Update to latest version (#2186901)

Signed-off-by: Fedora Release Engineering <releng@fedoraproject.org>

.gitignore

@@ -54,3 +54,19 @@
 /dask-2022.01.0.tar.gz
 /dask-2022.01.1.tar.gz
 /dask-2022.02.0.tar.gz
+/dask-2022.02.1.tar.gz
+/dask-2022.05.0.tar.gz
+/dask-2022.5.0.tar.gz
+/dask-2022.7.1.tar.gz
+/dask-2022.8.0.tar.gz
+/dask-2022.8.1.tar.gz
+/dask-2022.9.0.tar.gz
+/dask-2022.10.0.tar.gz
+/dask-2022.11.1.tar.gz
+/dask-2022.12.0.tar.gz
+/dask-2022.12.1.tar.gz
+/dask-2023.1.0.tar.gz
+/dask-2023.2.0.tar.gz
+/dask-2023.3.2.tar.gz
+/dask-2023.4.0.tar.gz
+/dask-2023.4.1.tar.gz

0001-Skip-test_encoding_gh601-on-big-endian-machines.patch

@@ -1,4 +1,4 @@
-From 5d9cdaa75e0547d8d5edf8c995f29279688f7e11 Mon Sep 17 00:00:00 2001
+From 5ee683719a19c8865a7b80725a0adf55baa3e9c9 Mon Sep 17 00:00:00 2001
 From: Elliott Sales de Andrade <quantum.analyst@gmail.com>
 Date: Sun, 7 Mar 2021 04:07:32 -0500
 Subject: [PATCH 1/2] Skip test_encoding_gh601 on big-endian machines.
@@ -9,18 +9,18 @@ Signed-off-by: Elliott Sales de Andrade <quantum.analyst@gmail.com>
  1 file changed, 2 insertions(+)
 
 diff --git a/dask/dataframe/io/tests/test_csv.py b/dask/dataframe/io/tests/test_csv.py
-index 48780ef8..f63d0e80 100644
+index b50a727b..0ac1df4c 100644
 --- a/dask/dataframe/io/tests/test_csv.py
 +++ b/dask/dataframe/io/tests/test_csv.py
-@@ -1,6 +1,7 @@
- import gzip
+@@ -2,6 +2,7 @@ import gzip
  import os
- from io import BytesIO
+ import warnings
+ from io import BytesIO, StringIO
 +import sys
  from unittest import mock
  
  import pytest
-@@ -1099,6 +1100,7 @@ def test_read_csv_with_datetime_index_partitions_n():
+@@ -1146,6 +1147,7 @@ def test_read_csv_with_datetime_index_partitions_n():
  xfail_pandas_100 = pytest.mark.xfail(reason="https://github.com/dask/dask/issues/5787")
  
  
@@ -29,5 +29,5 @@ index 48780ef8..f63d0e80 100644
      "encoding",
      [
 -- 
-2.31.1
+2.40.0
 

0002-Skip-coverage-testing.patch

@@ -0,0 +1,26 @@
+From 54487feeef6d7d41e3b6d4e4ece5b6342071068b Mon Sep 17 00:00:00 2001
+From: Elliott Sales de Andrade <quantum.analyst@gmail.com>
+Date: Sun, 7 May 2023 23:13:59 -0400
+Subject: [PATCH 2/2] Skip coverage testing
+
+Signed-off-by: Elliott Sales de Andrade <quantum.analyst@gmail.com>
+---
+ pyproject.toml | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/pyproject.toml b/pyproject.toml
+index a7d504ff..36dfca8e 100644
+--- a/pyproject.toml
++++ b/pyproject.toml
+@@ -121,7 +121,7 @@ markers = [
+     "skip_with_pyarrow_strings: Tests that should be skipped when pyarrow string conversion is turned on",
+     "xfail_with_pyarrow_strings: Tests that should be xfailed when pyarrow string conversion is turned on",
+ ]
+-addopts = "-v -rsxfE --durations=10 --color=yes --cov-config=pyproject.toml"
++addopts = "-v -rsxfE --durations=10 --color=yes"
+ filterwarnings = [
+     # From Cython-1753
+     "ignore:can't resolve:ImportWarning",
+-- 
+2.40.0
+

