Drop Python 3.9 add 3.14 and add a few more lints

.github/workflows/oldest-support.yaml

@@ -33,7 +33,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
-          python-version: '3.9'  # Find a way to globally define our minimum python once
+          python-version: '3.10'  # Find a way to globally define our minimum python once
 
       - name: Install dependencies with minimum versions
         run: |

.github/workflows/regression-tests.yml

@@ -16,7 +16,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix: # Keep these in ascending order for automagic with coverage
-        python-version: ["3.9", "3.10", "3.11", "3.12", "3.13"]
+        python-version: ["3.10", "3.11", "3.12", "3.13", "3.14"]
 
     steps:
     - uses: actions/checkout@v4

.readthedocs.yaml

@@ -8,7 +8,7 @@ version: 2
 build:
   os: ubuntu-22.04
   tools:
-    python: "3.9"
+    python: "3.10"
 
 # Build documentation in the "docs/" directory with Sphinx
 sphinx:

docs/source/conf.py

@@ -72,7 +72,7 @@
 
 intersphinx_mapping = {
     "numpy": ("https://numpy.org/doc/stable/", None),
-    "python": ("http://docs.python.org/3.9/", None),
+    "python": ("http://docs.python.org/3.10/", None),
     "matplotlib": ("https://matplotlib.org/stable/", None),
 }
 

profiling/algorithms_profiling.ipynb

@@ -22,9 +22,11 @@
     "import pstats\n",
     "import subprocess\n",
     "from pathlib import Path\n",
-    "from typing import Callable, Optional, Union\n",
     "\n",
-    "from pyttb import cp_als, cp_apr, hosvd, import_data, sptensor, tensor, tucker_als"
+    "from pyttb import cp_als, cp_apr, hosvd, import_data, sptensor, tensor, tucker_als\n",
+    "\n",
+    "if TYPE_CHECKING:\n",
+    "    from collections.abc import Callable"
    ]
   },
   {

pyproject.toml

@@ -7,22 +7,23 @@ authors = [
 ]
 license = { text="BSD 2-Clause License" }
 readme = "README.md"
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 
 dependencies = [
     "numpy<3.0,>=1.24",
     "numpy_groupies>0.11",
-    "scipy<1.15,>1.9",
+    "scipy<2.0,>1.9",
     "matplotlib>3.7",
 ]
 
 classifiers = [
     "License :: OSI Approved :: BSD License",
     "Operating System :: OS Independent",
-    "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
 ]
 
 [project.urls]
@@ -68,7 +69,7 @@ requires = ["setuptools>=61.0", "numpy", "numpy_groupies", "scipy", "wheel"]
 build-backend = "setuptools.build_meta"
 
 [tool.ruff.lint]
-select = ["E", "F", "PL", "W", "I", "N", "NPY", "RUF", "B", "D", "FA", "UP", "TC"]
+select = ["E", "F", "PL", "W", "I", "N", "NPY", "RUF", "B", "D", "FA", "UP", "TC", "ARG", "PIE"]
 ignore = [
     # Ignored in conversion to ruff since not previously enforced
     "PLR2004",
@@ -86,6 +87,8 @@ ignore = [
     "B028",
     # Personal preference on magic method
     "D105",
+    # Rule removed in newer version of ruff
+    "UP038",
 ]
 [tool.ruff.lint.pydocstyle]
 convention = "numpy"

pyttb/cp_apr.py

@@ -664,7 +664,7 @@ def tt_cp_apr_pdnr(  # noqa: PLR0912,PLR0913,PLR0915
             nInnerIters[iteration] += countInnerIters[n]
 
         # Save output items for the outer iteration.
-        num_zero = 0
+        num_zero = np.intp(0)
         for n in range(N):
             num_zero += np.count_nonzero(M.factor_matrices[n] == 0)  # [0].size
 
@@ -1065,7 +1065,7 @@ def tt_cp_apr_pqnr(  # noqa: PLR0912,PLR0913,PLR0915
             nInnerIters[iteration] += countInnerIters[n]
 
