Support Python 3.12 and drop 3.9. (#656)

Author: hameerabbasi

.github/workflows/ci.yml

@@ -6,15 +6,15 @@ jobs:
     strategy:
       matrix:
         os: [ubuntu-latest]
-        python: ['3.9', '3.10', '3.11']
+        python: ['3.10', '3.11', '3.12']
         numba_boundscheck: [0]
         include:
           - os: macos-latest
-            python: '3.9'
+            python: '3.10'
           - os: windows-latest
-            python: '3.9'
+            python: '3.10'
           - os: ubuntu-latest
-            python: '3.9'
+            python: '3.10'
             numba_boundscheck: 1
       fail-fast: false
     runs-on: ${{ matrix.os }}
@@ -113,7 +113,7 @@ jobs:
           activate-environment: sparse-dev
           allow-softlinks: true
           environment-file: ci/environment.yml
-          python-version: '3.9'
+          python-version: '3.10'
           miniforge-version: latest
       - name: Install asv
         run: |

pyproject.toml

@@ -9,7 +9,7 @@ description = "Sparse n-dimensional arrays for the PyData ecosystem"
 readme = "README.rst"
 dependencies = ["numpy>=1.17", "numba>=0.49"]
 maintainers = [{ name = "Hameer Abbasi", email = "[email protected]" }]
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 license = { file = "LICENSE" }
 keywords = ["sparse", "numpy", "scipy", "dask"]
 classifiers = [
@@ -18,9 +18,9 @@ classifiers = [
     "License :: OSI Approved :: BSD License",
     "Programming Language :: Python",
     "Programming Language :: Python :: 3",
-    "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
     "Programming Language :: Python :: 3 :: Only",
     "Intended Audience :: Developers",
     "Intended Audience :: Science/Research",

sparse/pydata_backend/_common.py

@@ -73,7 +73,7 @@ def check_class_nan(test):
     from ._compressed import GCXS
     from ._coo import COO
 
-    if isinstance(test, (GCXS, COO)):
+    if isinstance(test, GCXS | COO):
         return nan_check(test.fill_value, test.data)
     if _is_scipy_sparse_obj(test):
         return nan_check(test.data)
@@ -248,7 +248,7 @@ def matmul(a, b):
         a = a[(None,) * (b.ndim - a.ndim)]
     if a.ndim > b.ndim:
         b = b[(None,) * (a.ndim - b.ndim)]
-    for i, j in zip(a.shape[:-2], b.shape[:-2]):
+    for i, j in zip(a.shape[:-2], b.shape[:-2], strict=True):
         if i != 1 and j != 1 and i != j:
             raise ValueError("shapes of a and b are not broadcastable")
 
@@ -655,11 +655,11 @@ def _dot_csr_csr(out_shape, a_data, b_data, a_indices, b_indices, a_indptr, b_in
             head = -2
             length = 0
             next_[:] = -1
-            for j, av in zip(
+            for j, av in zip(  # noqa: B905
                 a_indices[a_indptr[i] : a_indptr[i + 1]],
                 a_data[a_indptr[i] : a_indptr[i + 1]],
             ):
-                for k, bv in zip(
+                for k, bv in zip(  # noqa: B905
                     b_indices[b_indptr[j] : b_indptr[j + 1]],
                     b_data[b_indptr[j] : b_indptr[j + 1]],
                 ):
@@ -921,11 +921,11 @@ def _dot_coo_coo(out_shape, a_coords, b_coords, a_data, b_data, a_indptr, b_indp
             head = -2
             length = 0
             next_[:] = -1
-            for j, av in zip(
+            for j, av in zip(  # noqa: B905
                 a_coords[1, a_indptr[i] : a_indptr[i + 1]],
                 a_data[a_indptr[i] : a_indptr[i + 1]],
             ):
-                for k, bv in zip(
+                for k, bv in zip(  # noqa: B905
                     b_coords[1, b_indptr[j] : b_indptr[j + 1]],
                     b_data[b_indptr[j] : b_indptr[j + 1]],
                 ):
@@ -1425,7 +1425,7 @@ def einsum(*operands, **kwargs):
     sizes = {}
     for t, term in enumerate(terms):
         shape = operands[t].shape
-        for ix, d in zip(term, shape):
+        for ix, d in zip(term, shape, strict=False):
             if d != sizes.setdefault(ix, d):
                 raise ValueError(f"Inconsistent shape for index '{ix}'.")
             total.setdefault(ix, set()).add(t)
@@ -1437,7 +1437,7 @@ def einsum(*operands, **kwargs):
     # we could identify and dispatch to tensordot here?
 
     parrays = []
-    for term, array in zip(terms, operands):
+    for term, array in zip(terms, operands, strict=True):
         # calc the target indices for this term
         pterm = "".join(ix for ix in aligned_term if ix in term)
         if pterm != term:
@@ -1924,7 +1924,7 @@ def moveaxis(a, source, destination):
 
     order = [n for n in range(a.ndim) if n not in source]
 
