dvc: bump to 2.5.2

* Update s3.py

pass verify flag when creating a session

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

.github/CODEOWNERS

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

.github/workflows/packages.yaml

@@ -7,7 +7,7 @@ on:
     types: [released, prereleased]
 jobs:
   deb:
-    runs-on: ubuntu-16.04
+    runs-on: ubuntu-18.04
     steps:
     - uses: actions/[email protected]
     - name: Set up Python 3.8
@@ -36,7 +36,7 @@ jobs:
         asset_name: dvc_${{ github.event.release.tag_name }}_amd64.deb
         asset_content_type: binary/octet-stream
   rpm:
-    runs-on: ubuntu-16.04
+    runs-on: ubuntu-18.04
     steps:
     - uses: actions/[email protected]
     - name: Set up Python 3.8
@@ -125,7 +125,7 @@ jobs:
         asset_name: dvc-${{ github.event.release.tag_name }}.exe
         asset_content_type: binary/octet-stream
   pip:
-    runs-on: ubuntu-16.04
+    runs-on: ubuntu-18.04
     steps:
     - uses: actions/[email protected]
     - name: Set up Python 3.7

.github/workflows/tests.yaml

@@ -105,7 +105,7 @@ jobs:
         --cov-report=xml --cov-report=term
         --all ${{ env.extra_test_args }}
     - name: upload coverage report
-      uses: codecov/[email protected].0
+      uses: codecov/[email protected].2
       with:
         file: ./coverage.xml
         fail_ci_if_error: false
@@ -117,7 +117,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - name: Slack Notification
-      uses: rtCamp/action-slack-notify@v2.1.3
+      uses: rtCamp/action-slack-notify@v2.2.0
       env:
         SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
         SLACK_COLOR: ${{ job.status }}

.pre-commit-config.yaml

@@ -6,12 +6,19 @@ repos:
       - id: black
         language_version: python3
     repo: https://github.com/ambv/black
-    rev: 21.5b1
+    rev: 21.6b0
+  - hooks:
+      - id: codespell
+        args: 
+          - --ignore-words-list
+          - ba,datas,fo,uptodate
+    repo: https://github.com/codespell-project/codespell
+    rev: v2.1.0
   - hooks:
       - id: isort
         language_version: python3
     repo: https://github.com/timothycrosley/isort
-    rev: 5.8.0
+    rev: 5.9.1
   - hooks:
       - id: flake8
         language_version: python3
@@ -26,8 +33,13 @@ repos:
       -  id: mypy
          language_version: python3
          files: ^dvc/
+         additional_dependencies:
+           - types-requests
+           - types-paramiko
+           - types-tabulate
+           - types-toml
     repo: https://github.com/pre-commit/mirrors-mypy
-    rev: 'v0.812'
+    rev: 'v0.910'
   - repo: local
     hooks:
     - id: pylint

README.rst

@@ -184,7 +184,7 @@ Comparison to related technologies
 
 #. *Data Engineering* tools such as `AirFlow <https://airflow.apache.org/>`_, `Luigi <https://github.com/spotify/luigi>`_ and
    others - in DVC data, model and ML pipelines represent a single ML project and focuses on data scientist' experience while
-   data engineering pipeline orchestrates multiple data projects and focuses on efficent execution. DVC project can be used
+   data engineering pipeline orchestrates multiple data projects and focuses on efficient execution. DVC project can be used
    from data pipeline as a single execution step. `AirFlow DVC <https://github.com/covid-genomics/airflow-dvc>`_ is an example
    of such integration.
 
@@ -204,6 +204,10 @@ Comparison to related technologies
 #. `DAGsHub <https://dagshub.com/>`_ - online service of Git+DVC repositories with pipeline and metrics visualization and
    DVC-specific cloud storage.
 
+#. `DVC Studio <https://studio.iterative.ai/>`_ - online service of DVC repositories visualisation from DVC team. Also,
+   integrated with `CML (CI/CD for ML) <https://cml.dev/>`_ for training models in clouds and Kubernetes.
+
+
 Contributing
 ============
 

dvc/checkout.py

@@ -134,7 +134,7 @@ def _cache_is_copy(cache, path_info):
         return True
 
     workspace_file = path_info.with_name("." + uuid())
-    test_cache_file = cache.fs.path_info / ".cache_type_test_file"
+    test_cache_file = cache.path_info / ".cache_type_test_file"
     if not cache.fs.exists(test_cache_file):
         cache.makedirs(test_cache_file.parent)
         with cache.fs.open(test_cache_file, "wb") as fobj:

dvc/command/config.py

@@ -3,6 +3,7 @@ import logging
 import os
 
 from dvc.command.base import CmdBaseNoRepo, append_doc_link
+from dvc.ui import ui
 
 logger = logging.getLogger(__name__)
 
@@ -71,7 +72,9 @@ class CmdConfig(CmdBaseNoRepo):
             prefix = self._config_file_prefix(
                 self.args.show_origin, self.config, level
             )
-            logger.info("\n".join(self._format_config(conf, prefix)))
+            configs = list(self._format_config(conf, prefix))
+            if configs:
+                ui.write("\n".join(configs))
 
         return 0
 
@@ -96,7 +99,7 @@ class CmdConfig(CmdBaseNoRepo):
                 prefix = self._config_file_prefix(
                     self.args.show_origin, self.config, level
                 )
-                logger.info("{}{}".format(prefix, conf[section][opt]))
+                ui.write(prefix, conf[section][opt], sep="")
                 break
 
         return 0

dvc/command/data_sync.py

@@ -9,10 +9,11 @@ logger = logging.getLogger(__name__)
 
 class CmdDataBase(CmdBase):
     def log_summary(self, stats):
+        from dvc.ui import ui
         from dvc.utils.humanize import get_summary
 
         default_msg = "Everything is up to date."
-        logger.info(get_summary(stats.items()) or default_msg)
+        ui.write(get_summary(stats.items()) or default_msg)
 
 
 class CmdDataPull(CmdDataBase):

dvc/command/destroy.py

@@ -1,7 +1,6 @@
 import argparse
 import logging
 
-from dvc import prompt
 from dvc.command.base import CmdBase, append_doc_link
 
 logger = logging.getLogger(__name__)
@@ -10,6 +9,7 @@ logger = logging.getLogger(__name__)
 class CmdDestroy(CmdBase):
     def run(self):
         from dvc.exceptions import DvcException
+        from dvc.ui import ui
 
         try:
             statement = (
@@ -19,7 +19,7 @@ class CmdDestroy(CmdBase):
                 "Are you sure you want to continue?"
             )
 
-            if not self.args.force and not prompt.confirm(statement):
+            if not self.args.force and not ui.confirm(statement):
                 raise DvcException(
                     "cannot destroy without a confirmation from the user."
                     " Use `-f` to force."

dvc/command/diff.py

@@ -3,10 +3,9 @@ import json
 import logging
 import os
 
-import colorama
-
 from dvc.command import completion
 from dvc.command.base import CmdBase, append_doc_link
+from dvc.ui import ui
 
 logger = logging.getLogger(__name__)
 
@@ -18,13 +17,12 @@ def _digest(checksum):
 
 
 def _show_md(diff, show_hash=False, hide_missing=False):
-    from dvc.ui import ui
-
     headers = ["Status", "Hash", "Path"] if show_hash else ["Status", "Path"]
     rows = []
     statuses = ["added", "deleted", "renamed", "modified"]
     if not hide_missing:
         statuses.append("not in cache")
+
     for status in statuses:
         entries = diff.get(status, [])
         if not entries:
@@ -44,7 +42,7 @@ def _show_md(diff, show_hash=False, hide_missing=False):
 
 class CmdDiff(CmdBase):
     @staticmethod
-    def _format(diff, hide_missing=False):
+    def _show_diff(diff, hide_missing=False):
         """
         Given a diff structure, generate a string of paths separated
         by new lines and grouped together by their state.
@@ -71,15 +69,14 @@ class CmdDiff(CmdBase):
         """
 
         colors = {
-            "added": colorama.Fore.GREEN,
-            "modified": colorama.Fore.YELLOW,
-            "deleted": colorama.Fore.RED,
-            "renamed": colorama.Fore.GREEN,
-            "not in cache": colorama.Fore.YELLOW,
+            "added": "green",
+            "modified": "yellow",
+            "deleted": "red",
+            "renamed": "green",
+            "not in cache": "yellow",
         }
 
         summary = {}
-        groups = []
 
         states = ["added", "deleted", "renamed", "modified"]
         if not hide_missing:
@@ -91,7 +88,8 @@ class CmdDiff(CmdBase):
             if not entries:
                 continue
 
-            content = []
+            header = state.capitalize()
+            ui.write(f"[{colors[state]}]{header}[/]:", styled=True)
 
             for entry in entries:
                 path = entry["path"]
@@ -99,7 +97,7 @@ class CmdDiff(CmdBase):
                     path = f"{path['old']} -> {path['new']}"
                 checksum = entry.get("hash")
                 summary[state] += 1 if not path.endswith(os.sep) else 0
-                content.append(
+                ui.write(
                     "{space}{checksum}{separator}{path}".format(
                         space="    ",
                         checksum=_digest(checksum) if checksum else "",
@@ -108,27 +106,17 @@ class CmdDiff(CmdBase):
                     )
                 )
 
-            groups.append(
-                "{color}{header}{nc}:\n{content}".format(
-                    color=colors[state],
-                    header=state.capitalize(),
-                    nc=colorama.Fore.RESET,
-                    content="\n".join(content),
-                )
-            )
+            ui.write()
 
         if not sum(summary.values()):
             return None
 
-        fmt = (
-            "files summary: {added} added, {deleted} deleted,"
-            " {renamed} renamed, {modified} modified"
+        states_summary = ", ".join(
+            f"{summary[state]} {state}"
+            for state in states
+            if summary[state] > 0
         )
-        if not hide_missing:
-            fmt += ", {not in cache} not in cache"
-        groups.append(fmt.format_map(summary))
-
-        return "\n\n".join(groups)
+        ui.write("files summary:", states_summary)
 
     def run(self):
         from dvc.exceptions import DvcException
@@ -155,13 +143,11 @@ class CmdDiff(CmdBase):
                 diff[key] = entries
 
             if self.args.show_json:
-                logger.info(json.dumps(diff))
+                ui.write(json.dumps(diff))
             elif self.args.show_md:
-                logger.info(_show_md(diff, show_hash, hide_missing))
+                _show_md(diff, show_hash, hide_missing)
             elif diff:
-                output = self._format(diff, hide_missing)
-                if output:
-                    logger.info(output)
+                self._show_diff(diff, hide_missing)
 
         except DvcException:
             logger.exception("failed to get diff")

dvc/command/experiments.py

@@ -14,6 +14,7 @@ from dvc.command.metrics import DEFAULT_PRECISION
 from dvc.command.repro import CmdRepro
 from dvc.command.repro import add_arguments as add_repro_arguments
 from dvc.exceptions import DvcException, InvalidArgumentError
+from dvc.ui import ui
 from dvc.utils.flatten import flatten
 
 if TYPE_CHECKING:
@@ -396,7 +397,7 @@ def show_experiments(
     styles.update(
         {
             header: {
-                "justify": "left" if typ == "metrics" else "params",
+                "justify": "right" if typ == "metrics" else "left",
                 "header_style": f"black on {header_bg_colors[typ]}",
                 "collapse": idx != 0,
                 "no_wrap": typ == "metrics",
@@ -450,7 +451,7 @@ class CmdExperimentsShow(CmdBase):
         if self.args.show_json:
             import json
 
-            logger.info(json.dumps(all_experiments, default=_format_json))
+            ui.write(json.dumps(all_experiments, default=_format_json))
         else:
             show_experiments(
                 all_experiments,
@@ -494,7 +495,7 @@ class CmdExperimentsDiff(CmdBase):
         if self.args.show_json:
             import json
 
-            logger.info(json.dumps(diff))
+            ui.write(json.dumps(diff))
         else:
             from dvc.compare import show_diff
 
@@ -502,8 +503,6 @@ class CmdExperimentsDiff(CmdBase):
             diffs = [("metrics", "Metric"), ("params", "Param")]
             for idx, (key, title) in enumerate(diffs):
                 if idx:
-                    from dvc.ui import ui
-
                     # we are printing tables even in `--quiet` mode
                     # so we should also be printing the "table" separator
                     ui.write(force=True)
@@ -537,7 +536,7 @@ class CmdExperimentsRun(CmdRepro):
                 )
 
         if self.args.reset:
-            logger.info("Any existing checkpoints will be reset and re-run.")
+            ui.write("Any existing checkpoints will be reset and re-run.")
 
         results = self.repo.experiments.run(
             name=self.args.name,
@@ -554,7 +553,7 @@ class CmdExperimentsRun(CmdRepro):
         if self.args.metrics and results:
             metrics = self.repo.metrics.show(revs=list(results))
             metrics.pop("workspace", None)
-            logger.info(show_metrics(metrics))
+            show_metrics(metrics)
 
         return 0
 
@@ -608,12 +607,13 @@ class CmdExperimentsGC(CmdRepro):
         )
 
         if removed:
-            logger.info(
-                f"Removed {removed} experiments. To remove unused cache files "
-                "use 'dvc gc'."
+            ui.write(
+                f"Removed {removed} experiments.",
+                "To remove unused cache files",
+                "use 'dvc gc'.",
             )
         else:
-            logger.info("No experiments to remove.")
+            ui.write("No experiments to remove.")
         return 0
 
 
@@ -662,15 +662,15 @@ class CmdExperimentsPush(CmdBase):
             run_cache=self.args.run_cache,
         )
 
-        logger.info(
-            "Pushed experiment '%s' to Git remote '%s'.",
-            self.args.experiment,
-            self.args.git_remote,
+        ui.write(
+            f"Pushed experiment '{self.args.experiment}'"
+            f"to Git remote '{self.args.git_remote}'."
         )
         if not self.args.push_cache:
-            logger.info(
-                "To push cached outputs for this experiment to DVC remote "
-                "storage, re-run this command without '--no-cache'."
+            ui.write(
+                "To push cached outputs",
+                "for this experiment to DVC remote storage,"
+                "re-run this command without '--no-cache'.",
             )
 
         return 0
@@ -689,15 +689,15 @@ class CmdExperimentsPull(CmdBase):
             run_cache=self.args.run_cache,
         )
 
-        logger.info(
-            "Pulled experiment '%s' from Git remote '%s'. ",
-            self.args.experiment,
-            self.args.git_remote,
+        ui.write(
+            f"Pulled experiment '{self.args.experiment}'",
+            f"from Git remote '{self.args.git_remote}'.",
         )
         if not self.args.pull_cache:
-            logger.info(
-                "To pull cached outputs for this experiment from DVC remote "
-                "storage, re-run this command without '--no-cache'."
+            ui.write(
+                "To pull cached outputs for this experiment"
+                "from DVC remote storage,"
+                "re-run this command without '--no-cache'."
             )
 
         return 0
@@ -937,9 +937,7 @@ def add_parser(subparsers, parent_parser):
     )
     experiments_diff_parser.set_defaults(func=CmdExperimentsDiff)
 
-    EXPERIMENTS_RUN_HELP = (
-        "Reproduce complete or partial experiment pipelines."
-    )
+    EXPERIMENTS_RUN_HELP = "Run or resume an experiment."
     experiments_run_parser = experiments_subparsers.add_parser(
         "run",
         parents=[parent_parser],

dvc/command/gc.py

@@ -2,8 +2,8 @@ import argparse
 import logging
 import os
 
-from dvc import prompt
 from dvc.command.base import CmdBase, append_doc_link
+from dvc.ui import ui
 
 logger = logging.getLogger(__name__)
 
@@ -44,7 +44,7 @@ class CmdGC(CmdBase):
         logger.warning(msg)
 
         msg = "Are you sure you want to proceed?"
-        if not self.args.force and not prompt.confirm(msg):
+        if not self.args.force and not ui.confirm(msg):
             return 1
 
         self.repo.gc(

dvc/command/get.py

@@ -12,12 +12,13 @@ logger = logging.getLogger(__name__)
 class CmdGet(CmdBaseNoRepo):
     def _show_url(self):
         from dvc.api import get_url
+        from dvc.ui import ui
 
         try:
             url = get_url(
                 self.args.path, repo=self.args.url, rev=self.args.rev
             )
-            logger.info(url)
+            ui.write(url, force=True)
         except DvcException:
             logger.exception("failed to show URL")
             return 1

dvc/command/init.py

@@ -12,8 +12,10 @@ logger = logging.getLogger(__name__)
 
 
 def _welcome_message():
+    from dvc.ui import ui
+
     if analytics.is_enabled():
-        logger.info(
+        ui.write(
             boxify(
                 "DVC has enabled anonymous aggregate usage analytics.\n"
                 "Read the analytics documentation (and how to opt-out) here:\n"
@@ -30,7 +32,7 @@ def _welcome_message():
         f"- Star us on GitHub: {fmt_link('https://github.com/iterative/dvc')}"
     ).format(yellow=colorama.Fore.YELLOW, nc=colorama.Fore.RESET)
 
-    logger.info(msg)
+    ui.write(msg)
 
 
 class CmdInit(CmdBaseNoRepo):

dvc/command/live.py

@@ -1,11 +1,9 @@
 import argparse
-import logging
 import os
 
 from dvc.command import completion
 from dvc.command.base import CmdBase, fix_subparsers
-
-logger = logging.getLogger(__name__)
+from dvc.ui import ui
 
 
 class CmdLive(CmdBase):
@@ -17,7 +15,7 @@ class CmdLive(CmdBase):
         html_path = self.args.target + ".html"
         self.repo.plots.write_html(html_path, plots, metrics)
 
-        logger.info(f"\nfile://{os.path.abspath(html_path)}")
+        ui.write("\nfile://", os.path.abspath(html_path), sep="")
 
         return 0
 

dvc/command/metrics.py

@@ -4,6 +4,7 @@ import logging
 from dvc.command import completion
 from dvc.command.base import CmdBase, append_doc_link, fix_subparsers
 from dvc.exceptions import BadMetricError, DvcException
+from dvc.ui import ui
 
 logger = logging.getLogger(__name__)
 
@@ -32,7 +33,7 @@ class CmdMetricsShow(CmdMetricsBase):
         if self.args.show_json:
             import json
 
-            logger.info(json.dumps(metrics))
+            ui.write(json.dumps(metrics))
         else:
             from dvc.compare import show_metrics
 
@@ -73,7 +74,7 @@ class CmdMetricsDiff(CmdMetricsBase):
         if self.args.show_json:
             import json
 
-            logger.info(json.dumps(diff))
+            ui.write(json.dumps(diff))
         else:
             from dvc.compare import show_diff
 

dvc/command/params.py

@@ -4,6 +4,7 @@ import logging
 from dvc.command import completion
 from dvc.command.base import CmdBase, append_doc_link, fix_subparsers
 from dvc.exceptions import DvcException
+from dvc.ui import ui
 
 logger = logging.getLogger(__name__)
 
@@ -27,7 +28,7 @@ class CmdParamsDiff(CmdBase):
         if self.args.show_json:
             import json
 
-            logger.info(json.dumps(diff))
+            ui.write(json.dumps(diff))
         else:
             from dvc.compare import show_diff
 

dvc/command/plots.py

@@ -4,6 +4,7 @@ import logging
 from dvc.command import completion
 from dvc.command.base import CmdBase, append_doc_link, fix_subparsers
 from dvc.exceptions import DvcException
+from dvc.ui import ui
 from dvc.utils import format_link
 
 logger = logging.getLogger(__name__)
@@ -38,7 +39,7 @@ class CmdPlots(CmdBase):
 
             if self.args.show_vega:
                 target = self.args.targets[0]
-                logger.info(plots[target])
+                ui.write(plots[target])
                 return 0
 
             rel: str = self.args.out or "plots.html"
@@ -53,13 +54,15 @@ class CmdPlots(CmdBase):
 
         assert path.is_absolute()  # as_uri throws ValueError if not absolute
         url = path.as_uri()
-        logger.info(url)
+        ui.write(url)
         if self.args.open:
             import webbrowser
 
             opened = webbrowser.open(rel)
             if not opened:
-                logger.error("Failed to open. Please try opening it manually.")
+                ui.error_write(
+                    "Failed to open. Please try opening it manually."
+                )
                 return 1
 
         return 0

dvc/command/remote.py

@@ -1,12 +1,10 @@
 import argparse
-import logging
 
 from dvc.command.base import append_doc_link, fix_subparsers
 from dvc.command.config import CmdConfig
+from dvc.ui import ui
 from dvc.utils import format_link
 
-logger = logging.getLogger(__name__)
-
 
 class CmdRemote(CmdConfig):
     def __init__(self, args):
@@ -27,7 +25,7 @@ class CmdRemoteAdd(CmdRemote):
         from dvc.config import ConfigError
 
         if self.args.default:
-            logger.info(f"Setting '{self.args.name}' as a default remote.")
+            ui.write(f"Setting '{self.args.name}' as a default remote.")
 
         with self.config.edit(self.args.level) as conf:
             if self.args.name in conf["remote"] and not self.args.force:
@@ -90,7 +88,7 @@ class CmdRemoteDefault(CmdRemote):
             try:
                 print(conf["core"]["remote"])
             except KeyError:
-                logger.info("No default remote set")
+                ui.write("No default remote set")
                 return 1
         else:
             with self.config.edit(self.args.level) as conf:
@@ -116,7 +114,7 @@ class CmdRemoteList(CmdRemote):
     def run(self):
         conf = self.config.read(self.args.level)
         for name, conf in conf["remote"].items():
-            logger.info("{}\t{}".format(name, conf["url"]))
+            ui.write(name, conf["url"], sep="\t")
         return 0
 
 

dvc/command/repro.py

@@ -1,20 +1,19 @@
 import argparse
-import logging
 
 from dvc.command import completion
 from dvc.command.base import CmdBase, append_doc_link
 from dvc.command.status import CmdDataStatus
 
-logger = logging.getLogger(__name__)
-
 
 class CmdRepro(CmdBase):
     def run(self):
+        from dvc.ui import ui
+
         stages = self.repo.reproduce(**self._repro_kwargs)
         if len(stages) == 0:
-            logger.info(CmdDataStatus.UP_TO_DATE_MSG)
+            ui.write(CmdDataStatus.UP_TO_DATE_MSG)
         else:
-            logger.info(
+            ui.write(
                 "Use `dvc push` to send your updates to " "remote storage."
             )
 

dvc/command/root.py

@@ -10,8 +10,9 @@ logger = logging.getLogger(__name__)
 class CmdRoot(CmdBaseNoRepo):
     def run(self):
         from dvc.repo import Repo
+        from dvc.ui import ui
 
-        logger.info(relpath(Repo.find_root()))
+        ui.write(relpath(Repo.find_root()))
         return 0
 
 

dvc/command/status.py

@@ -2,6 +2,7 @@ import logging
 
 from dvc.command.data_sync import CmdDataBase
 from dvc.exceptions import DvcException
+from dvc.ui import ui
 from dvc.utils import format_link
 
 logger = logging.getLogger(__name__)
@@ -26,7 +27,7 @@ class CmdDataStatus(CmdDataBase):
         ind = indent * self.STATUS_INDENT
 
         if isinstance(status, str):
-            logger.info(f"{ind}{status}")
+            ui.write(f"{ind}{status}")
             return
 
         if isinstance(status, list):
@@ -38,14 +39,13 @@ class CmdDataStatus(CmdDataBase):
 
         for key, value in status.items():
             if isinstance(value, str):
-                logger.info("{}{}{}".format(ind, self._normalize(value), key))
+                ui.write("{}{}{}".format(ind, self._normalize(value), key))
             elif value:
-                logger.info(f"{ind}{key}:")
+                ui.write(f"{ind}{key}:")
                 self._show(value, indent + 1)
 
     def run(self):
         from dvc.repo import lock_repo
-        from dvc.ui import ui
 
         indent = 1 if self.args.cloud else 0
 
@@ -72,7 +72,7 @@ class CmdDataStatus(CmdDataBase):
             if self.args.show_json:
                 import json
 
-                logger.info(json.dumps(st))
+                ui.write(json.dumps(st))
                 return 0
 
             if st:

dvc/command/version.py

@@ -2,6 +2,7 @@ import argparse
 import logging
 
 from dvc.command.base import CmdBaseNoRepo, append_doc_link
+from dvc.ui import ui
 
 logger = logging.getLogger(__name__)
 
@@ -11,7 +12,7 @@ class CmdVersion(CmdBaseNoRepo):
         from dvc.info import get_dvc_info
 
         dvc_info = get_dvc_info()
-        logger.info(dvc_info)
+        ui.write(dvc_info, force=True)
         return 0
 
 

dvc/config_schema.py

@@ -93,7 +93,7 @@ HTTP_COMMON = {
     "user": str,
     "password": str,
     "ask_password": Bool,
-    "ssl_verify": Bool,
+    "ssl_verify": Any(Bool, str),
     "method": str,
 }
 WEBDAV_COMMON = {
@@ -104,6 +104,7 @@ WEBDAV_COMMON = {
     "cert_path": str,
     "key_path": str,
     "timeout": Coerce(int),
+    "ssl_verify": Any(Bool, str),
 }
 
 SCHEMA = {
@@ -188,6 +189,7 @@ SCHEMA = {
                     "tenant_id": str,
                     "client_id": str,
                     "client_secret": str,
+                    "allow_anonymous_login": bool,
                     **REMOTE_COMMON,
                 },
                 "oss": {

dvc/dagascii.py

@@ -163,7 +163,7 @@ class AsciiCanvas:
         self.point(x0 + width, y0 + height, "+")
 
 
-def _build_sugiyama_layout(vertexes, edges):
+def _build_sugiyama_layout(vertices, edges):
     #
     # Just a reminder about naming conventions:
     # +------------X
@@ -174,17 +174,17 @@ def _build_sugiyama_layout(vertexes, edges):
     # Y
     #
 
-    vertexes = {v: Vertex(f" {v} ") for v in vertexes}
+    vertices = {v: Vertex(f" {v} ") for v in vertices}
     # NOTE: reverting edges to correctly orientate the graph
-    edges = [Edge(vertexes[e], vertexes[s]) for s, e in edges]
-    vertexes = vertexes.values()
-    graph = Graph(vertexes, edges)
+    edges = [Edge(vertices[e], vertices[s]) for s, e in edges]
+    vertices = vertices.values()
+    graph = Graph(vertices, edges)
 
-    for vertex in vertexes:
+    for vertex in vertices:
         vertex.view = VertexViewer(vertex.data)
 
     # NOTE: determine min box length to create the best layout
-    minw = min(v.view.w for v in vertexes)
+    minw = min(v.view.w for v in vertices)
 
     for edge in edges:
         edge.view = EdgeViewer()
@@ -204,11 +204,11 @@ def _build_sugiyama_layout(vertexes, edges):
     return sug
 
 
-def draw(vertexes, edges):
+def draw(vertices, edges):
     """Build a DAG and draw it in ASCII.
 
