignore_file(tmp_dir, dvc, monkeypatch):
     tmp_dir.gen(DvcIgnore.DVCIGNORE_FILE, "dir/subdir/*_ignore")
     tmp_dir.gen({"dir": {DvcIgnore.DVCIGNORE_FILE: "!subdir/not_ignore"}})
     dvc._reset()
-    result = walk_files(dvc, dvc.fs, tmp_dir / "dir")
+    result = dvc.dvcignore.find(dvc.fs, tmp_dir / "dir")
     assert set(result) == {
         (tmp_dir / "dir" / "subdir" / "not_ignore").fs_path,
         (tmp_dir / "dir" / DvcIgnore.DVCIGNORE_FILE).fs_path,
@@ -386,7 +437,7 @@ def test_ignore_in_added_dir(tmp_dir, dvc):
     dvc._reset()
 
     ignored_path = tmp_dir / "dir" / "sub" / "ignored"
-    result = walk_files(dvc, dvc.fs, ignored_path)
+    result = dvc.dvcignore.find(dvc.fs, ignored_path)
     assert set(result) == set()
     assert ignored_path.exists()
 

tests/func/test_import.py

@@ -217,6 +217,68 @@ def test_pull_imported_stage(tmp_dir, dvc, erepo_dir):
     assert os.path.isfile(dst_cache)
 
 
+def test_import_no_download(tmp_dir, scm, dvc, erepo_dir):
+    with erepo_dir.chdir():
+        erepo_dir.dvc_gen("foo", "foo content", commit="create foo")
+
+    dvc.imp(os.fspath(erepo_dir), "foo", "foo_imported", no_download=True)
+
+    assert not os.path.exists("foo_imported")
+
+    dst_stage = Dvcfile(dvc, "foo_imported.dvc").stage
+
+    assert dst_stage.deps[0].def_repo == {
+        "url": os.fspath(erepo_dir),
+        "rev_lock": erepo_dir.scm.get_rev(),
+    }
+    assert scm.is_ignored("foo_imported")
+
+
+def test_pull_import_no_download(tmp_dir, scm, dvc, erepo_dir):
+    with erepo_dir.chdir():
+        erepo_dir.scm_gen(os.path.join("foo", "bar"), b"bar", commit="add bar")
+        erepo_dir.dvc_gen(
+            os.path.join("foo", "baz"), b"baz contents", commit="add baz"
+        )
+        size = (
+            len(b"bar")
+            + len(b"baz contents")
+            + len((erepo_dir / "foo" / ".gitignore").read_bytes())
+        )
+
+    dvc.imp(os.fspath(erepo_dir), "foo", "foo_imported", no_download=True)
+
+    dvc.pull(["foo_imported.dvc"])
+    assert os.path.exists("foo_imported")
+
+    stage = Dvcfile(dvc, "foo_imported.dvc").stage
+
+    assert (
+        stage.outs[0].hash_info.value == "bdb8641831d8fcb03939637e09011c21.dir"
+    )
+
+    assert stage.outs[0].meta.size == size
+    assert stage.outs[0].meta.nfiles == 3
+    assert stage.outs[0].meta.isdir
+
+
+def test_pull_import_no_download_rev_lock(
+    tmp_dir,
+    dvc,
+    erepo_dir,
+):
+    with erepo_dir.chdir():
+        erepo_dir.dvc_gen("foo", "foo content", commit="add")
+
+    dvc.imp(os.fspath(erepo_dir), "foo", "foo_imported", no_download=True)
+
+    with erepo_dir.chdir():
+        erepo_dir.dvc_gen("foo", "modified foo content", commit="modify foo")
+
+    dvc.pull(["foo_imported.dvc"])
+    assert (tmp_dir / "foo_imported").read_text() == "foo content"
+
+
 def test_cache_type_is_properly_overridden(tmp_dir, scm, dvc, erepo_dir):
     with erepo_dir.chdir():
         with erepo_dir.dvc.config.edit() as conf:

tests/func/test_import_url.py

@@ -8,9 +8,10 @@ import pytest
 from dvc.annotations import Annotation
 from dvc.cli import main
 from dvc.dependency.base import Dependency, DependencyDoesNotExistError
+from dvc.dvcfile import Dvcfile
 from dvc.exceptions import InvalidArgumentError
 from dvc.stage import Stage
-from dvc.testing.test_workspace import TestImport as _TestImport
+from dvc.testing.workspace_tests import TestImport as _TestImport
 from dvc.utils.fs import makedirs
 from tests.basic_env import TestDvc
 