-    for dest, src in sorted(zip(destination, source)):
+    for dest, src in sorted(zip(destination, source, strict=True)):
         order.insert(dest, src)
 
     return a.transpose(order)
@@ -2042,7 +2042,7 @@ def asarray(obj, /, *, dtype=None, format="coo", device=None, copy=False):
 
     format_dict = {"coo": COO, "dok": DOK, "gcxs": GCXS}
 
-    if isinstance(obj, (COO, DOK, GCXS)):
+    if isinstance(obj, COO | DOK | GCXS):
         return obj.asformat(format)
 
     if _is_scipy_sparse_obj(obj):
@@ -2051,7 +2051,7 @@ def asarray(obj, /, *, dtype=None, format="coo", device=None, copy=False):
             dtype = sparse_obj.dtype
         return sparse_obj.astype(dtype=dtype, copy=copy)
 
-    if np.isscalar(obj) or isinstance(obj, (np.ndarray, Iterable)):
+    if np.isscalar(obj) or isinstance(obj, np.ndarray | Iterable):
         sparse_obj = format_dict[format].from_numpy(np.asarray(obj))
         if dtype is None:
             dtype = sparse_obj.dtype
@@ -2068,7 +2068,7 @@ def _support_numpy(func):
 
     def wrapper_func(*args, **kwargs):
         x = args[0]
-        if isinstance(x, (np.ndarray, np.number)):
+        if isinstance(x, np.ndarray | np.number):
             warnings.warn(
                 f"Sparse {func.__name__} received dense NumPy array instead "
                 "of sparse array. Dispatching to NumPy function.",

sparse/pydata_backend/_compressed/indexing.py

@@ -185,7 +185,7 @@ def get_slicing_selection(arr_data, arr_indices, indptr, starts, ends, col):  #
     """
     indices = []
     ind_list = []
-    for i, (start, end) in enumerate(zip(starts, ends)):
+    for i, (start, end) in enumerate(zip(starts, ends)):  # noqa: B905
         inds = []
         current_row = arr_indices[start:end]
         if current_row.size < col.size:  # linear filtering
@@ -246,7 +246,7 @@ def get_array_selection(arr_data, arr_indices, indptr, starts, ends, col):  # pr
     """
     indices = []
     ind_list = []
-    for i, (start, end) in enumerate(zip(starts, ends)):
+    for i, (start, end) in enumerate(zip(starts, ends)):  # noqa: B905
         inds = []
         current_row = arr_indices[start:end]
         if len(current_row) == 0:

sparse/pydata_backend/_coo/common.py

@@ -2,7 +2,7 @@
 import warnings
 from collections.abc import Iterable
 from functools import reduce
-from typing import Any, NamedTuple, Optional
+from typing import Any, NamedTuple
 
 import numba
 
@@ -125,7 +125,7 @@ def kron(a, b):
     b_expanded_coords = b.coords[:, b_idx]
     o_coords = a_expanded_coords * np.asarray(b.shape)[:, None] + b_expanded_coords
     o_data = a.data[a_idx] * b.data[b_idx]
-    o_shape = tuple(i * j for i, j in zip(a.shape, b.shape))
+    o_shape = tuple(i * j for i, j in zip(a.shape, b.shape, strict=True))
 
     return COO(o_coords, o_data, shape=o_shape, has_duplicates=False)
 
@@ -783,7 +783,7 @@ def roll(a, shift, axis=None):
     # shift elements
     coords, data = np.copy(a.coords), np.copy(a.data)
     try:
-        for sh, ax in zip(shift, axis):
+        for sh, ax in zip(shift, axis, strict=True):
             coords[ax] += sh
             coords[ax] %= a.shape[ax]
     except UFuncTypeError as e:
@@ -1462,7 +1462,7 @@ def _compute_minmax_args(
 
 def _arg_minmax_common(
     x: SparseArray,
-    axis: Optional[int],
+    axis: int | None,
     keepdims: bool,
     mode: str,
 ):
@@ -1474,7 +1474,7 @@ def _arg_minmax_common(
 
     x = _validate_coo_input(x)
 
-    if not isinstance(axis, (int, type(None))):
+    if not isinstance(axis, int | type(None)):
         raise ValueError(f"`axis` must be `int` or `None`, but it's: {type(axis)}.")
     if isinstance(axis, int) and axis >= x.ndim:
         raise ValueError(f"`axis={axis}` is out of bounds for array of dimension {x.ndim}.")
@@ -1483,7 +1483,7 @@ def _arg_minmax_common(
 
     # If `axis` is None then we need to flatten the input array and memorize
     # the original dimensionality for the final reshape operation.
-    axis_none_original_ndim: Optional[int] = None
+    axis_none_original_ndim: int | None = None
     if axis is None:
         axis_none_original_ndim = x.ndim
         x = x.reshape(-1)[:, None]

sparse/pydata_backend/_coo/core.py

@@ -1577,7 +1577,7 @@ def as_coo(x, shape=None, fill_value=None, idx_dtype=None):
     if _is_scipy_sparse_obj(x):
         return COO.from_scipy_sparse(x)
 
