Merge remote-tracking branch 'upstream/branch-24.08' into bug/nunique

.github/workflows/external_issue_labeler.yml

@@ -20,36 +20,41 @@ on:
     types:
       - opened
 
-  pull_request:
+  pull_request_target:
     types:
       - opened
 
 env:
   GITHUB_TOKEN: ${{ github.token }}
 
-permissions:
-  issues: write
-  pull-requests: write
-
 jobs:
   Label-Issue:
     runs-on: ubuntu-latest
-    # Only run if the issue author is not part of RAPIDS
-    if: ${{ ! contains(fromJSON('["OWNER", "MEMBER", "COLLABORATOR"]'), github.event.issue.author_association)}}
+    permissions:
+      issues: write
+    if: github.event_name == 'issues'
     steps:
       - name: add-external-labels
+        # Only run if the issue author is not part of RAPIDS
+        if: ${{ ! contains(fromJSON('["OWNER", "MEMBER", "COLLABORATOR"]'), github.event.issue.author_association)}}
         run: |
+          echo ${{ github.event.issue.author_association }}
           issue_url=${{ github.event.issue.html_url }}
           gh issue edit ${issue_url} --add-label "External"
         continue-on-error: true
 
   Label-PR:
     runs-on: ubuntu-latest
-    # Only run if the issue author is not part of RAPIDS
-    if: ${{ ! contains(fromJSON('["OWNER", "MEMBER", "COLLABORATOR"]'), github.event.pull_request.author_association)}}
+    permissions:
+      pull-requests: write
+      issues: write
+    if: github.event_name == 'pull_request_target'
     steps:
       - name: add-external-labels
+        # Only run if the issue author is not part of RAPIDS
+        if: ${{ ! contains(fromJSON('["OWNER", "MEMBER", "COLLABORATOR"]'), github.event.pull_request.author_association)}}
         run: |
+            echo ${{ github.event.pull_request.author_association }}
             pr_url=${{ github.event.pull_request.html_url }}
             gh issue edit ${pr_url} --add-label "External"
-    continue-on-error: true
+        continue-on-error: true

.github/workflows/pr_issue_status_automation.yml

@@ -50,7 +50,7 @@ jobs:
 
     update-sprint:
       # This job sets the PR and its linked issues to the current "Weekly Sprint"
-      uses: jarmak-nv/shared-workflows/.github/workflows/[email protected]
+      uses: rapidsai/shared-workflows/.github/workflows/[email protected]
       if: github.event.pull_request.state == 'open'
       needs: get-project-id
       with:

docs/cudf/source/developer_guide/cudf_pandas.md

@@ -20,6 +20,7 @@ The "wrapped" types/classes are the Pandas and cuDF specific types that have bee
 Wrapped objects and proxy objects are instances of wrapped types and proxy types, respectively.
 In the snippet below `s1` and `s2` are wrapped objects and `s3` is a fast-slow proxy object.
 Also note that the module `xpd` is a wrapped module and contains cuDF and Pandas modules as attributes.
+To check if an object is a proxy type, we can use `cudf.pandas.is_proxy_object`.
   ```python
   import cudf.pandas
   cudf.pandas.install()
@@ -31,6 +32,14 @@ Also note that the module `xpd` is a wrapped module and contains cuDF and Pandas
   s1 = cudf.Series([1,2])
   s2 = pd.Series([1,2])
   s3 = xpd.Series([1,2])
+
+  from cudf.pandas import is_proxy_object
+
+  is_proxy_object(s1) # returns False
+
+  is_proxy_object(s2) # returns False
+
+  is_proxy_object(s3) # returns True
   ```
 
 ```{note}

python/cudf/cudf/pandas/__init__.py

@@ -1,11 +1,12 @@
-# SPDX-FileCopyrightText: Copyright (c) 2023 NVIDIA CORPORATION & AFFILIATES.
+# SPDX-FileCopyrightText: Copyright (c) 2023-2024, NVIDIA CORPORATION & AFFILIATES.
 # All rights reserved.
 # SPDX-License-Identifier: Apache-2.0
 
+from .fast_slow_proxy import is_proxy_object
 from .magics import load_ipython_extension
 from .profiler import Profiler
 
-__all__ = ["Profiler", "load_ipython_extension", "install"]
+__all__ = ["Profiler", "load_ipython_extension", "install", "is_proxy_object"]
 
 
 LOADED = False

python/cudf/cudf/pandas/fast_slow_proxy.py

@@ -1185,6 +1185,20 @@ def _replace_closurevars(
     )
 
 
+def is_proxy_object(obj: Any) -> bool:
+    """Determine if an object is proxy object
+
+    Parameters
+    ----------
+    obj : object
+        Any python object.
+
+    """
+    if _FastSlowProxyMeta in type(type(obj)).__mro__:
+        return True
+    return False
+
+
 NUMPY_TYPES: Set[str] = set(np.sctypeDict.values())
 
 

python/cudf/cudf_pandas_tests/test_cudf_pandas.py

@@ -20,7 +20,7 @@
 from pytz import utc
 
 from cudf.pandas import LOADED, Profiler
-from cudf.pandas.fast_slow_proxy import _Unusable
+from cudf.pandas.fast_slow_proxy import _Unusable, is_proxy_object
 
 if not LOADED:
     raise ImportError("These tests must be run with cudf.pandas loaded")
@@ -1488,3 +1488,17 @@ def mock_mean_none(self, *args, **kwargs):
 
 def test_excelwriter_pathlike():
     assert isinstance(pd.ExcelWriter("foo.xlsx"), os.PathLike)
+
+
+def test_is_proxy_object():
+    np_arr = np.array([1])
+
+    s1 = xpd.Series([1])
+    s2 = pd.Series([1])
+
+    np_arr_proxy = s1.to_numpy()
+
+    assert not is_proxy_object(np_arr)
+    assert is_proxy_object(np_arr_proxy)
+    assert is_proxy_object(s1)
+    assert not is_proxy_object(s2)

python/dask_cudf/dask_cudf/io/parquet.py

@@ -6,6 +6,7 @@
 from io import BufferedWriter, BytesIO, IOBase
 
 import numpy as np
+import pandas as pd
 from pyarrow import dataset as pa_ds, parquet as pq
 
 from dask import dataframe as dd
@@ -41,6 +42,10 @@ def _create_dd_meta(cls, dataset_info, **kwargs):
         meta_pd = super()._create_dd_meta(dataset_info, **kwargs)
 
         # Convert to cudf
+        # (drop unsupported timezone information)
+        for k, v in meta_pd.dtypes.items():
+            if isinstance(v, pd.DatetimeTZDtype) and v.tz is not None:
+                meta_pd[k] = meta_pd[k].dt.tz_localize(None)
         meta_cudf = cudf.from_pandas(meta_pd)
 
         # Re-set "object" dtypes to align with pa schema

python/dask_cudf/dask_cudf/io/tests/test_parquet.py

@@ -610,9 +610,8 @@ def test_timezone_column(tmpdir):
         }
     )
     pdf.to_parquet(path)
+
+    # Check that `cudf` and `dask_cudf` results match
     got = dask_cudf.read_parquet(path)
-    # cudf.read_parquet does not support reading timezone aware types yet
-    assert got["time"].dtype == pd.DatetimeTZDtype("ns", "UTC")
-    got["time"] = got["time"].astype("datetime64[ns]")
-    expected = cudf.read_parquet(path)
-    dd.assert_eq(got, expected)
+    expect = cudf.read_parquet(path)
+    dd.assert_eq(got, expect)