add missing stat() methods to DBFSPath and WorkspacePath (#144)

Fixes #142 and #143

---------

Co-authored-by: Eric Vergnaud <[email protected]>

Author: ericvergnaud

src/databricks/labs/blueprint/paths.py

@@ -11,6 +11,7 @@
 import posixpath
 import re
 import shutil
+import stat
 from abc import abstractmethod
 from collections.abc import Generator, Iterable, Sequence
 from io import BytesIO, StringIO
@@ -121,7 +122,6 @@ class _DatabricksPath(Path, abc.ABC):  # pylint: disable=too-many-public-methods
     # Public APIs that we don't support.
     as_uri = _na("as_uri")
     cwd = _na("cwd")
-    stat = _na("stat")
     chmod = _na("chmod")
     lchmod = _na("lchmod")
     lstat = _na("lstat")
@@ -138,6 +138,7 @@ def __new__(cls, *args, **kwargs):
         # Force all initialisation to go via __init__() irrespective of the (Python-specific) base version.
         return object.__new__(cls)
 
+    # pylint: disable=super-init-not-called
     def __init__(self, ws: WorkspaceClient, *args: str | bytes | os.PathLike) -> None:
         # We deliberately do _not_ call the super initializer because we're taking over complete responsibility for the
         # implementation of the public API.
@@ -385,6 +386,7 @@ def with_suffix(self: P, suffix: str) -> P:
             raise ValueError(msg)
         return self.with_name(stem + suffix)
 
+    # pylint: disable=arguments-differ
     def relative_to(self: P, *other: str | bytes | os.PathLike, walk_up: bool = False) -> P:
         normalized = self.with_segments(*other)
         if self.anchor != normalized.anchor:
@@ -691,6 +693,14 @@ def _file_info(self) -> FileInfo:
             self._cached_file_info = self._ws.dbfs.get_status(self.as_posix())
             return self._cached_file_info
 
+    def stat(self, *, follow_symlinks=True) -> os.stat_result:
+        seq: list[float] = [-1.0] * 10
+        seq[stat.ST_SIZE] = self._file_info.file_size or -1  # 6
+        seq[stat.ST_MTIME] = (
+            float(self._file_info.modification_time) / 1000.0 if self._file_info.modification_time else -1.0
+        )  # 8
+        return os.stat_result(seq)
+
     def is_dir(self) -> bool:
         """Return True if the path points to a DBFS directory."""
         try:
@@ -841,6 +851,15 @@ def _object_info(self) -> ObjectInfo:
             self._cached_object_info = self._ws.workspace.get_status(self.as_posix())
             return self._object_info
 
+    def stat(self, *, follow_symlinks=True) -> os.stat_result:
+        seq: list[float] = [-1.0] * 10
+        seq[stat.ST_SIZE] = self._object_info.size or -1  # 6
+        seq[stat.ST_MTIME] = (
+            float(self._object_info.modified_at) / 1000.0 if self._object_info.modified_at else -1.0
+        )  # 8
+        seq[stat.ST_CTIME] = float(self._object_info.created_at) / 1000.0 if self._object_info.created_at else -1.0  # 9
+        return os.stat_result(seq)
+
     def is_dir(self) -> bool:
         """Return True if the path points to a directory in Databricks Workspace."""
         try:

tests/integration/test_paths.py

@@ -1,4 +1,5 @@
 import codecs
+from datetime import datetime
 from pathlib import Path
 
 import pytest
@@ -67,6 +68,21 @@ def test_open_text_io(ws, make_random, cls):
     assert not hello_txt.exists()
 
 
+@pytest.mark.parametrize("cls", DATABRICKS_PATHLIKE)
+def test_stat(ws, make_random, cls):
+    now = datetime.now().timestamp()
+    name = make_random()
+    wsp = cls(ws, f"~/{name}/a/b/c")
+    with_user = wsp.expanduser()
+    with_user.mkdir(parents=True)
+
+    hello_txt = with_user / "hello.txt"
+    hello_txt.write_text("Hello, World!")
+    if cls is WorkspacePath:  # DBFSPath has no st_ctime
+        assert hello_txt.stat().st_ctime >= now
+    assert hello_txt.stat().st_mtime >= now
+
+
 @pytest.mark.parametrize("cls", DATABRICKS_PATHLIKE)
 def test_unlink(ws, make_random, cls):
     name = make_random()

tests/unit/test_paths.py

@@ -7,14 +7,15 @@
 from databricks.sdk import WorkspaceClient
 from databricks.sdk.errors import NotFound, ResourceDoesNotExist
 from databricks.sdk.mixins.workspace import WorkspaceExt
+from databricks.sdk.service.files import FileInfo
 from databricks.sdk.service.workspace import (
     ImportFormat,
     Language,
     ObjectInfo,
     ObjectType,
 )
 
-from databricks.labs.blueprint.paths import WorkspacePath
+from databricks.labs.blueprint.paths import DBFSPath, WorkspacePath
 
 
 def test_empty_init() -> None:
@@ -1007,3 +1008,24 @@ def test_rglob() -> None:
         WorkspacePath(ws, "/test/path/dir1/file1.json"),
         WorkspacePath(ws, "/test/path/dir2/file2.json"),
     }
+
+
+def test_workspace_path_stat_has_fields():
+    info = ObjectInfo(created_at=1234, modified_at=2345, size=3456)
+    ws = create_autospec(WorkspaceClient)
+    ws.workspace.get_status.return_value = info
+    workspace_path = WorkspacePath(ws, "/test/path")
+    stats = workspace_path.stat()
+    assert stats.st_ctime == info.created_at / 1000.0
+    assert stats.st_mtime == info.modified_at / 1000.0
+    assert stats.st_size == info.size
+
+
+def test_dbfs_path_stat_has_fields():
+    info = FileInfo(modification_time=2345, file_size=3456)
+    ws = create_autospec(WorkspaceClient)
+    ws.dbfs.get_status.return_value = info
+    dbfs_path = DBFSPath(ws, "/test/path")
+    stats = dbfs_path.stat()
+    assert stats.st_mtime == info.modification_time / 1000.0
+    assert stats.st_size == info.file_size