     Args:
-        vertexes (list): list of graph vertexes.
+        vertices (list): list of graph vertices.
         edges (list): list of graph edges.
     """
     # pylint: disable=too-many-locals
@@ -217,7 +217,7 @@ def draw(vertexes, edges):
     Xs = []  # pylint: disable=invalid-name
     Ys = []  # pylint: disable=invalid-name
 
-    sug = _build_sugiyama_layout(vertexes, edges)
+    sug = _build_sugiyama_layout(vertices, edges)
 
     for vertex in sug.g.sV:
         # NOTE: moving boxes w/2 to the left

dvc/data_cloud.py

@@ -1,6 +1,11 @@
 """Manages dvc remotes that user can use with push/pull/status commands."""
 
 import logging
+from typing import TYPE_CHECKING, Iterable, Optional
+
+if TYPE_CHECKING:
+    from dvc.objects.file import HashFile
+    from dvc.remote.base import Remote
 
 logger = logging.getLogger(__name__)
 
@@ -19,7 +24,11 @@ class DataCloud:
     def __init__(self, repo):
         self.repo = repo
 
-    def get_remote(self, name=None, command="<command>"):
+    def get_remote(
+        self,
+        name: Optional[str] = None,
+        command: str = "<command>",
+    ) -> "Remote":
         from dvc.config import NoRemoteError
 
         if not name:
@@ -48,76 +57,88 @@ class DataCloud:
 
         return get_remote(self.repo, name=name)
 
-    def push(self, cache, jobs=None, remote=None, show_checksums=False):
+    def push(
+        self,
+        objs: Iterable["HashFile"],
+        jobs: Optional[int] = None,
+        remote: Optional[str] = None,
+        show_checksums: bool = False,
+    ):
         """Push data items in a cloud-agnostic way.
 
         Args:
-            cache (NamedCache): named checksums to push to the cloud.
-            jobs (int): number of jobs that can be running simultaneously.
-            remote (dvc.remote.base.BaseRemote): optional remote to push to.
+            objs: objects to push to the cloud.
+            jobs: number of jobs that can be running simultaneously.
+            remote: optional remote to push to.
                 By default remote from core.remote config option is used.
-            show_checksums (bool): show checksums instead of file names in
+            show_checksums: show checksums instead of file names in
                 information messages.
         """
-        remote = self.get_remote(remote, "push")
+        remote_obj = self.get_remote(remote, "push")
 
-        return remote.push(
+        return remote_obj.push(
             self.repo.odb.local,
-            cache,
+            objs,
             jobs=jobs,
             show_checksums=show_checksums,
         )
 
-    def pull(self, cache, jobs=None, remote=None, show_checksums=False):
+    def pull(
+        self,
+        objs: Iterable["HashFile"],
+        jobs: Optional[int] = None,
+        remote: Optional[str] = None,
+        show_checksums: bool = False,
+    ):
         """Pull data items in a cloud-agnostic way.
 
         Args:
-            cache (NamedCache): named checksums to pull from the cloud.
-            jobs (int): number of jobs that can be running simultaneously.
-            remote (dvc.remote.base.BaseRemote): optional remote to pull from.
+            objs: objects to pull from the cloud.
+            jobs: number of jobs that can be running simultaneously.
+            remote: optional remote to pull from.
                 By default remote from core.remote config option is used.
-            show_checksums (bool): show checksums instead of file names in
+            show_checksums: show checksums instead of file names in
                 information messages.
         """
-        remote = self.get_remote(remote, "pull")
+        remote_obj = self.get_remote(remote, "pull")
 
-        return remote.pull(
+        return remote_obj.pull(
             self.repo.odb.local,
-            cache,
+            objs,
             jobs=jobs,
             show_checksums=show_checksums,
         )
 
     def status(
         self,
-        cache,
-        jobs=None,
-        remote=None,
-        show_checksums=False,
-        log_missing=True,
+        objs: Iterable["HashFile"],
+        jobs: Optional[int] = None,
+        remote: Optional[str] = None,
+        show_checksums: bool = False,
+        log_missing: bool = True,
     ):
         """Check status of data items in a cloud-agnostic way.
 
         Args:
-            cache (NamedCache): named checksums to check status for.
-            jobs (int): number of jobs that can be running simultaneously.
-            remote (dvc.remote.base.BaseRemote): optional remote to compare
+            objs: objects to check status for.
+            jobs: number of jobs that can be running simultaneously.
+            remote: optional remote to compare
                 cache to. By default remote from core.remote config option
                 is used.
-            show_checksums (bool): show checksums instead of file names in
+            show_checksums: show checksums instead of file names in
                 information messages.
-            log_missing (bool): log warning messages if file doesn't exist
+            log_missing: log warning messages if file doesn't exist
                 neither in cache, neither in cloud.
         """
-        remote = self.get_remote(remote, "status")
-        return remote.status(
+        remote_obj = self.get_remote(remote, "status")
+        return remote_obj.status(
             self.repo.odb.local,
-            cache,
+            objs,
             jobs=jobs,
             show_checksums=show_checksums,
             log_missing=log_missing,
         )
 
     def get_url_for(self, remote, checksum):
-        remote = self.get_remote(remote)
-        return str(remote.odb.hash_to_path_info(checksum))
+        remote_obj = self.get_remote(remote)
+        return str(remote_obj.odb.hash_to_path_info(checksum))

dvc/dependency/repo.py

@@ -1,9 +1,17 @@
+import os
+from collections import defaultdict
+from typing import TYPE_CHECKING, Dict, Optional, Set
+
 from voluptuous import Required
 
 from dvc.path_info import PathInfo
 
 from .base import Dependency
 
+if TYPE_CHECKING:
+    from dvc.objects.db.base import ObjectDB
+    from dvc.objects.file import HashFile
+
 
 class RepoDependency(Dependency):
     PARAM_REPO = "repo"
@@ -21,6 +29,7 @@ class RepoDependency(Dependency):
 
     def __init__(self, def_repo, stage, *args, **kwargs):
         self.def_repo = def_repo
+        self._staged_objs: Dict[str, "HashFile"] = {}
         super().__init__(stage, *args, **kwargs)
 
     def _parse_path(self, fs, path_info):
@@ -30,37 +39,12 @@ class RepoDependency(Dependency):
     def is_in_repo(self):
         return False
 
-    @property
-    def repo_pair(self):
-        d = self.def_repo
-        rev = d.get(self.PARAM_REV_LOCK) or d.get(self.PARAM_REV)
-        return d[self.PARAM_URL], rev
-
     def __str__(self):
         return "{} ({})".format(self.def_path, self.def_repo[self.PARAM_URL])
 
-    def _make_repo(self, *, locked=True, **kwargs):
-        from dvc.external_repo import external_repo
-
-        d = self.def_repo
-        rev = (d.get("rev_lock") if locked else None) or d.get("rev")
-        return external_repo(d["url"], rev=rev, **kwargs)
-
-    def _get_hash(self, locked=True):
-        from dvc.objects.stage import stage
-
-        with self._make_repo(locked=locked) as repo:
-            path_info = PathInfo(repo.root_dir) / self.def_path
-            return stage(
-                self.repo.odb.local,
-                path_info,
-                repo.repo_fs,
-                self.repo.odb.local.fs.PARAM_CHECKSUM,
-            ).hash_info
-
     def workspace_status(self):
-        current = self._get_hash(locked=True)
-        updated = self._get_hash(locked=False)
+        current = self.get_obj(locked=True).hash_info
+        updated = self.get_obj(locked=False).hash_info
 
         if current != updated:
             return {str(self): "update available"}
@@ -78,31 +62,21 @@ class RepoDependency(Dependency):
 
     def download(self, to, jobs=None):
         from dvc.checkout import checkout
-        from dvc.config import NoRemoteError
-        from dvc.exceptions import NoOutputOrStageError
         from dvc.objects import save
-        from dvc.objects.stage import stage
+        from dvc.objects.db.git import GitObjectDB
+        from dvc.repo.fetch import fetch_from_odb
 
-        odb = self.repo.odb.local
-
-        with self._make_repo(cache_dir=odb.cache_dir) as repo:
-            if self.def_repo.get(self.PARAM_REV_LOCK) is None:
-                self.def_repo[self.PARAM_REV_LOCK] = repo.get_rev()
-            path_info = PathInfo(repo.root_dir) / self.def_path
-            try:
-                repo.fetch([path_info.fspath], jobs=jobs, recursive=True)
-            except (NoOutputOrStageError, NoRemoteError):
-                pass
-            obj = stage(
-                odb, path_info, repo.repo_fs, odb.fs.PARAM_CHECKSUM, jobs=jobs
-            )
-            save(odb, obj, jobs=jobs)
+        for odb, objs in self.get_used_objs().items():
+            if not isinstance(odb, GitObjectDB):
+                fetch_from_odb(self.repo, odb, objs, jobs=jobs)
 
+        obj = self.get_obj()
+        save(self.repo.odb.local, obj, jobs=jobs)
         checkout(
             to.path_info,
             to.fs,
             obj,
-            odb,
+            self.repo.odb.local,
             dvcignore=None,
             state=self.repo.state,
         )
@@ -110,7 +84,6 @@ class RepoDependency(Dependency):
     def update(self, rev=None):
         if rev:
             self.def_repo[self.PARAM_REV] = rev
-
         with self._make_repo(locked=False) as repo:
             self.def_repo[self.PARAM_REV_LOCK] = repo.get_rev()
 
@@ -119,3 +92,100 @@ class RepoDependency(Dependency):
         # origin project url and rev_lock, and it makes RepoDependency
         # immutable, hence its impossible for checksum to change.
         return False
+
+    def get_used_objs(
+        self, **kwargs
+    ) -> Dict[Optional["ObjectDB"], Set["HashFile"]]:
+        from dvc.config import NoRemoteError
+        from dvc.exceptions import NoOutputOrStageError, PathMissingError
+        from dvc.objects.db.git import GitObjectDB
+        from dvc.objects.stage import stage
+
+        local_odb = self.repo.odb.local
+        locked = kwargs.pop("locked", True)
+        with self._make_repo(
+            locked=locked, cache_dir=local_odb.cache_dir
+        ) as repo:
+            used_objs = defaultdict(set)
+            rev = repo.get_rev()
+            if locked and self.def_repo.get(self.PARAM_REV_LOCK) is None:
+                self.def_repo[self.PARAM_REV_LOCK] = rev
+
+            path_info = PathInfo(repo.root_dir) / str(self.def_path)
+            try:
+                for odb, objs in repo.used_objs(
+                    [os.fspath(path_info)],
+                    force=True,
+                    jobs=kwargs.get("jobs"),
+                    recursive=True,
+                ).items():
+                    if odb is None:
+                        odb = repo.cloud.get_remote().odb
+                    self._check_circular_import(odb)
+                    used_objs[odb].update(objs)
+            except (NoRemoteError, NoOutputOrStageError):
+                pass
+
+            try:
+                staged_obj = stage(
+                    local_odb,
+                    path_info,
+                    repo.repo_fs,
+                    local_odb.fs.PARAM_CHECKSUM,
+                )
+            except FileNotFoundError as exc:
+                raise PathMissingError(
+                    self.def_path, self.def_repo[self.PARAM_URL]
+                ) from exc
+
+            self._staged_objs[rev] = staged_obj
+            git_odb = GitObjectDB(repo.repo_fs, repo.root_dir)
+            used_objs[git_odb].add(staged_obj)
+            return used_objs
+
+    def _check_circular_import(self, odb):
+        from dvc.exceptions import CircularImportError
+        from dvc.fs.repo import RepoFileSystem
+
+        if not odb or not isinstance(odb.fs, RepoFileSystem):
+            return
+
+        self_url = self.repo.url or self.repo.root_dir
+        if odb.fs.repo_url is not None and odb.fs.repo_url == self_url:
+            raise CircularImportError(self, odb.fs.repo_url, self_url)
+
+    def get_obj(self, filter_info=None, **kwargs):
+        from dvc.objects.stage import stage
+
+        odb = self.repo.odb.local
+        locked = kwargs.pop("locked", True)
+        with self._make_repo(locked=locked, cache_dir=odb.cache_dir) as repo:
+            rev = repo.get_rev()
+            if locked and self.def_repo.get(self.PARAM_REV_LOCK) is None:
+                self.def_repo[self.PARAM_REV_LOCK] = rev
+            obj = self._staged_objs.get(rev)
+            if obj is not None:
+                return obj
+
+            path_info = PathInfo(repo.root_dir) / str(self.def_path)
+            obj = stage(
+                odb,
+                path_info,
+                repo.repo_fs,
+                odb.fs.PARAM_CHECKSUM,
+            )
+            self._staged_objs[rev] = obj
+            return obj
+
+    def _make_repo(self, locked=True, **kwargs):
+        from dvc.external_repo import external_repo
+
+        d = self.def_repo
+        rev = self._get_rev(locked=locked)
+        return external_repo(d[self.PARAM_URL], rev=rev, **kwargs)
+
+    def _get_rev(self, locked=True):
+        d = self.def_repo
+        return (d.get(self.PARAM_REV_LOCK) if locked else None) or d.get(
+            self.PARAM_REV
+        )

dvc/env.py

@@ -6,3 +6,4 @@ DVCLIVE_PATH = "DVCLIVE_PATH"
 DVCLIVE_SUMMARY = "DVCLIVE_SUMMARY"
 DVCLIVE_HTML = "DVCLIVE_HTML"
 DVCLIVE_RESUME = "DVCLIVE_RESUME"
+DVC_IGNORE_ISATTY = "DVC_IGNORE_ISATTY"

dvc/exceptions.py

@@ -384,3 +384,11 @@ class CacheLinkError(DvcException):
         )
         super().__init__(msg)
         self.path_infos = path_infos
+
+
+class CircularImportError(DvcException):
+    def __init__(self, dep, a, b):
+        super().__init__(
+            f"'{dep}' contains invalid circular import. "
+            f"DVC repo '{a}' already imports from '{b}'."
+        )

dvc/fs/__init__.py

@@ -12,8 +12,7 @@ from .local import LocalFileSystem
 from .oss import OSSFileSystem
 from .s3 import S3FileSystem
 from .ssh import SSHFileSystem
-from .webdav import WebDAVFileSystem
-from .webdavs import WebDAVSFileSystem
+from .webdav import WebDAVFileSystem, WebDAVSFileSystem
 from .webhdfs import WebHDFSFileSystem
 
 FS_MAP = {
@@ -106,4 +105,11 @@ def get_cloud_fs(repo, **kwargs):
     if cls == GDriveFileSystem and repo:
         remote_conf["gdrive_credentials_tmp_dir"] = repo.tmp_dir
 
-    return cls, remote_conf
+    url = remote_conf.pop("url")
+    path_info = cls.PATH_CLS(
+        cls._strip_protocol(url)  # pylint:disable=protected-access
+    )
+
+    extras = cls._get_kwargs_from_urls(url)  # pylint:disable=protected-access
+    conf = {**extras, **remote_conf}  # remote config takes priority
+    return cls, conf, path_info

dvc/fs/azure.py

@@ -5,7 +5,8 @@ import sys
 import threading
 from contextlib import contextmanager
 
-from funcy import cached_property, wrap_prop
+from fsspec.utils import infer_storage_options
+from funcy import cached_property, memoize, wrap_prop
 
 from dvc.exceptions import DvcException
 from dvc.path_info import CloudURLInfo
@@ -24,7 +25,7 @@ _DEFAULT_CREDS_STEPS = (
 
 @contextmanager
 def _temp_event_loop():
-    """When trying to initalize azure filesystem instances
+    """When trying to initialize azure filesystem instances
     with DefaultCredentials, the authentication process requires
     to have an access to a separate event loop. The normal calls
     are run in a separate loop managed by the fsspec, but the
@@ -67,6 +68,19 @@ class AzureAuthError(DvcException):
     pass
 
 
+@memoize
+def _az_config():
+    # NOTE: ideally we would've used get_default_cli().config from
+    # azure.cli.core, but azure-cli-core has a lot of conflicts with other
+    # dependencies. So instead we are just use knack directly
+    from knack.config import CLIConfig
+
+    config_dir = os.getenv(
+        "AZURE_CONFIG_DIR", os.path.expanduser(os.path.join("~", ".azure"))
+    )
+    return CLIConfig(config_dir=config_dir, config_env_var_prefix="AZURE")
+
+
 # pylint:disable=abstract-method
 class AzureFileSystem(ObjectFSWrapper):
     scheme = Schemes.AZURE
@@ -79,17 +93,21 @@ class AzureFileSystem(ObjectFSWrapper):
         "azure-identity": "azure.identity",
     }
 
-    def __init__(self, **config):
-        super().__init__(**config)
+    @classmethod
+    def _strip_protocol(cls, path: str):
+        bucket = infer_storage_options(path).get("host")
+        if bucket:
+            return path
 
-        url = config.get("url")
-        self.path_info = self.PATH_CLS(url)
+        bucket = _az_config().get("storage", "container_name", None)
+        return f"azure://{bucket}"
 
-        if not self.path_info.bucket:
-            container = self._az_config.get("storage", "container_name", None)
-            url = f"azure://{container}"
-
-        self.path_info = self.PATH_CLS(url)
+    @staticmethod
+    def _get_kwargs_from_urls(urlpath):
+        ops = infer_storage_options(urlpath)
+        if "host" in ops:
+            return {"bucket": ops["host"]}
+        return {}
 
     def _prepare_credentials(self, **config):
         from azure.identity.aio import DefaultAzureCredential
@@ -100,16 +118,16 @@ class AzureFileSystem(ObjectFSWrapper):
         login_info = {}
         login_info["connection_string"] = config.get(
             "connection_string",
-            self._az_config.get("storage", "connection_string", None),
+            _az_config().get("storage", "connection_string", None),
         )
         login_info["account_name"] = config.get(
-            "account_name", self._az_config.get("storage", "account", None)
+            "account_name", _az_config().get("storage", "account", None)
         )
         login_info["account_key"] = config.get(
-            "account_key", self._az_config.get("storage", "key", None)
+            "account_key", _az_config().get("storage", "key", None)
         )
         login_info["sas_token"] = config.get(
-            "sas_token", self._az_config.get("storage", "sas_token", None)
+            "sas_token", _az_config().get("storage", "sas_token", None)
         )
         login_info["tenant_id"] = config.get("tenant_id")
         login_info["client_id"] = config.get("client_id")
@@ -129,7 +147,7 @@ class AzureFileSystem(ObjectFSWrapper):
         if (
             login_info["account_name"]
             and not any_secondary
-            and not config.get("allow_anonymous_login")
+            and not config.get("allow_anonymous_login", False)
         ):
             login_info["credential"] = DefaultAzureCredential(
                 exclude_interactive_browser_credential=False
@@ -157,18 +175,6 @@ class AzureFileSystem(ObjectFSWrapper):
         self.login_method = login_method
         return login_info
 
-    @cached_property
-    def _az_config(self):
-        # NOTE: ideally we would've used get_default_cli().config from
-        # azure.cli.core, but azure-cli-core has a lot of conflicts with other
-        # dependencies. So instead we are just use knack directly
-        from knack.config import CLIConfig
-
-        config_dir = os.getenv(
-            "AZURE_CONFIG_DIR", os.path.expanduser(os.path.join("~", ".azure"))
-        )
-        return CLIConfig(config_dir=config_dir, config_env_var_prefix="AZURE")
-
     @wrap_prop(threading.Lock())
     @cached_property
     def fs(self):

dvc/fs/base.py

@@ -33,6 +33,8 @@ class RemoteMissingDepsError(DvcException):
 
 
 class BaseFileSystem:
+    sep = "/"
+
     scheme = "base"
     REQUIRES: ClassVar[Dict[str, str]] = {}
     PATH_CLS = URLInfo  # type: Any
@@ -55,11 +57,17 @@ class BaseFileSystem:
     def __init__(self, **kwargs):
         self._check_requires(**kwargs)
 
-        self.path_info = None
-
         self.jobs = kwargs.get("jobs") or self._JOBS
         self.hash_jobs = kwargs.get("checksum_jobs") or self.HASH_JOBS
 
+    @classmethod
+    def _strip_protocol(cls, path: str):
+        return path
+
+    @staticmethod
+    def _get_kwargs_from_urls(urlpath):  # pylint:disable=unused-argument
+        return {}
+
     @classmethod
     def get_missing_deps(cls):
         import importlib
@@ -164,7 +172,7 @@ class BaseFileSystem:
     def ls(self, path_info, detail=False):
         raise RemoteActionNotImplemented("ls", self.scheme)
 
-    def find(self, path_info, detail=False):
+    def find(self, path_info, detail=False, prefix=None):
         raise RemoteActionNotImplemented("find", self.scheme)
 
     def is_empty(self, path_info):
@@ -229,14 +237,23 @@ class BaseFileSystem:
             no_progress_bar=no_progress_bar,
         )
 
-    def upload_fobj(self, fobj, to_info, no_progress_bar=False, **pbar_args):
+    def upload_fobj(
+        self, fobj, to_info, no_progress_bar=False, size=None, **pbar_args
+    ):
         if not hasattr(self, "_upload_fobj"):
             raise RemoteActionNotImplemented("upload_fobj", self.scheme)
 
         with Tqdm.wrapattr(
-            fobj, "read", disable=no_progress_bar, bytes=True, **pbar_args
+            fobj,
+            "read",
+            disable=no_progress_bar,
+            bytes=True,
+            total=size,
+            **pbar_args,
         ) as wrapped:
-            self._upload_fobj(wrapped, to_info)  # pylint: disable=no-member
+            self._upload_fobj(  # pylint: disable=no-member
+                wrapped, to_info, size=size
+            )
 
     def download(
         self,

dvc/fs/dvc.py

@@ -23,6 +23,8 @@ class DvcFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         repo: DVC repo.
     """
 
+    sep = os.sep
+
     scheme = "local"
     PARAM_CHECKSUM = "md5"
 
@@ -231,7 +233,7 @@ class DvcFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
 
     def info(self, path_info):
         meta = self.metadata(path_info)
-        ret = {"type": "dir" if meta.isdir else "file"}
+        ret = {"type": "directory" if meta.isdir else "file"}
         if meta.is_output and len(meta.outs) == 1 and meta.outs[0].hash_info:
             hash_info = meta.outs[0].hash_info
             ret["size"] = hash_info.size

dvc/fs/fsspec_wrapper.py

@@ -48,7 +48,7 @@ class FSSpecWrapper(BaseFileSystem):
             yield entry
 
     def _entry_hook(self, entry):
-        """Simple hook method to be overriden when wanted to process
+        """Simple hook method to be overridden when wanted to process
         entries within info() and ls(detail=True) calls"""
         return entry
 
@@ -96,7 +96,8 @@ class FSSpecWrapper(BaseFileSystem):
         files = self.fs.ls(path, detail=detail)
         yield from self._strip_buckets(files, detail=detail)
 
-    def find(self, path_info, detail=False):
+    # pylint: disable=unused-argument
+    def find(self, path_info, detail=False, prefix=None):
         path = self._with_bucket(path_info)
         files = self.fs.find(path, detail=detail)
         if detail:
@@ -105,7 +106,7 @@ class FSSpecWrapper(BaseFileSystem):
         yield from self._strip_buckets(files, detail=detail)
 
     def walk_files(self, path_info, **kwargs):
-        for file in self.find(path_info):
+        for file in self.find(path_info, **kwargs):
             yield path_info.replace(path=file)
 
     def remove(self, path_info):
@@ -117,20 +118,20 @@ class FSSpecWrapper(BaseFileSystem):
         info["name"] = self._strip_bucket(info["name"])
         return info
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, size=None):
         with self.open(to_info, "wb") as fdest:
             shutil.copyfileobj(fobj, fdest, length=fdest.blocksize)
 
     def _upload(
         self, from_file, to_info, name=None, no_progress_bar=False, **kwargs
     ):
-        total = os.path.getsize(from_file)
+        size = os.path.getsize(from_file)
         with open(from_file, "rb") as fobj:
             self.upload_fobj(
                 fobj,
                 self._with_bucket(to_info),
+                size=size,
                 desc=name,
-                total=total,
                 no_progress_bar=no_progress_bar,
             )
         self.fs.invalidate_cache(self._with_bucket(to_info.parent))
@@ -155,6 +156,8 @@ class FSSpecWrapper(BaseFileSystem):
 
 # pylint: disable=abstract-method
 class ObjectFSWrapper(FSSpecWrapper):
+    TRAVERSE_PREFIX_LEN = 3
+
     def _isdir(self, path_info):
         # Directory in object storages are interpreted differently
         # among different fsspec providers, so this logic is a temporary
@@ -169,9 +172,16 @@ class ObjectFSWrapper(FSSpecWrapper):
             and entry["name"].endswith("/")
         )
 
-    def find(self, path_info, detail=False):
-        path = self._with_bucket(path_info)
-        files = self.fs.find(path, detail=detail)
+    def find(self, path_info, detail=False, prefix=None):
+        if prefix:
+            path = self._with_bucket(path_info.parent)
+            files = self.fs.find(
+                path, detail=detail, prefix=path_info.parts[-1]
+            )
+        else:
+            path = self._with_bucket(path_info)
+            files = self.fs.find(path, detail=detail)
+
         if detail:
             files = files.values()
 

dvc/fs/gdrive.py

@@ -1,3 +1,4 @@
+import errno
 import io
 import logging
 import os
@@ -11,7 +12,7 @@ from urllib.parse import urlparse
 from funcy import cached_property, retry, wrap_prop, wrap_with
 from funcy.py3 import cat
 
-from dvc.exceptions import DvcException, FileMissingError
+from dvc.exceptions import DvcException
 from dvc.path_info import CloudURLInfo
 from dvc.progress import Tqdm
 from dvc.scheme import Schemes
@@ -142,6 +143,10 @@ class GDriveFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             )
         )
 
+    @staticmethod
+    def _get_kwargs_from_urls(urlpath):
+        return {"url": urlpath}
+
     def _validate_config(self):
         # Validate Service Account configuration
         if (
@@ -517,12 +522,14 @@ class GDriveFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             return min(item_ids)
 
         assert not create
-        raise FileMissingError(path_info, hint)
+        raise FileNotFoundError(
+            errno.ENOENT, os.strerror(errno.ENOENT), hint or path_info
+        )
 
     def exists(self, path_info) -> bool:
         try:
             self._get_item_id(path_info)
-        except FileMissingError:
+        except FileNotFoundError:
             return False
         else:
             return True
@@ -534,7 +541,7 @@ class GDriveFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         query = f"({query}) and trashed=false"
         return self._gdrive_list(query)
 
-    def find(self, path_info, detail=False):
+    def find(self, path_info, detail=False, prefix=None):
         root_path = path_info.path
         seen_paths = set()
 