0002-fix-index_col-duplication-if-index_col-is-type-str.patch

@@ -1,30 +0,0 @@
-From 41306c1829a22f0e8654f8d75dc23ce766c26ea8 Mon Sep 17 00:00:00 2001
-From: McToel <theo.doellmann@gmx.de>
-Date: Sun, 16 May 2021 11:11:06 +0200
-Subject: [PATCH 2/2] fix index_col duplication if index_col is type str
-
-Signed-off-by: Elliott Sales de Andrade <quantum.analyst@gmail.com>
----
- dask/dataframe/io/sql.py | 6 ++----
- 1 file changed, 2 insertions(+), 4 deletions(-)
-
-diff --git a/dask/dataframe/io/sql.py b/dask/dataframe/io/sql.py
-index 6573a325..9c15360c 100644
---- a/dask/dataframe/io/sql.py
-+++ b/dask/dataframe/io/sql.py
-@@ -125,10 +125,8 @@ def read_sql_table(
-         if columns
-         else list(table.columns)
-     )
--    if index_col not in columns:
--        columns.append(
--            table.columns[index_col] if isinstance(index_col, str) else index_col
--        )
-+    if index not in columns:
-+        columns.append(index)
- 
-     if isinstance(index_col, str):
-         kwargs["index_col"] = index_col
--- 
-2.31.1
-

_version.py

@@ -1,21 +0,0 @@
-
-# This file was generated by 'versioneer.py' (0.16) from
-# revision-control system data, or from the parent directory name of an
-# unpacked source archive. Distribution tarballs contain a pre-generated copy
-# of this file.
-
-import json
-import sys
-
-version_json = '''
-{
- "dirty": false,
- "error": null,
- "full-revisionid": "a3299586b05d422800ff8a618a2ba35e00e62665",
- "version": "2022.01.0"
-}
-'''  # END VERSION_JSON
-
-
-def get_versions():
-    return json.loads(version_json)

python-dask.spec

@@ -5,21 +5,23 @@
 # Force bootstrap for package review.
 %bcond_without bootstrap
 
+# We have an arched package to detect arch-dependent issues in dependencies,
+# but all of the installable RPMs are noarch and there is no compiled code.
+%global debug_package %{nil}
+
 Name:           python-%{srcname}
-Version:        2022.2.0
-%global tag     2022.02.0
+Version:        2023.4.1
+%global tag     2023.4.1
 Release:        %autorelease
 Summary:        Parallel PyData with Task Scheduling
 
-License:        BSD
-URL:            https://github.com/dask/dask/
-Source0:        https://github.com/dask/dask/archive/%{tag}/%{srcname}-%{tag}.tar.gz
-# Grab this from the sdist until we switch back to it.
-Source1:        _version.py
+License:        BSD-3-Clause
+URL:            https://github.com/dask/dask
+Source0:        %{pypi_source %{srcname}}
 # https://github.com/dask/dask/issues/6725
-Patch0001:      0001-Skip-test_encoding_gh601-on-big-endian-machines.patch
-
-BuildArch:      noarch
+Patch:          0001-Skip-test_encoding_gh601-on-big-endian-machines.patch
+# Fedora-specific patch.
+Patch:          0002-Skip-coverage-testing.patch
 
 %description
 Dask is a flexible parallel computing library for analytics.
@@ -28,6 +30,8 @@ Dask is a flexible parallel computing library for analytics.
 %package -n     python3-%{srcname}
 Summary:        %{summary}
 
+BuildArch:      noarch
+
 BuildRequires:  python3-devel
 BuildRequires:  python3dist(graphviz)
 BuildRequires:  python3dist(ipython)
@@ -36,27 +40,54 @@ BuildRequires:  python3dist(scikit-image)
 BuildRequires:  python3dist(xarray)
 %endif
 # Optional test requirements.
+# Fastavro does not support 32 bit architectures and is ExcludeArch:
+# https://bugzilla.redhat.com/show_bug.cgi?id=1943932
+%ifnarch %{arm32} %{ix86}
 BuildRequires:  python3dist(fastavro)
+%endif
 BuildRequires:  python3dist(h5py)
 BuildRequires:  python3dist(psutil)
+# libarrow does not support 32 bit architectures and is ExcludeArch.
+# Tests don't pass on s390x either.
+%ifnarch %{arm} %{ix86} s390x
+BuildRequires:  python3dist(pyarrow)
+%endif
 BuildRequires:  python3dist(requests)
 BuildRequires:  python3dist(sqlalchemy)
+BuildRequires:  python3dist(tables)
 BuildRequires:  python3dist(zarr)
 
