Py upgrade

.github/workflows/test.yml

@@ -13,15 +13,15 @@ jobs:
         python-version: ['3.12']
         include:
           - BACKEND: 'numpy'
-            python-version: '3.9'
+            python-version: '3.10'
           - BACKEND: 'numpy'
-            python-version: '3.11'
+            python-version: '3.13'
             TENSORLY_TENALG_BACKEND: ['einsum']
           - BACKEND: 'jax'
             python-version: '3.13'
             TENSORLY_TENALG_BACKEND: ['einsum']
           - BACKEND: 'tensorflow'
-            python-version: '3.12'
+            python-version: '3.13'
             TENSORLY_TENALG_BACKEND: ['einsum']
 
     steps:

examples/applications/plot_image_compression.py

@@ -12,7 +12,6 @@
 from scipy.ndimage import zoom
 from tensorly.decomposition import parafac
 from tensorly.decomposition import tucker
-from math import ceil
 
 
 random_state = 12345

examples/decomposition/plot_permute_factors.py

@@ -13,7 +13,6 @@
 # The permuted tensor (or list of tensors) and list of permutation for each permuted tensors are returned.
 # Tensorly CPTensor should be used as an input to permute their factors and weights simultaneously.
 
-import tensorly as tl
 from tensorly.random import random_cp
 from tensorly.cp_tensor import cp_permute_factors
 import matplotlib.pyplot as plt

tensorly/__init__.py

@@ -31,7 +31,6 @@
     pad_tt_rank,
 )
 from .tt_matrix import (
-    tt_matrix_to_tensor,
     tt_matrix_to_tensor,
     validate_tt_matrix_rank,
     tt_matrix_to_unfolded,
@@ -72,7 +71,6 @@
     where,
     conj,
     index,
-    index_update,
     clip,
     max,
     min,

tensorly/backend/core.py

@@ -1,9 +1,3 @@
-import inspect
-import importlib
-import os
-import sys
-import threading
-import types
 import warnings
 import math
 
@@ -304,6 +298,11 @@ def to_numpy(tensor):
     def copy(tensor):
         """Return a copy of the given tensor"""
         raise NotImplementedError
+
+    @staticmethod
+    def ndim(tensor):
+        """Return the number of dimensions of a tensor"""
+        return tensor.ndim
 
     @staticmethod
     def concatenate(tensors, axis=0):

tensorly/backend/cupy_backend.py

@@ -10,7 +10,6 @@
     )
     raise ImportError(message) from error
 
-import warnings
 
 from .core import (
     Backend,
@@ -39,10 +38,6 @@ def to_numpy(tensor):
             return cp.asnumpy(tensor)
         return tensor
 
-    @staticmethod
-    def ndim(tensor):
-        return tensor.ndim
-
     @staticmethod
     def clip(tensor, a_min=None, a_max=None):
         return cp.clip(tensor, a_min, a_max)

tensorly/backend/jax_backend.py

@@ -51,10 +51,6 @@ def copy(self, tensor):
         return self.tensor(tensor.copy(), **self.context(tensor))
         # return copy.copy(tensor)
 
-    @staticmethod
-    def ndim(tensor):
-        return tensor.ndim
-
     @staticmethod
     def lstsq(a, b, rcond=None):
         return np.linalg.lstsq(a, b, rcond=rcond, numpy_resid=True)

tensorly/backend/numpy_backend.py

@@ -25,10 +25,6 @@ def is_tensor(tensor):
     def to_numpy(tensor):
         return np.copy(tensor)
 
-    @staticmethod
-    def ndim(tensor):
-        return tensor.ndim
-
     @staticmethod
     def clip(tensor, a_min=None, a_max=None):
         return np.clip(tensor, a_min, a_max)

tensorly/backend/paddle_backend.py

@@ -1,5 +1,5 @@
 from __future__ import annotations
-from typing import Sequence
+from collections.abc import Sequence
 
 from packaging.version import Version
 import warnings
@@ -158,10 +158,6 @@ def to_numpy(tensor: paddle.Tensor):
     def shape(tensor: paddle.Tensor):
         return tuple(tensor.shape)
 
-    @staticmethod
-    def ndim(tensor: paddle.Tensor):
-        return tensor.ndim
-
     @staticmethod
     def clip(tensor: paddle.Tensor, a_min=None, a_max=None, inplace=False):
         if inplace:
@@ -176,7 +172,7 @@ def all(tensor: paddle.Tensor):
 
     def transpose(self, tensor: paddle.Tensor, axes: int | Sequence[int] | None = None):
         axes = axes or list(range(self.ndim(tensor)))[::-1]
-        if not isinstance(axes, (tuple, list)):
+        if not isinstance(axes, tuple | list):
             axes = list(axes)
         return tensor.transpose(axes)
 

tensorly/backend/pytorch_backend.py

@@ -85,10 +85,6 @@ def to_numpy(tensor):
     def shape(tensor):
         return tuple(tensor.shape)
 