@@ -612,24 +619,29 @@ class GDriveFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         item_id = self._get_item_id(path_info)
         gdrive_file = self._drive.CreateFile({"id": item_id})
         gdrive_file.FetchMetadata(fields="fileSize")
-        return {"size": gdrive_file.get("fileSize")}
+        return {"size": gdrive_file.get("fileSize"), "type": "file"}
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, **kwargs):
         dirname = to_info.parent
         assert dirname
         parent_id = self._get_item_id(dirname, create=True)
         self._gdrive_upload_fobj(fobj, parent_id, to_info.name)
 
     def _upload(
-        self, from_file, to_info, name=None, no_progress_bar=False, **_kwargs
+        self,
+        from_file,
+        to_info,
+        name=None,
+        no_progress_bar=False,
+        **_kwargs,
     ):
         with open(from_file, "rb") as fobj:
             self.upload_fobj(
                 fobj,
                 to_info,
+                size=os.path.getsize(from_file),
                 no_progress_bar=no_progress_bar,
                 desc=name or to_info.name,
-                total=os.path.getsize(from_file),
             )
 
     def _download(self, from_info, to_file, name=None, no_progress_bar=False):

dvc/fs/git.py

@@ -1,5 +1,6 @@
 import errno
 import os
+import stat
 
 from dvc.utils import is_exec, relpath
 
@@ -9,6 +10,8 @@ from .base import BaseFileSystem
 class GitFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     """Proxies the repo file access methods to Git objects"""
 
+    sep = os.sep
+
     scheme = "local"
 
     def __init__(self, root_dir, trie):
@@ -104,7 +107,10 @@ class GitFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             raise FileNotFoundError(
                 errno.ENOENT, os.strerror(errno.ENOENT), path_info
             )
-        return {"size": st.st_size}
+        return {
+            "size": st.st_size,
+            "type": "directory" if stat.S_ISDIR(st.st_mode) else "file",
+        }
 
     def stat(self, path):
         key = self._get_key(path)

dvc/fs/gs.py

@@ -16,12 +16,7 @@ class GSFileSystem(ObjectFSWrapper):
     REQUIRES = {"gcsfs": "gcsfs"}
     PARAM_CHECKSUM = "etag"
     DETAIL_FIELDS = frozenset(("etag", "size"))
-
-    def __init__(self, **config):
-        super().__init__(**config)
-
-        url = config.get("url", "gs://")
-        self.path_info = self.PATH_CLS(url)
+    TRAVERSE_PREFIX_LEN = 2
 
     def _prepare_credentials(self, **config):
         login_info = {"consistency": None}

dvc/fs/hdfs.py

@@ -1,3 +1,4 @@
+import errno
 import io
 import logging
 import os
@@ -6,7 +7,6 @@ import shutil
 import subprocess
 from collections import deque
 from contextlib import closing, contextmanager
-from urllib.parse import urlparse
 
 from dvc.hash_info import HashInfo
 from dvc.progress import Tqdm
@@ -68,24 +68,19 @@ class HDFSFileSystem(BaseFileSystem):
     def __init__(self, **config):
         super().__init__(**config)
 
-        self.path_info = None
-        url = config.get("url")
-        if not url:
-            return
-
-        parsed = urlparse(url)
-        user = parsed.username or config.get("user")
-
-        self.path_info = self.PATH_CLS.from_parts(
-            scheme=self.scheme,
-            host=parsed.hostname,
-            user=user,
-            port=parsed.port,
-            path=parsed.path,
-        )
+        self.host = config["host"]
+        self.user = config.get("user")
+        self.port = config.get("port")
 
         self.kerb_ticket = config.get("kerb_ticket")
 
+    @staticmethod
+    def _get_kwargs_from_urls(urlpath):
+        from fsspec.implementations.hdfs import PyArrowHDFS
+
+        # pylint:disable=protected-access
+        return PyArrowHDFS._get_kwargs_from_urls(urlpath)
+
     def hdfs(self, path_info):
         import pyarrow.fs
 
@@ -226,8 +221,19 @@ class HDFSFileSystem(BaseFileSystem):
 
     def info(self, path_info):
         with self.hdfs(path_info) as hdfs:
-            finfo = hdfs.get_file_info(path_info.path)
-            return {"size": finfo.size}
+            from pyarrow.fs import FileType
+
+            file_info = hdfs.get_file_info(path_info.path)
+            if file_info.type is FileType.Directory:
+                kind = "directory"
+            elif file_info.type is FileType.File:
+                kind = "file"
+            else:
+                raise FileNotFoundError(
+                    errno.ENOENT, os.strerror(errno.ENOENT), path_info
+                )
+
+            return {"size": file_info.size, "type": kind}
 
     def checksum(self, path_info):
         return HashInfo(
@@ -236,7 +242,7 @@ class HDFSFileSystem(BaseFileSystem):
             size=self.getsize(path_info),
         )
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, **kwargs):
         with self.hdfs(to_info) as hdfs:
             with hdfs.open_output_stream(to_info.path) as fdest:
                 shutil.copyfileobj(fobj, fdest)

dvc/fs/http.py

@@ -2,6 +2,7 @@ import logging
 import os.path
 import threading
 from typing import Optional
+from urllib.parse import urlparse
 
 from funcy import cached_property, memoize, wrap_prop, wrap_with
 
@@ -30,6 +31,7 @@ class HTTPFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     PATH_CLS = HTTPURLInfo
     PARAM_CHECKSUM = "etag"
     CAN_TRAVERSE = False
+    REQUIRES = {"requests": "requests"}
 
     SESSION_RETRIES = 5
     SESSION_BACKOFF_FACTOR = 0.1
@@ -39,13 +41,7 @@ class HTTPFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     def __init__(self, **config):
         super().__init__(**config)
 
-        url = config.get("url")
-        if url:
-            self.path_info = self.PATH_CLS(url)
-            self.user = config.get("user", None)
-            self.host = self.path_info.host
-        else:
-            self.path_info = None
+        self.user = config.get("user", None)
 
         self.auth = config.get("auth", None)
         self.custom_auth_header = config.get("custom_auth_header", None)
@@ -55,12 +51,12 @@ class HTTPFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         self.ssl_verify = config.get("ssl_verify", True)
         self.method = config.get("method", "POST")
 
-    def _auth_method(self):
+    def _auth_method(self, url):
         from requests.auth import HTTPBasicAuth, HTTPDigestAuth
 
         if self.auth:
             if self.ask_password and self.password is None:
-                self.password = ask_password(self.host, self.user)
+                self.password = ask_password(urlparse(url).hostname, self.user)
             if self.auth == "basic":
                 return HTTPBasicAuth(self.user, self.password)
             if self.auth == "digest":
@@ -103,7 +99,7 @@ class HTTPFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             res = self._session.request(
                 method,
                 url,
-                auth=self._auth_method(),
+                auth=self._auth_method(url),
                 headers=self.headers,
                 **kwargs,
             )
@@ -150,9 +146,9 @@ class HTTPFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         resp = self._head(path_info.url)
         etag = resp.headers.get("ETag") or resp.headers.get("Content-MD5")
         size = self._content_length(resp)
-        return {"etag": etag, "size": size}
+        return {"etag": etag, "size": size, "type": "file"}
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, **kwargs):
         def chunks(fobj):
             while True:
                 chunk = fobj.read(self.CHUNK_SIZE)
@@ -189,9 +185,9 @@ class HTTPFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             self.upload_fobj(
                 fobj,
                 to_info,
+                size=None if no_progress_bar else os.path.getsize(from_file),
                 no_progress_bar=no_progress_bar,
                 desc=name or to_info.url,
-                total=None if no_progress_bar else os.path.getsize(from_file),
             )
 
     @staticmethod

dvc/fs/local.py

@@ -14,6 +14,8 @@ logger = logging.getLogger(__name__)
 
 
 class LocalFileSystem(BaseFileSystem):
+    sep = os.sep
+
     scheme = Schemes.LOCAL
     PATH_CLS = PathInfo
     PARAM_CHECKSUM = "md5"
@@ -25,8 +27,6 @@ class LocalFileSystem(BaseFileSystem):
 
         super().__init__(**config)
         self.fs = LocalFS()
-        url = config.get("url")
-        self.path_info = self.PATH_CLS(url) if url else None
 
     @staticmethod
     def open(path_info, mode="r", encoding=None, **kwargs):
@@ -105,7 +105,7 @@ class LocalFileSystem(BaseFileSystem):
             self.remove(tmp_info)
             raise
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, **kwargs):
         self.makedirs(to_info.parent)
         tmp_info = to_info.parent / tmp_fname("")
         try:
@@ -157,7 +157,7 @@ class LocalFileSystem(BaseFileSystem):
         st = os.stat(path_info)
         return {
             "size": st.st_size,
-            "type": "dir" if stat.S_ISDIR(st.st_mode) else "file",
+            "type": "directory" if stat.S_ISDIR(st.st_mode) else "file",
         }
 
     def _upload(

dvc/fs/oss.py

@@ -42,9 +42,6 @@ class OSSFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     def __init__(self, **config):
         super().__init__(**config)
 
-        url = config.get("url")
-        self.path_info = self.PATH_CLS(url) if url else None
-
         self.endpoint = config.get("oss_endpoint") or os.getenv("OSS_ENDPOINT")
 
         self.key_id = (
@@ -107,7 +104,7 @@ class OSSFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         logger.debug(f"Removing oss://{path_info}")
         self._get_bucket(path_info.bucket).delete_object(path_info.path)
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, **kwargs):
         self._get_bucket(to_info.bucket).put_object(to_info.path, fobj)
 
     def _upload(

dvc/fs/repo.py

@@ -31,6 +31,8 @@ class RepoFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         kwargs: Additional keyword arguments passed to the `DvcFileSystem()`.
     """
 
+    sep = os.sep
+
     scheme = "local"
     PARAM_CHECKSUM = "md5"
 
@@ -64,6 +66,12 @@ class RepoFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         if hasattr(repo, "dvc_dir"):
             self._dvcfss[repo.root_dir] = DvcFileSystem(repo=repo)
 
+    @property
+    def repo_url(self):
+        if self._main_repo is None:
+            return None
+        return self._main_repo.url
+
     def _get_repo(self, path: str) -> Optional["Repo"]:
         """Returns repo that the path falls in, using prefix.
 

dvc/fs/s3.py

@@ -29,12 +29,6 @@ class BaseS3FileSystem(ObjectFSWrapper):
         "grant_write_acp": "GrantWriteACP",
     }
 
-    def __init__(self, **config):
-        super().__init__(**config)
-
-        url = config.get("url", "s3://")
-        self.path_info = self.PATH_CLS(url)
-
     _TRANSFER_CONFIG_ALIASES = {
         "max_queue_size": "max_io_queue",
         "max_concurrent_requests": "max_concurrency",
@@ -194,6 +188,7 @@ class S3FileSystem(BaseS3FileSystem):  # pylint:disable=abstract-method
             "s3",
             endpoint_url=client_kwargs.get("endpoint_url"),
             use_ssl=login_info["use_ssl"],
+            verify=client_kwargs.get("verify"),
         )
 
     def _get_obj(self, path_info):

dvc/fs/ssh/__init__.py

@@ -5,7 +5,6 @@ import os
 import shutil
 import threading
 from contextlib import closing, contextmanager
-from urllib.parse import urlparse
 
 from funcy import first, memoize, silent, wrap_with
 
@@ -48,34 +47,19 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
 
     def __init__(self, **config):
         super().__init__(**config)
-        url = config.get("url")
-        if url:
-            parsed = urlparse(url)
-            user_ssh_config = self._load_user_ssh_config(parsed.hostname)
+        user_ssh_config = self._load_user_ssh_config(config["host"])
 
-            host = user_ssh_config.get("hostname", parsed.hostname)
-            user = (
-                config.get("user")
-                or parsed.username
-                or user_ssh_config.get("user")
-                or getpass.getuser()
-            )
-            port = (
-                config.get("port")
-                or parsed.port
-                or self._try_get_ssh_config_port(user_ssh_config)
-                or self.DEFAULT_PORT
-            )
-            self.path_info = self.PATH_CLS.from_parts(
-                scheme=self.scheme,
-                host=host,
-                user=user,
-                port=port,
-                path=parsed.path,
-            )
-        else:
-            self.path_info = None
-            user_ssh_config = {}
+        self.host = user_ssh_config.get("hostname", config["host"])
+        self.user = (
+            config.get("user")
+            or user_ssh_config.get("user")
+            or getpass.getuser()
+        )
+        self.port = (
+            config.get("port")
+            or self._try_get_ssh_config_port(user_ssh_config)
+            or self.DEFAULT_PORT
+        )
 
         self.keyfile = config.get(
             "keyfile"
@@ -93,6 +77,16 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             self.sock = None
         self.allow_agent = config.get("allow_agent", True)
 
+    @staticmethod
+    def _get_kwargs_from_urls(urlpath):
+        from fsspec.implementations.sftp import SFTPFileSystem
+
+        # pylint:disable=protected-access
+        kwargs = SFTPFileSystem._get_kwargs_from_urls(urlpath)
+        if "username" in kwargs:
+            kwargs["user"] = kwargs.pop("username")
+        return kwargs
+
     @staticmethod
     def ssh_config_filename():
         return os.path.expanduser(os.path.join("~", ".ssh", "config"))
@@ -120,25 +114,21 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     def _try_get_ssh_config_keyfile(user_ssh_config):
         return first(user_ssh_config.get("identityfile") or ())
 
-    def ensure_credentials(self, path_info=None):
-        if path_info is None:
-            path_info = self.path_info
-
+    def ensure_credentials(self):
         # NOTE: we use the same password regardless of the server :(
         if self.ask_password and self.password is None:
-            host, user, port = path_info.host, path_info.user, path_info.port
-            self.password = ask_password(host, user, port)
+            self.password = ask_password(self.host, self.user, self.port)
 
-    def ssh(self, path_info):
-        self.ensure_credentials(path_info)
+    def ssh(self):
+        self.ensure_credentials()
 
         from .connection import SSHConnection
 
         return get_connection(
             SSHConnection,
-            path_info.host,
-            username=path_info.user,
-            port=path_info.port,
+            self.host,
+            username=self.user,
+            port=self.port,
             key_filename=self.keyfile,
             timeout=self.timeout,
             password=self.password,
@@ -151,7 +141,7 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     def open(self, path_info, mode="r", encoding=None, **kwargs):
         assert mode in {"r", "rt", "rb", "wb"}
 
-        with self.ssh(path_info) as ssh, closing(
+        with self.ssh() as ssh, closing(
             ssh.sftp.open(path_info.path, mode)
         ) as fd:
             if "b" in mode:
@@ -160,19 +150,19 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
                 yield io.TextIOWrapper(fd, encoding=encoding)
 
     def exists(self, path_info) -> bool:
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             return ssh.exists(path_info.path)
 
     def isdir(self, path_info):
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             return ssh.isdir(path_info.path)
 
     def isfile(self, path_info):
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             return ssh.isfile(path_info.path)
 
     def walk_files(self, path_info, **kwargs):
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             for fname in ssh.walk_files(path_info.path):
                 yield path_info.replace(path=fname)
 
@@ -180,32 +170,32 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         if path_info.scheme != self.scheme:
             raise NotImplementedError
 
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             ssh.remove(path_info.path)
 
     def makedirs(self, path_info):
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             ssh.makedirs(path_info.path)
 
     def move(self, from_info, to_info):
         if from_info.scheme != self.scheme or to_info.scheme != self.scheme:
             raise NotImplementedError
 
-        with self.ssh(from_info) as ssh:
+        with self.ssh() as ssh:
             ssh.move(from_info.path, to_info.path)
 
     def copy(self, from_info, to_info):
         if not from_info.scheme == to_info.scheme == self.scheme:
             raise NotImplementedError
 
-        with self.ssh(from_info) as ssh:
+        with self.ssh() as ssh:
             ssh.atomic_copy(from_info.path, to_info.path)
 
     def symlink(self, from_info, to_info):
         if not from_info.scheme == to_info.scheme == self.scheme:
             raise NotImplementedError
 
-        with self.ssh(from_info) as ssh:
+        with self.ssh() as ssh:
             ssh.symlink(from_info.path, to_info.path)
 
     def hardlink(self, from_info, to_info):
@@ -215,7 +205,7 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         # See dvc/remote/local/__init__.py - hardlink()
         if self.getsize(from_info) == 0:
 
-            with self.ssh(to_info) as ssh:
+            with self.ssh() as ssh:
                 ssh.sftp.open(to_info.path, "w").close()
 
             logger.debug(
@@ -225,18 +215,18 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             )
             return
 
-        with self.ssh(from_info) as ssh:
+        with self.ssh() as ssh:
             ssh.hardlink(from_info.path, to_info.path)
 
     def reflink(self, from_info, to_info):
         if from_info.scheme != self.scheme or to_info.scheme != self.scheme:
             raise NotImplementedError
 
-        with self.ssh(from_info) as ssh:
+        with self.ssh() as ssh:
             ssh.reflink(from_info.path, to_info.path)
 
     def md5(self, path_info):
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             return HashInfo(
                 "md5",
                 ssh.md5(path_info.path),
@@ -244,16 +234,16 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
             )
 
     def info(self, path_info):
-        with self.ssh(path_info) as ssh:
+        with self.ssh() as ssh:
             return ssh.info(path_info.path)
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, **kwargs):
         self.makedirs(to_info.parent)
         with self.open(to_info, mode="wb") as fdest:
             shutil.copyfileobj(fobj, fdest)
 
     def _download(self, from_info, to_file, name=None, no_progress_bar=False):
-        with self.ssh(from_info) as ssh:
+        with self.ssh() as ssh:
             ssh.download(
                 from_info.path,
                 to_file,
@@ -264,7 +254,7 @@ class SSHFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     def _upload(
         self, from_file, to_info, name=None, no_progress_bar=False, **_kwargs
     ):
-        with self.ssh(to_info) as ssh:
+        with self.ssh() as ssh:
             ssh.upload(
                 from_file,
                 to_info.path,

dvc/fs/ssh/connection.py

@@ -82,7 +82,7 @@ class SSHConnection:
         st = self.sftp.stat(path)
         return {
             "size": st.st_size,
-            "type": "dir" if stat.S_ISDIR(st.st_mode) else "file",
+            "type": "directory" if stat.S_ISDIR(st.st_mode) else "file",
         }
 
     def getsize(self, path):

dvc/fs/webdav.py

@@ -1,270 +1,101 @@
-import io
 import logging
 import os
 import threading
-from collections import deque
+from functools import lru_cache
 
-from funcy import cached_property, nullcontext, wrap_prop
+from funcy import cached_property, wrap_prop
 
-from dvc.config import ConfigError
-from dvc.exceptions import DvcException
-from dvc.path_info import HTTPURLInfo, WebDAVURLInfo
-from dvc.progress import Tqdm
+from dvc.path_info import WebDAVURLInfo
 from dvc.scheme import Schemes
 
-from .base import BaseFileSystem
+from .fsspec_wrapper import FSSpecWrapper
 from .http import ask_password
 
 logger = logging.getLogger(__name__)
 
 
-class WebDAVConnectionError(DvcException):
-    def __init__(self, host):
-        super().__init__(f"Unable to connect to WebDAV {host}.")
-
-
-class WebDAVFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
-    # Use webdav scheme
+class WebDAVFileSystem(FSSpecWrapper):  # pylint:disable=abstract-method
     scheme = Schemes.WEBDAV
-
-    # URLInfo for Webdav ~ replaces webdav -> http
     PATH_CLS = WebDAVURLInfo
-
-    # Traversable as walk_files is implemented
     CAN_TRAVERSE = True
-
-    # Length of walk_files prefix
     TRAVERSE_PREFIX_LEN = 2
-
-    # Implementation based on webdav3.client
-    REQUIRES = {"webdavclient3": "webdav3.client"}
-
-    # Chunk size for buffered upload/download with progress bar
-    CHUNK_SIZE = 2 ** 16
-
+    REQUIRES = {"webdav4": "webdav4"}
     PARAM_CHECKSUM = "etag"
     DETAIL_FIELDS = frozenset(("etag", "size"))
 
-    # Constructor
     def __init__(self, **config):
-        # Call BaseFileSystem constructor
         super().__init__(**config)
 
-        # Get username from configuration
-        self.user = config.get("user", None)
+        cert_path = config.get("cert_path", None)
+        key_path = config.get("key_path", None)
+        self.prefix = config.get("prefix", "")
+        self.fs_args.update(
+            {
+                "base_url": config["url"],
+                "cert": cert_path if not key_path else (cert_path, key_path),
+                "verify": config.get("ssl_verify", True),
+                "timeout": config.get("timeout", 30),
+            }
+        )
 
-        # Get password from configuration (might be None ~ not set)
-        self.password = config.get("password", None)
-
-        # Whether to ask for password if it is not set
-        self.ask_password = config.get("ask_password", False)
-
-        # Use token for webdav auth
-        self.token = config.get("token", None)
-
-        # Path to certificate
-        self.cert_path = config.get("cert_path", None)
-
-        # Path to private key
-        self.key_path = config.get("key_path", None)
-
-        # Connection timeout
-        self.timeout = config.get("timeout", 30)
-
-        # Get URL from configuration
-        self.url = config.get("url", None)
-
-        # If URL in config parse path_info
-        if self.url:
-            self.path_info = self.PATH_CLS(self.url)
-
-            # If username not specified try to use from URL
-            if self.user is None and self.path_info.user is not None:
-                self.user = self.path_info.user
-
-            # Construct hostname from path_info by stripping path
-            http_info = HTTPURLInfo(self.path_info.url)
-            self.hostname = http_info.replace(path="").url
-
-    # Webdav client
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def _client(self):
-        from webdav3.client import Client
-
-        # Set password or ask for it
-        if self.ask_password and self.password is None and self.token is None:
-            self.password = ask_password(self.hostname, self.user)
-
-        # Setup webdav client options dictionary
-        options = {
-            "webdav_hostname": self.hostname,
-            "webdav_login": self.user,
-            "webdav_password": self.password,
-            "webdav_token": self.token,
-            "webdav_cert_path": self.cert_path,
-            "webdav_key_path": self.key_path,
-            "webdav_timeout": self.timeout,
-            "webdav_chunk_size": self.CHUNK_SIZE,
+    @staticmethod
+    def _get_kwargs_from_urls(urlpath):
+        path_info = WebDAVURLInfo(urlpath)
+        return {
+            "prefix": path_info.path.rstrip("/"),
+            "host": path_info.replace(path="").url,
+            "url": path_info.url.rstrip("/"),
+            "user": path_info.user,
         }
 
-        client = Client(options)
+    def _prepare_credentials(self, **config):
+        user = config.get("user", None)
+        password = config.get("password", None)
 
-        # Check whether client options are valid
-        if not client.valid():
-            raise ConfigError(
-                f"Configuration for WebDAV {self.hostname} is invalid."
-            )
+        headers = {}
+        token = config.get("token")
+        auth = None
+        if token:
+            headers.update({"Authorization": f"Bearer {token}"})
+        elif user:
+            if not password and config.get("ask_password"):
+                password = ask_password(config["host"], user)
+            auth = (user, password)
 
-        # Check whether connection is valid (root should always exist)
-        if not client.check(self.path_info.path):
-            raise WebDAVConnectionError(self.hostname)
+        return {"headers": headers, "auth": auth}
 
-        return client
+    @wrap_prop(threading.Lock())
+    @cached_property
+    def fs(self):
+        from webdav4.fsspec import WebdavFileSystem
 
-    def open(self, path_info, mode="r", encoding=None, **kwargs):
-        from webdav3.exceptions import RemoteResourceNotFound
+        return WebdavFileSystem(**self.fs_args)
 
-        assert mode in {"r", "rt", "rb"}
+    def _upload_fobj(self, fobj, to_info, size: int = None):
+        rpath = self.translate_path_info(to_info)
+        self.makedirs(os.path.dirname(rpath))
+        # using upload_fileobj to directly upload fileobj rather than buffering
+        # and using overwrite=True to avoid check for an extra exists call,
+        # as caller should ensure that the file does not exist beforehand.
+        return self.fs.client.upload_fileobj(
+            fobj, rpath, overwrite=True, size=size
+        )
 
-        fobj = io.BytesIO()
-
-        try:
-            self._client.download_from(buff=fobj, remote_path=path_info.path)
-        except RemoteResourceNotFound as exc:
-            raise FileNotFoundError from exc
-
-        fobj.seek(0)
-
-        if "mode" == "rb":
-            return fobj
-
-        return io.TextIOWrapper(fobj, encoding=encoding)
-
-    # Checks whether file/directory exists at remote
-    def exists(self, path_info) -> bool:
-        # Use webdav check to test for file existence
-        return self._client.check(path_info.path)
-
-    # Checks whether path points to directory
-    def isdir(self, path_info):
-        # Use webdav is_dir to test whether path points to a directory
-        return self._client.is_dir(path_info.path)
-
-    def walk_files(self, path_info, **kwargs):
-        if not self.exists(path_info):
-            return
-
-        for file in self.find(path_info):
-            yield path_info.replace(path=file)
-
-    def ls(self, path_info, detail=False):
-        for entry in self._client.list(path_info.path):
-            path = entry["path"]
-            if detail:
-                if entry["isdir"]:
-                    yield {"type": "directory", "name": path}
-                else:
-                    yield {
-                        "type": "file",
-                        "name": path,
-                        "size": entry["size"],
-                        "etag": entry["etag"],
-                    }
-            else:
-                yield path
-
-    def find(self, path_info, detail=False):
-        dirs = deque([path_info.path])
-
-        while dirs:
-            for entry in self._client.list(dirs.pop(), get_info=True):
-                path = entry["path"]
-                if entry["isdir"]:
-                    dirs.append(path)
-                    continue
-
-                if detail:
-                    yield {
-                        "type": "file",
-                        "name": path,
-                        "size": entry["size"],
-                        "etag": entry["etag"],
-                    }
-                else:
-                    yield path
-
-    # Removes file/directory
-    def remove(self, path_info):
-        # Use webdav client clean (DELETE) method to remove file/directory
-        self._client.clean(path_info.path)
-
-    # Creates directories
     def makedirs(self, path_info):
-        # Terminate recursion
-        if path_info.path == self.path_info.path or self.exists(path_info):
-            return
+        path = self.translate_path_info(path_info)
+        return self.fs.makedirs(path, exist_ok=True)
 
-        # Recursively descent to root
-        self.makedirs(path_info.parent)
+    @lru_cache(512)
+    def translate_path_info(self, path):
+        if isinstance(path, self.PATH_CLS):
+            return path.path[len(self.prefix) :].lstrip("/")
+        return path
 
-        # Construct directory at current recursion depth
-        self._client.mkdir(path_info.path)
+    _with_bucket = translate_path_info
 