 Recommends:     python3-%{srcname}+array = %{version}-%{release}
 Recommends:     python3-%{srcname}+bag = %{version}-%{release}
 Recommends:     python3-%{srcname}+dataframe = %{version}-%{release}
 Recommends:     python3-%{srcname}+delayed = %{version}-%{release}
-Recommends:     python3-%{srcname}+diagnostics = %{version}-%{release}
 %if %{without bootstrap}
 Recommends:     python3-%{srcname}+distributed = %{version}-%{release}
 %endif
+# No recent enough Bokeh is packaged
+Obsoletes:      python3-%{srcname}+diagnostics < 2022.5.0-1
+
+# There is nothing that can be unbundled; there are some some snippets forked
+# or copied from unspecified versions of numpy, under a BSD-3-Clause license
+# similar to that of dask itself.
+#
+# - dask/array/numpy_compat.py:
+#     _Recurser, moveaxis, rollaxis, sliding_window_view
+# - dask/array/backends.py:
+#     _tensordot
+# - dask/array/core.py:
+#     block
+# - dask/array/einsumfuncs.py:
+#     parse_einsum_input
+# - dask/array/routines.py:
+#     cov, _average
+Provides:       bundled(numpy)
 
 %description -n python3-%{srcname}
 Dask is a flexible parallel computing library for analytics.
 
 
-%pyproject_extras_subpkg -n python3-%{srcname} array bag dataframe delayed diagnostics
+%pyproject_extras_subpkg -n python3-%{srcname} array bag dataframe delayed
 %if %{without bootstrap}
 %pyproject_extras_subpkg distributed
 %endif
@@ -66,6 +97,8 @@ Dask is a flexible parallel computing library for analytics.
 %package -n python-%{srcname}-doc
 Summary:        dask documentation
 
+BuildArch:      noarch
+
 BuildRequires:  python3dist(dask_sphinx_theme) >= 1.3.5
 BuildRequires:  python3dist(numpydoc)
 BuildRequires:  python3dist(sphinx) >= 4
@@ -76,12 +109,13 @@ Documentation for dask.
 
 
 %prep
-%autosetup -n %{srcname}-%{tag} -p1
-cp %SOURCE1 %{srcname}/
+%autosetup -n %{srcname}-%{version} -p1
+# we don't use pre-commit when running tests
+sed -i '/"pre-commit"/d' setup.py
 
 
 %generate_buildrequires
-%pyproject_buildrequires -r -x test,array,bag,dataframe,delayed,diagnostics
+%pyproject_buildrequires -r -x test,array,bag,dataframe,delayed
 %if %{without bootstrap}
 %pyproject_buildrequires -x distributed
 %endif
@@ -110,28 +144,93 @@ rm -rf html/.{doctrees,buildinfo}
 %global have_arm 1
 %endif
 