         # Save output items for the outer iteration.
-        num_zero = 0
+        num_zero = np.intp(0)
         for n in range(N):
             num_zero += np.count_nonzero(M.factor_matrices[n] == 0)  # [0].size
 

pyttb/create_problem.py

@@ -4,8 +4,9 @@
 
 import logging
 import math
+from collections.abc import Callable
 from dataclasses import dataclass, field
-from typing import Callable, Union, cast, overload
+from typing import cast, overload
 
 import numpy as np
 from numpy_groupies import aggregate as accumarray
@@ -14,9 +15,7 @@
 from pyttb.pyttb_utils import Shape, parse_shape
 
 solution_generator = Callable[[tuple[int, ...]], np.ndarray]
-core_generator_t = Callable[
-    [tuple[int, ...]], Union[ttb.tensor, ttb.sptensor, np.ndarray]
-]
+core_generator_t = Callable[[tuple[int, ...]], ttb.tensor | ttb.sptensor | np.ndarray]
 
 
 def randn(shape: tuple[int, ...]) -> np.ndarray:
@@ -471,7 +470,7 @@ def generate_solution_factors(base_params: BaseProblem) -> list[np.ndarray]:
             f"{nfactors} and {shape}"
         )
     factor_matrices = []
-    for shape_i, nfactors_i in zip(shape, nfactors):
+    for shape_i, nfactors_i in zip(shape, nfactors, strict=False):
         factor_matrices.append(base_params.factor_generator((shape_i, nfactors_i)))
 
     if base_params.symmetric is not None:

pyttb/gcp/fg_setup.py

@@ -6,8 +6,8 @@
 
 from __future__ import annotations
 
+from collections.abc import Callable
 from functools import partial
-from typing import Callable
 
 import numpy as np
 

pyttb/gcp/optimizers.py

@@ -10,7 +10,7 @@
 import time
 from abc import ABC, abstractmethod
 from math import inf
-from typing import TYPE_CHECKING, Callable, TypedDict
+from typing import TYPE_CHECKING, TypedDict
 
 import numpy as np
 from scipy.optimize import fmin_l_bfgs_b
@@ -21,6 +21,8 @@
 from pyttb.gcp.samplers import GCPSampler
 
 if TYPE_CHECKING:
+    from collections.abc import Callable
+
     from pyttb.gcp.fg_setup import function_type
 
 
@@ -252,7 +254,7 @@ def update_step(  # noqa: D102
         step = self._decay**self._nfails * self._rate
         factor_matrices = [
             np.maximum(lower_bound, factor - step * grad)
-            for factor, grad in zip(model.factor_matrices, gradient)
+            for factor, grad in zip(model.factor_matrices, gradient, strict=False)
         ]
         return factor_matrices, step
 
@@ -348,19 +350,21 @@ def update_step(  # noqa: D102
         self._v_prev = self._v.copy()
         self._m = [
             self._beta_1 * mk + (1 - self._beta_1) * gk
-            for mk, gk in zip(self._m, gradient)
+            for mk, gk in zip(self._m, gradient, strict=False)
         ]
         self._v = [
             self._beta_2 * vk + (1 - self._beta_2) * gk**2
-            for vk, gk in zip(self._v, gradient)
+            for vk, gk in zip(self._v, gradient, strict=False)
         ]
         mhat = [mk / (1 - self._beta_1**self._total_iterations) for mk in self._m]
         vhat = [vk / (1 - self._beta_2**self._total_iterations) for vk in self._v]
         factor_matrices = [
             np.maximum(
                 lower_bound, factor_k - step * mhk / (np.sqrt(vhk) + self._epsilon)
             )
-            for factor_k, mhk, vhk in zip(model.factor_matrices, mhat, vhat)
+            for factor_k, mhk, vhk in zip(
+                model.factor_matrices, mhat, vhat, strict=False
+            )
         ]
         return factor_matrices, step
 
@@ -401,7 +405,7 @@ def update_step(  # noqa: D102
         step = 1.0 / np.sqrt(self._gnormsum)
         factor_matrices = [
             np.maximum(lower_bound, factor_k - step * gk)
-            for factor_k, gk in zip(model.factor_matrices, gradient)
+            for factor_k, gk in zip(model.factor_matrices, gradient, strict=False)
         ]
         return factor_matrices, step
 

pyttb/gcp/samplers.py

@@ -7,11 +7,12 @@
 from __future__ import annotations
 
 import logging
+from collections.abc import Callable
 from dataclasses import dataclass
 from enum import Enum
 from functools import partial
 from math import ceil
-from typing import Callable, Union, cast
+from typing import cast
 
 import numpy as np
 
@@ -21,7 +22,7 @@
 from pyttb.tensor import tensor
 
 sample_type = tuple[np.ndarray, np.ndarray, np.ndarray]
-sampler_type = Callable[[Union[tensor, sptensor]], sample_type]
+sampler_type = Callable[[tensor | sptensor], sample_type]
 
 
 @dataclass

pyttb/ktensor.py

@@ -8,11 +8,10 @@
 
 import logging
 import warnings
-from collections.abc import Sequence
+from collections.abc import Callable, Sequence
 from math import prod
 from typing import (
     TYPE_CHECKING,
-    Callable,
     Literal,
     cast,
     overload,

pyttb/matlab/matlab_support.py

@@ -5,15 +5,13 @@
 # U.S. Government retains certain rights in this software.
 from __future__ import annotations
 