@@ -282,7 +283,7 @@ def test_import_url_to_remote_status(tmp_dir, dvc, local_cloud, local_remote):
     assert len(status) == 0
 
 
-def test_import_url_no_download(tmp_dir, dvc, local_workspace):
+def test_import_url_no_download(tmp_dir, scm, dvc, local_workspace):
     local_workspace.gen("file", "file content")
     dst = tmp_dir / "file"
     stage = dvc.imp_url(
@@ -292,6 +293,7 @@ def test_import_url_no_download(tmp_dir, dvc, local_workspace):
     assert stage.deps[0].hash_info.value == "d10b4c3ff123b26dc068d43a8bef2d23"
 
     assert not dst.exists()
+    assert scm.is_ignored(dst)
 
     out = stage.outs[0]
     assert not out.hash_info
@@ -303,6 +305,21 @@ def test_import_url_no_download(tmp_dir, dvc, local_workspace):
     ]
 
 
+def test_partial_import_pull(tmp_dir, scm, dvc, local_workspace):
+    local_workspace.gen("file", "file content")
+    dst = tmp_dir / "file"
+    dvc.imp_url("remote://workspace/file", os.fspath(dst), no_download=True)
+
+    dvc.pull(["file.dvc"])
+
+    assert dst.exists()
+
+    stage = Dvcfile(dvc, "file.dvc").stage
+
+    assert stage.outs[0].hash_info.value == "d10b4c3ff123b26dc068d43a8bef2d23"
+    assert stage.outs[0].meta.size == 12
+
+
 def test_imp_url_with_annotations(M, tmp_dir, dvc, local_workspace):
     local_workspace.gen("foo", "foo")
     annot = {

tests/func/test_ls_url.py

@@ -0,0 +1,5 @@
+from dvc.testing.workspace_tests import TestLsUrl as _TestLsUrl
+
+
+class TestLsUrl(_TestLsUrl):
+    pass

tests/func/test_merge_driver.py

@@ -39,6 +39,22 @@ def _gen(tmp_dir, struct, name):
         (None, {"foo": "foo"}, {"bar": "bar"}, {"foo": "foo", "bar": "bar"}),
         (None, None, {"bar": "bar"}, {"bar": "bar"}),
         (None, {"foo": "foo"}, None, {"foo": "foo"}),
+        (
+            {"foo": "foo"},
+            {"foo": "bar"},
+            {"foo": "foo", "baz": "baz"},
+            {"foo": "bar", "baz": "baz"},
+        ),
+        ({"foo": "foo"}, {}, {"foo": "foo", "bar": "bar"}, {"bar": "bar"}),
+        (
+            {"common": "common", "subdir": {"foo": "foo", "bar": "bar"}},
+            {"common": "common", "subdir": {"foo": "foo", "bar": "baz"}},
+            {"common": "common", "subdir": {"bar": "bar", "bizz": "bizz"}},
+            {
+                "common": "common",
+                "subdir": {"bar": "baz", "bizz": "bizz"},
+            },
+        ),
     ],
 )
 def test_merge(tmp_dir, dvc, ancestor, our, their, merged):
@@ -74,18 +90,15 @@ def test_merge(tmp_dir, dvc, ancestor, our, their, merged):
             {"foo": "foo"},
             {"foo": "bar"},
             {"foo": "baz"},
-            (
-                "unable to auto-merge directories with "
-                "diff that contains 'change'ed files"
-            ),
+            ("unable to auto-merge the following paths:\nfoo"),
         ),
         (
             {"common": "common", "foo": "foo"},
             {"common": "common", "bar": "bar"},
             {"baz": "baz"},
             (
-                "unable to auto-merge directories with "
-                "diff that contains 'remove'ed files"
+                "unable to auto-merge the following paths:\n"
+                "both deleted: ('foo',)"
             ),
         ),
     ],

tests/func/test_stage.py

@@ -265,6 +265,23 @@ def test_collect_repo_ignored_dir_unignored_pattern(tmp_dir, dvc, scm):
     assert dvc.stage.collect_repo() == [stage]
 