-    # Moves file/directory at remote
-    def move(self, from_info, to_info):
-        # Webdav client move
-        self._client.move(from_info.path, to_info.path)
+    def _strip_bucket(self, entry):
+        return entry
 
-    def _upload_fobj(self, fobj, to_info):
-        # In contrast to other upload_fobj implementations, this one does not
-        # exactly do a chunked-upload but rather put everything in one request.
-        self.makedirs(to_info.parent)
-        self._client.upload_to(buff=fobj, remote_path=to_info.path)
 
-    # Downloads file from remote to file
-    def _download(self, from_info, to_file, name=None, no_progress_bar=False):
-        # Progress from HTTPFileSystem
-        with open(to_file, "wb") as fd:
-            with Tqdm.wrapattr(
-                fd,
-                "write",
-                total=None if no_progress_bar else self.getsize(from_info),
-                leave=False,
-                desc=from_info.url if name is None else name,
-                disable=no_progress_bar,
-            ) as fd_wrapped:
-                # Download from WebDAV via buffer
-                self._client.download_from(
-                    buff=fd_wrapped, remote_path=from_info.path
-                )
-
-    # Uploads file to remote
-    def _upload(
-        self, from_file, to_info, name=None, no_progress_bar=False, **_kwargs
-    ):
-        # First try to create parent directories
-        self.makedirs(to_info.parent)
-
-        file_size = os.path.getsize(from_file)
-        with open(from_file, "rb") as fd:
-            progress_context = (
-                nullcontext(fd)
-                if file_size == 0
-                else Tqdm.wrapattr(
-                    fd,
-                    "read",
-                    total=None if no_progress_bar else file_size,
-                    leave=False,
-                    desc=to_info.url if name is None else name,
-                    disable=no_progress_bar,
-                )
-            )
-            with progress_context as fd_wrapped:
-                self._client.upload_to(
-                    buff=fd_wrapped, remote_path=to_info.path
-                )
-
-    def info(self, path_info):
-        info = self._client.info(path_info.path)
-        return {"size": int(info["size"]), "etag": info["etag"]}
+class WebDAVSFileSystem(WebDAVFileSystem):  # pylint:disable=abstract-method
+    scheme = Schemes.WEBDAVS

dvc/fs/webdavs.py

@@ -1,7 +0,0 @@
-from dvc.scheme import Schemes
-
-from .webdav import WebDAVFileSystem
-
-
-class WebDAVSFileSystem(WebDAVFileSystem):  # pylint:disable=abstract-method
-    scheme = Schemes.WEBDAVS

dvc/fs/webhdfs.py

@@ -4,7 +4,6 @@ import posixpath
 import shutil
 import threading
 from contextlib import contextmanager
-from urllib.parse import urlparse
 
 from funcy import cached_property, wrap_prop
 
@@ -40,23 +39,24 @@ class WebHDFSFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
     def __init__(self, **config):
         super().__init__(**config)
 
-        self.path_info = None
-        url = config.get("url")
-        if not url:
-            return
-
-        self.path_info = self.PATH_CLS(url)
-
-        parsed = urlparse(url)
-
-        self.host = parsed.hostname
-        self.user = parsed.username or config.get("user")
-        self.port = parsed.port
+        self.host = config["host"]
+        self.user = config.get("user")
+        self.port = config.get("port")
 
         self.hdfscli_config = config.get("hdfscli_config")
         self.token = config.get("webhdfs_token")
         self.alias = config.get("webhdfs_alias")
 
+    @staticmethod
+    def _get_kwargs_from_urls(urlpath):
+        from fsspec.implementations.webhdfs import WebHDFS
+
+        return (
+            WebHDFS._get_kwargs_from_urls(  # pylint:disable=protected-access
+                urlpath
+            )
+        )
+
     @wrap_prop(threading.Lock())
     @cached_property
     def hdfs_client(self):
@@ -123,7 +123,7 @@ class WebHDFSFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
 
     def info(self, path_info):
         st = self.hdfs_client.status(path_info.path)
-        return {"size": st["length"]}
+        return {"size": st["length"], "type": "file"}
 
     def checksum(self, path_info):
         return HashInfo(
@@ -141,7 +141,7 @@ class WebHDFSFileSystem(BaseFileSystem):  # pylint:disable=abstract-method
         self.hdfs_client.makedirs(to_info.parent.path)
         self.hdfs_client.rename(from_info.path, to_info.path)
 
-    def _upload_fobj(self, fobj, to_info):
+    def _upload_fobj(self, fobj, to_info, **kwargs):
         with self.hdfs_client.write(to_info.path) as fdest:
             shutil.copyfileobj(fobj, fdest)
 

dvc/hash_info.py

@@ -21,6 +21,9 @@ class HashInfo:
     def __str__(self):
         return f"{self.name}: {self.value}"
 
+    def __hash__(self):
+        return hash((self.name, self.value))
+
     @classmethod
     def from_dict(cls, d):
         _d = d.copy() if d else {}

dvc/info.py

@@ -20,6 +20,11 @@ if PKG is None:
 else:
     package = f"({PKG})"
 
+try:
+    import importlib.metadata as importlib_metadata
+except ImportError:  # < 3.8
+    import importlib_metadata  # type: ignore[no-redef]
+
 
 def get_dvc_info():
     info = [
@@ -27,7 +32,7 @@ def get_dvc_info():
         "---------------------------------",
         f"Platform: Python {platform.python_version()} on "
         f"{platform.platform()}",
-        f"Supports: {_get_supported_remotes()}",
+        f"Supports:{_get_supported_remotes()}",
     ]
 
     try:
@@ -113,19 +118,23 @@ def _get_linktype_support_info(repo):
 
 
 def _get_supported_remotes():
-
     supported_remotes = []
     for scheme, fs_cls in FS_MAP.items():
         if not fs_cls.get_missing_deps():
-            supported_remotes.append(scheme)
+            dependencies = []
+            for requirement in fs_cls.REQUIRES:
+                dependencies.append(
+                    f"{requirement} = "
+                    f"{importlib_metadata.version(requirement)}"
+                )
 
-    if len(supported_remotes) == len(FS_MAP):
-        return "All remotes"
+            remote_info = scheme
+            if dependencies:
+                remote_info += " (" + ", ".join(dependencies) + ")"
+            supported_remotes.append(remote_info)
 
-    if len(supported_remotes) == 1:
-        return supported_remotes
-
-    return ", ".join(supported_remotes)
+    assert len(supported_remotes) >= 1
+    return "\n\t" + ",\n\t".join(supported_remotes)
 
 
 def get_fs_type(path):

dvc/objects/__init__.py

@@ -1,14 +1,24 @@
 import logging
 from concurrent.futures import ThreadPoolExecutor, as_completed
+from typing import TYPE_CHECKING, Optional
 
 from dvc.progress import Tqdm
 
 from .tree import Tree
 
+if TYPE_CHECKING:
+    from .db.base import ObjectDB
+    from .file import HashFile
+
 logger = logging.getLogger(__name__)
 
 
-def save(odb, obj, jobs=None, **kwargs):
+def save(
+    odb: "ObjectDB",
+    obj: "HashFile",
+    jobs: Optional[int] = None,
+    **kwargs,
+):
     if isinstance(obj, Tree):
         with ThreadPoolExecutor(max_workers=jobs) as executor:
             for future in Tqdm(
@@ -18,7 +28,7 @@ def save(odb, obj, jobs=None, **kwargs):
                         entry.path_info,
                         entry.fs,
                         entry.hash_info,
-                        **kwargs
+                        **kwargs,
                     )
                     for _, entry in obj
                 ),

dvc/objects/db/__init__.py

@@ -1,24 +1,22 @@
-from collections import defaultdict
-
 from dvc.scheme import Schemes
 
 
-def get_odb(fs, **config):
+def get_odb(fs, path_info, **config):
     from .base import ObjectDB
     from .gdrive import GDriveObjectDB
     from .local import LocalObjectDB
     from .ssh import SSHObjectDB
 
     if fs.scheme == Schemes.LOCAL:
-        return LocalObjectDB(fs, **config)
+        return LocalObjectDB(fs, path_info, **config)
 
     if fs.scheme == Schemes.SSH:
-        return SSHObjectDB(fs, **config)
+        return SSHObjectDB(fs, path_info, **config)
 
     if fs.scheme == Schemes.GDRIVE:
-        return GDriveObjectDB(fs, **config)
+        return GDriveObjectDB(fs, path_info, **config)
 
-    return ObjectDB(fs, **config)
+    return ObjectDB(fs, path_info, **config)
 
 
 def _get_odb(repo, settings):
@@ -27,8 +25,8 @@ def _get_odb(repo, settings):
     if not settings:
         return None
 
-    cls, config = get_cloud_fs(repo, **settings)
-    return get_odb(cls(**config), state=repo.state, **config)
+    cls, config, path_info = get_cloud_fs(repo, **settings)
+    return get_odb(cls(**config), path_info, state=repo.state, **config)
 
 
 class ODBManager:
@@ -80,99 +78,3 @@ class ODBManager:
     def by_scheme(self):
         self._init_odb(self.CLOUD_SCHEMES)
         yield from self._odb.items()
-
-
-class NamedCacheItem:
-    def __init__(self):
-        self.names = set()
-        self.children = defaultdict(NamedCacheItem)
-
-    def __eq__(self, other):
-        return self.names == other.names and self.children == other.children
-
-    def child_keys(self):
-        for key, child in self.children.items():
-            yield key
-            yield from child.child_keys()
-
-    def child_names(self):
-        for key, child in self.children.items():
-            yield key, child.names
-            yield from child.child_names()
-
-    def add(self, checksum, item):
-        self.children[checksum].update(item)
-
-    def update(self, item, suffix=""):
-        if suffix:
-            self.names.update(n + suffix for n in item.names)
-        else:
-            self.names.update(item.names)
-        for checksum, child_item in item.children.items():
-            self.children[checksum].update(child_item)
-
-
-class NamedCache:
-    # pylint: disable=protected-access
-    def __init__(self):
-        self._items = defaultdict(lambda: defaultdict(NamedCacheItem))
-        self.external = defaultdict(set)
-
-    @classmethod
-    def make(cls, scheme, checksum, name):
-        cache = cls()
-        cache.add(scheme, checksum, name)
-        return cache
-
-    def __getitem__(self, key):
-        return self._items[key]
-
-    def add(self, scheme, checksum, name):
-        """Add a mapped name for the specified checksum."""
-        self._items[scheme][checksum].names.add(name)
-
-    def add_child_cache(self, checksum, cache, suffix=""):
-        """Add/update child cache for the specified checksum."""
-        for scheme, src in cache._items.items():
-            dst = self._items[scheme][checksum].children
-            for child_checksum, item in src.items():
-                dst[child_checksum].update(item, suffix=suffix)
-
-        for repo_pair, files in cache.external.items():
-            self.external[repo_pair].update(files)
-
-    def add_external(self, url, rev, path):
-        self.external[url, rev].add(path)
-
-    def update(self, cache, suffix=""):
-        for scheme, src in cache._items.items():
-            dst = self._items[scheme]
-            for checksum, item in src.items():
-                dst[checksum].update(item, suffix=suffix)
-
-        for repo_pair, files in cache.external.items():
-            self.external[repo_pair].update(files)
-
-    def scheme_keys(self, scheme):
-        """Iterate over a flat list of all keys for the specified scheme,
-        including children.
-        """
-        for key, item in self._items[scheme].items():
-            yield key
-            yield from item.child_keys()
-
-    def scheme_names(self, scheme):
-        """Iterate over a flat list of checksum, names items for the specified
-        scheme, including children.
-        """
-        for key, item in self._items[scheme].items():
-            yield key, item.names
-            yield from item.child_names()
-
-    def dir_keys(self, scheme):
-        return (
-            key for key, item in self._items[scheme].items() if item.children
-        )
-
-    def child_keys(self, scheme, checksum):
-        return self._items[scheme][checksum].child_keys()

dvc/objects/db/base.py

@@ -7,6 +7,7 @@ from typing import Optional
 
 from dvc.objects.errors import ObjectFormatError
 from dvc.objects.file import HashFile
+from dvc.objects.tree import Tree
 from dvc.progress import Tqdm
 
 logger = logging.getLogger(__name__)
@@ -18,15 +19,33 @@ class ObjectDB:
     DEFAULT_CACHE_TYPES = ["copy"]
     CACHE_MODE: Optional[int] = None
 
-    def __init__(self, fs, **config):
+    def __init__(self, fs, path_info, **config):
         from dvc.state import StateNoop
 
         self.fs = fs
+        self.path_info = path_info
         self.state = config.get("state", StateNoop())
         self.verify = config.get("verify", self.DEFAULT_VERIFY)
         self.cache_types = config.get("type") or copy(self.DEFAULT_CACHE_TYPES)
         self.cache_type_confirmed = False
         self.slow_link_warning = config.get("slow_link_warning", True)
+        self.tmp_dir = config.get("tmp_dir")
+
+    @property
+    def config(self):
+        return {
+            "state": self.state,
+            "verify": self.verify,
+            "type": self.cache_types,
+            "slow_link_warning": self.slow_link_warning,
+            "tmp_dir": self.tmp_dir,
+        }
+
+    def __eq__(self, other):
+        return self.fs == other.fs and self.path_info == other.path_info
+
+    def __hash__(self):
+        return hash((self.fs.scheme, self.path_info))
 
     def move(self, from_info, to_info):
         self.fs.move(from_info, to_info)
@@ -87,7 +106,7 @@ class ObjectDB:
             callback(1)
 
     def hash_to_path_info(self, hash_):
-        return self.fs.path_info / hash_[0:2] / hash_[2:]
+        return self.path_info / hash_[0:2] / hash_[2:]
 
     # Override to return path as a string instead of PathInfo for clouds
     # which support string paths (see local)
@@ -143,12 +162,12 @@ class ObjectDB:
     def _list_paths(self, prefix=None, progress_callback=None):
         if prefix:
             if len(prefix) > 2:
-                path_info = self.fs.path_info / prefix[:2] / prefix[2:]
+                path_info = self.path_info / prefix[:2] / prefix[2:]
             else:
-                path_info = self.fs.path_info / prefix[:2]
+                path_info = self.path_info / prefix[:2]
             prefix = True
         else:
-            path_info = self.fs.path_info
+            path_info = self.path_info
             prefix = False
         if progress_callback:
             for file_info in self.fs.walk_files(path_info, prefix=prefix):
@@ -316,14 +335,22 @@ class ObjectDB:
         pass
 
     def gc(self, used, jobs=None):
+        used_hashes = set()
+        for obj in used:
+            used_hashes.add(obj.hash_info.value)
+            if isinstance(obj, Tree):
+                used_hashes.update(
+                    entry_obj.hash_info.value for _, entry_obj in obj
+                )
+
         removed = False
         # hashes must be sorted to ensure we always remove .dir files first
         for hash_ in sorted(
-            self.all(jobs, str(self.fs.path_info)),
+            self.all(jobs, str(self.path_info)),
             key=self.fs.is_dir_hash,
             reverse=True,
         ):
-            if hash_ in used:
+            if hash_ in used_hashes:
                 continue
             path_info = self.hash_to_path_info(hash_)
             if self.fs.is_dir_hash(hash_):

dvc/objects/db/git.py

@@ -0,0 +1,24 @@
+import logging
+
+from .base import ObjectDB
+
+logger = logging.getLogger(__name__)
+
+
+class GitObjectDB(ObjectDB):
+    """Dummy read-only ODB for uncached objects in external Git repos."""
+
+    def __init__(self, fs, path_info, **config):
+        from dvc.fs.repo import RepoFileSystem
+
+        assert isinstance(fs, RepoFileSystem)
+        super().__init__(fs, path_info)
+
+    def get(self, hash_info):
+        raise NotImplementedError
+
+    def add(self, path_info, fs, hash_info, move=True, **kwargs):
+        raise NotImplementedError
+
+    def gc(self, used, jobs=None):
+        raise NotImplementedError

dvc/objects/db/local.py

@@ -22,9 +22,9 @@ class LocalObjectDB(ObjectDB):
     CACHE_MODE = 0o444
     UNPACKED_DIR_SUFFIX = ".unpacked"
 
-    def __init__(self, fs, **config):
-        super().__init__(fs, **config)
-        self.cache_dir = config.get("url")
+    def __init__(self, fs, path_info, **config):
+        super().__init__(fs, path_info, **config)
+        self.cache_dir = path_info.fspath
 
         shared = config.get("shared")
         if shared:
@@ -36,11 +36,11 @@ class LocalObjectDB(ObjectDB):
 
     @property
     def cache_dir(self):
-        return self.fs.path_info.fspath if self.fs.path_info else None
+        return self.path_info.fspath if self.path_info else None
 
     @cache_dir.setter
     def cache_dir(self, value):
-        self.fs.path_info = PathInfo(value) if value else None
+        self.path_info = PathInfo(value) if value else None
 
     @cached_property
     def cache_path(self):
@@ -81,13 +81,13 @@ class LocalObjectDB(ObjectDB):
         return ret
 
     def _list_paths(self, prefix=None, progress_callback=None):
-        assert self.fs.path_info is not None
+        assert self.path_info is not None
         if prefix:
-            path_info = self.fs.path_info / prefix[:2]
+            path_info = self.path_info / prefix[:2]
             if not self.fs.exists(path_info):
                 return
         else:
-            path_info = self.fs.path_info
+            path_info = self.path_info
         # NOTE: use utils.fs walk_files since fs.walk_files will not follow
         # symlinks
         if progress_callback:

dvc/objects/db/ssh.py

@@ -28,7 +28,7 @@ class SSHObjectDB(ObjectDB):
                 callback(path)
             return ret
 
-        with self.fs.ssh(path_infos[0]) as ssh:
+        with self.fs.ssh() as ssh:
             channels = ssh.open_max_sftp_channels()
             max_workers = len(channels)
 
@@ -75,10 +75,10 @@ class SSHObjectDB(ObjectDB):
 
     def _list_paths(self, prefix=None, progress_callback=None):
         if prefix:
-            root = posixpath.join(self.fs.path_info.path, prefix[:2])
+            root = posixpath.join(self.path_info.path, prefix[:2])
         else:
-            root = self.fs.path_info.path
-        with self.fs.ssh(self.fs.path_info) as ssh:
+            root = self.path_info.path
+        with self.fs.ssh() as ssh:
             if prefix and not ssh.exists(root):
                 return
             # If we simply return an iterator then with above closes instantly

dvc/objects/file.py

@@ -1,23 +1,34 @@
 import errno
 import logging
 import os
+from typing import TYPE_CHECKING, Optional
 
 from .errors import ObjectFormatError
 
+if TYPE_CHECKING:
+    from dvc.fs.base import BaseFileSystem
+    from dvc.hash_info import HashInfo
+    from dvc.types import DvcPath
+
 logger = logging.getLogger(__name__)
 
 
 class HashFile:
-    def __init__(self, path_info, fs, hash_info):
+    def __init__(
+        self,
+        path_info: Optional["DvcPath"],
+        fs: Optional["BaseFileSystem"],
+        hash_info: "HashInfo",
+        name: Optional[str] = None,
+    ):
         self.path_info = path_info
         self.fs = fs
         self.hash_info = hash_info
+        self.name = name
 
     @property
     def size(self):
-        if not (self.path_info and self.fs):
-            return None
-        return self.fs.getsize(self.path_info)
+        return self.hash_info.size
 
     def __len__(self):
         return 1
@@ -37,6 +48,15 @@ class HashFile:
             and self.hash_info == other.hash_info
         )
 
+    def __hash__(self):
+        return hash(
+            (
+                self.hash_info,
+                self.path_info,
+                self.fs.scheme if self.fs else None,
+            )
+        )
+
     def check(self, odb, check_hash=True):
         from .stage import get_file_hash
 

dvc/objects/stage.py

@@ -1,5 +1,3 @@
-import errno
-import os
 from concurrent.futures import ThreadPoolExecutor
 from functools import partial
 
@@ -15,11 +13,11 @@ def _upload_file(path_info, fs, odb):
     from dvc.utils import tmp_fname
     from dvc.utils.stream import HashedStreamReader
 
-    tmp_info = odb.fs.path_info / tmp_fname()
+    tmp_info = odb.path_info / tmp_fname()
     with fs.open(path_info, mode="rb", chunk_size=fs.CHUNK_SIZE) as stream:
         stream = HashedStreamReader(stream)
         odb.fs.upload_fobj(
-            stream, tmp_info, desc=path_info.name, total=fs.getsize(path_info)
+            stream, tmp_info, desc=path_info.name, size=fs.getsize(path_info)
         )
 
     obj = HashFile(tmp_info, odb.fs, stream.hash_info)
@@ -52,11 +50,6 @@ def get_file_hash(path_info, fs, name, state=None):
         if hash_info:
             return hash_info
 
-    if not fs.exists(path_info):
-        raise FileNotFoundError(
-            errno.ENOENT, os.strerror(errno.ENOENT), path_info
-        )
-
     hash_info = _get_file_hash(path_info, fs, name)
 
     if state:
@@ -179,11 +172,7 @@ def stage(odb, path_info, fs, name, upload=False, **kwargs):
         isinstance(path_info, str) or path_info.scheme == fs.scheme
     )
 
-    if not fs.exists(path_info):
-        raise FileNotFoundError(
-            errno.ENOENT, os.strerror(errno.ENOENT), path_info
-        )
-
+    details = fs.info(path_info)
     state = odb.state
     # pylint: disable=assignment-from-none
     hash_info = state.get(path_info, fs)
@@ -215,12 +204,12 @@ def stage(odb, path_info, fs, name, upload=False, **kwargs):
         obj.hash_info.size = hash_info.size
         return obj
 
-    if fs.isdir(path_info):
+    if details["type"] == "directory":
         obj = _get_tree_obj(path_info, fs, name, odb, state, upload, **kwargs)
     else:
         _, obj = _get_file_obj(path_info, fs, name, odb, state, upload)
 
-    if obj.hash_info and fs.exists(path_info):
+    if obj.hash_info:
         state.save(path_info, fs, obj.hash_info)
 
     return obj

dvc/objects/tree.py

@@ -1,6 +1,7 @@
 import json
 import logging
 import posixpath
+from typing import TYPE_CHECKING, Optional, Tuple
 
 from funcy import cached_property
 
@@ -8,6 +9,9 @@ from .errors import ObjectFormatError
 from .file import HashFile
 from .stage import get_file_hash
 
+if TYPE_CHECKING:
+    from .db.base import ObjectDB
+
 logger = logging.getLogger(__name__)
 
 
@@ -24,13 +28,6 @@ class Tree(HashFile):
 
         return Trie(self._dict)
 
-    @property
-    def size(self):
-        try:
-            return sum(obj.size for _, obj in self)
-        except TypeError:
-            return None
-
     def add(self, key, obj):
         self.__dict__.pop("trie", None)
         self._dict[key] = obj
@@ -48,7 +45,10 @@ class Tree(HashFile):
         self.path_info = path_info
         self.hash_info = get_file_hash(path_info, memfs, "md5")
         self.hash_info.value += ".dir"
-        self.hash_info.size = self.size
+        try:
+            self.hash_info.size = sum(obj.size for _, obj in self)
+        except TypeError:
+            self.hash_info.size = None
         self.hash_info.nfiles = len(self)
 
     def __len__(self):
@@ -115,24 +115,45 @@ class Tree(HashFile):
         tree = cls.from_list(raw)
         tree.path_info = obj.path_info
         tree.fs = obj.fs
+        for _, entry_obj in tree:
+            entry_obj.fs = obj.fs
         tree.hash_info = hash_info
 
         return tree
 
-    def filter(self, odb, prefix):
+    def filter(
+        self, odb: "ObjectDB", prefix: Tuple[str], copy: bool = False
+    ) -> Optional[HashFile]:
+        """Return filter object(s) for this tree.
+
+        If copy is True, returned object will be a Tree containing
+        filtered entries, but with hash_info copied from the original tree.
+
+        If copy is False, returned object will be a raw HashFile or Tree with
+        newly computed hash_info for the filtered object.
+        """
         obj = self._dict.get(prefix)
         if obj:
+            if copy:
+                tree = Tree(self.path_info, self.fs, self.hash_info)
+                tree.add(prefix, obj)
+                return tree
             return obj
 
-        depth = len(prefix)
-        tree = Tree(None, None, None)
+        if copy:
+            tree = Tree(self.path_info, self.fs, self.hash_info)
+            depth = 0
+        else:
+            tree = Tree(None, None, None)
+            depth = len(prefix)
         try:
             for key, obj in self.trie.items(prefix):
                 tree.add(key[depth:], obj)
         except KeyError:
             return None
-        tree.digest()
-        odb.add(tree.path_info, tree.fs, tree.hash_info)
+        if not copy:
+            tree.digest()
+            odb.add(tree.path_info, tree.fs, tree.hash_info)
         return tree
 
 

dvc/output.py

@@ -2,7 +2,7 @@ import logging
 import os
 from collections import defaultdict
 from copy import copy
-from typing import Type
+from typing import TYPE_CHECKING, Dict, Optional, Set, Type
 from urllib.parse import urlparse
 
 from funcy import collecting, project
@@ -24,14 +24,18 @@ from .fs.local import LocalFileSystem
 from .fs.s3 import S3FileSystem
 from .hash_info import HashInfo
 from .istextfile import istextfile
+from .objects import Tree
 from .objects import save as osave
-from .objects.db import NamedCache
 from .objects.errors import ObjectFormatError
 from .objects.stage import stage as ostage
 from .scheme import Schemes
 from .utils import relpath
 from .utils.fs import path_isin
 
+if TYPE_CHECKING:
+    from .objects.db.base import ObjectDB
+    from .objects.file import HashFile
+
 logger = logging.getLogger(__name__)
 
 
@@ -41,18 +45,24 @@ CHECKSUM_SCHEMA = Any(
     And(Any(str, And(int, Coerce(str))), Length(min=3), Lower),
 )
 
+CASE_SENSITIVE_CHECKSUM_SCHEMA = Any(
+    None,
+    And(str, Length(max=0), SetTo(None)),
+    And(Any(str, And(int, Coerce(str))), Length(min=3)),
+)
+
 # NOTE: currently there are only 3 possible checksum names:
 #
 #    1) md5 (LOCAL, SSH);
-#    2) etag (S3);
+#    2) etag (S3, GS, OSS, AZURE, HTTP);
 #    3) checksum (HDFS);
 #
 # so when a few types of outputs share the same name, we only need
 # specify it once.
 CHECKSUMS_SCHEMA = {
     LocalFileSystem.PARAM_CHECKSUM: CHECKSUM_SCHEMA,
-    S3FileSystem.PARAM_CHECKSUM: CHECKSUM_SCHEMA,
     HDFSFileSystem.PARAM_CHECKSUM: CHECKSUM_SCHEMA,
+    S3FileSystem.PARAM_CHECKSUM: CASE_SENSITIVE_CHECKSUM_SCHEMA,
 }
 