-from typing import Union
-
 import numpy as np
 
 from pyttb.tensor import tensor
 
 from .matlab_utilities import _matlab_array_str
 
-PRINT_CLASSES = Union[tensor, np.ndarray]
+PRINT_CLASSES = tensor | np.ndarray
 
 
 def matlab_print(

pyttb/pyttb_utils.py

@@ -12,7 +12,6 @@
 from typing import (
     Any,
     Literal,
-    Union,
     cast,
     get_args,
     overload,
@@ -23,16 +22,16 @@
 
 import pyttb as ttb
 
-Shape = Union[int, Iterable[int]]
+Shape = int | Iterable[int]
 """Shape represents the object size or dimensions. It can be specified as
 either a single integer or an iterable of integers, which will be normalized
 to a tuple internally."""
 
-OneDArray = Union[int, float, Iterable[int], Iterable[float], np.ndarray]
+OneDArray = int | float | Iterable[int] | Iterable[float] | np.ndarray
 """OneDArray represents any one-dimensional array, which can be a single
 integer or float, and iterable of integerss or floats, or a NumPy array."""
 
-MemoryLayout = Union[Literal["F"], Literal["C"]]
+MemoryLayout = Literal["F"] | Literal["C"]
 """MemoryLayout is the set of options for the layout of a tensor.
 It can be "F", meaning Fortran ordered and analogous to column-major for matrices,
 or "C", meaning C ordered and analogous to row-major for matrices.
@@ -347,7 +346,7 @@ def tt_renumber(
     """
     newshape = np.array(shape)
     newsubs = subs
-    for i in range(0, len(shape)):
+    for i in range(len(shape)):
         if not number_range[i] == slice(None, None, None):
             if subs.size == 0:
                 if not isinstance(number_range[i], slice):
@@ -365,7 +364,7 @@ def tt_renumber(
                     #   without assert
                     number_range_i = number_range[i]
                     assert isinstance(number_range_i, slice)
-                    newshape[i] = len(range(0, shape[i])[number_range_i])
+                    newshape[i] = len(range(shape[i])[number_range_i])
             else:
                 newsubs[:, i], newshape[i] = tt_renumberdim(
                     subs[:, i], shape[i], number_range[i]
@@ -397,7 +396,7 @@ def tt_renumberdim(
         number_range = [int(number_range)]
         newshape = 0
     elif isinstance(number_range, slice):
-        number_range = list(range(0, shape))[number_range]
+        number_range = list(range(shape))[number_range]
         newshape = len(number_range)
     elif isinstance(number_range, (Sequence, np.ndarray)):
         newshape = len(number_range)
@@ -406,7 +405,7 @@ def tt_renumberdim(
 
     # Create map from old range to the new range
     idx_map = np.zeros(shape=shape)
-    for i in range(0, newshape):
+    for i in range(newshape):
         idx_map[number_range[i]] = int(i)
 
     # Do the mapping
@@ -492,7 +491,7 @@ def tt_ind2sub(
     return np.array(np.unravel_index(idx, shape, order=order)).transpose()
 
 
-def tt_subsubsref(obj: np.ndarray, s: Any) -> float | np.ndarray:
+def tt_subsubsref(obj: np.ndarray, s: Any) -> float | np.ndarray:  # noqa: ARG001
     """Helper function for tensor toolbox subsref.
 
     Parameters
@@ -759,8 +758,8 @@ class IndexVariant(Enum):
 
 
 # We probably want to create a specific file for utility types
-LinearIndexType = Union[int, np.integer, slice]
-IndexType = Union[LinearIndexType, Sequence[int], np.ndarray]
+LinearIndexType = int | np.integer | slice
+IndexType = LinearIndexType | Sequence[int] | np.ndarray
 
 
 def get_index_variant(indices: IndexType) -> IndexVariant:

pyttb/sptenmat.py

@@ -328,7 +328,7 @@ def shape(self) -> tuple[int, ...]:
             n = np.prod(np.array(self.tshape)[self.cdims])
             return int(m), int(n)
 
-    def double(self, immutable: bool = False) -> sparse.coo_matrix:
+    def double(self, immutable: bool = False) -> sparse.coo_matrix:  # noqa: ARG002
         """
         Convert a :class:`pyttb.sptenmat` to a COO :class:`scipy.sparse.coo_matrix`.
 
@@ -621,7 +621,7 @@ def __repr__(self):
 
         # An empty ndarray with minimum dimensions still has a shape
         if self.subs.size > 0:
-            for i in range(0, self.subs.shape[0]):
+            for i in range(self.subs.shape[0]):
                 s += "\t"
                 s += "["
                 idx = self.subs[i, :]