dvcfs: implement basic du

dvc/cli/parser.py

@@ -19,6 +19,7 @@ from dvc.commands import (
     data_sync,
     destroy,
     diff,
+    du,
     experiments,
     freeze,
     gc,
@@ -95,6 +96,7 @@ COMMANDS = [
     data,
     artifacts,
     studio,
+    du,
 ]
 
 

dvc/commands/du.py

@@ -0,0 +1,80 @@
+import argparse
+import logging
+
+from dvc.cli import completion
+from dvc.cli.command import CmdBaseNoRepo
+from dvc.cli.utils import DictAction, append_doc_link
+from dvc.ui import ui
+
+logger = logging.getLogger(__name__)
+
+
+class CmdDU(CmdBaseNoRepo):
+    def run(self):
+        from dvc.repo import Repo
+        from dvc.utils.humanize import naturalsize
+
+        entries = Repo.du(
+            self.args.url,
+            self.args.path,
+            rev=self.args.rev,
+            summarize=self.args.summarize,
+            config=self.args.config,
+            remote=self.args.remote,
+            remote_config=self.args.remote_config,
+        )
+        ui.table([(naturalsize(size), path) for path, size in entries])
+        return 0
+
+
+def add_parser(subparsers, parent_parser):
+    DU_HELP = "Show disk usage."
+    du_parser = subparsers.add_parser(
+        "du",
+        parents=[parent_parser],
+        description=append_doc_link(DU_HELP, "du"),
+        help=DU_HELP,
+        formatter_class=argparse.RawTextHelpFormatter,
+    )
+    du_parser.add_argument("url", help="Location of DVC repository")
+    du_parser.add_argument(
+        "--rev",
+        nargs="?",
+        help="Git revision (e.g. SHA, branch, tag)",
+        metavar="<commit>",
+    )
+    du_parser.add_argument(
+        "-s",
+        "--summarize",
+        action="store_true",
+        help="Show total disk usage.",
+    )
+    du_parser.add_argument(
+        "--config",
+        type=str,
+        help=(
+            "Path to a config file that will be merged with the config "
+            "in the target repository."
+        ),
+    )
+    du_parser.add_argument(
+        "--remote",
+        type=str,
+        help="Remote name to set as a default in the target repository.",
+    )
+    du_parser.add_argument(
+        "--remote-config",
+        type=str,
+        nargs="*",
+        action=DictAction,
+        help=(
+            "Remote config options to merge with a remote's config (default or one "
+            "specified by '--remote') in the target repository."
+        ),
+    )
+    du_parser.add_argument(
+        "path",
+        nargs="?",
+        help="Path to directory within the repository",
+    ).complete = completion.DIR
+    du_parser.set_defaults(func=CmdDU)

dvc/fs/dvc.py

@@ -4,6 +4,7 @@ import ntpath
 import os
 import posixpath
 import threading
+from collections import deque
 from contextlib import ExitStack, suppress
 from typing import TYPE_CHECKING, Any, Callable, Dict, Optional, Tuple, Type, Union
 
@@ -60,6 +61,7 @@ def _merge_info(repo, key, fs_info, dvc_info):
     if fs_info:
         ret["type"] = fs_info["type"]
         ret["size"] = fs_info["size"]
+        ret["fs_info"] = fs_info
         isexec = False
         if fs_info["type"] == "file":
             isexec = utils.is_exec(fs_info["mode"])
@@ -421,6 +423,45 @@ class _DVCFileSystem(AbstractFileSystem):
         dvc_path = _get_dvc_path(dvc_fs, subkey)
         return dvc_fs.get_file(dvc_path, lpath, **kwargs)
 
+    def du(self, path, total=True, maxdepth=None, withdirs=False, **kwargs):
+        if maxdepth is not None:
+            raise NotImplementedError
+
+        sizes = {}
+        dus = {}
+        todo = deque([self.info(path)])
+        while todo:
+            info = todo.popleft()
+            isdir = info["type"] == "directory"
+            size = info["size"] or 0
+            name = info["name"]
+
+            if not isdir:
+                sizes[name] = size
+                continue
+
+            dvc_info = info.get("dvc_info") or {}
+            fs_info = info.get("fs_info")
+            entry = dvc_info.get("entry")
+            if (
+                dvc_info
+                and not fs_info
+                and entry is not None
+                and entry.size is not None
+            ):
+                dus[name] = entry.size
+                continue
+
+            if withdirs:
+                sizes[name] = size
+
+            todo.extend(self.ls(info["name"], detail=True))
+
+        if total:
+            return sum(sizes.values()) + sum(dus.values())
+
+        return sizes
+
     def close(self):
         self._repo_stack.close()
 