 
@@ -261,8 +271,6 @@ class Output:
     IsStageFileError = OutputIsStageFileError  # type: Type[DvcException]
     IsIgnoredError = OutputIsIgnoredError  # type: Type[DvcException]
 
-    sep = "/"
-
     def __init__(
         self,
         stage,
@@ -279,9 +287,8 @@ class Output:
     ):
         self.repo = stage.repo if stage else None
 
-        fs_cls, fs_config = get_cloud_fs(self.repo, url=path)
+        fs_cls, fs_config, path_info = get_cloud_fs(self.repo, url=path)
         self.fs = fs_cls(**fs_config)
-        path_info = self.fs.path_info
 
         if (
             self.fs.scheme == "local"
@@ -320,6 +327,8 @@ class Output:
         self.obj = None
         self.isexec = False if self.IS_DEPENDENCY else isexec
 
+        self.def_remote = None
+
     def _parse_path(self, fs, path_info):
         if fs.scheme != "local":
             return path_info
@@ -648,7 +657,7 @@ class Output:
     def download(self, to, jobs=None):
         self.fs.download(self.path_info, to.path_info, jobs=jobs)
 
-    def get_obj(self, filter_info=None):
+    def get_obj(self, filter_info=None, **kwargs):
         if self.obj:
             obj = self.obj
         elif self.hash_info:
@@ -661,7 +670,7 @@ class Output:
 
         if filter_info and filter_info != self.path_info:
             prefix = filter_info.relative_to(self.path_info).parts
-            obj = obj.filter(self.odb, prefix)
+            obj = obj.filter(self.odb, prefix, **kwargs)
 
         return obj
 
@@ -741,9 +750,8 @@ class Output:
         if odb is None:
             odb = self.odb
 
-        cls, config = get_cloud_fs(self.repo, url=source)
+        cls, config, from_info = get_cloud_fs(self.repo, url=source)
         from_fs = cls(**config)
-        from_info = from_fs.path_info
 
         # When running import-url --to-remote / add --to-remote/-o ... we
         # assume that it is unlikely that the odb will contain majority of the
@@ -788,18 +796,14 @@ class Output:
             self.odb.unprotect(self.path_info)
 
     def get_dir_cache(self, **kwargs):
-
         if not self.is_dir_checksum:
             raise DvcException("cannot get dir cache for file checksum")
 
+        obj = self.odb.get(self.hash_info)
         try:
-            objects.check(self.odb, self.odb.get(self.hash_info))
+            objects.check(self.odb, obj)
         except (FileNotFoundError, ObjectFormatError):
-            self.repo.cloud.pull(
-                NamedCache.make("local", self.hash_info.value, str(self)),
-                show_checksums=False,
-                **kwargs,
-            )
+            self.repo.cloud.pull([obj], show_checksums=False, **kwargs)
 
         try:
             self.obj = objects.load(self.odb, self.hash_info)
@@ -810,27 +814,8 @@ class Output:
 
     def collect_used_dir_cache(
         self, remote=None, force=False, jobs=None, filter_info=None
-    ):
-        """Get a list of `info`s related to the given directory.
-
-        - Pull the directory entry from the remote cache if it was changed.
-
-        Example:
-
-            Given the following commands:
-
-            $ echo "foo" > directory/foo
-            $ echo "bar" > directory/bar
-            $ dvc add directory
-
-            It will return a NamedCache like:
-
-            nc = NamedCache()
-            nc.add(self.scheme, 'c157a79031e1', 'directory/foo')
-            nc.add(self.scheme, 'd3b07384d113', 'directory/bar')
-        """
-
-        cache = NamedCache()
+    ) -> Dict[Optional["ObjectDB"], Set["HashFile"]]:
+        """Fetch dir cache and return used objects for this out."""
 
         try:
             self.get_dir_cache(jobs=jobs, remote=remote)
@@ -850,38 +835,22 @@ class Output:
                     "unable to fully collect used cache"
                     " without cache for directory '{}'".format(self)
                 )
-            return cache
+            return {}
 
-        path = str(self.path_info)
-        filter_path = str(filter_info) if filter_info else None
-        for entry_key, entry_obj in self.obj:
-            entry_path = os.path.join(path, *entry_key)
-            if (
-                not filter_path
-                or entry_path == filter_path
-                or entry_path.startswith(filter_path + os.sep)
-            ):
-                cache.add(self.scheme, entry_obj.hash_info.value, entry_path)
+        obj = self.get_obj(filter_info=filter_info, copy=True)
+        self._set_obj_names(obj)
+        return {None: {obj}}
 
-        return cache
-
-    def get_used_cache(self, **kwargs):
-        """Get a dumpd of the given `out`, with an entry including the branch.
-
-        The `used_cache` of an output is no more than its `info`.
-
-        In case that the given output is a directory, it will also
-        include the `info` of its files.
-        """
+    def get_used_objs(
+        self, **kwargs
+    ) -> Dict[Optional["ObjectDB"], Set["HashFile"]]:
+        """Return filtered set of used objects for this out."""
 
         if not self.use_cache:
-            return NamedCache()
+            return {}
 
         if self.stage.is_repo_import:
-            cache = NamedCache()
-            (dep,) = self.stage.deps
-            cache.external[dep.repo_pair].add(dep.def_path)
-            return cache
+            return self.get_used_external(**kwargs)
 
         if not self.hash_info:
             msg = (
@@ -900,18 +869,32 @@ class Output:
                     )
                 )
             logger.warning(msg)
-            return NamedCache()
+            return {}
 
-        ret = NamedCache.make(self.scheme, self.hash_info.value, str(self))
+        if self.is_dir_checksum:
+            return self.collect_used_dir_cache(**kwargs)
 
-        if not self.is_dir_checksum:
-            return ret
+        obj = self.get_obj(filter_info=kwargs.get("filter_info"))
+        if not obj:
+            obj = self.odb.get(self.hash_info)
+        self._set_obj_names(obj)
 
-        ret.add_child_cache(
-            self.hash_info.value, self.collect_used_dir_cache(**kwargs)
-        )
+        return {None: {obj}}
 
-        return ret
+    def _set_obj_names(self, obj):
+        obj.name = str(self)
+        if isinstance(obj, Tree):
+            for key, entry_obj in obj:
+                entry_obj.name = self.fs.sep.join([obj.name, *key])
+
+    def get_used_external(
+        self, **kwargs
+    ) -> Dict[Optional["ObjectDB"], Set["HashFile"]]:
+        if not self.use_cache or not self.stage.is_repo_import:
+            return {}
+
+        (dep,) = self.stage.deps
+        return dep.get_used_objs()
 
     def _validate_output_path(self, path, stage=None):
         from dvc.dvcfile import is_valid_filename

dvc/progress.py

@@ -6,6 +6,7 @@ from threading import RLock
 
 from tqdm import tqdm
 
+from dvc.env import DVC_IGNORE_ISATTY
 from dvc.utils import env2bool
 
 logger = logging.getLogger(__name__)
@@ -50,7 +51,7 @@ class Tqdm(tqdm):
         file=None,
         total=None,
         postfix=None,
-        **kwargs
+        **kwargs,
     ):
         """
         bytes   : shortcut for
@@ -77,7 +78,7 @@ class Tqdm(tqdm):
         # auto-disable based on TTY
         if (
             not disable
-            and not env2bool("DVC_IGNORE_ISATTY")
+            and not env2bool(DVC_IGNORE_ISATTY)
             and hasattr(file, "isatty")
         ):
             disable = not file.isatty()
@@ -89,7 +90,7 @@ class Tqdm(tqdm):
             bar_format="!",
             lock_args=(False,),
             total=total,
-            **kwargs
+            **kwargs,
         )
         self.postfix = postfix or {"info": ""}
         if bar_format is None:
@@ -105,13 +106,16 @@ class Tqdm(tqdm):
             self.bar_format = bar_format
         self.refresh()
 
-    def update_msg(self, msg, n=1):
+    def update_msg(self, msg: str, n: int = 1) -> None:
         """
         Sets `msg` as a postfix and calls `update(n)`.
         """
-        self.postfix["info"] = " %s |" % msg
+        self.set_msg(msg)
         self.update(n)
 
+    def set_msg(self, msg: str) -> None:
+        self.postfix["info"] = f" {msg} |"
+
     def update_to(self, current, total=None):
         if total:
             self.total = total

dvc/remote/__init__.py

@@ -4,8 +4,8 @@ from .local import LocalRemote
 
 
 def get_remote(repo, **kwargs):
-    cls, config = get_cloud_fs(repo, **kwargs)
+    cls, config, path_info = get_cloud_fs(repo, **kwargs)
     fs = cls(**config)
     if fs.scheme == "local":
-        return LocalRemote(fs, repo.tmp_dir, **config)
-    return Remote(fs, repo.tmp_dir, **config)
+        return LocalRemote(fs, path_info, repo.tmp_dir, **config)
+    return Remote(fs, path_info, repo.tmp_dir, **config)

dvc/remote/base.py

@@ -8,9 +8,10 @@ from functools import partial, wraps
 
 from dvc.exceptions import DownloadError, UploadError
 from dvc.hash_info import HashInfo
+from dvc.objects.tree import Tree
 
 from ..progress import Tqdm
-from .index import RemoteIndex, RemoteIndexNoop
+from .index import RemoteIndex
 
 logger = logging.getLogger(__name__)
 
@@ -67,25 +68,39 @@ class Remote:
 
     INDEX_CLS = RemoteIndex
 
-    def __init__(self, fs, tmp_dir, **config):
+    def __init__(self, fs, path_info, tmp_dir, **config):
         from dvc.objects.db import get_odb
 
         self.fs = fs
-        self.odb = get_odb(self.fs, **config)
+        config["tmp_dir"] = tmp_dir
+        self.odb = get_odb(self.fs, path_info, **config)
 
-        url = config.get("url")
-        if url:
-            index_name = hashlib.sha256(url.encode("utf-8")).hexdigest()
-            self.index = self.INDEX_CLS(
-                tmp_dir, index_name, dir_suffix=self.fs.CHECKSUM_DIR_SUFFIX
-            )
-        else:
-            self.index = RemoteIndexNoop()
+        url = getattr(path_info, "fspath", path_info.url)
+        index_name = hashlib.sha256(url.encode("utf-8")).hexdigest()
+        self.index = self.INDEX_CLS(
+            tmp_dir, index_name, dir_suffix=self.fs.CHECKSUM_DIR_SUFFIX
+        )
 
     def __repr__(self):
         return "{class_name}: '{path_info}'".format(
             class_name=type(self).__name__,
-            path_info=self.fs.path_info or "No path",
+            path_info=self.odb.path_info or "No path",
+        )
+
+    def __eq__(self, other):
+        return self.odb == other.odb
+
+    def __hash__(self):
+        return hash(self.odb)
+
+    @classmethod
+    def from_odb(cls, odb):
+        config = odb.config
+        return cls(
+            odb.fs,
+            odb.path_info,
+            config.pop("tmp_dir"),
+            **config,
         )
 
     @index_locked
@@ -101,7 +116,7 @@ class Remote:
     def status(
         self,
         cache,
-        named_cache,
+        objs,
         jobs=None,
         show_checksums=False,
         download=False,
@@ -110,7 +125,7 @@ class Remote:
         # Return flattened dict containing all status info
         dir_status, file_status, _ = self._status(
             cache,
-            named_cache,
+            objs,
             jobs=jobs,
             show_checksums=show_checksums,
             download=download,
@@ -129,10 +144,10 @@ class Remote:
 
         return indexed_hashes + self.odb.hashes_exist(list(hashes), **kwargs)
 
-    def _status(
+    def _status(  # pylint: disable=unused-argument
         self,
         cache,
-        named_cache,
+        objs,
         jobs=None,
         show_checksums=False,
         download=False,
@@ -145,8 +160,18 @@ class Remote:
         {dir_hash: set(file_hash, ...)} which can be used to map
         a .dir file to its file contents.
         """
-        logger.debug(f"Preparing to collect status from {self.fs.path_info}")
-        md5s = set(named_cache.scheme_keys(cache.fs.scheme))
+
+        logger.debug(f"Preparing to collect status from {self.odb.path_info}")
+        md5s = set()
+        dir_objs = {}
+        for obj in objs:
+            if obj.fs.scheme != cache.fs.scheme:
+                continue
+            md5s.add(obj.hash_info.value)
+            if isinstance(obj, Tree):
+                dir_objs[obj.hash_info.value] = obj
+                for _, entry_obj in obj:
+                    md5s.add(entry_obj.hash_info.value)
 
         logger.debug("Collecting information from local cache...")
         local_exists = frozenset(
@@ -162,21 +187,18 @@ class Remote:
         else:
             logger.debug("Collecting information from remote cache...")
             remote_exists = set()
-            dir_md5s = set(named_cache.dir_keys(cache.fs.scheme))
-            if dir_md5s:
-                remote_exists.update(
-                    self._indexed_dir_hashes(cache, named_cache, dir_md5s)
-                )
+            if dir_objs:
+                remote_exists.update(self._indexed_dir_hashes(dir_objs))
                 md5s.difference_update(remote_exists)
             if md5s:
                 remote_exists.update(
                     self.hashes_exist(
-                        md5s, jobs=jobs, name=str(self.fs.path_info)
+                        md5s, jobs=jobs, name=str(self.odb.path_info)
                     )
                 )
         return self._make_status(
             cache,
-            named_cache,
+            objs,
             show_checksums,
             local_exists,
             remote_exists,
@@ -186,29 +208,36 @@ class Remote:
     def _make_status(
         self,
         cache,
-        named_cache,
+        objs,
         show_checksums,
         local_exists,
         remote_exists,
         log_missing,
     ):
-        def make_names(hash_, names):
-            return {"name": hash_ if show_checksums else " ".join(names)}
+        def make_names(obj):
+            if show_checksums or obj.name is None:
+                name = obj.hash_info.value
+            else:
+                name = obj.name
+            return {"name": name}
 
         dir_status = {}
         file_status = {}
         dir_contents = {}
-        for hash_, item in named_cache[cache.fs.scheme].items():
-            if item.children:
-                dir_status[hash_] = make_names(hash_, item.names)
+        for obj in objs:
+            if obj.fs.scheme != cache.fs.scheme:
+                continue
+
+            hash_ = obj.hash_info.value
+            if isinstance(obj, Tree):
+                dir_status[hash_] = make_names(obj)
                 dir_contents[hash_] = set()
-                for child_hash, child in item.children.items():
-                    file_status[child_hash] = make_names(
-                        child_hash, child.names
-                    )
+                for _, child_obj in obj:
+                    child_hash = child_obj.hash_info.value
+                    file_status[child_hash] = make_names(child_obj)
                     dir_contents[hash_].add(child_hash)
             else:
-                file_status[hash_] = make_names(hash_, item.names)
+                file_status[hash_] = make_names(obj)
 
         self._fill_statuses(dir_status, local_exists, remote_exists)
         self._fill_statuses(file_status, local_exists, remote_exists)
@@ -218,9 +247,10 @@ class Remote:
 
         return dir_status, file_status, dir_contents
 
-    def _indexed_dir_hashes(self, cache, named_cache, dir_md5s):
+    def _indexed_dir_hashes(self, dir_objs):
         # Validate our index by verifying all indexed .dir hashes
         # still exist on the remote
+        dir_md5s = set(dir_objs.keys())
         indexed_dirs = set(self.index.dir_hashes())
         indexed_dir_exists = set()
         if indexed_dirs:
@@ -242,9 +272,9 @@ class Remote:
         # If .dir hash exists on the remote, assume directory contents
         # still exists on the remote
         for dir_hash in dir_exists:
-            file_hashes = list(
-                named_cache.child_keys(cache.fs.scheme, dir_hash)
-            )
+            file_hashes = [
+                obj.hash_info.value for _, obj in dir_objs[dir_hash]
+            ]
             if dir_hash not in self.index:
                 logger.debug(
                     "Indexing new .dir '{}' with '{}' nested files".format(
@@ -294,7 +324,7 @@ class Remote:
     def _process(
         self,
         cache,
-        named_cache,
+        objs,
         jobs=None,
         show_checksums=False,
         download=False,
@@ -302,7 +332,7 @@ class Remote:
         logger.debug(
             "Preparing to {} '{}'".format(
                 "download data from" if download else "upload data to",
-                self.fs.path_info,
+                self.odb.path_info,
             )
         )
 
@@ -320,7 +350,7 @@ class Remote:
 
         dir_status, file_status, dir_contents = self._status(
             cache,
-            named_cache,
+            objs,
             jobs=jobs,
             show_checksums=show_checksums,
             download=download,
@@ -460,17 +490,20 @@ class Remote:
             self.index.update([dir_hash], file_hashes)
 
     @index_locked
-    def push(self, cache, named_cache, jobs=None, show_checksums=False):
+    def push(self, cache, objs, jobs=None, show_checksums=False):
         ret = self._process(
             cache,
-            named_cache,
+            objs,
             jobs=jobs,
             show_checksums=show_checksums,
             download=False,
         )
 
         if self.fs.scheme == "local":
-            for checksum in named_cache.scheme_keys("local"):
+            for obj in objs:
+                if obj.fs.scheme != "local":
+                    continue
+                checksum = obj.hash_info.value
                 cache_file = self.odb.hash_to_path_info(checksum)
                 if self.fs.exists(cache_file):
                     hash_info = HashInfo(self.fs.PARAM_CHECKSUM, checksum)
@@ -480,17 +513,19 @@ class Remote:
         return ret
 
     @index_locked
-    def pull(self, cache, named_cache, jobs=None, show_checksums=False):
+    def pull(self, cache, objs, jobs=None, show_checksums=False):
         ret = self._process(
             cache,
-            named_cache,
+            objs,
             jobs=jobs,
             show_checksums=show_checksums,
             download=True,
         )
 
         if not self.odb.verify:
-            for checksum in named_cache.scheme_keys("local"):
+
+            def save_state(obj):
+                checksum = obj.hash_info.value
                 cache_file = cache.hash_to_path_info(checksum)
                 if cache.fs.exists(cache_file):
                     # We can safely save here, as existing corrupted files
@@ -501,6 +536,14 @@ class Remote:
                     cache.state.save(cache_file, cache.fs, hash_info)
                     cache.protect(cache_file)
 
+            for obj in objs:
+                if obj.fs.scheme != "local":
+                    continue
+                save_state(obj)
+                if isinstance(obj, Tree):
+                    for _, entry_obj in obj:
+                        save_state(entry_obj)
+
         return ret
 
     @staticmethod

dvc/repo/__init__.py

@@ -1,5 +1,6 @@
 import logging
 import os
+from collections import defaultdict
 from contextlib import contextmanager
 from functools import wraps
 from typing import TYPE_CHECKING, Callable, Optional
@@ -340,7 +341,7 @@ class Repo:
             existing_stages = self.stages if old_stages is None else old_stages
             build_graph(existing_stages + new_stages)
 
-    def used_cache(
+    def used_objs(
         self,
         targets=None,
         all_branches=False,
@@ -366,13 +367,22 @@ class Repo:
         the scope.
 
         Returns:
-            A dictionary with Schemes (representing output's location) mapped
-            to items containing the output's `dumpd` names and the output's
-            children (if the given output is a directory).
+            A dict mapping (remote) ODB instances to sets of objects that
+            belong to each ODB. If the ODB instance is None, the objects
+            are naive and do not belong to a specific remote ODB.
         """
-        from dvc.objects.db import NamedCache
+        used = defaultdict(set)
 
-        cache = NamedCache()
+        def _add_suffix(objs, suffix):
+            from dvc.objects.tree import Tree
+
+            for obj in objs:
+                if obj.name is not None:
+                    obj.name += suffix
+                if isinstance(obj, Tree):
+                    for _, entry_obj in obj:
+                        if entry_obj.name is not None:
+                            entry_obj.name += suffix
 
         for branch in self.brancher(
             revs=revs,
@@ -390,23 +400,24 @@ class Repo:
                 for target in targets
             )
 
-            suffix = f"({branch})" if branch else ""
             for stage, filter_info in pairs:
-                used_cache = stage.get_used_cache(
+                for odb, objs in stage.get_used_objs(
                     remote=remote,
                     force=force,
                     jobs=jobs,
                     filter_info=filter_info,
-                )
-                cache.update(used_cache, suffix=suffix)
+                ).items():
+                    if branch:
+                        _add_suffix(objs, f" ({branch})")
+                    used[odb].update(objs)
 
         if used_run_cache:
-            used_cache = self.stage_cache.get_used_cache(
+            for odb, objs in self.stage_cache.get_used_objs(
                 used_run_cache, remote=remote, force=force, jobs=jobs
-            )
-            cache.update(used_cache)
+            ).items():
+                used[odb].update(objs)
 
-        return cache
+        return used
 
     @cached_property
     def outs_trie(self):

dvc/repo/add.py

@@ -1,9 +1,13 @@
 import logging
 import os
-from typing import TYPE_CHECKING
+from contextlib import contextmanager
+from itertools import tee
+from typing import TYPE_CHECKING, Any, Iterator, List
 
 import colorama
 
+from dvc.ui import ui
+
 from ..exceptions import (
     CacheLinkError,
     InvalidArgumentError,
@@ -11,46 +15,51 @@ from ..exceptions import (
     OverlappingOutputPathsError,
     RecursiveAddingWhileUsingFilename,
 )
-from ..progress import Tqdm
 from ..repo.scm_context import scm_context
 from ..utils import LARGE_DIR_SIZE, glob_targets, resolve_output, resolve_paths
+from ..utils.collections import ensure_list, validate
 from . import locked
 
 if TYPE_CHECKING:
+    from dvc.repo import Repo
+    from dvc.stage import Stage
     from dvc.types import TargetType
 
-
+Stages = List["Stage"]
 logger = logging.getLogger(__name__)
 
 
-@locked
-@scm_context
-def add(  # noqa: C901
-    repo,
-    targets: "TargetType",
-    recursive=False,
-    no_commit=False,
-    fname=None,
-    to_remote=False,
-    **kwargs,
-):
-    from dvc.utils.collections import ensure_list
+OVERLAPPING_OUTPUT_FMT = (
+    "Cannot add '{out}', because it is overlapping with other "
+    "DVC tracked output: '{parent}'.\n"
+    "To include '{out}' in '{parent}', run "
+    "'dvc commit {parent_stage}'"
+)
 
-    if recursive and fname:
+
+def check_recursive_and_fname(args):
+    if args.recursive and args.fname:
         raise RecursiveAddingWhileUsingFilename()
 
-    targets = ensure_list(targets)
 
-    to_cache = kwargs.get("out") and not to_remote
+def transform_targets(args):
+    args.targets = ensure_list(args.targets)
+
+
+def check_arg_combinations(args):
+    kwargs = args.kwargs
     invalid_opt = None
+    to_remote = args.to_remote
+    to_cache = kwargs.get("out") and not to_remote
+
     if to_remote or to_cache:
         message = "{option} can't be used with "
         message += "--to-remote" if to_remote else "-o"
-        if len(targets) != 1:
+        if len(args.targets) != 1:
             invalid_opt = "multiple targets"
-        elif no_commit:
+        elif args.no_commit:
             invalid_opt = "--no-commit option"
-        elif recursive:
+        elif args.recursive:
             invalid_opt = "--recursive option"
         elif kwargs.get("external"):
             invalid_opt = "--external option"
@@ -64,214 +73,192 @@ def add(  # noqa: C901
     if invalid_opt is not None:
         raise InvalidArgumentError(message.format(option=invalid_opt))
 
-    link_failures = []
-    stages_list = []
-    num_targets = len(targets)
-    with Tqdm(total=num_targets, desc="Add", unit="file", leave=True) as pbar:
-        if num_targets == 1:
-            # clear unneeded top-level progress bar for single target
-            pbar.bar_format = "Adding..."
-            pbar.refresh()
-        for target in targets:
-            sub_targets = _find_all_targets(repo, target, recursive)
-            pbar.total += len(sub_targets) - 1
 
-            if os.path.isdir(target) and len(sub_targets) > LARGE_DIR_SIZE:
-                logger.warning(
-                    "You are adding a large directory '{target}' recursively,"
-                    " consider tracking it as a whole instead.\n"
-                    "{purple}HINT:{nc} Remove the generated DVC file and then"
-                    " run `{cyan}dvc add {target}{nc}`".format(
-                        purple=colorama.Fore.MAGENTA,
-                        cyan=colorama.Fore.CYAN,
-                        nc=colorama.Style.RESET_ALL,
-                        target=target,
-                    )
-                )
-
-            stages = _create_stages(
-                repo,
-                sub_targets,
-                fname,
-                pbar=pbar,
-                transfer=to_remote or to_cache,
-                **kwargs,
-            )
-
-            try:
-                repo.check_modified_graph(stages)
-            except OverlappingOutputPathsError as exc:
-                msg = (
-                    "Cannot add '{out}', because it is overlapping with other "
-                    "DVC tracked output: '{parent}'.\n"
-                    "To include '{out}' in '{parent}', run "
-                    "'dvc commit {parent_stage}'"
-                ).format(
-                    out=exc.overlapping_out.path_info,
-                    parent=exc.parent.path_info,
-                    parent_stage=exc.parent.stage.addressing,
-                )
-                raise OverlappingOutputPathsError(
-                    exc.parent, exc.overlapping_out, msg
-                )
-            except OutputDuplicationError as exc:
-                raise OutputDuplicationError(
-                    exc.output, list(set(exc.stages) - set(stages))
-                )
-
-            link_failures.extend(
-                _process_stages(
-                    repo,
-                    sub_targets,
-                    stages,
-                    no_commit,
-                    pbar,
-                    to_remote,
-                    to_cache,
-                    **kwargs,
-                )
-            )
-            stages_list += stages
-
-        if num_targets == 1:  # restore bar format for stats
-            pbar.bar_format = pbar.BAR_FMT_DEFAULT
-
-    if link_failures:
-        msg = (
-            "Some targets could not be linked from cache to workspace.\n{}\n"
-            "To re-link these targets, reconfigure cache types and then run:\n"
-            "\n\tdvc checkout {}"
-        ).format(
-            CacheLinkError.SUPPORT_LINK,
-            " ".join([str(stage.relpath) for stage in link_failures]),
+@contextmanager
+def translate_graph_error(stages: Stages) -> Iterator[None]:
+    try:
+        yield
+    except OverlappingOutputPathsError as exc:
+        msg = OVERLAPPING_OUTPUT_FMT.format(
+            out=exc.overlapping_out.path_info,
+            parent=exc.parent.path_info,
+            parent_stage=exc.parent.stage.addressing,
+        )
+        raise OverlappingOutputPathsError(exc.parent, exc.overlapping_out, msg)
+    except OutputDuplicationError as exc:
+        raise OutputDuplicationError(
+            exc.output, list(set(exc.stages) - set(stages))
         )
-        logger.warning(msg)
-
-    return stages_list
 