-    if isinstance(x, (Iterable, Iterator)):
+    if isinstance(x, Iterable | Iterator):
         return COO.from_iter(x, shape=shape, fill_value=fill_value)
 
     raise NotImplementedError(

sparse/pydata_backend/_coo/indexing.py

@@ -238,7 +238,7 @@ def _prune_indices(indices, shape, prune_none=True):
         indices = [idx for idx in indices if idx is not None]
 
     i = 0
-    for idx, sh in zip(indices[::-1], shape[::-1]):
+    for idx, sh in zip(indices[::-1], shape[::-1], strict=True):
         if not isinstance(idx, slice):
             break
 

sparse/pydata_backend/_dok.py

@@ -181,7 +181,7 @@ def from_coo(cls, x):
         """
         ar = cls(x.shape, dtype=x.dtype, fill_value=x.fill_value)
 
-        for c, d in zip(x.coords.T, x.data):
+        for c, d in zip(x.coords.T, x.data, strict=True):
             ar.data[tuple(c)] = d
 
         return ar
@@ -235,7 +235,7 @@ def from_numpy(cls, x):
         coords = np.nonzero(x)
         data = x[coords]
 
-        for c in zip(data, *coords):
+        for c in zip(data, *coords, strict=True):
             d, c = c[0], c[1:]
             ar.data[c] = d
 
@@ -338,7 +338,7 @@ def __getitem__(self, key):
     def _fancy_getitem(self, key):
         """Subset of fancy indexing, when all dimensions are accessed"""
         new_data = {}
-        for i, k in enumerate(zip(*key)):
+        for i, k in enumerate(zip(*key, strict=True)):
             if k in self.data:
                 new_data[i] = self.data[k]
         return DOK(
@@ -352,7 +352,7 @@ def __setitem__(self, key, value):
         value = np.asarray(value, dtype=self.dtype)
 
         # 1D fancy indexing
-        if self.ndim == 1 and isinstance(key, Iterable) and all(isinstance(i, (int, np.integer)) for i in key):
+        if self.ndim == 1 and isinstance(key, Iterable) and all(isinstance(i, int | np.integer) for i in key):
             key = (key,)
 
         if isinstance(key, tuple) and all(isinstance(k, Iterable) for k in key):
@@ -383,7 +383,7 @@ def _fancy_setitem(self, idxs, values):
             raise ValueError(f"Shape mismatch of indices ({idxs[0].shape}) and values ({values.shape})!")
         fill_value = self.fill_value
         data = self.data
-        for idx, value in zip(zip(*idxs), values):
+        for idx, value in zip(zip(*idxs, strict=True), values, strict=True):
             if value != fill_value:
                 data[idx] = value
             elif idx in data:

sparse/pydata_backend/_slicing.py

@@ -53,7 +53,7 @@ def normalize_index(idx, shape):
         else:
             none_shape.append(None)
 
-    for i, d in zip(idx, none_shape):
+    for i, d in zip(idx, none_shape, strict=True):
         if d is not None:
             check_index(i, d)
     idx = tuple(map(sanitize_index, idx))
@@ -206,7 +206,7 @@ def posify_index(shape, ind):
             return ind + shape
 
         return ind
-    if isinstance(ind, (np.ndarray, list)) and not math.isnan(shape):
+    if isinstance(ind, np.ndarray | list) and not math.isnan(shape):
         ind = np.asanyarray(ind)
         return np.where(ind < 0, ind + shape, ind)
     if isinstance(ind, slice):

sparse/pydata_backend/_sparse_array.py

@@ -2,10 +2,9 @@
 import operator
 import warnings
 from abc import ABCMeta, abstractmethod
-from collections.abc import Iterable
+from collections.abc import Callable, Iterable
 from functools import reduce
 from numbers import Integral
-from typing import Callable
 
 import numpy as np
 
@@ -695,7 +694,7 @@ def mean(self, axis=None, keepdims=False, dtype=None, out=None):
         den = reduce(operator.mul, (self.shape[i] for i in axis), 1)
 
         if dtype is None:
-            if issubclass(self.dtype.type, (np.integer, np.bool_)):
+            if issubclass(self.dtype.type, np.integer | np.bool_):
                 dtype = inter_dtype = np.dtype("f8")
             else:
                 dtype = self.dtype
@@ -786,7 +785,7 @@ def var(self, axis=None, dtype=None, out=None, ddof=0, keepdims=False):
             warnings.warn("Degrees of freedom <= 0 for slice", RuntimeWarning, stacklevel=1)
 
         # Cast bool, unsigned int, and int to float64 by default
-        if dtype is None and issubclass(self.dtype.type, (np.integer, np.bool_)):
+        if dtype is None and issubclass(self.dtype.type, np.integer | np.bool_):
             dtype = np.dtype("f8")
 
         arrmean = self.sum(axis, dtype=dtype, keepdims=True)