 
+@pytest.mark.parametrize("with_deps", (False, True))
+def test_collect_symlink(tmp_dir, dvc, with_deps):
+    tmp_dir.gen({"data": {"foo": "foo contents"}})
+    foo_path = os.path.join("data", "foo")
+    dvc.add(foo_path)
+
+    data_link = tmp_dir / "data_link"
+    data_link.symlink_to("data")
+    stage = list(
+        dvc.stage.collect(
+            target=str(data_link / "foo.dvc"), with_deps=with_deps
+        )
+    )[0]
+
+    assert stage.addressing == f"{foo_path}.dvc"
+
+
 def test_stage_strings_representation(tmp_dir, dvc, run_copy):
     tmp_dir.dvc_gen("foo", "foo")
     stage1 = run_copy("foo", "bar", single_stage=True)

tests/func/utils/test_hydra.py

@@ -183,6 +183,22 @@ def test_compose_and_dump(tmp_dir, suffix, overrides, expected):
 @pytest.mark.parametrize(
     "overrides, expected",
     [
+        (
+            {"params.yaml": ["defaults/foo=1,2"]},
+            [
+                {"params.yaml": ["defaults/foo=1"]},
+                {"params.yaml": ["defaults/foo=2"]},
+            ],
+        ),
+        (
+            {"params.yaml": ["+foo=1,2", "~bar", "++foobar=5,6"]},
+            [
+                {"params.yaml": ["+foo=1", "~bar=null", "++foobar=5"]},
+                {"params.yaml": ["+foo=1", "~bar=null", "++foobar=6"]},
+                {"params.yaml": ["+foo=2", "~bar=null", "++foobar=5"]},
+                {"params.yaml": ["+foo=2", "~bar=null", "++foobar=6"]},
+            ],
+        ),
         (
             {"params.yaml": ["foo=1,2", "bar=3,4"]},
             [

tests/integration/plots/test_plots.py

@@ -11,6 +11,7 @@ from bs4 import BeautifulSoup
 from funcy import first
 
 from dvc.cli import main
+from dvc.render import REVISION_FIELD, VERSION_FIELD
 
 JSON_OUT = "vis_data"
 
@@ -60,17 +61,6 @@ def extract_vega_specs(html_path, plots_ids):
     return result
 
 
-def filter_fields(datapoints: List[Dict], fields: List[str]):
-
-    tmp = deepcopy(datapoints)
-    for datapoint in tmp:
-        keys = set(datapoint.keys())
-        for key in keys:
-            if key not in fields:
-                datapoint.pop(key)
-    return tmp
-
-
 def drop_fields(datapoints: List[Dict], fields: List[str]):
     tmp = deepcopy(datapoints)
     for datapoint in tmp:
@@ -219,10 +209,35 @@ def test_repo_with_plots(
     html_result = extract_vega_specs(
         html_path, ["linear.json", "confusion.json"]
     )
+
+    assert json_result["linear.json"][0]["content"]["data"][
+        "values"
+    ] == _update_datapoints(
+        linear_v1,
+        {
+            VERSION_FIELD: {
+                "revision": "workspace",
+                "filename": "linear.json",
+                "field": "y",
+            },
+        },
+    )
     assert html_result["linear.json"]["data"]["values"] == _update_datapoints(
         linear_v1,
         {
-            "rev": "workspace",
+            REVISION_FIELD: "workspace",
+        },
+    )
+    assert json_result["confusion.json"][0]["content"]["data"][
+        "values"
+    ] == _update_datapoints(
+        confusion_v1,
+        {
+            VERSION_FIELD: {
+                "revision": "workspace",
+                "filename": "confusion.json",
+                "field": "actual",
+            },
         },
     )
     assert html_result["confusion.json"]["data"][
@@ -230,10 +245,9 @@ def test_repo_with_plots(
     ] == _update_datapoints(
         confusion_v1,
         {
-            "rev": "workspace",
+            REVISION_FIELD: "workspace",
         },
     )
-
     verify_image(
         tmp_dir, "workspace", "image.png", image_v1, html_path, json_result
     )
@@ -280,17 +294,59 @@ def test_repo_with_plots(
             html_path,
             ["../linear.json", "../confusion.json"],
         )
+        assert json_result["../linear.json"][0]["content"]["data"][
+            "values"
+        ] == _update_datapoints(
+            linear_v2,
+            {
+                VERSION_FIELD: {
+                    "revision": "workspace",
+                    "filename": "../linear.json",
+                    "field": "y",
+                },
+            },
+        ) + _update_datapoints(
+            linear_v1,
+            {
+                VERSION_FIELD: {
+                    "revision": "HEAD",
+                    "filename": "../linear.json",
+                    "field": "y",
+                },
+            },
+        )
         assert html_result["../linear.json"]["data"][
             "values"
         ] == _update_datapoints(
             linear_v2,
             {
-                "rev": "workspace",
+                REVISION_FIELD: "workspace",
             },
         ) + _update_datapoints(
             linear_v1,
             {
-                "rev": "HEAD",
+                REVISION_FIELD: "HEAD",
+            },
+        )
+        assert json_result["../confusion.json"][0]["content"]["data"][
+            "values"
+        ] == _update_datapoints(
+            confusion_v2,
+            {
+                VERSION_FIELD: {
+                    "revision": "workspace",
+                    "filename": "../confusion.json",
+                    "field": "actual",
+                },
+            },
+        ) + _update_datapoints(
+            confusion_v1,
+            {
+                VERSION_FIELD: {
+                    "revision": "HEAD",
+                    "filename": "../confusion.json",
+                    "field": "actual",
+                },
             },
         )
         assert html_result["../confusion.json"]["data"][
@@ -298,12 +354,12 @@ def test_repo_with_plots(
         ] == _update_datapoints(
             confusion_v2,
             {
-                "rev": "workspace",
+                REVISION_FIELD: "workspace",
             },
         ) + _update_datapoints(
             confusion_v1,
             {
-                "rev": "HEAD",
+                REVISION_FIELD: "HEAD",
             },
         )
 