 
-def _process_stages(
-    repo, sub_targets, stages, no_commit, pbar, to_remote, to_cache, **kwargs
+def progress_iter(stages: Stages) -> Iterator["Stage"]:
+    total = len(stages)
+    desc = "Adding..."
+    with ui.progress(
+        stages, total=total, desc=desc, unit="file", leave=True
+    ) as pbar:
+        if total == 1:
+            pbar.bar_format = desc
+            pbar.refresh()
+
+        for stage in pbar:
+            if total > 1:
+                pbar.set_msg(f"{stage.outs[0]}")
+            yield stage
+            if total == 1:  # restore bar format for stats
+                # pylint: disable=no-member
+                pbar.bar_format = pbar.BAR_FMT_DEFAULT
+
+
+LINK_FAILURE_MESSAGE = (
+    "\nSome targets could not be linked from cache to workspace.\n{}\n"
+    "To re-link these targets, reconfigure cache types and then run:\n"
+    "\n\tdvc checkout {}"
+)
+
+
+@contextmanager
+def warn_link_failures() -> Iterator[List[str]]:
+    link_failures: List[str] = []
+    try:
+        yield link_failures
+    finally:
+        if link_failures:
+            msg = LINK_FAILURE_MESSAGE.format(
+                CacheLinkError.SUPPORT_LINK,
+                " ".join(link_failures),
+            )
+            ui.error_write(msg)
+
+
+VALIDATORS = (
+    check_recursive_and_fname,
+    transform_targets,
+    check_arg_combinations,
+)
+
+
+@validate(*VALIDATORS)
+@locked
+@scm_context
+def add(  # noqa: C901
+    repo: "Repo",
+    targets: "TargetType",
+    recursive: bool = False,
+    no_commit: bool = False,
+    fname: str = None,
+    to_remote: bool = False,
+    **kwargs: Any,
 ):
-    link_failures = []
-    from dvc.dvcfile import Dvcfile
+    to_cache = bool(kwargs.get("out")) and not to_remote
+    transfer = to_remote or to_cache
 
-    from ..output import OutputDoesNotExistError
+    glob = kwargs.get("glob", False)
+    add_targets = collect_targets(repo, targets, recursive, glob)
+    # pass one for creating stages, other one is used for iterating here
+    add_targets, sources = tee(add_targets)
 
-    if to_remote or to_cache:
-        # Already verified in the add()
-        (stage,) = stages
-        (target,) = sub_targets
-        (out,) = stage.outs
+    # collect targets and build stages as we go
+    desc = "Collecting targets"
+    stages_it = create_stages(repo, add_targets, fname, transfer, **kwargs)
+    stages = list(ui.progress(stages_it, desc=desc, unit="file"))
 
-        odb = None
-        if to_remote:
-            remote = repo.cloud.get_remote(kwargs.get("remote"), "add")
-            odb = remote.odb
+    msg = "Collecting stages from the workspace"
+    with translate_graph_error(stages), ui.status(msg) as status:
+        # remove existing stages that are to-be replaced with these
+        # new stages for the graph checks.
+        old_stages = set(repo.stages) - set(stages)
+        status.update("Checking graph")
+        repo.check_modified_graph(stages, list(old_stages))
 
-        out.transfer(target, odb=odb, jobs=kwargs.get("jobs"))
+    odb = None
+    if to_remote:
+        remote = repo.cloud.get_remote(kwargs.get("remote"), "add")
+        odb = remote.odb
 
-        if to_cache:
-            out.checkout()
-
-        Dvcfile(repo, stage.path).dump(stage)
-        return link_failures
-
-    with Tqdm(
-        total=len(stages),
-        desc="Processing",
-        unit="file",
-        disable=len(stages) == 1,
-    ) as pbar_stages:
-        for stage in stages:
-            try:
-                stage.save()
-            except OutputDoesNotExistError:
-                pbar.n -= 1
-                raise
-
-            try:
-                if not no_commit:
-                    stage.commit()
-            except CacheLinkError:
-                link_failures.append(stage)
-
-            Dvcfile(repo, stage.path).dump(stage)
-            pbar_stages.update()
-
-    return link_failures
+    with warn_link_failures() as link_failures:
+        for stage, source in zip(progress_iter(stages), sources):
+            if to_remote or to_cache:
+                stage.transfer(source, to_remote=to_remote, odb=odb, **kwargs)
+            else:
+                try:
+                    stage.save()
+                    if not no_commit:
+                        stage.commit()
+                except CacheLinkError:
+                    link_failures.append(str(stage.relpath))
+            stage.dump()
+    return stages
 
 
-def _find_all_targets(repo, target, recursive):
+LARGE_DIR_RECURSIVE_ADD_WARNING = (
+    "You are adding a large directory '{target}' recursively.\n"
+    "Consider tracking it as a whole instead with "
+    "`{cyan}dvc add {target}{nc}`."
+)
+
+
+def collect_targets(
+    repo: "Repo",
+    targets: "TargetType",
+    recursive: bool = False,
+    glob: bool = False,
+) -> Iterator[str]:
+    for target in glob_targets(ensure_list(targets), glob=glob):
+        expanded_targets = _find_all_targets(repo, target, recursive=recursive)
+        for index, path in enumerate(expanded_targets):
+            if index == LARGE_DIR_SIZE:
+                msg = LARGE_DIR_RECURSIVE_ADD_WARNING.format(
+                    cyan=colorama.Fore.CYAN,
+                    nc=colorama.Style.RESET_ALL,
+                    target=target,
+                )
+                ui.error_write(msg)
+            yield path
+
+
+def _find_all_targets(
+    repo: "Repo", target: str, recursive: bool = False
+) -> Iterator[str]:
     from dvc.dvcfile import is_dvc_file
 
     if os.path.isdir(target) and recursive:
-        return [
-            os.fspath(path)
-            for path in Tqdm(
-                repo.dvcignore.walk_files(repo.fs, target),
-                desc="Searching " + target,
-                bar_format=Tqdm.BAR_FMT_NOTOTAL,
-                unit="file",
-            )
-            if not repo.is_dvc_internal(os.fspath(path))
-            if not is_dvc_file(os.fspath(path))
-            if not repo.scm.belongs_to_scm(os.fspath(path))
-            if not repo.scm.is_tracked(os.fspath(path))
-        ]
-    return [target]
-
-
-def _create_stages(
-    repo,
-    targets,
-    fname,
-    pbar=None,
-    external=False,
-    glob=False,
-    desc=None,
-    transfer=False,
-    **kwargs,
-):
-    from dvc.dvcfile import Dvcfile
-    from dvc.stage import Stage, create_stage, restore_meta
-
-    expanded_targets = glob_targets(targets, glob=glob)
-
-    stages = []
-    for out in Tqdm(
-        expanded_targets,
-        desc="Creating DVC files",
-        disable=len(expanded_targets) < LARGE_DIR_SIZE,
-        unit="file",
-    ):
-        if kwargs.get("out"):
-            out = resolve_output(out, kwargs["out"])
-        path, wdir, out = resolve_paths(
-            repo, out, always_local=transfer and not kwargs.get("out")
+        files = map(os.fspath, repo.dvcignore.walk_files(repo.fs, target))
+        yield from (
+            path
+            for path in files
+            if not repo.is_dvc_internal(path)
+            if not is_dvc_file(path)
+            if not repo.scm.belongs_to_scm(path)
+            if not repo.scm.is_tracked(path)
         )
-        stage = create_stage(
-            Stage,
-            repo,
-            fname or path,
+    else:
+        yield target
+
+
+def create_stages(
+    repo: "Repo",
+    targets: Iterator[str],
+    fname: str = None,
+    transfer: bool = False,
+    external: bool = False,
+    **kwargs: Any,
+) -> Iterator["Stage"]:
+    for target in targets:
+        if kwargs.get("out"):
+            target = resolve_output(target, kwargs["out"])
+        path, wdir, out = resolve_paths(
+            repo, target, always_local=transfer and not kwargs.get("out")
+        )
+
+        stage = repo.stage.create(
+            single_stage=True,
+            validate=False,
+            fname=fname or path,
             wdir=wdir,
             outs=[out],
             external=external,
         )
-        restore_meta(stage)
-        Dvcfile(repo, stage.path).remove()
-        if desc:
-            stage.outs[0].desc = desc
-
-        repo._reset()  # pylint: disable=protected-access
-
-        if not stage:
-            if pbar is not None:
-                pbar.total -= 1
-            continue
-
-        stages.append(stage)
-        if pbar is not None:
-            pbar.update_msg(out)
-
-    return stages
+        if kwargs.get("desc"):
+            stage.outs[0].desc = kwargs["desc"]
+        yield stage

dvc/repo/experiments/__init__.py

@@ -753,7 +753,7 @@ class Experiments:
         assert entry.index == 0
 
         # NOTE: the stash commit to be popped already contains all the current
-        # workspace changes plus CLI modifed --params changes.
+        # workspace changes plus CLI modified --params changes.
         # `checkout --force` here will not lose any data (popping stash commit
         # will result in conflict between workspace params and stashed CLI
         # params, but we always want the stashed version).

dvc/repo/fetch.py

@@ -1,9 +1,9 @@
 import logging
-import os
 
 from dvc.config import NoRemoteError
-from dvc.exceptions import DownloadError, NoOutputOrStageError
+from dvc.exceptions import DownloadError
 
+from ..scheme import Schemes
 from . import locked
 
 logger = logging.getLogger(__name__)
@@ -36,11 +36,12 @@ def fetch(
         config.NoRemoteError: thrown when downloading only local files and no
             remote is configured
     """
+    from dvc.objects.db.git import GitObjectDB
 
     if isinstance(targets, str):
         targets = [targets]
 
-    used = self.used_cache(
+    used = self.used_objs(
         targets,
         all_branches=all_branches,
         all_tags=all_tags,
@@ -59,17 +60,39 @@ def fetch(
     try:
         if run_cache:
             self.stage_cache.pull(remote)
-        downloaded += self.cloud.pull(
-            used, jobs, remote=remote, show_checksums=show_checksums
-        )
-    except NoRemoteError:
-        if not used.external and used["local"]:
-            raise
     except DownloadError as exc:
         failed += exc.amount
 
-    for (repo_url, repo_rev), files in used.external.items():
-        d, f = _fetch_external(self, repo_url, repo_rev, files, jobs)
+    external = set()
+    for odb, objs in used.items():
+        if odb is None:
+            # objs contains naive objects to be pulled from specified remote
+            d, f = _fetch_naive_objs(
+                self,
+                objs,
+                jobs=jobs,
+                remote=remote,
+                show_checksums=show_checksums,
+            )
+            downloaded += d
+            failed += f
+        elif isinstance(odb, GitObjectDB):
+            # objs contains staged import objects which should be saved
+            # last (after all other objects have been pulled)
+            external.update(objs)
+        else:
+            d, f = fetch_from_odb(
+                self,
+                odb,
+                objs,
+                jobs=jobs,
+                show_checksums=show_checksums,
+            )
+            downloaded += d
+            failed += f
+
+    if external:
+        d, f = _fetch_external(self, external, jobs=jobs)
         downloaded += d
         failed += f
 
@@ -79,44 +102,51 @@ def fetch(
     return downloaded
 
 
-def _fetch_external(self, repo_url, repo_rev, files, jobs):
-    from dvc.external_repo import external_repo
-    from dvc.objects import save
-    from dvc.objects.stage import stage
-    from dvc.path_info import PathInfo
-    from dvc.scm.base import CloneError
-
+def _fetch_naive_objs(repo, objs, **kwargs):
+    # objs contains naive objects to be pulled from specified remote
+    downloaded = 0
     failed = 0
+    try:
+        downloaded += repo.cloud.pull(objs, **kwargs)
+    except NoRemoteError:
+        if any(obj.fs.scheme == Schemes.LOCAL for obj in objs):
+            raise
+    except DownloadError as exc:
+        failed += exc.amount
+    return downloaded, failed
+
+
+def fetch_from_odb(repo, odb, objs, **kwargs):
+    from dvc.remote.base import Remote
+
+    downloaded = 0
+    failed = 0
+    remote = Remote.from_odb(odb)
+    try:
+        downloaded += remote.pull(
+            repo.odb.local,
+            objs,
+            **kwargs,
+        )
+    except DownloadError as exc:
+        failed += exc.amount
+    return downloaded, failed
+
+
+def _fetch_external(repo, objs, **kwargs):
+    from dvc.objects import save
+    from dvc.objects.errors import ObjectError
 
     results = []
+    failed = 0
 
-    def cb(result):
+    def callback(result):
         results.append(result)
 
-    odb = self.odb.local
-    try:
-        with external_repo(
-            repo_url, repo_rev, cache_dir=odb.cache_dir
-        ) as repo:
-            root = PathInfo(repo.root_dir)
-            for path in files:
-                path_info = root / path
-                try:
-                    used = repo.used_cache(
-                        [os.fspath(path_info)],
-                        force=True,
-                        jobs=jobs,
-                        recursive=True,
-                    )
-                    cb(repo.cloud.pull(used, jobs))
-                except (NoOutputOrStageError, NoRemoteError):
-                    pass
-                obj = stage(odb, path_info, repo.repo_fs, "md5", jobs=jobs)
-                save(odb, obj, jobs=jobs, download_callback=cb)
-    except CloneError:
-        failed += 1
-        logger.exception(
-            "failed to fetch data for '{}'".format(", ".join(files))
-        )
+    for obj in objs:
+        try:
+            save(repo.odb.local, obj, download_callback=callback, **kwargs)
+        except ObjectError:
+            failed += 1
 
     return sum(results), failed

dvc/repo/gc.py

@@ -1,11 +1,14 @@
 import logging
+from typing import TYPE_CHECKING, Set
 
 from dvc.exceptions import InvalidArgumentError
-from dvc.objects.db import NamedCache
 
 from ..scheme import Schemes
 from . import locked
 
+if TYPE_CHECKING:
+    from dvc.objects.file import HashFile
+
 logger = logging.getLogger(__name__)
 
 
@@ -52,30 +55,32 @@ def gc(
         repos = []
     all_repos = [Repo(path) for path in repos]
 
+    used_objs: Set["HashFile"] = set()
     with ExitStack() as stack:
         for repo in all_repos:
             stack.enter_context(repo.lock)
 
-        used = NamedCache()
         for repo in all_repos + [self]:
-            used.update(
-                repo.used_cache(
-                    all_branches=all_branches,
-                    with_deps=with_deps,
-                    all_tags=all_tags,
-                    all_commits=all_commits,
-                    all_experiments=all_experiments,
-                    remote=remote,
-                    force=force,
-                    jobs=jobs,
-                )
-            )
+            for objs in repo.used_objs(
+                all_branches=all_branches,
+                with_deps=with_deps,
+                all_tags=all_tags,
+                all_commits=all_commits,
+                all_experiments=all_experiments,
+                remote=remote,
+                force=force,
+                jobs=jobs,
+            ).values():
+                used_objs.update(objs)
 
     for scheme, odb in self.odb.by_scheme():
         if not odb:
             continue
 
-        removed = odb.gc(set(used.scheme_keys(scheme)), jobs=jobs)
+        removed = odb.gc(
+            {obj for obj in used_objs if obj.fs.scheme == scheme},
+            jobs=jobs,
+        )
         if not removed:
             logger.info(f"No unused '{scheme}' cache to remove.")
 
@@ -83,6 +88,9 @@ def gc(
         return
 
     remote = self.cloud.get_remote(remote, "gc -c")
-    removed = remote.gc(set(used.scheme_keys(Schemes.LOCAL)), jobs=jobs)
+    removed = remote.gc(
+        {obj for obj in used_objs if obj.fs.scheme == Schemes.LOCAL},
+        jobs=jobs,
+    )
     if not removed:
         logger.info("No unused cache to remove from remote.")

dvc/repo/ls.py

@@ -28,7 +28,7 @@ def ls(url, path=None, rev=None, recursive=None, dvc_only=False):
     """
     from . import Repo
 
-    with Repo.open(url, rev, subrepos=True, uninitialized=True) as repo:
+    with Repo.open(url, rev=rev, subrepos=True, uninitialized=True) as repo:
         path_info = PathInfo(repo.root_dir)
         if path:
             path_info /= path

dvc/repo/move.py

@@ -62,7 +62,11 @@ def move(self, from_path, to_path):
         )
         to_path = os.path.relpath(to_path, new_wdir)
         new_stage = self.stage.create(
-            single_stage=True, fname=new_fname, wdir=new_wdir, outs=[to_path]
+            single_stage=True,
+            fname=new_fname,
+            wdir=new_wdir,
+            outs=[to_path],
+            meta=stage.meta,
         )
 
         os.unlink(stage.path)

dvc/repo/push.py

@@ -24,7 +24,7 @@ def push(
 
     expanded_targets = glob_targets(targets, glob=glob)
 
-    used = self.used_cache(
+    used = self.used_objs(
         expanded_targets,
         all_branches=all_branches,
         all_tags=all_tags,
@@ -38,4 +38,8 @@ def push(
         revs=revs,
     )
 
-    return len(used_run_cache) + self.cloud.push(used, jobs, remote=remote)
+    pushed = len(used_run_cache)
+    for odb, objs in used.items():
+        if odb is None:
+            pushed += self.cloud.push(objs, jobs, remote=remote)
+    return pushed

dvc/repo/stage.py

@@ -462,7 +462,8 @@ class StageLoad:
 
         def is_out_or_ignored(root, directory):
             dir_path = f"{root}{sep}{directory}"
-            return dir_path in outs or is_ignored(dir_path)
+            # trailing slash needed to check if a directory is gitignored
+            return dir_path in outs or is_ignored(f"{dir_path}{sep}")
 
         stages = []
         for root, dirs, files in self.repo.dvcignore.walk(

dvc/repo/status.py

@@ -78,7 +78,7 @@ def _cloud_status(
     """
     import dvc.remote.base as cloud
 
-    used = self.used_cache(
+    used = self.used_objs(
         targets,
         all_branches=all_branches,
         all_tags=all_tags,
@@ -91,22 +91,26 @@ def _cloud_status(
     )
 
     ret = {}
-    status_info = self.cloud.status(
-        used, jobs, remote=remote, log_missing=False
-    )
-    for info in status_info.values():
-        name = info["name"]
-        status_ = info["status"]
-        if status_ == cloud.STATUS_OK:
+    for odb, objs in used.items():
+        if odb is not None:
+            # ignore imported objects
             continue
+        status_info = self.cloud.status(
+            objs, jobs, remote=remote, log_missing=False
+        )
+        for info in status_info.values():
+            name = info["name"]
+            status_ = info["status"]
+            if status_ == cloud.STATUS_OK:
+                continue
 
-        prefix_map = {
-            cloud.STATUS_DELETED: "deleted",
-            cloud.STATUS_NEW: "new",
-            cloud.STATUS_MISSING: "missing",
-        }
+            prefix_map = {
+                cloud.STATUS_DELETED: "deleted",
+                cloud.STATUS_NEW: "new",
+                cloud.STATUS_MISSING: "missing",
+            }
 
-        ret[name] = prefix_map[status_]
+            ret[name] = prefix_map[status_]
 
     return ret
 

dvc/scm/git/backend/base.py

@@ -187,7 +187,7 @@ class BaseGitBackend(ABC):
 
     @abstractmethod
     def get_refs_containing(self, rev: str, pattern: Optional[str] = None):
-        """Iterate over all git refs containing the specfied revision."""
+        """Iterate over all git refs containing the specified revision."""
 
     @abstractmethod
     def push_refspec(

dvc/scm/git/backend/dulwich.py

@@ -27,6 +27,8 @@ from ..objects import GitObject
 from .base import BaseGitBackend
 
 if TYPE_CHECKING:
+    from dvc.types import StrPath
+
     from ..objects import GitCommit
 
 logger = logging.getLogger(__name__)
@@ -278,12 +280,14 @@ class DulwichBackend(BaseGitBackend):  # pylint:disable=abstract-method
 
         return IgnoreFilterManager.from_repo(self.repo)
 
-    def is_ignored(self, path: str) -> bool:
+    def is_ignored(self, path: "StrPath") -> bool:
         # `is_ignored` returns `false` if excluded in `.gitignore` and
         # `None` if it's not mentioned at all. `True` if it is ignored.
-        return bool(
-            self.ignore_manager.is_ignored(relpath(path, self.root_dir))
-        )
+        relative_path = relpath(path, self.root_dir)
+        # if checking a directory, a trailing slash must be included
+        if str(path)[-1] == os.sep:
+            relative_path += os.sep
+        return bool(self.ignore_manager.is_ignored(relative_path))
 
     def set_ref(
         self,

dvc/scm/git/backend/gitpython.py

@@ -3,7 +3,16 @@ import locale
 import logging
 import os
 from functools import partial
-from typing import Callable, Iterable, List, Mapping, Optional, Tuple, Union
+from typing import (
+    TYPE_CHECKING,
+    Callable,
+    Iterable,
+    List,
+    Mapping,
+    Optional,
+    Tuple,
+    Union,
+)
 
 from funcy import ignore
 
@@ -14,6 +23,9 @@ from dvc.utils import fix_env, is_binary, relpath
 from ..objects import GitCommit, GitObject
 from .base import BaseGitBackend
 
+if TYPE_CHECKING:
+    from dvc.types import StrPath
+
 logger = logging.getLogger(__name__)
 
 
@@ -111,11 +123,11 @@ class GitPythonBackend(BaseGitBackend):  # pylint:disable=abstract-method
     def git(self):
         return self.repo.git
 
-    def is_ignored(self, path: str) -> bool:
+    def is_ignored(self, path: "StrPath") -> bool:
         from git.exc import GitCommandError
 
         func = ignore(GitCommandError)(self.repo.git.check_ignore)
-        return bool(func(path))
+        return bool(func(str(path)))
 
     @property
     def root_dir(self) -> str:

dvc/scm/git/backend/pygit2.py

@@ -4,7 +4,16 @@ import os
 import stat
 from contextlib import contextmanager
 from io import BytesIO, StringIO
-from typing import Callable, Iterable, List, Mapping, Optional, Tuple, Union
+from typing import (
+    TYPE_CHECKING,
+    Callable,
+    Iterable,
+    List,
+    Mapping,
+    Optional,
+    Tuple,
+    Union,
+)
 
 from funcy import cached_property
 
@@ -17,6 +26,10 @@ from .base import BaseGitBackend
 logger = logging.getLogger(__name__)
 
 
+if TYPE_CHECKING:
+    from dvc.types import StrPath
+
+
 class Pygit2Object(GitObject):
     def __init__(self, obj):
         self.obj = obj
@@ -81,6 +94,15 @@ class Pygit2Backend(BaseGitBackend):  # pylint:disable=abstract-method
 
         return RefdbFsBackend(self.repo)
 
+    def _resolve_refish(self, refish: str):
+        from pygit2 import GIT_OBJ_COMMIT, Tag
+
+        commit, ref = self.repo.resolve_refish(refish)
+        if isinstance(commit, Tag):
+            ref = commit
+            commit = commit.peel(GIT_OBJ_COMMIT)
+        return commit, ref
+
     @property
     def default_signature(self):
         try:
@@ -140,7 +162,7 @@ class Pygit2Backend(BaseGitBackend):  # pylint:disable=abstract-method
                 if branch == "-":
                     branch = "@{-1}"
                 try:
-                    commit, ref = self.repo.resolve_refish(branch)
+                    commit, ref = self._resolve_refish(branch)
                 except (KeyError, GitError):
                     raise RevError(f"unknown Git revision '{branch}'")
                 self.repo.checkout_tree(commit, strategy=checkout_strategy)
@@ -200,7 +222,7 @@ class Pygit2Backend(BaseGitBackend):  # pylint:disable=abstract-method
         from pygit2 import GitError
 
         try:
-            commit, _ref = self.repo.resolve_refish(rev)
+            commit, _ref = self._resolve_refish(rev)
             return str(commit.id)
         except (KeyError, GitError):
             pass
@@ -220,7 +242,7 @@ class Pygit2Backend(BaseGitBackend):  # pylint:disable=abstract-method
         from pygit2 import GitError
 
         try:
-            commit, _ref = self.repo.resolve_refish(rev)
+            commit, _ref = self._resolve_refish(rev)
         except (KeyError, GitError):
             raise SCMError(f"Invalid commit '{rev}'")
         return GitCommit(
@@ -234,7 +256,7 @@ class Pygit2Backend(BaseGitBackend):  # pylint:disable=abstract-method
     def _get_stash(self, ref: str):
         raise NotImplementedError
 
-    def is_ignored(self, path: str) -> bool:
+    def is_ignored(self, path: "StrPath") -> bool:
         rel = relpath(path, self.root_dir)
         if os.name == "nt":
             rel.replace("\\", "/")
@@ -294,12 +316,12 @@ class Pygit2Backend(BaseGitBackend):  # pylint:disable=abstract-method
         from pygit2 import GitError
 
         def _contains(repo, ref, search_commit):
-            commit, _ref = self.repo.resolve_refish(ref)
+            commit, _ref = self._resolve_refish(ref)
             base = repo.merge_base(search_commit.id, commit.id)
             return base == search_commit.id
 
         try:
-            search_commit, _ref = self.repo.resolve_refish(rev)
+            search_commit, _ref = self._resolve_refish(rev)
         except (KeyError, GitError):
             raise SCMError(f"Invalid rev '{rev}'")
 
@@ -384,7 +406,7 @@ class Pygit2Backend(BaseGitBackend):  # pylint:disable=abstract-method
 
         # libgit2 stash apply only accepts refs/stash items by index. If rev is
         # not in refs/stash, we will push it onto the stash, and then pop it
-        commit, _ref = self.repo.resolve_refish(rev)
+        commit, _ref = self._resolve_refish(rev)
         stash = self.repo.references.get(Stash.DEFAULT_STASH)
         if stash:
             for i, entry in enumerate(stash.log()):

dvc/stage/__init__.py

@@ -4,7 +4,7 @@ import string
 from collections import defaultdict
 from contextlib import suppress
 from dataclasses import dataclass
-from typing import TYPE_CHECKING, Dict, Optional
+from typing import TYPE_CHECKING, Any, Dict, Optional, Set
 
 from funcy import cached_property, project
 
@@ -36,6 +36,8 @@ from .utils import (
 
 if TYPE_CHECKING:
     from dvc.dvcfile import DVCFile
+    from dvc.objects.db.base import ObjectDB
+    from dvc.objects.file import HashFile
 
 logger = logging.getLogger(__name__)
 # Disallow all punctuation characters except hyphen and underscore
@@ -388,6 +390,19 @@ class Stage(params.StageParams):
         if purge:
             self.dvcfile.remove_stage(self)
 