-    @staticmethod
-    def ndim(tensor):
-        return tensor.dim()
-
     @staticmethod
     def arange(start, stop=None, step=1.0, *args, **kwargs):
         if stop is None:

tensorly/backend/tensorflow_backend.py

@@ -180,7 +180,6 @@ def logsumexp(tensor, axis=0):
         "mean",
         "sum",
         "moveaxis",
-        "ndim",
         "arange",
         "sort",
         "argsort",

tensorly/contrib/sparse/tests/test_decomposition.py

@@ -1,5 +1,4 @@
 from ..decomposition import parafac
-from ..tenalg import multi_mode_dot
 from ..cp_tensor import cp_to_tensor
 from .... import backend as tl
 

tensorly/cp_tensor.py

@@ -177,7 +177,7 @@ def _validate_cp_tensor(cp_tensor):
     if isinstance(cp_tensor, CPTensor):
         # it's already been validated at creation
         return cp_tensor.shape, cp_tensor.rank
-    elif isinstance(cp_tensor, (float, int)):  # 0-order tensor
+    elif isinstance(cp_tensor, float | int):  # 0-order tensor
         return 0, 0
 
     weights, factors = cp_tensor

tensorly/datasets/data_imports.py

@@ -4,10 +4,6 @@
 
 from os.path import dirname
 import numpy as np
-from urllib.request import urlopen
-import scipy.io
-from zipfile import ZipFile
-from io import BytesIO
 import tensorly as tl
 
 

tensorly/datasets/tests/test_imports.py

@@ -1,4 +1,3 @@
-import pytest
 
 from ..data_imports import (
     load_IL2data,

tensorly/decomposition/_constrained_cp.py

@@ -121,7 +121,7 @@ def initialize_constrained_parafac(
 
             factors.append(U[:, :rank])
 
-    elif isinstance(init, (tuple, list, CPTensor)):
+    elif isinstance(init, tuple | list | CPTensor):
         try:
             weights, factors = CPTensor(init)
 

tensorly/decomposition/_cp.py

@@ -99,7 +99,7 @@ def initialize_cp(
 
         kt = CPTensor((None, factors))
 
-    elif isinstance(init, (tuple, list, CPTensor)):
+    elif isinstance(init, tuple | list | CPTensor):
         # TODO: Test this
         try:
             if normalize_factors is True:

tensorly/decomposition/_parafac2.py

@@ -1,5 +1,6 @@
 from warnings import warn
-from typing import Iterable, Optional, Sequence, Literal, Union
+from typing import Literal
+from collections.abc import Iterable, Sequence
 
 import tensorly as tl
 from ._base_decomposition import DecompositionMixin
@@ -140,7 +141,7 @@ def initialize_decomposition(
         projections = _compute_projections(tensor_slices, (A, B, C), svd)
         return Parafac2Tensor((None, [A, B, C], projections))
 
-    elif isinstance(init, (tuple, list, Parafac2Tensor, CPTensor)):
+    elif isinstance(init, tuple | list | Parafac2Tensor | CPTensor):
         try:
             decomposition = Parafac2Tensor.from_CPTensor(init, parafac2_tensor_ok=True)
         except ValueError:
@@ -404,9 +405,9 @@ def parafac2(
     svd: SVD_TYPES = "truncated_svd",
     normalize_factors: bool = False,
     tol: float = 1.0e-8,
-    nn_modes: Optional[Union[Sequence[int], Literal["all"]]] = None,
+    nn_modes: Sequence[int] | Literal["all"] | None = None,
     random_state=None,
-    verbose: Union[bool, int] = False,
+    verbose: bool | int = False,
     return_errors: bool = False,
     n_iter_parafac: int = 5,
     linesearch: bool = True,

tensorly/decomposition/_tt.py

@@ -1,7 +1,7 @@
 import tensorly as tl
 from ._base_decomposition import DecompositionMixin
 from ..tt_tensor import validate_tt_rank, TTTensor
-from ..tt_matrix import validate_tt_matrix_rank, TTMatrix
+from ..tt_matrix import TTMatrix
 from ..tenalg.svd import svd_interface
 
 

tensorly/decomposition/_tucker.py

@@ -8,9 +8,6 @@
     validate_tucker_rank,
     tucker_normalize,
 )
-from ..solvers.penalizations import (
-    process_regularization_weights,
-)
 from ..solvers.nnls import hals_nnls, fista, active_set_nnls
 from math import sqrt
 import warnings

tensorly/decomposition/tests/test_constrained_parafac.py

@@ -11,7 +11,6 @@
     assert_array_almost_equal,
     assert_class_wrapper_correctly_passes_arguments,
 )
-from ...random import random_cp
 
 
 def test_constrained_parafac_nonnegative(monkeypatch):

tensorly/decomposition/tests/test_cp.py

@@ -18,7 +18,6 @@
     CP_NN_HALS,
 )
 from ...cp_tensor import cp_to_tensor
-from ...cp_tensor import cp_to_tensor
 from ...random import random_cp
 from ...tenalg import khatri_rao
 from ... import backend as T

tensorly/decomposition/tests/test_tr_svd.py

@@ -1,13 +1,10 @@
-import pytest
-import tensorly as tl
 
-from .._tr_svd import tensor_ring, TensorRing
+from .._tr_svd import tensor_ring
 from ...random import random_tr
 from ...testing import (
     assert_,
     assert_array_almost_equal,
     assert_raises,
-    assert_class_wrapper_correctly_passes_arguments,
 )
 