@@ -390,11 +446,9 @@ def test_repo_with_config_plots(tmp_dir, capsys, repo_with_config_plots):
     repo_state = repo_with_config_plots()
     plots = next(repo_state)
 
-    html_path, json_result, split_json_result = call(capsys)
+    html_path, _, __ = call(capsys)
 
     assert os.path.exists(html_path)
-    assert json_result == {}
-    assert split_json_result == {}
     html_result = extract_vega_specs(
         html_path,
         [
@@ -405,9 +459,15 @@ def test_repo_with_config_plots(tmp_dir, capsys, repo_with_config_plots):
     assert html_result["dvc.yaml::linear_train_vs_test"]["data"][
         "values"
     ] == _update_datapoints(
-        plots["data"]["linear_train.json"], {"rev": "linear_train.json"}
+        plots["data"]["linear_train.json"],
+        {
+            REVISION_FIELD: "linear_train.json",
+        },
     ) + _update_datapoints(
-        plots["data"]["linear_test.json"], {"rev": "linear_test.json"}
+        plots["data"]["linear_test.json"],
+        {
+            REVISION_FIELD: "linear_test.json",
+        },
     )
 
     # TODO check json results once vscode is able to handle flexible plots

tests/unit/command/test_debug.py

@@ -1,19 +0,0 @@
-from dvc.cli import main
-
-
-def test_viztracer(tmp_dir, dvc, mocker):
-    viztracer_profile = mocker.patch("dvc._debug.viztracer_profile")
-
-    assert main(["status", "--viztracer"]) == 0
-
-    args = viztracer_profile.call_args[1]
-    assert callable(args["path"])
-    assert args["max_stack_depth"] == -1
-
-    assert main(["status", "--viztracer", "--viztracer-depth", "5"]) == 0
-    args = viztracer_profile.call_args[1]
-    assert args["max_stack_depth"] == 5
-
-    assert main(["status", "--viztracer-depth", "2"]) == 0
-    args = viztracer_profile.call_args[1]
-    assert args["max_stack_depth"] == 2