+    def transfer(
+        self,
+        source: str,
+        odb: "ObjectDB" = None,
+        to_remote: bool = False,
+        **kwargs: Any,
+    ) -> None:
+        assert len(self.outs) == 1
+        (out,) = self.outs
+        out.transfer(source, odb=odb, jobs=kwargs.get("jobs"))
+        if not to_remote:
+            out.checkout()
+
     @rwlocked(read=["deps"], write=["outs"])
     def reproduce(self, interactive=False, **kwargs):
         if not (kwargs.get("force", False) or self.changed()):
@@ -622,14 +637,15 @@ class Stage(params.StageParams):
             for out in self.filter_outs(filter_info)
         )
 
-    def get_used_cache(self, *args, **kwargs):
-        from dvc.objects.db import NamedCache
-
-        cache = NamedCache()
+    def get_used_objs(
+        self, *args, **kwargs
+    ) -> Dict[Optional["ObjectDB"], Set["HashFile"]]:
+        """Return set of objects used by this stage."""
+        used_objs = defaultdict(set)
         for out in self.filter_outs(kwargs.get("filter_info")):
-            cache.update(out.get_used_cache(*args, **kwargs))
-
-        return cache
+            for odb, objs in out.get_used_objs(*args, **kwargs).items():
+                used_objs[odb].update(objs)
+        return used_objs
 
     @staticmethod
     def _check_can_merge(stage, ancestor_out=None):

dvc/stage/cache.py

@@ -196,7 +196,8 @@ class StageCache:
         cached_stage = self._create_stage(cache, wdir=stage.wdir)
 
         if pull:
-            self.repo.cloud.pull(cached_stage.get_used_cache())
+            for objs in cached_stage.get_used_objs().values():
+                self.repo.cloud.pull(objs)
 
         if not cached_stage.outs_cached():
             raise RunCacheNotFoundError(stage)
@@ -212,15 +213,17 @@ class StageCache:
         ret = []
 
         runs = from_remote.path_info / "runs"
-        if not from_remote.exists(runs):
+        if not from_remote.fs.exists(runs):
             return []
 
-        for src in from_remote.walk_files(runs):
+        for src in from_remote.fs.walk_files(runs):
             rel = src.relative_to(from_remote.path_info)
             dst = to_remote.path_info / rel
             key = dst.parent
             # check if any build cache already exists for this key
-            if to_remote.exists(key) and first(to_remote.walk_files(key)):
+            if to_remote.fs.exists(key) and first(
+                to_remote.fs.walk_files(key)
+            ):
                 continue
             func(src, dst)
             ret.append((src.parent.name, src.name))
@@ -233,8 +236,8 @@ class StageCache:
         remote = self.repo.cloud.get_remote(remote)
         return self._transfer(
             _log_exceptions(remote.fs.upload, "upload"),
-            self.repo.odb.local.fs,
-            remote.fs,
+            self.repo.odb.local,
+            remote.odb,
         )
 
     def pull(self, remote):
@@ -243,19 +246,20 @@ class StageCache:
         remote = self.repo.cloud.get_remote(remote)
         return self._transfer(
             _log_exceptions(remote.fs.download, "download"),
-            remote.fs,
-            self.repo.odb.local.fs,
+            remote.odb,
+            self.repo.odb.local,
         )
 
-    def get_used_cache(self, used_run_cache, *args, **kwargs):
-        from dvc.objects.db import NamedCache
-
-        cache = NamedCache()
+    def get_used_objs(self, used_run_cache, *args, **kwargs):
+        """Return used cache for the specified run-cached stages."""
+        from collections import defaultdict
 
+        used_objs = defaultdict(set)
         for key, value in used_run_cache:
             entry = self._load_cache(key, value)
             if not entry:
                 continue
             stage = self._create_stage(entry)
-            cache.update(stage.get_used_cache(*args, **kwargs))
-        return cache
+            for odb, objs in stage.get_used_objs(*args, **kwargs).items():
+                used_objs[odb].update(objs)
+        return used_objs

dvc/stage/monitor.py

@@ -75,6 +75,7 @@ class CheckpointTask(MonitorTask):
     def _run_callback(stage, callback_func):
         stage.save(allow_missing=True)
         stage.commit(allow_missing=True)
+        stage.unprotect_outs()
         logger.debug("Running checkpoint callback for stage '%s'", stage)
         callback_func()
 

dvc/ui/__init__.py

@@ -1,5 +1,4 @@
 import sys
-from collections import defaultdict
 from typing import (
     TYPE_CHECKING,
     Any,
@@ -7,20 +6,23 @@ from typing import (
     Iterable,
     Optional,
     Sequence,
+    TextIO,
     Union,
 )
 
 from funcy import cached_property
 
-from dvc.progress import Tqdm
-from dvc.utils import colorize
-
 if TYPE_CHECKING:
+    from rich.status import Status
+
+    from dvc.progress import Tqdm
     from dvc.ui.table import Headers, Styles, TableData
 
 
 class Formatter:
     def __init__(self, theme: Dict = None, defaults: Dict = None) -> None:
+        from collections import defaultdict
+
         theme = theme or {
             "success": {"color": "green", "style": "bold"},
             "warn": {"color": "yellow"},
@@ -29,6 +31,8 @@ class Formatter:
         self.theme = defaultdict(lambda: defaults or {}, theme)
 
     def format(self, message: str, style: str = None, **kwargs) -> str:
+        from dvc.utils import colorize
+
         return colorize(message, **self.theme[style])
 
 
@@ -75,22 +79,30 @@ class Console:
         stderr: bool = False,
         force: bool = False,
         styled: bool = False,
+        file: TextIO = None,
     ) -> None:
+        from dvc.progress import Tqdm
+
         sep = " " if sep is None else sep
         end = "\n" if end is None else end
         if not self._enabled and not force:
             return
 
-        if styled:
-            console = self.error_console if stderr else self.rich_console
-            return console.print(*objects, sep=sep, end=end)
+        file = file or (sys.stderr if stderr else sys.stdout)
+        with Tqdm.external_write_mode(file=file):
+            if styled:
+                console = self.error_console if stderr else self.rich_console
+                return console.print(*objects, sep=sep, end=end)
 
-        file = sys.stderr if stderr else sys.stdout
-        values = (self.formatter.format(obj, style=style) for obj in objects)
-        return print(*values, sep=sep, end=end, file=file)
+            values = (
+                self.formatter.format(obj, style=style) for obj in objects
+            )
+            return print(*values, sep=sep, end=end, file=file)
 
     @staticmethod
-    def progress(*args, **kwargs) -> Tqdm:
+    def progress(*args, **kwargs) -> "Tqdm":
+        from dvc.progress import Tqdm
+
         return Tqdm(*args, **kwargs)
 
     def prompt(
@@ -176,6 +188,9 @@ class Console:
             self, data, headers, markdown=markdown, pager=pager, force=force
         )
 
+    def status(self, status: str, **kwargs: Any) -> "Status":
+        return self.error_console.status(status, **kwargs)
+
 
 ui = Console()
 

dvc/utils/__init__.py

@@ -319,10 +319,19 @@ def relpath(path, start=os.curdir):
     start = os.path.abspath(os.fspath(start))
 
     # Windows path on different drive than curdir doesn't have relpath
-    if os.name == "nt" and not os.path.commonprefix(
-        [start, os.path.abspath(path)]
-    ):
-        return path
+    if os.name == "nt":
+        # Since python 3.8 os.realpath resolves network shares to their UNC
+        # path. So, to be certain that relative paths correctly captured,
+        # we need to resolve to UNC path first. We resolve only the drive
+        # name so that we don't follow any 'real' symlinks on the path
+        def resolve_network_drive_windows(path_to_resolve):
+            drive, tail = os.path.splitdrive(path_to_resolve)
+            return os.path.join(os.path.realpath(drive), tail)
+
+        path = resolve_network_drive_windows(os.path.abspath(path))
+        start = resolve_network_drive_windows(start)
+        if not os.path.commonprefix([start, path]):
+            return path
     return os.path.relpath(path, start)
 
 

dvc/utils/collections.py

@@ -1,6 +1,8 @@
+import inspect
 import os
 from collections.abc import Mapping
-from typing import Dict, Iterable, List, TypeVar, Union
+from functools import wraps
+from typing import Callable, Dict, Iterable, List, TypeVar, Union
 
 from pygtrie import StringTrie as _StringTrie
 
@@ -76,3 +78,59 @@ def chunk_dict(d: Dict[_KT, _VT], size: int = 1) -> List[Dict[_KT, _VT]]:
     from funcy import chunks
 
     return [{key: d[key] for key in chunk} for chunk in chunks(size, d)]
+
+
+class _NamespacedDict(dict):
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.__dict__ = self
+
+
+def validate(*validators: Callable, post: bool = False):
+    """
+    Validate and transform arguments and results from function calls.
+
+    The validators functions are passed a dictionary of arguments, which
+    supports dot notation access too.
+
+    The key is derived from the function signature, and hence is the name of
+    the argument, whereas the value is the one passed to the function
+    (if it is not passed, default value from keyword arguments are provided).
+
+    >>> def validator(args):
+    ...    assert args["l"] >= 0 and args.b >= 0 and args.h >= 0
+
+    >>> @validate(validator)
+    ... def cuboid_area(l, b, h=1):
+    ...   return 2*(l*b + l*h + b*h)
+
+    >>> cuboid_area(5, 20)
+    250
+    >>> cuboid_area(-1, -2)
+    Traceback (most recent call last):
+      ...
+    AssertionError
+    """
+
+    def wrapped(func: Callable):
+        sig = inspect.signature(func)
+
+        @wraps(func)
+        def inner(*args, **kwargs):
+            ba = sig.bind(*args, **kwargs)
+            ba.apply_defaults()
+            ba.arguments = _NamespacedDict(ba.arguments)
+
+            if not post:
+                for validator in validators:
+                    validator(ba.arguments)
+
+            result = func(*ba.args, **ba.kwargs)
+            if post:
+                for validator in validators:
+                    result = validator(result)
+            return result
+
+        return inner
+
+    return wrapped

dvc/utils/fs.py

@@ -4,11 +4,15 @@ import os
 import shutil
 import stat
 import sys
+from typing import TYPE_CHECKING
 
 from dvc.exceptions import DvcException
 from dvc.system import System
 from dvc.utils import dict_md5
 
+if TYPE_CHECKING:
+    from dvc.types import StrPath
+
 logger = logging.getLogger(__name__)
 
 LOCAL_CHUNK_SIZE = 2 ** 20  # 1 MB
@@ -73,9 +77,9 @@ class BasePathNotInCheckedPathException(DvcException):
         super().__init__(msg)
 
 
-def contains_symlink_up_to(path, base_path):
-    base_path = os.fspath(base_path)
-    path = os.fspath(path)
+def contains_symlink_up_to(path: "StrPath", base_path: "StrPath"):
+    base_path = os.path.normcase(os.fspath(base_path))
+    path = os.path.normcase(os.fspath(path))
 
     if base_path not in path:
         raise BasePathNotInCheckedPathException(path, base_path)
@@ -144,11 +148,11 @@ def remove(path):
             raise
 
 
-def path_isin(child, parent):
+def path_isin(child: "StrPath", parent: "StrPath"):
     """Check if given `child` path is inside `parent`."""
 
-    def normalize_path(path):
-        return os.path.normpath(path)
+    def normalize_path(path) -> str:
+        return os.path.normcase(os.path.normpath(path))
 
     parent = os.path.join(normalize_path(parent), "")
     child = normalize_path(child)

dvc/utils/serialize/_common.py

@@ -3,6 +3,7 @@ import os
 from contextlib import contextmanager
 from typing import TYPE_CHECKING, Any, Callable, ContextManager, Dict, Union
 
+from funcy import reraise
 from typing_extensions import Protocol
 
 from dvc.exceptions import DvcException
@@ -42,13 +43,24 @@ class ParseError(DvcException):
         from dvc.utils import relpath
 
         path = relpath(path)
+        self.path = path
         super().__init__(f"unable to read: '{path}', {message}")
 
 
+class EncodingError(ParseError):
+    """We could not read a file with the given encoding"""
+
+    def __init__(self, path: "AnyPath", encoding: str):
+        self.encoding = encoding
+        super().__init__(path, f"is not valid {encoding}")
+
+
 def _load_data(path: "AnyPath", parser: ParserFn, fs: "BaseFileSystem" = None):
     open_fn = fs.open if fs else open
-    with open_fn(path, encoding="utf-8") as fd:  # type: ignore
-        return parser(fd.read(), path)
+    encoding = "utf-8"
+    with open_fn(path, encoding=encoding) as fd:  # type: ignore
+        with reraise(UnicodeDecodeError, EncodingError(path, encoding)):
+            return parser(fd.read(), path)
 
 
 def _dump_data(path, data: Any, dumper: DumperFn, fs: "BaseFileSystem" = None):

dvc/utils/stream.py

@@ -70,6 +70,7 @@ class HashedStreamReader(io.IOBase):
     def __init__(self, fobj):
         self.fobj = fobj
         self.md5 = hashlib.md5()
+        self.total_read = 0
         self.is_text_file = None
         super().__init__()
 
@@ -95,9 +96,12 @@ class HashedStreamReader(io.IOBase):
         else:
             data = chunk
         self.md5.update(data)
+        self.total_read += len(data)
 
         return chunk
 
     @property
     def hash_info(self):
-        return HashInfo(self.PARAM_CHECKSUM, self.md5.hexdigest(), nfiles=1)
+        return HashInfo(
+            self.PARAM_CHECKSUM, self.md5.hexdigest(), size=self.total_read
+        )

dvc/version.py

@@ -6,7 +6,7 @@
 import os
 import subprocess
 
-_BASE_VERSION = "2.3.0"
+_BASE_VERSION = "2.5.2"
 
 
 def _generate_version(base_version):

scripts/innosetup/setup.iss

@@ -34,7 +34,7 @@ Source: "{#MyAppDir}\*"; DestDir: "{app}"; Flags: ignoreversion recursesubdirs c
 
 [Tasks]
 Name: modifypath; Description: Adds dvc's application directory to environmental path; Flags: checkablealone;
-Name: modifypath\system; Description: Adds dvc's application directory to enviromental path for all users;
+Name: modifypath\system; Description: Adds dvc's application directory to environmental path for all users;
 Name: addsymlinkpermissions; Description: Add permission for creating symbolic links; Flags: checkablealone;
 Name: addsymlinkpermissions\system; Description: Add permissions for creating symbolic links for all users;
 

setup.py

@@ -78,9 +78,10 @@ install_requires = [
     "pydot>=1.2.4",
     "speedcopy>=2.0.1; python_version < '3.8' and sys_platform == 'win32'",
     "dataclasses==0.7; python_version < '3.7'",
+    "importlib-metadata==1.4; python_version < '3.8'",
     "flatten_dict>=0.3.0,<1",
     "tabulate>=0.8.7",
-    "pygtrie==2.3.2",
+    "pygtrie>=2.3.2",
     "dpath>=2.0.1,<3",
     "shtab>=1.3.4,<2",
     "rich>=10.0.0",
@@ -88,16 +89,16 @@ install_requires = [
     "python-benedict>=0.21.1",
     "pyparsing==2.4.7",
     "typing_extensions>=3.7.4",
-    "fsspec==2021.5.0",
+    "fsspec>=2021.6.1",
     "diskcache>=5.2.1",
 ]
 
 
 # Extra dependencies for remote integrations
 
-gs = ["gcsfs==2021.5.0"]
+gs = ["gcsfs==2021.6.1"]
 gdrive = ["pydrive2>=1.8.1", "six >= 1.13.0"]
-s3 = ["s3fs==2021.5.0", "aiobotocore[boto3]==1.3.0"]
+s3 = ["s3fs==2021.6.1", "aiobotocore[boto3]==1.3.0"]
 azure = ["adlfs==0.7.1", "azure-identity>=1.4.0", "knack"]
 # https://github.com/Legrandin/pycryptodome/issues/465
 oss = ["oss2==2.6.1", "pycryptodome>=3.10"]
@@ -106,7 +107,7 @@ ssh = ["paramiko[invoke]>=2.7.0"]
 # Remove the env marker if/when pyarrow is available for Python3.9
 hdfs = ["pyarrow>=2.0.0"]
 webhdfs = ["hdfs==2.5.8"]
-webdav = ["webdavclient3>=3.14.5"]
+webdav = ["webdav4>=0.8.1"]
 # gssapi should not be included in all_remotes, because it doesn't have wheels
 # for linux and mac, so it will fail to compile if user doesn't have all the
 # requirements, including kerberos itself. Once all the wheels are available,

test_requirements.txt

@@ -3,11 +3,11 @@ wheel==0.36.2
 # Test requirements
 pytest==6.2.4
 pytest-cov==2.12.1
-pytest-xdist==2.2.1
+pytest-xdist==2.3.0
 pytest-mock==3.6.1
 pytest-lazy-fixture==0.6.3
 pytest-timeout==1.4.2
-pytest-docker==0.10.1
+pytest-docker==0.10.3
 
 flaky==3.7.0
 mock==4.0.3
@@ -17,12 +17,12 @@ beautifulsoup4==4.9.3
 wget==3.2
 filelock==3.0.12
 wsgidav==3.1.0
-crc32c==2.2
+crc32c==2.2.post0
 xmltodict==0.12.0
 google-compute-engine==2.8.13
-google-cloud-storage==1.38.0
+google-cloud-storage==1.40.0
 
-urllib3==1.26.5
+urllib3==1.26.6
 
 # required by collective.checkdocs
 Pygments==2.9.0
@@ -39,7 +39,11 @@ pylint-pytest==1.0.3
 pylint-plugin-utils==0.6
 
 # type-checking
-mypy==0.812
+mypy==0.910
+types-requests
+types-paramiko
+types-tabulate
+types-toml
 
 # pypi doesn't allow for direct dependencies
 # gdrivefs @ git+https://github.com/intake/gdrivefs.git

tests/func/experiments/test_checkpoints.py

@@ -7,9 +7,13 @@ from dvc.repo.experiments.base import EXEC_APPLY, EXEC_CHECKPOINT
 
 
 @pytest.mark.parametrize("workspace", [True, False])
+@pytest.mark.parametrize("links", ["reflink,copy", "hardlink,symlink"])
 def test_new_checkpoint(
-    tmp_dir, scm, dvc, checkpoint_stage, mocker, workspace
+    tmp_dir, scm, dvc, checkpoint_stage, mocker, workspace, links
 ):
+    with dvc.config.edit() as conf:
+        conf["cache"]["type"] = links
+
     new_mock = mocker.spy(dvc.experiments, "new")
     results = dvc.experiments.run(
         checkpoint_stage.addressing, params=["foo=2"], tmp_dir=not workspace

tests/func/plots/test_show.py

@@ -571,7 +571,7 @@ def test_raise_on_wrong_field(tmp_dir, scm, dvc, run_copy_metrics):
 
 
 def test_load_metric_from_dict_json(tmp_dir):
-    metric = [{"acccuracy": 1, "loss": 2}, {"accuracy": 3, "loss": 4}]
+    metric = [{"accuracy": 1, "loss": 2}, {"accuracy": 3, "loss": 4}]
     dmetric = {"train": metric}
 
     plot_data = JSONPlotData("-", "revision", json.dumps(dmetric))
@@ -584,7 +584,7 @@ def test_load_metric_from_dict_json(tmp_dir):
 
 
 def test_load_metric_from_dict_yaml(tmp_dir):
-    metric = [{"acccuracy": 1, "loss": 2}, {"accuracy": 3, "loss": 4}]
+    metric = [{"accuracy": 1, "loss": 2}, {"accuracy": 3, "loss": 4}]
     dmetric = {"train": metric}
 
     plot_data = YAMLPlotData("-", "revision", dumps_yaml(dmetric))
@@ -683,17 +683,18 @@ def test_show_no_repo(tmp_dir):
     dvc.plots.show(["metric.json"])
 
 
-def test_show_from_subdir(tmp_dir, dvc, caplog):
+def test_show_from_subdir(tmp_dir, dvc, capsys):
     subdir = tmp_dir / "subdir"
 
     subdir.mkdir()
     metric = [{"first_val": 100, "val": 2}, {"first_val": 200, "val": 3}]
     _write_json(subdir, metric, "metric.json")
 
-    with subdir.chdir(), caplog.at_level(logging.INFO, "dvc"):
+    with subdir.chdir():
         assert main(["plots", "show", "metric.json"]) == 0
 
-    assert subdir.as_uri() in caplog.text
+    out, _ = capsys.readouterr()
+    assert subdir.as_uri() in out
     assert (subdir / "plots.html").exists()
 
 

tests/func/test_add.py

@@ -1,6 +1,5 @@
 import errno
 import filecmp
-import logging
 import os
 import shutil
 import stat
@@ -24,7 +23,11 @@ from dvc.fs.local import LocalFileSystem
 from dvc.hash_info import HashInfo
 from dvc.main import main
 from dvc.objects.db import ODBManager
-from dvc.output import OutputAlreadyTrackedError, OutputIsStageFileError
+from dvc.output import (
+    OutputAlreadyTrackedError,
+    OutputDoesNotExistError,
+    OutputIsStageFileError,
+)
 from dvc.stage import Stage
 from dvc.stage.exceptions import (
     StageExternalOutputsError,
@@ -127,14 +130,12 @@ class TestAddCmdDirectoryRecursive(TestDvc):
 
     def test_warn_about_large_directories(self):
         warning = (
-            "You are adding a large directory 'large-dir' recursively,"
-            " consider tracking it as a whole instead.\n"
-            "{purple}HINT:{nc} Remove the generated DVC file and then"
-            " run `{cyan}dvc add large-dir{nc}`".format(
-                purple=colorama.Fore.MAGENTA,
-                cyan=colorama.Fore.CYAN,
-                nc=colorama.Style.RESET_ALL,
-            )
+            "You are adding a large directory 'large-dir' recursively."
+            "\nConsider tracking it as a whole instead with "
+            "`{cyan}dvc add large-dir{nc}`"
+        ).format(
+            cyan=colorama.Fore.CYAN,
+            nc=colorama.Style.RESET_ALL,
         )
 
         os.mkdir("large-dir")
@@ -145,9 +146,8 @@ class TestAddCmdDirectoryRecursive(TestDvc):
             with open(path, "w") as fobj:
                 fobj.write(path)
 
-        with self._caplog.at_level(logging.WARNING, logger="dvc"):
-            assert main(["add", "--recursive", "large-dir"]) == 0
-            assert warning in self._caplog.messages
+        assert main(["add", "--recursive", "large-dir"]) == 0
+        assert warning in self._capsys.readouterr()[1]
 
 
 class TestAddDirectoryWithForwardSlash(TestDvc):
@@ -639,7 +639,7 @@ def test_failed_add_cleanup(tmp_dir, scm, dvc):
 
 
 def test_should_not_track_git_internal_files(mocker, dvc, tmp_dir):
-    stage_creator_spy = mocker.spy(dvc_module.repo.add, "_create_stages")
+    stage_creator_spy = mocker.spy(dvc_module.repo.add, "create_stages")
 
     ret = main(["add", "-R", dvc.root_dir])
     assert ret == 0
@@ -947,15 +947,15 @@ def test_add_file_in_symlink_dir(make_tmp_dir, tmp_dir, dvc, external):
         dvc.add(os.path.join("dir", "foo"))
 
 
-def test_add_with_cache_link_error(tmp_dir, dvc, mocker, caplog):
+def test_add_with_cache_link_error(tmp_dir, dvc, mocker, capsys):
     tmp_dir.gen("foo", "foo")
 
     mocker.patch(
         "dvc.checkout._do_link", side_effect=DvcException("link failed")
     )
-    with caplog.at_level(logging.WARNING, logger="dvc"):
-        dvc.add("foo")
-        assert "reconfigure cache types" in caplog.text
+    dvc.add("foo")
+    err = capsys.readouterr()[1]
+    assert "reconfigure cache types" in err
 
     assert not (tmp_dir / "foo").exists()
     assert (tmp_dir / "foo.dvc").exists()
@@ -1024,6 +1024,7 @@ def test_add_to_remote(tmp_dir, dvc, local_cloud, local_remote):
 
     hash_info = stage.outs[0].hash_info
     assert local_remote.hash_to_path_info(hash_info.value).read_text() == "foo"
+    assert hash_info.size == len("foo")
 
 
 def test_add_to_remote_absolute(tmp_dir, make_tmp_dir, dvc, local_remote):
@@ -1067,6 +1068,8 @@ def test_add_to_cache_dir(tmp_dir, dvc, local_cloud):
     (stage,) = dvc.add(str(local_cloud / "data"), out="data")
     assert len(stage.deps) == 0
     assert len(stage.outs) == 1
+    assert stage.outs[0].hash_info.size == len("foo") + len("bar")
+    assert stage.outs[0].hash_info.nfiles == 2
 
     data = tmp_dir / "data"
     assert data.read_text() == {"foo": "foo", "bar": "bar"}
@@ -1187,3 +1190,42 @@ def test_add_ignored(tmp_dir, scm, dvc):
     assert str(exc.value) == ("bad DVC file name '{}' is git-ignored.").format(
         os.path.join("dir", "subdir.dvc")
     )
+
+
+def test_add_on_not_existing_file_should_not_remove_stage_file(tmp_dir, dvc):
+    (stage,) = tmp_dir.dvc_gen("foo", "foo")
+    (tmp_dir / "foo").unlink()
+    dvcfile_contents = (tmp_dir / stage.path).read_text()
+
+    with pytest.raises(OutputDoesNotExistError):
+        dvc.add("foo")
+    assert (tmp_dir / "foo.dvc").exists()
+    assert (tmp_dir / stage.path).read_text() == dvcfile_contents
+
+
+@pytest.mark.parametrize(
+    "target",
+    [
+        "dvc.repo.Repo.check_modified_graph",
+        "dvc.stage.Stage.save",
+        "dvc.stage.Stage.commit",
+    ],
+)
+def test_add_does_not_remove_stage_file_on_failure(
+    tmp_dir, dvc, mocker, target
+):
+    (stage,) = tmp_dir.dvc_gen("foo", "foo")
+    tmp_dir.gen("foo", "foobar")  # update file
+    dvcfile_contents = (tmp_dir / stage.path).read_text()
+
+    exc_msg = f"raising error from mocked '{target}'"
+    mocker.patch(
+        target,
+        side_effect=DvcException(exc_msg),
+    )
+
+    with pytest.raises(DvcException) as exc_info:
+        dvc.add("foo")
+    assert str(exc_info.value) == exc_msg
+    assert (tmp_dir / "foo.dvc").exists()
+    assert (tmp_dir / stage.path).read_text() == dvcfile_contents

tests/func/test_config.py

@@ -98,7 +98,7 @@ def test_config_set_local(tmp_dir, dvc):
         ),
     ],
 )
-def test_config_get(tmp_dir, dvc, caplog, args, ret, msg):
+def test_config_get(tmp_dir, dvc, capsys, caplog, args, ret, msg):
     (tmp_dir / ".dvc" / "config").write_text(
         textwrap.dedent(
             """\
@@ -123,12 +123,12 @@ def test_config_get(tmp_dir, dvc, caplog, args, ret, msg):
         )
     )
 