dvc/repo/__init__.py

@@ -70,6 +70,7 @@ class Repo:
     from dvc.repo.commit import commit  # type: ignore[misc]
     from dvc.repo.destroy import destroy  # type: ignore[misc]
     from dvc.repo.diff import diff  # type: ignore[misc]
+    from dvc.repo.du import du as _du  # type: ignore[misc]
     from dvc.repo.fetch import fetch  # type: ignore[misc]
     from dvc.repo.freeze import freeze, unfreeze  # type: ignore[misc]
     from dvc.repo.gc import gc  # type: ignore[misc]
@@ -93,6 +94,7 @@ class Repo:
     from .cache import check_missing as cache_check_missing  # type: ignore[misc]
     from .data import status as data_status  # type: ignore[misc]
 
+    du = staticmethod(_du)
     ls = staticmethod(_ls)
     ls_url = staticmethod(_ls_url)
     get = staticmethod(_get)

dvc/repo/du.py

@@ -0,0 +1,42 @@
+from typing import Any, Dict, Optional, Union
+
+
+def du(
+    url: str,
+    path: Optional[str] = None,
+    rev: Optional[str] = None,
+    summarize: bool = False,
+    config: Union[None, Dict[str, Any], str] = None,
+    remote: Optional[str] = None,
+    remote_config: Optional[dict] = None,
+):
+    from dvc.config import Config
+
+    from . import Repo
+
+    if config and not isinstance(config, dict):
+        config_dict = Config.load_file(config)
+    else:
+        config_dict = None
+
+    with Repo.open(
+        url,
+        rev=rev,
+        subrepos=True,
+        uninitialized=True,
+        config=config_dict,
+        remote=remote,
+        remote_config=remote_config,
+    ) as repo:
+        path = path or ""
+
+        fs = repo.dvcfs
+
+        if summarize or not fs.isdir(path):
+            return [(path, fs.du(path, total=True))]
+
+        ret = [
+            (entry_path, fs.du(entry_path, total=True)) for entry_path in fs.ls(path)
+        ]
+        ret.append((path, sum(entry[1] for entry in ret)))
+        return ret

tests/func/test_du.py

@@ -0,0 +1,49 @@
+import os
+
+
+def test_du(tmp_dir, dvc):
+    tmp_dir.gen(
+        {
+            "file": b"file",
+            "dvcfile": b"dvcfile",
+            "dir": {
+                "dirfile": b"dirfile",
+                "subdir": {
+                    "subdirfile": b"subdirfile",
+                },
+                "dvcsubdir": {
+                    "dvcsubdirfile": b"dvcsubdirfile",
+                },
+            },
+        }
+    )
+
+    dvc.add("dvcfile")
+    dvc.add(os.path.join("dir", "dvcsubdir"))
+
+    assert dvc.du(".", "file") == [("file", 4)]
+    assert dvc.du(".", "dvcfile") == [("dvcfile", 7)]
+    assert set(dvc.du(".", "dir/subdir")) == {
+        ("dir/subdir/subdirfile", 10),
+        ("dir/subdir", 10),
+    }
+    assert dvc.du(".", "dir/subdir", summarize=True) == [("dir/subdir", 10)]
+    assert set(dvc.du(".", "dir/dvcsubdir")) == {
+        ("dir/dvcsubdir/dvcsubdirfile", 13),
+        ("dir/dvcsubdir", 13),
+    }
+    assert dvc.du(".", "dir/dvcsubdir", summarize=True) == [("dir/dvcsubdir", 13)]
+    assert set(dvc.du(".", "dir")) == {
+        ("dir/dvcsubdir", 13),
+        ("dir/subdir", 10),
+        ("dir/dirfile", 7),
+        ("dir", 30),
+    }
+    assert dvc.du(".", "dir", summarize=True) == [("dir", 30)]
+    assert set(dvc.du(".", "/")) == {
+        ("/dvcfile", 7),
+        ("/dir", 30),
+        ("/file", 4),
+        ("/", 41),
+    }
+    assert dvc.du(".", "/", summarize=True) == [("/", 41)]

tests/unit/command/test_du.py

@@ -0,0 +1,21 @@
+from dvc.cli import parse_args
+from dvc.commands.du import CmdDU
+
+
+def test_du(mocker):
+    cli_args = parse_args(["du", "myurl", "mypath", "--summarize", "--rev", "myrev"])
+    assert cli_args.func == CmdDU
+
+    cmd = cli_args.func(cli_args)
+    mock_du = mocker.patch("dvc.repo.Repo.du")
+
+    assert cmd.run() == 0
+    mock_du.assert_called_once_with(
+        "myurl",
+        "mypath",
+        rev="myrev",
+        summarize=True,
+        config=None,
+        remote=None,
+        remote_config=None,
+    )