tests/unit/command/test_experiments.py

@@ -95,6 +95,8 @@ def test_experiments_show(dvc, scm, mocker):
             "--all-tags",
             "--all-branches",
             "--all-commits",
+            "--hide-queued",
+            "--hide-failed",
             "--sha",
             "--param-deps",
             "-n",
@@ -115,6 +117,8 @@ def test_experiments_show(dvc, scm, mocker):
         all_tags=True,
         all_branches=True,
         all_commits=True,
+        hide_queued=True,
+        hide_failed=True,
         num=1,
         revs="foo",
         sha_only=True,
@@ -433,8 +437,7 @@ def test_show_experiments_csv(capsys):
                             }
                         }
                     },
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {
                         "scores.json": {
@@ -470,8 +473,7 @@ def test_show_experiments_csv(capsys):
                             }
                         }
                     },
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {
                         "scores.json": {
@@ -506,8 +508,7 @@ def test_show_experiments_csv(capsys):
                             }
                         }
                     },
-                    "queued": True,
-                    "running": True,
+                    "status": "Running",
                     "executor": None,
                     "metrics": {
                         "scores.json": {
@@ -561,8 +562,7 @@ def test_show_experiments_md(capsys):
                 "data": {
                     "timestamp": None,
                     "params": {"params.yaml": {"data": {"foo": 1}}},
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {
                         "scores.json": {"data": {"bar": 0.9544670443829399}}
@@ -599,8 +599,7 @@ def test_show_experiments_sort_by(capsys, sort_order):
                             }
                         }
                     },
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {},
                 }
@@ -617,8 +616,7 @@ def test_show_experiments_sort_by(capsys, sort_order):
                             }
                         }
                     },
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {},
                     "name": "master",
@@ -634,8 +632,7 @@ def test_show_experiments_sort_by(capsys, sort_order):
                             }
                         }
                     },
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {},
                     "name": "exp-89140",
@@ -651,8 +648,7 @@ def test_show_experiments_sort_by(capsys, sort_order):
                             }
                         }
                     },
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {},
                     "name": "exp-43537",
@@ -668,8 +664,7 @@ def test_show_experiments_sort_by(capsys, sort_order):
                             }
                         }
                     },
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {},
                     "name": "exp-4f89e",
@@ -915,8 +910,7 @@ def test_show_experiments_pcp(tmp_dir, mocker):
                 "data": {
                     "timestamp": None,
                     "params": {"params.yaml": {"data": {"foo": 1}}},
-                    "queued": False,
-                    "running": False,
+                    "status": "Success",
                     "executor": None,
                     "metrics": {
                         "scores.json": {"data": {"bar": 0.9544670443829399}}

tests/unit/command/test_imp_url.py

@@ -41,6 +41,7 @@ def test_import_url(mocker):
         labels=None,
         meta=None,
         jobs=4,
+        version_aware=False,
     )
 
 
@@ -98,6 +99,7 @@ def test_import_url_no_exec_download_flags(mocker, flag, expected):
         type=None,
         labels=None,
         meta=None,
+        version_aware=False,
         **expected
     )
 
@@ -135,6 +137,7 @@ def test_import_url_to_remote(mocker):
         labels=None,
         meta=None,
         jobs=None,
+        version_aware=False,
     )
 
 

tests/unit/command/test_ls_url.py

@@ -0,0 +1,24 @@
+from dvc.cli import parse_args
+from dvc.commands.ls_url import CmdListUrl
+
+
+def test_ls_url(mocker):
+    cli_args = parse_args(["ls-url", "src"])
+    assert cli_args.func == CmdListUrl
+    cmd = cli_args.func(cli_args)
+    m = mocker.patch("dvc.repo.Repo.ls_url", autospec=True)
+
+    assert cmd.run() == 0
+
+    m.assert_called_once_with("src", recursive=False)
+
+
+def test_recursive(mocker):
+    cli_args = parse_args(["ls-url", "-R", "src"])
+    assert cli_args.func == CmdListUrl
+    cmd = cli_args.func(cli_args)
+    m = mocker.patch("dvc.repo.Repo.ls_url", autospec=True)
+
+    assert cmd.run() == 0
+
+    m.assert_called_once_with("src", recursive=True)

tests/unit/command/test_plots.py

@@ -334,10 +334,6 @@ def test_plots_diff_json(dvc, mocker, capsys):
     )
     render_mock = mocker.patch("dvc_render.render_html")
 