 

tensorly/random/base.py

@@ -1,4 +1,3 @@
-import numpy as np
 from ..cp_tensor import cp_to_tensor, CPTensor, cp_normalize, validate_cp_rank
 from ..tucker_tensor import tucker_to_tensor, TuckerTensor, validate_tucker_rank
 from ..tt_tensor import tt_to_tensor, TTTensor, validate_tt_rank

tensorly/solvers/penalizations.py

@@ -1,6 +1,3 @@
-import tensorly as tl
-import numpy as np
-import copy
 import warnings
 
 
@@ -24,10 +21,10 @@ def process_regularization_weights(ridge_coefficients, sparsity_coefficients, n_
     list of floats
         sparsity coefficients, processed
     """
-    if ridge_coefficients is None or isinstance(ridge_coefficients, (int, float)):
+    if ridge_coefficients is None or isinstance(ridge_coefficients, int | float):
         # Populate None or the input float in a list for all modes
         ridge_coefficients = [ridge_coefficients] * n_modes
-    if sparsity_coefficients is None or isinstance(sparsity_coefficients, (int, float)):
+    if sparsity_coefficients is None or isinstance(sparsity_coefficients, int | float):
         # Populate None or the input float in a list for all modes
         sparsity_coefficients = [sparsity_coefficients] * n_modes
     # Convert None to 0

tensorly/solvers/tests/test_admm.py

@@ -2,14 +2,13 @@
 import tensorly as tl
 
 from tensorly.solvers.admm import admm
-from tensorly.testing import assert_, assert_array_equal, assert_array_almost_equal
-from tensorly import tensor_to_vec, truncated_svd
+from tensorly.testing import assert_array_almost_equal
 import pytest
 
 # Author: Jean Kossaifi
 skip_tensorflow = pytest.mark.skipif(
     (tl.get_backend() == "tensorflow"),
-    reason=f"Indexing with list not supported in TensorFlow",
+    reason="Indexing with list not supported in TensorFlow",
 )
 
 

tensorly/solvers/tests/test_nnls.py

@@ -6,14 +6,14 @@
     fista,
     active_set_nnls,
 )
-from tensorly.testing import assert_, assert_array_equal, assert_array_almost_equal
-from tensorly import tensor_to_vec, truncated_svd
+from tensorly.testing import assert_array_almost_equal
+from tensorly import tensor_to_vec
 import pytest
 
 # Author: Jean Kossaifi
 skip_tensorflow = pytest.mark.skipif(
     (tl.get_backend() == "tensorflow"),
-    reason=f"Indexing with list not supported in TensorFlow",
+    reason="Indexing with list not supported in TensorFlow",
 )
 
 

tensorly/solvers/tests/test_penalizations.py

@@ -1,4 +1,3 @@
-import numpy as np
 import tensorly as tl
 
 from tensorly.solvers.penalizations import process_regularization_weights

tensorly/tenalg/einsum_tenalg/generalised_inner_product.py

@@ -1,5 +1,4 @@
 from ... import backend as T
-import numpy as np
 
 # Author: Jean Kossaifi
 # License: BSD 3 clause

tensorly/tenalg/proximal.py

@@ -391,8 +391,8 @@ def unimodality_prox(tensor):
     )
     # Next line finds mutual peak points
     values = tl.tensor(
-        tl.to_numpy((tensor - monotone_decreasing >= 0))
-        * tl.to_numpy((tensor - monotone_increasing >= 0)),
+        tl.to_numpy(tensor - monotone_decreasing >= 0)
+        * tl.to_numpy(tensor - monotone_increasing >= 0),
         **tl.context(tensor)
     )
 

tensorly/tenalg/tenalg_utils.py

@@ -34,7 +34,7 @@ def _validate_contraction_modes(shape1, shape2, modes, batched_modes=False):
 
     if len(modes1) != len(modes2):
         if batched_modes:
-            message = f"Both tensors must have the same number of batched modes"
+            message = "Both tensors must have the same number of batched modes"
         else:
             message = (
                 "Both tensors must have the same number of modes to contract along. "

tensorly/tenalg/tests/test_batched_tensordot.py

@@ -3,7 +3,6 @@
 from ...testing import assert_array_almost_equal, assert_raises, assert_
 from ... import random
 from .. import tensordot
-import pytest
 
 
 def test_batched_tensordot():

tensorly/tenalg/tests/test_outer_product.py

@@ -1,6 +1,5 @@
 import tensorly as tl
 from tensorly import testing
-from tensorly import random
 from tensorly import tenalg
 
 from .. import outer, batched_outer