+%if 0%{?__isa_bits} == 32
+# read_sql_query with meta converts dtypes from 32 to 64.
+# https://github.com/dask/dask/issues/8620
+
+# >           tm.assert_frame_equal(
+#                 a, b, check_names=check_names, check_dtype=check_dtype, **kwargs
+# E               AssertionError: Attributes of DataFrame.iloc[:, 1] (column name="age") are different
+# E
+# E               Attribute "dtype" are different
+# E               [left]:  int32
+# E               [right]: int64
+# dask/dataframe/utils.py:555: AssertionError
+k="${k-}${k+ and }not test_query_with_meta"
+%endif
+
+%ifarch ppc64le
+# TODO: Should this be reported upstream? Is it a dask issue, or a numpy one?
+# Possibly related to
+# https://fedoraproject.org/wiki/Changes/PPC64LE_Float128_Transition?
+
+# >           assert allclose(a, b, equal_nan=equal_nan, **kwargs), msg
+# E           AssertionError: found values in 'a' and 'b' which differ by more than the allowed amount
+# E           assert False
+# E            +  where False = allclose(array([0.12586355-0.09957204j, 0.20256483+0.04098342j,\n       0.05781123-0.03588671j, 0.01135963-0.03334219j,\n       0.03747771+0.07495994j, 0.2106574 -0.0363521j ,\n       0.16352091+0.03782915j, 0.1381678 -0.06815128j,\n       0.03781295-0.04011523j, 0.01493269+0.07780643j]), array([0.12559072-0.07164038j, 0.20256483+0.05438578j,\n       0.05781123-0.03588671j, 0.01135963-0.03334219j,\n       0.03747771+0.07495994j, 0.2106574 -0.0363521j ,\n       0.16352091+0.03782915j, 0.1381678 -0.06815128j,\n       0.03781295-0.04011523j, 0.01493269+0.07780643j]), equal_nan=True, **{})
+# dask/array/utils.py:361: AssertionError
+k="${k-}${k+ and }not test_lstsq[100-10-10-True]"
+# >           assert allclose(a, b, equal_nan=equal_nan, **kwargs), msg
+# E           AssertionError: found values in 'a' and 'b' which differ by more than the allowed amount
+# E           assert False
+# E            +  where False = allclose(array([ 0.20168675+0.08857556j,  0.144233  -0.19173091j,\n       -0.03367557-0.08053959j,  0.04108325-0.24648308j,\n       -0.01844576+0.00841932j,  0.29652375+0.05682199j,\n        0.05551828+0.20156798j, -0.08409592+0.02354949j,\n        0.09848743-0.00748637j,  0.22889193-0.07372773j]), array([ 0.20067551+0.2642591j ,  0.144233  -0.18573336j,\n       -0.03367557-0.08053959j,  0.04108325-0.24648308j,\n       -0.01844576+0.00841932j,  0.29652375+0.05682199j,\n        0.05551828+0.20156798j, -0.08409592+0.02354949j,\n        0.09848743-0.00748637j,  0.22889193-0.07372773j]), equal_nan=True, **{})
+# dask/array/utils.py:361: AssertionError
+k="${k-}${k+ and }not test_lstsq[20-10-5-True]"
+
+# test_vdot fails with NumPy 1.19.0
+# https://github.com/dask/dask/issues/6406
+#
+# vdot returns incorrect results on ppc64le
+# https://github.com/numpy/numpy/issues/17087
+
+# >           assert allclose(a, b, equal_nan=equal_nan, **kwargs), msg
+# E           AssertionError: found values in 'a' and 'b' which differ by more than the allowed amount
+# E           assert False
+# E            +  where False = allclose((0.38772781971416226-0.6851997484294434j), (0.38772781971416226-0.306563166009585j), equal_nan=True, **{})
+# dask/array/utils.py:361: AssertionError
+k="${k-}${k+ and }not test_vdot[shape0-chunks0]"
+# >           assert allclose(a, b, equal_nan=equal_nan, **kwargs), msg
+# E           AssertionError: found values in 'a' and 'b' which differ by more than the allowed amount
+# E           assert False
+# E            +  where False = allclose((0.38772781971416226-0.6851997484294434j), (0.38772781971416226-0.306563166009585j), equal_nan=True, **{})
+# dask/array/utils.py:361: AssertionError
+k="${k-}${k+ and }not test_vdot[shape1-chunks1]"
+%endif
+
+# This test compares against files in .github/. It does not work on the PyPI
+# sdist, and is only relevant to upstream CI anyway.
+#
+# test_development_guidelines_matches_ci fails from sdist
+# https://github.com/dask/dask/issues/8499
+k="${k-}${k+ and }not test_development_guidelines_matches_ci"
+
 pytest_args=(
   -m 'not network'
 
-  # https://bugzilla.redhat.com/show_bug.cgi?id=1968947#c4
-  --deselect=dask/dataframe/io/tests/test_sql.py::test_select_from_select
-
   -n %[0%{?have_arm}?"2":"auto"]
 
-  # Ignore https://github.com/numpy/numpy/issues/20225 and Pandas issues
-  -W ignore::DeprecationWarning
+%ifarch %{ix86}
+  # Ignore 32-bit warning
+  -W 'ignore:invalid value encountered in cast:RuntimeWarning'
+%endif
+
+  -k "${k-}"
+
+  --pyargs dask
 )
 
+cd docs
 %{pytest} "${pytest_args[@]}"
 
 %files -n python3-%{srcname} -f %{pyproject_files}
 %doc README.rst
-%license LICENSE.txt
+%license LICENSE.txt dask/array/NUMPY_LICENSE.txt
+%{_bindir}/dask
 
 %if %{without bootstrap}
 %files -n python-%{srcname}-doc
 %doc html
-%license LICENSE.txt
+%license LICENSE.txt dask/array/NUMPY_LICENSE.txt
 %endif
 
 

sources

@@ -1 +1 @@
-SHA512 (dask-2022.02.0.tar.gz) = 397dabd3caabac0657808ff031299a1461c7acf4a123db4d1876d8e72a0ae971266dcb256e29498b1988b1b4bea506db3aec962b876f280ec532b96da35d356c
+SHA512 (dask-2023.4.1.tar.gz) = 2fad1317aad845f7b11d8efceafb11eb9e945aca649bcf7ccd71e8692c3ca18c6f595c82b4e2930cba4e9622dd384965b3cf0ea3f68bdbc2f6f18c7a2d806b26