-    show_json_mock = mocker.patch(
-        "dvc.commands.plots._filter_unhandled_renderers",
-        return_value=renderers,
-    )
     show_json_mock = mocker.patch("dvc.commands.plots._show_json")
 
     assert cmd.run() == 0

tests/unit/command/test_run.py

@@ -8,7 +8,7 @@ def test_run(mocker, dvc):
         [
             "run",
             "--name",
-            "nam",
+            "name",
             "--deps",
             "deps",
             "--outs",
@@ -86,7 +86,7 @@ def test_run(mocker, dvc):
         no_commit=True,
         always_changed=True,
         cmd="command",
-        name="nam",
+        name="name",
         single_stage=False,
         external=True,
         desc="description",

tests/unit/fs/test_fs.py

@@ -1,10 +1,12 @@
+from dataclasses import dataclass
+
 import pytest
 from dvc_http import HTTPFileSystem, HTTPSFileSystem
 from dvc_s3 import S3FileSystem
 from dvc_ssh import SSHFileSystem
 
 from dvc.config import RemoteNotFoundError
-from dvc.fs import LocalFileSystem, get_fs_cls, get_fs_config
+from dvc.fs import LocalFileSystem, get_cloud_fs, get_fs_cls, get_fs_config
 
 url_cls_pairs = [
     ("s3://bucket/path", S3FileSystem),
@@ -36,5 +38,39 @@ def test_get_fs_cls(url, cls):
 
 
 def test_get_fs_config():
+    result = get_fs_config({}, url="ssh://example.com:/dir/path")
+    assert result == {"url": "ssh://example.com:/dir/path"}
+
+
+def test_get_fs_config_error():
     with pytest.raises(RemoteNotFoundError):
-        get_fs_config(None, {"remote": {}}, name="myremote")
+        get_fs_config({"remote": {}}, name="myremote")
+
+
+def test_remote_url():
+    config = {
+        "remote": {
+            "base": {"url": "http://example.com"},
+            "r1": {"url": "remote://base/r1", "user": "user"},
+            "r2": {"url": "remote://r1/r2", "password": "123"},
+        }
+    }
+    result = get_fs_config(config, url="remote://r2/foo")
+    assert result == {
+        "password": "123",
+        "user": "user",
+        "url": "http://example.com/r1/r2/foo",
+    }
+
+
+@dataclass
+class FakeRepo:
+    config: dict
+
+
+def test_get_cloud_fs():
+    repo = FakeRepo({})
+    cls, config, path = get_cloud_fs(repo, url="ssh://example.com:/dir/path")
+    assert cls is SSHFileSystem
+    assert config == {"host": "example.com", "verify": False}
+    assert path == "/dir/path"