-    caplog.clear()
     assert main(["config"] + args) == ret
-    assert msg in caplog.text
+    text = caplog.text if ret else capsys.readouterr()[0]
+    assert msg in text
 
 
-def test_config_list(tmp_dir, dvc, caplog):
+def test_config_list(tmp_dir, dvc, capsys):
     (tmp_dir / ".dvc" / "config").write_text(
         textwrap.dedent(
             """\
@@ -155,17 +155,18 @@ def test_config_list(tmp_dir, dvc, caplog):
         )
     )
 
-    caplog.clear()
     assert main(["config", "--list"]) == 0
-    assert "remote.myremote.url=s3://bucket/path" in caplog.text
-    assert "remote.myremote.region=us-east-2" in caplog.text
-    assert "remote.myremote.profile=iterative" in caplog.text
-    assert "remote.myremote.access_key_id=abcde" in caplog.text
-    assert "remote.myremote.secret_access_key=123456" in caplog.text
-    assert "remote.other.url=gs://bucket/path" in caplog.text
-    assert "core.analytics=False" in caplog.text
-    assert "core.no_scm=true" in caplog.text
-    assert "core.remote=myremote" in caplog.text
+
+    out, _ = capsys.readouterr()
+    assert "remote.myremote.url=s3://bucket/path" in out
+    assert "remote.myremote.region=us-east-2" in out
+    assert "remote.myremote.profile=iterative" in out
+    assert "remote.myremote.access_key_id=abcde" in out
+    assert "remote.myremote.secret_access_key=123456" in out
+    assert "remote.other.url=gs://bucket/path" in out
+    assert "core.analytics=False" in out
+    assert "core.no_scm=true" in out
+    assert "core.remote=myremote" in out
 
 
 @pytest.mark.parametrize(
@@ -241,57 +242,56 @@ def test_load_relative_paths(dvc, field, remote_url):
     )
 
 
-def test_config_remote(tmp_dir, dvc, caplog):
+def test_config_remote(tmp_dir, dvc, capsys):
     (tmp_dir / ".dvc" / "config").write_text(
         "['remote \"myremote\"']\n"
         "  url = s3://bucket/path\n"
         "  region = myregion\n"
     )
 
-    caplog.clear()
     assert main(["config", "remote.myremote.url"]) == 0
-    assert "s3://bucket/path" in caplog.text
+    out, _ = capsys.readouterr()
+    assert "s3://bucket/path" in out
 
-    caplog.clear()
     assert main(["config", "remote.myremote.region"]) == 0
-    assert "myregion" in caplog.text
+    out, _ = capsys.readouterr()
+    assert "myregion" in out
 
 
-def test_config_show_origin_single(tmp_dir, dvc, caplog):
+def test_config_show_origin_single(tmp_dir, dvc, capsys):
     (tmp_dir / ".dvc" / "config").write_text(
         "['remote \"myremote\"']\n"
         "  url = s3://bucket/path\n"
         "  region = myregion\n"
     )
 
-    caplog.clear()
     assert (
         main(["config", "--show-origin", "--project", "remote.myremote.url"])
         == 0
     )
+    out, _ = capsys.readouterr()
     assert (
         "{}\t{}\n".format(os.path.join(".dvc", "config"), "s3://bucket/path")
-        in caplog.text
+        in out
     )
 
-    caplog.clear()
     assert (
         main(["config", "--show-origin", "--local", "remote.myremote.url"])
         == 251
     )
 
-    caplog.clear()
     assert main(["config", "--list", "--project", "--show-origin"]) == 0
+    out, _ = capsys.readouterr()
     assert (
         "{}\t{}\n".format(
             os.path.join(".dvc", "config"),
             "remote.myremote.url=s3://bucket/path",
         )
-        in caplog.text
+        in out
     )
 
 
-def test_config_show_origin_merged(tmp_dir, dvc, caplog):
+def test_config_show_origin_merged(tmp_dir, dvc, capsys):
     (tmp_dir / ".dvc" / "config").write_text(
         "['remote \"myremote\"']\n"
         "  url = s3://bucket/path\n"
@@ -302,18 +302,19 @@ def test_config_show_origin_merged(tmp_dir, dvc, caplog):
         "['remote \"myremote\"']\n  timeout = 100\n"
     )
 
-    caplog.clear()
     assert main(["config", "--list", "--show-origin"]) == 0
+    out, _ = capsys.readouterr()
     assert (
         "{}\t{}\n".format(
             os.path.join(".dvc", "config"),
             "remote.myremote.url=s3://bucket/path",
         )
-        in caplog.text
+        in out
     )
+
     assert (
         "{}\t{}\n".format(
             os.path.join(".dvc", "config.local"), "remote.myremote.timeout=100"
         )
-        in caplog.text
+        in out
     )

tests/func/test_data_cloud.py

@@ -8,7 +8,6 @@ from flaky.flaky_decorator import flaky
 import dvc as dvc_module
 from dvc.external_repo import clean_repos
 from dvc.main import main
-from dvc.objects.db import NamedCache
 from dvc.remote.base import (
     STATUS_DELETED,
     STATUS_MISSING,
@@ -57,7 +56,7 @@ def test_cloud(tmp_dir, dvc, remote):  # pylint:disable=unused-argument
     out = stage.outs[0]
     cache = out.cache_path
     md5 = out.hash_info.value
-    info = out.get_used_cache()
+    foo_objs = out.get_used_objs().get(None, set())
 
     (stage_dir,) = tmp_dir.dvc_gen(
         {
@@ -70,29 +69,39 @@ def test_cloud(tmp_dir, dvc, remote):  # pylint:disable=unused-argument
     )
     out_dir = stage_dir.outs[0]
     cache_dir = out_dir.cache_path
-    name_dir = str(out_dir)
     md5_dir = out_dir.hash_info.value
-    info_dir = NamedCache.make(out_dir.scheme, md5_dir, name_dir)
+    dir_objs = {out_dir.obj}
+    entry_md5s = [entry_obj.hash_info.value for _, entry_obj in out_dir.obj]
+
+    def _make_dir_status(status):
+        expected = {md5_dir: {"name": md5_dir, "status": status}}
+        expected.update(
+            {
+                entry_md5: {"name": entry_md5, "status": status}
+                for entry_md5 in entry_md5s
+            }
+        )
+        return expected
 
     # Check status
-    status = dvc.cloud.status(info, show_checksums=True)
+    status = dvc.cloud.status(foo_objs, show_checksums=True)
     expected = {md5: {"name": md5, "status": STATUS_NEW}}
     assert status == expected
 
-    status_dir = dvc.cloud.status(info_dir, show_checksums=True)
-    expected = {md5_dir: {"name": md5_dir, "status": STATUS_NEW}}
+    status_dir = dvc.cloud.status(dir_objs, show_checksums=True)
+    expected = _make_dir_status(STATUS_NEW)
     assert status_dir == expected
 
     # Move cache and check status
     # See issue https://github.com/iterative/dvc/issues/4383 for details
     backup_dir = dvc.odb.local.cache_dir + ".backup"
     move(dvc.odb.local.cache_dir, backup_dir)
-    status = dvc.cloud.status(info, show_checksums=True)
+    status = dvc.cloud.status(foo_objs, show_checksums=True)
     expected = {md5: {"name": md5, "status": STATUS_MISSING}}
     assert status == expected
 
-    status_dir = dvc.cloud.status(info_dir, show_checksums=True)
-    expected = {md5_dir: {"name": md5_dir, "status": STATUS_MISSING}}
+    status_dir = dvc.cloud.status(dir_objs, show_checksums=True)
+    expected = _make_dir_status(STATUS_MISSING)
     assert status_dir == expected
 
     # Restore original cache:
@@ -100,48 +109,48 @@ def test_cloud(tmp_dir, dvc, remote):  # pylint:disable=unused-argument
     move(backup_dir, dvc.odb.local.cache_dir)
 
     # Push and check status
-    dvc.cloud.push(info)
+    dvc.cloud.push(foo_objs)
     assert os.path.exists(cache)
     assert os.path.isfile(cache)
 
-    dvc.cloud.push(info_dir)
+    dvc.cloud.push(dir_objs)
     assert os.path.isfile(cache_dir)
 
-    status = dvc.cloud.status(info, show_checksums=True)
+    status = dvc.cloud.status(foo_objs, show_checksums=True)
     expected = {md5: {"name": md5, "status": STATUS_OK}}
     assert status == expected
 
-    status_dir = dvc.cloud.status(info_dir, show_checksums=True)
-    expected = {md5_dir: {"name": md5_dir, "status": STATUS_OK}}
+    status_dir = dvc.cloud.status(dir_objs, show_checksums=True)
+    expected = _make_dir_status(STATUS_OK)
     assert status_dir == expected
 
     # Remove and check status
     remove(dvc.odb.local.cache_dir)
 
-    status = dvc.cloud.status(info, show_checksums=True)
+    status = dvc.cloud.status(foo_objs, show_checksums=True)
     expected = {md5: {"name": md5, "status": STATUS_DELETED}}
     assert status == expected
 
-    status_dir = dvc.cloud.status(info_dir, show_checksums=True)
-    expected = {md5_dir: {"name": md5_dir, "status": STATUS_DELETED}}
+    status_dir = dvc.cloud.status(dir_objs, show_checksums=True)
+    expected = _make_dir_status(STATUS_DELETED)
     assert status_dir == expected
 
     # Pull and check status
-    dvc.cloud.pull(info)
+    dvc.cloud.pull(foo_objs)
     assert os.path.exists(cache)
     assert os.path.isfile(cache)
     with open(cache) as fd:
         assert fd.read() == "foo"
 
-    dvc.cloud.pull(info_dir)
+    dvc.cloud.pull(dir_objs)
     assert os.path.isfile(cache_dir)
 
-    status = dvc.cloud.status(info, show_checksums=True)
+    status = dvc.cloud.status(foo_objs, show_checksums=True)
     expected = {md5: {"name": md5, "status": STATUS_OK}}
     assert status == expected
 
-    status_dir = dvc.cloud.status(info_dir, show_checksums=True)
-    expected = {md5_dir: {"name": md5_dir, "status": STATUS_OK}}
+    status_dir = dvc.cloud.status(dir_objs, show_checksums=True)
+    expected = _make_dir_status(STATUS_OK)
     assert status_dir == expected
 
 
@@ -239,7 +248,7 @@ def test_warn_on_outdated_stage(tmp_dir, dvc, local_remote, caplog):
 
 def test_hash_recalculation(mocker, dvc, tmp_dir, local_remote):
     tmp_dir.gen({"foo": "foo"})
-    test_get_file_hash = mocker.spy(dvc_module.objects.stage, "_get_file_hash")
+    test_file_md5 = mocker.spy(dvc_module.objects.stage, "file_md5")
     ret = main(["config", "cache.type", "hardlink"])
     assert ret == 0
     ret = main(["add", "foo"])
@@ -248,7 +257,7 @@ def test_hash_recalculation(mocker, dvc, tmp_dir, local_remote):
     assert ret == 0
     ret = main(["run", "--single-stage", "-d", "foo", "echo foo"])
     assert ret == 0
-    assert test_get_file_hash.mock.call_count == 1
+    assert test_file_md5.mock.call_count == 1
 
 
 def test_missing_cache(tmp_dir, dvc, local_remote, caplog):
@@ -277,7 +286,10 @@ def test_missing_cache(tmp_dir, dvc, local_remote, caplog):
     assert bar in caplog.text
 
     caplog.clear()
-    assert dvc.status(cloud=True) == {"bar": "missing", "foo": "missing"}
+    assert dvc.status(cloud=True) == {
+        "bar": "missing",
+        "foo": "missing",
+    }
     assert header not in caplog.text
     assert foo not in caplog.text
     assert bar not in caplog.text
@@ -295,7 +307,7 @@ def test_verify_hashes(
     remove("dir")
     remove(dvc.odb.local.cache_dir)
 
-    hash_spy = mocker.spy(dvc_module.objects.stage, "_get_file_hash")
+    hash_spy = mocker.spy(dvc_module.objects.stage, "file_md5")
 
     dvc.pull()
     assert hash_spy.call_count == 0
@@ -463,13 +475,12 @@ def test_pipeline_file_target_ops(tmp_dir, dvc, run_copy, local_remote):
         ({}, "Everything is up to date"),
     ],
 )
-def test_push_stats(tmp_dir, dvc, fs, msg, caplog, local_remote):
+def test_push_stats(tmp_dir, dvc, fs, msg, capsys, local_remote):
     tmp_dir.dvc_gen(fs)
 
-    caplog.clear()
-    with caplog.at_level(level=logging.INFO, logger="dvc"):
-        main(["push"])
-    assert msg in caplog.text
+    main(["push"])
+    out, _ = capsys.readouterr()
+    assert msg in out
 
 
 @pytest.mark.parametrize(
@@ -480,17 +491,17 @@ def test_push_stats(tmp_dir, dvc, fs, msg, caplog, local_remote):
         ({}, "Everything is up to date."),
     ],
 )
-def test_fetch_stats(tmp_dir, dvc, fs, msg, caplog, local_remote):
+def test_fetch_stats(tmp_dir, dvc, fs, msg, capsys, local_remote):
     tmp_dir.dvc_gen(fs)
     dvc.push()
     clean(list(fs.keys()), dvc)
-    caplog.clear()
-    with caplog.at_level(level=logging.INFO, logger="dvc"):
-        main(["fetch"])
-    assert msg in caplog.text
+
+    main(["fetch"])
+    out, _ = capsys.readouterr()
+    assert msg in out
 
 
-def test_pull_stats(tmp_dir, dvc, capsys, caplog, local_remote):
+def test_pull_stats(tmp_dir, dvc, capsys, local_remote):
     tmp_dir.dvc_gen({"foo": "foo", "bar": "bar"})
     dvc.push()
     clean(["foo", "bar"], dvc)
@@ -501,12 +512,13 @@ def test_pull_stats(tmp_dir, dvc, capsys, caplog, local_remote):
     out, _ = capsys.readouterr()
     assert "M\tbar".expandtabs() in out
     assert "A\tfoo".expandtabs() in out
-    assert "2 files fetched" in caplog.text
-    assert "1 file added" in caplog.text
-    assert "1 file modified" in caplog.text
-    with caplog.at_level(level=logging.INFO, logger="dvc"):
-        main(["pull"])
-    assert "Everything is up to date." in caplog.text
+    assert "2 files fetched" in out
+    assert "1 file added" in out
+    assert "1 file modified" in out
+
+    main(["pull"])
+    out, _ = capsys.readouterr()
+    assert "Everything is up to date." in out
 
 
 @pytest.mark.parametrize(
@@ -576,3 +588,23 @@ def test_pull_00_prefix(tmp_dir, dvc, remote, monkeypatch):
     stats = dvc.pull()
     assert stats["fetched"] == 2
     assert set(stats["added"]) == {"foo", "bar"}
+
+
+@pytest.mark.parametrize("remote", full_clouds, indirect=True)
+def test_pull_no_00_prefix(tmp_dir, dvc, remote, monkeypatch):
+    # Related: https://github.com/iterative/dvc/issues/6244
+
+    fs_type = type(dvc.cloud.get_remote("upstream").fs)
+    monkeypatch.setattr(fs_type, "_ALWAYS_TRAVERSE", True, raising=False)
+    monkeypatch.setattr(fs_type, "LIST_OBJECT_PAGE_SIZE", 256, raising=False)
+
+    # foo's md5 checksum is 14ffd92a6cbf5f2f657067df0d5881a6
+    # bar's md5 checksum is 64020400f00960c0ef04052547b134b3
+    tmp_dir.dvc_gen({"foo": "dvc", "bar": "cml"})
+
+    dvc.push()
+    clean(["foo", "bar"], dvc)
+
+    stats = dvc.pull()
+    assert stats["fetched"] == 2
+    assert set(stats["added"]) == {"foo", "bar"}

tests/func/test_fs.py

@@ -244,9 +244,8 @@ def test_walk_dont_ignore_subrepos(tmp_dir, scm, dvc):
 )
 def test_fs_getsize(dvc, cloud):
     cloud.gen({"data": {"foo": "foo"}, "baz": "baz baz"})
-    cls, config = get_cloud_fs(dvc, **cloud.config)
+    cls, config, path_info = get_cloud_fs(dvc, **cloud.config)
     fs = cls(**config)
-    path_info = fs.path_info
 
     assert fs.getsize(path_info / "baz") == 7
     assert fs.getsize(path_info / "data" / "foo") == 3
@@ -270,11 +269,11 @@ def test_fs_getsize(dvc, cloud):
 )
 def test_fs_upload_fobj(dvc, tmp_dir, cloud):
     tmp_dir.gen("foo", "foo")
-    cls, config = get_cloud_fs(dvc, **cloud.config)
+    cls, config, path_info = get_cloud_fs(dvc, **cloud.config)
     fs = cls(**config)
 
     from_info = tmp_dir / "foo"
-    to_info = fs.path_info / "foo"
+    to_info = path_info / "foo"
 
     with open(from_info, "rb") as stream:
         fs.upload_fobj(stream, to_info)
@@ -297,9 +296,9 @@ def test_fs_ls(dvc, cloud):
             }
         }
     )
-    cls, config = get_cloud_fs(dvc, **cloud.config)
+    cls, config, path_info = get_cloud_fs(dvc, **cloud.config)
     fs = cls(**config)
-    path_info = cloud / "directory"
+    path_info /= "directory"
 
     assert {os.path.basename(file_key) for file_key in fs.ls(path_info)} == {
         "foo",
@@ -327,9 +326,8 @@ def test_fs_ls(dvc, cloud):
 )
 def test_fs_find_recursive(dvc, cloud):
     cloud.gen({"data": {"foo": "foo", "bar": {"baz": "baz"}, "quux": "quux"}})
-    cls, config = get_cloud_fs(dvc, **cloud.config)
+    cls, config, path_info = get_cloud_fs(dvc, **cloud.config)
     fs = cls(**config)
-    path_info = fs.path_info
 
     assert {
         os.path.basename(file_key) for file_key in fs.find(path_info / "data")
@@ -349,9 +347,8 @@ def test_fs_find_recursive(dvc, cloud):
 )
 def test_fs_find_with_etag(dvc, cloud):
     cloud.gen({"data": {"foo": "foo", "bar": {"baz": "baz"}, "quux": "quux"}})
-    cls, config = get_cloud_fs(dvc, **cloud.config)
+    cls, config, path_info = get_cloud_fs(dvc, **cloud.config)
     fs = cls(**config)
-    path_info = fs.path_info
 
     for details in fs.find(path_info / "data", detail=True):
         assert (
@@ -371,7 +368,7 @@ def test_fs_find_with_etag(dvc, cloud):
 )
 def test_fs_fsspec_path_management(dvc, cloud):
     cloud.gen({"foo": "foo", "data": {"bar": "bar", "baz": {"foo": "foo"}}})
-    cls, config = get_cloud_fs(dvc, **cloud.config)
+    cls, config, _ = get_cloud_fs(dvc, **cloud.config)
     fs = cls(**config)
 
     root = cloud.parents[len(cloud.parents) - 1]

tests/func/test_import.py

@@ -8,7 +8,7 @@ from mock import patch
 import dvc.data_cloud as cloud
 from dvc.config import NoRemoteError
 from dvc.dvcfile import Dvcfile
-from dvc.exceptions import DownloadError
+from dvc.exceptions import DownloadError, PathMissingError
 from dvc.objects.db import ODBManager
 from dvc.stage.exceptions import StagePathNotFoundError
 from dvc.system import System
@@ -289,7 +289,7 @@ def test_push_wildcard_from_bare_git_repo(
     dvc_repo = make_tmp_dir("dvc-repo", scm=True, dvc=True)
     with dvc_repo.chdir():
         dvc_repo.dvc.imp(os.fspath(tmp_dir), "dirextra")
-        with pytest.raises(FileNotFoundError):
+        with pytest.raises(PathMissingError):
             dvc_repo.dvc.imp(os.fspath(tmp_dir), "dir123")
 
 
@@ -347,11 +347,11 @@ def test_pull_non_workspace(tmp_dir, scm, dvc, erepo_dir):
 
 
 def test_import_non_existing(erepo_dir, tmp_dir, dvc):
-    with pytest.raises(FileNotFoundError):
+    with pytest.raises(PathMissingError):
         tmp_dir.dvc.imp(os.fspath(erepo_dir), "invalid_output")
 
     # https://github.com/iterative/dvc/pull/2837#discussion_r352123053
-    with pytest.raises(FileNotFoundError):
+    with pytest.raises(PathMissingError):
         tmp_dir.dvc.imp(os.fspath(erepo_dir), "/root/", "root")
 
 
@@ -530,7 +530,7 @@ def test_import_with_no_exec(tmp_dir, dvc, erepo_dir):
 
 
 def test_import_with_jobs(mocker, dvc, erepo_dir):
-    from dvc.data_cloud import DataCloud
+    import dvc as dvc_module
 
     with erepo_dir.chdir():
         erepo_dir.dvc_gen(
@@ -545,7 +545,57 @@ def test_import_with_jobs(mocker, dvc, erepo_dir):
             commit="init",
         )
 
-    spy = mocker.spy(DataCloud, "pull")
+    spy = mocker.spy(dvc_module.objects, "save")
     dvc.imp(os.fspath(erepo_dir), "dir1", jobs=3)
     run_jobs = tuple(spy.call_args_list[0])[1].get("jobs")
     assert run_jobs == 3
+
+
+def test_chained_import(tmp_dir, dvc, make_tmp_dir, erepo_dir, local_cloud):
+    erepo_dir.add_remote(config=local_cloud.config)
+    with erepo_dir.chdir():
+        erepo_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}}, commit="init")
+    erepo_dir.dvc.push()
+    remove(erepo_dir.dvc.odb.local.cache_dir)
+    remove(os.fspath(erepo_dir / "dir"))
+
+    erepo2 = make_tmp_dir("erepo2", scm=True, dvc=True)
+    with erepo2.chdir():
+        erepo2.dvc.imp(os.fspath(erepo_dir), "dir")
+        erepo2.scm.add("dir.dvc")
+        erepo2.scm.commit("import")
+    remove(erepo2.dvc.odb.local.cache_dir)
+    remove(os.fspath(erepo2 / "dir"))
+
+    dvc.imp(os.fspath(erepo2), "dir", "dir_imported")
+    dst = tmp_dir / "dir_imported"
+    assert (dst / "foo").read_text() == "foo"
+    assert (dst / "bar").read_text() == "bar"
+
+    remove(dvc.odb.local.cache_dir)
+    remove("dir_imported")
+
+    # pulled objects should come from the original upstream repo's remote,
+    # no cache or remote should be needed from the intermediate repo
+    dvc.pull(["dir_imported.dvc"])
+    assert not os.path.exists(erepo_dir.dvc.odb.local.cache_dir)
+    assert not os.path.exists(erepo2.dvc.odb.local.cache_dir)
+    assert (dst / "foo").read_text() == "foo"
+    assert (dst / "bar").read_text() == "bar"
+
+
+def test_circular_import(tmp_dir, dvc, scm, erepo_dir):
+    from dvc.exceptions import CircularImportError
+
+    with erepo_dir.chdir():
+        erepo_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}}, commit="init")
+
+    dvc.imp(os.fspath(erepo_dir), "dir", "dir_imported")
+    scm.add("dir_imported.dvc")
+    scm.commit("import")
+
+    with erepo_dir.chdir():
+        with pytest.raises(CircularImportError):
+            erepo_dir.dvc.imp(
+                os.fspath(tmp_dir), "dir_imported", "circular_import"
+            )

tests/func/test_import_url.py

@@ -123,6 +123,7 @@ def test_import_url_with_no_exec(tmp_dir, dvc, erepo_dir):
     [
         pytest.lazy_fixture("local_cloud"),
         pytest.lazy_fixture("s3"),
+        pytest.lazy_fixture("azure"),
         pytest.param(
             pytest.lazy_fixture("gs"), marks=pytest.mark.needs_internet
         ),
@@ -282,6 +283,7 @@ def test_import_url_to_remote_single_file(
 
     hash_info = stage.outs[0].hash_info
     assert local_remote.hash_to_path_info(hash_info.value).read_text() == "foo"
+    assert hash_info.size == len("foo")
 
 
 @pytest.mark.parametrize(

tests/func/test_ls.py

@@ -273,6 +273,24 @@ def test_ls_repo_with_removed_dvc_dir_with_path_file(tmp_dir, dvc, scm):
     match_files(files, ((("file",), True),))
 
 
+def test_ls_repo_with_rev(erepo_dir):
+    with erepo_dir.chdir():
+        erepo_dir.scm_gen(FS_STRUCTURE, commit="init")
+        erepo_dir.dvc_gen(DVC_STRUCTURE, commit="dvc")
+
+    rev = erepo_dir.scm.list_all_commits()[1]
+    files = Repo.ls(os.fspath(erepo_dir), rev=rev)
+    match_files(
+        files,
+        (
+            ((".dvcignore",), False),
+            ((".gitignore",), False),
+            (("README.md",), False),
+            (("model",), False),
+        ),
+    )
+
+
 def test_ls_remote_repo(erepo_dir):
     with erepo_dir.chdir():
         erepo_dir.scm_gen(FS_STRUCTURE, commit="init")

tests/func/test_move.py

@@ -1,11 +1,12 @@
 import os
+import textwrap
 
 import pytest
 
 from dvc.dvcfile import DVC_FILE_SUFFIX
 from dvc.exceptions import DvcException, MoveNotDataSourceError
 from dvc.main import main
-from dvc.utils.serialize import load_yaml
+from dvc.utils.serialize import dump_yaml, load_yaml
 from tests.basic_env import TestDvc, TestDvcGit
 from tests.func.test_repro import TestRepro
 from tests.utils import cd
@@ -244,3 +245,24 @@ def test_move_output_overlap(tmp_dir, dvc):
     assert (tmp_dir / "foo.dvc").exists()
     assert not (tmp_dir / "dir" / "foo").exists()
     assert not (tmp_dir / "dir" / "foo.dvc").exists()
+
+
+def test_move_meta(tmp_dir, dvc):
+    (stage,) = tmp_dir.dvc_gen("foo", "foo")
+    data = load_yaml(stage.path)
+    data["meta"] = {"custom_key": 42}
+    dump_yaml(stage.path, data)
+
+    dvc.move("foo", "bar")
+    res = (tmp_dir / "bar.dvc").read_text()
+    print(res)
+    assert res == textwrap.dedent(
+        """\
+        outs:
+        - md5: acbd18db4cc2f85cedef654fccc4a4d8
+          size: 3
+          path: bar
+        meta:
+          custom_key: 42
+    """
+    )