tests/unit/fs/test_ssh.py

@@ -1,184 +0,0 @@
-# pylint: disable=W0613
-
-import getpass
-import os
-from unittest.mock import mock_open, patch
-
-import pytest
-from dvc_ssh import DEFAULT_PORT, SSHFileSystem
-
-
-def test_get_kwargs_from_urls():
-    user = "test"
-    host = "123.45.67.89"
-    port = 1234
-    path = "/path/to/dir"
-
-    # URL ssh://[user@]host.xz[:port]/path
-    url = f"ssh://{user}@{host}:{port}{path}"
-
-    assert SSHFileSystem._get_kwargs_from_urls(url) == {
-        "username": user,
-        "host": host,
-        "port": port,
-    }
-
-    # SCP-like URL ssh://[user@]host.xz:/absolute/path
-    url = f"ssh://{user}@{host}:{path}"
-
-    assert SSHFileSystem._get_kwargs_from_urls(url) == {
-        "username": user,
-        "host": host,
-    }
-
-
-def test_init():
-    fs = SSHFileSystem(
-        user="test", host="12.34.56.78", port="1234", password="xxx"
-    )
-    assert fs.fs_args["username"] == "test"
-    assert fs.fs_args["host"] == "12.34.56.78"
-    assert fs.fs_args["port"] == "1234"
-    assert fs.fs_args["password"] == fs.fs_args["passphrase"] == "xxx"
-
-
-def test_ssh_ask_password(mocker):
-    mocker.patch("dvc_ssh.ask_password", return_value="fish")
-    fs = SSHFileSystem(user="test", host="2.2.2.2", ask_password=True)
-    assert fs.fs_args["password"] == fs.fs_args["passphrase"] == "fish"
-
-
-mock_ssh_config = """
-Host example.com
-   User ubuntu
-   HostName 1.2.3.4
-   Port 1234
-   IdentityFile ~/.ssh/not_default.key
-"""
-
-
-@pytest.mark.parametrize(
-    "config,expected_host",
-    [
-        ({"host": "example.com"}, "1.2.3.4"),
-        ({"host": "not_in_ssh_config.com"}, "not_in_ssh_config.com"),
-    ],
-)
-@patch(
-    "builtins.open",
-    new_callable=mock_open,
-    read_data=mock_ssh_config,
-)
-def test_ssh_host_override_from_config(mock_file, config, expected_host):
-    fs = SSHFileSystem(**config)
-    assert fs.fs_args["host"] == expected_host
-
-
-@pytest.mark.parametrize(
-    "config,expected_user",
-    [
-        ({"host": "example.com", "user": "test1"}, "test1"),
-        ({"host": "example.com"}, "ubuntu"),
-        ({"host": "not_in_ssh_config.com", "user": "test1"}, "test1"),
-        ({"host": "not_in_ssh_config.com"}, getpass.getuser()),
-    ],
-)
-@patch(
-    "builtins.open",
-    new_callable=mock_open,
-    read_data=mock_ssh_config,
-)
-def test_ssh_user(mock_file, config, expected_user):
-    fs = SSHFileSystem(**config)
-    assert fs.fs_args["username"] == expected_user
-
-
-@pytest.mark.parametrize(
-    "config,expected_port",
-    [
-        ({"host": "example.com"}, 1234),
-        ({"host": "example.com", "port": 4321}, 4321),
-        ({"host": "not_in_ssh_config.com"}, DEFAULT_PORT),
-        ({"host": "not_in_ssh_config.com", "port": 2222}, 2222),
-    ],
-)
-@patch(
-    "builtins.open",
-    new_callable=mock_open,
-    read_data=mock_ssh_config,
-)
-def test_ssh_port(mock_file, config, expected_port):
-    fs = SSHFileSystem(**config)
-    assert fs.fs_args["port"] == expected_port
-
-
-@pytest.mark.parametrize(
-    "config,expected_keyfile",
-    [
-        (
-            {"host": "example.com", "keyfile": "dvc_config.key"},
-            ["dvc_config.key"],
-        ),
-        (
-            {"host": "example.com"},
-            ["~/.ssh/not_default.key"],
-        ),
-        (
-            {
-                "host": "not_in_ssh_config.com",
-                "keyfile": "dvc_config.key",
-            },
-            ["dvc_config.key"],
-        ),
-        ({"host": "not_in_ssh_config.com"}, None),
-    ],
-)
-@patch(
-    "builtins.open",
-    new_callable=mock_open,
-    read_data=mock_ssh_config,
-)
-def test_ssh_keyfile(mock_file, config, expected_keyfile):
-    fs = SSHFileSystem(**config)
-    expected_keyfiles = (
-        [os.path.expanduser(path) for path in expected_keyfile]
-        if expected_keyfile
-        else expected_keyfile
-    )
-    assert fs.fs_args.get("client_keys") == expected_keyfiles
-
-
-mock_ssh_multi_key_config = """
-IdentityFile file_1
-
-Host example.com
-    IdentityFile file_2
-"""
-
-
-@patch(
-    "builtins.open",
-    new_callable=mock_open,
-    read_data=mock_ssh_multi_key_config,
-)
-def test_ssh_multi_identity_files(mock_file):
-    fs = SSHFileSystem(host="example.com")
-    assert fs.fs_args.get("client_keys") == ["file_1", "file_2"]
-
-
-@pytest.mark.parametrize(
-    "config,expected_gss_auth",
-    [
-        ({"host": "example.com", "gss_auth": True}, True),
-        ({"host": "example.com", "gss_auth": False}, False),
-        ({"host": "not_in_ssh_config.com"}, False),
-    ],
-)
-@patch(
-    "builtins.open",
-    new_callable=mock_open,
-    read_data=mock_ssh_config,
-)
-def test_ssh_gss_auth(mock_file, config, expected_gss_auth):
-    fs = SSHFileSystem(**config)
-    assert fs.fs_args["gss_auth"] == expected_gss_auth