(chore): adopt scientific python deprecation schedule

.azure-pipelines.yml

@@ -14,18 +14,18 @@ jobs:
       vmImage: "ubuntu-22.04"
     strategy:
       matrix:
-        Python3.12:
-          python.version: "3.12"
+        Python3.13:
+          python.version: "3.13"
           RUN_COVERAGE: yes
           TEST_TYPE: "coverage"
-        Python3.10:
-          python.version: "3.10"
+        Python3.11:
+          python.version: "3.11"
         PreRelease:
-          python.version: "3.12"
+          python.version: "3.13"
           DEPENDENCIES_VERSION: "pre-release"
           TEST_TYPE: "strict-warning"
         minimum_versions:
-          python.version: "3.10"
+          python.version: "3.11"
           DEPENDENCIES_VERSION: "minimum"
           TEST_TYPE: "coverage"
     steps:
@@ -57,7 +57,7 @@ jobs:
           set -e
           uv pip install --system --compile tomli packaging
           deps=`python3 ci/scripts/min-deps.py pyproject.toml --extra dev test`
-          uv pip install --system --compile $deps pytest-cov "anndata @ ."
+          uv pip install --system --compile $deps pytest-cov "anndata[test,dev] @ ."
         displayName: "Install minimum dependencies"
         condition: eq(variables['DEPENDENCIES_VERSION'], 'minimum')
 
@@ -104,8 +104,8 @@ jobs:
     steps:
       - task: UsePythonVersion@0
         inputs:
-          versionSpec: "3.12"
-        displayName: "Use Python 3.12"
+          versionSpec: "3.13"
+        displayName: "Use Python 3.13"
 
       - script: |
           set -e

.readthedocs.yml

@@ -2,7 +2,7 @@ version: 2
 build:
   os: ubuntu-20.04
   tools:
-    python: "3.12"
+    python: "3.13"
   jobs:
     post_checkout:
       # unshallow so version can be derived from tag

ci/scripts/min-deps.py

@@ -10,17 +10,13 @@
 
 import argparse
 import sys
+import tomllib
 from collections import deque
 from contextlib import ExitStack
 from functools import cached_property
 from pathlib import Path
 from typing import TYPE_CHECKING
 
-if sys.version_info >= (3, 11):
-    import tomllib
-else:
-    import tomli as tomllib
-
 from packaging.requirements import Requirement
 from packaging.version import Version
 

docs/conf.py

@@ -137,6 +137,7 @@ def setup(app: Sphinx):
     "pandas.DataFrame.loc": ("py:attr", "pandas.DataFrame.loc"),
     # should be fixed soon: https://github.com/tox-dev/sphinx-autodoc-typehints/pull/516
     "types.EllipsisType": ("py:data", "types.EllipsisType"),
+    "pathlib._local.Path": "pathlib.Path",
 }
 autodoc_type_aliases = dict(
     NDArray=":data:`~numpy.typing.NDArray`",

docs/release-notes/1767.breaking.md

@@ -0,0 +1 @@
+Tighten usage of {class}`scipy.sparse.spmatrix` for describing sparse matrices in types and instance checks to only {class}`scipy.sparse.csr_matrix` and {class}`scipy.sparse.csc_matrix` {user}`ilan-gold`

hatch.toml

@@ -26,8 +26,8 @@ overrides.matrix.deps.pre-install-commands = [
     { if = ["min"], value = "uv run ci/scripts/min-deps.py pyproject.toml --all-extras -o ci/min-deps.txt" },
 ]
 overrides.matrix.deps.python = [
-    { if = ["min"], value = "3.10" },
-    { if = ["stable", "pre"], value = "3.12" },
+    { if = ["min"], value = "3.11" },
+    { if = ["stable", "pre"], value = "3.13" },
 ]
 
 [[envs.hatch-test.matrix]]

pyproject.toml

@@ -5,7 +5,7 @@ requires = ["hatchling", "hatch-vcs"]
 [project]
 name = "anndata"
 description = "Annotated data."
-requires-python = ">=3.10"
+requires-python = ">=3.11"
 license = "BSD-3-Clause"
 authors = [
     { name = "Philipp Angerer" },
@@ -29,20 +29,20 @@ classifiers = [
     "Operating System :: Microsoft :: Windows",
     "Operating System :: POSIX :: Linux",
     "Programming Language :: Python :: 3",
-    "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
     "Topic :: Scientific/Engineering :: Bio-Informatics",
     "Topic :: Scientific/Engineering :: Visualization",
 ]
 dependencies = [
     # pandas <1.4 has pandas/issues/35446
     # pandas 2.1.0rc0 has pandas/issues/54622
-    "pandas >=1.4, !=2.1.0rc0, !=2.1.2",
-    "numpy>=1.23",
+    "pandas >=2.0.0, !=2.1.0rc0, !=2.1.2",
+    "numpy>=1.25",
     # https://github.com/scverse/anndata/issues/1434
-    "scipy >1.8",
-    "h5py>=3.7",
+    "scipy >1.11",
+    "h5py>=3.8",
     "exceptiongroup; python_version<'3.11'",
     "natsort",
     "packaging>=24.2",
@@ -115,6 +115,8 @@ source = "vcs"
 raw-options.version_scheme = "release-branch-semver"
 [tool.hatch.build.targets.wheel]
 packages = ["src/anndata", "src/testing"]
+[tool.hatch.metadata]
+allow-direct-references = true
 
 [tool.coverage.run]
 data_file = "test-data/coverage"

src/anndata/_core/aligned_mapping.py

@@ -9,10 +9,9 @@
 
 import numpy as np
 import pandas as pd
-from scipy.sparse import spmatrix
 
 from .._warnings import ExperimentalFeatureWarning, ImplicitModificationWarning
-from ..compat import AwkArray
+from ..compat import AwkArray, SpMatrix
 from ..utils import (
     axis_len,
     convert_to_dict,
@@ -36,7 +35,7 @@
 OneDIdx = Sequence[int] | Sequence[bool] | slice
 TwoDIdx = tuple[OneDIdx, OneDIdx]
 # TODO: pd.DataFrame only allowed in AxisArrays?
-Value = pd.DataFrame | spmatrix | np.ndarray
+Value = pd.DataFrame | SpMatrix | np.ndarray
 
 P = TypeVar("P", bound="AlignedMappingBase")
 """Parent mapping an AlignedView is based on."""

src/anndata/_core/anndata.py

@@ -231,14 +231,14 @@ class AnnData(metaclass=utils.DeprecationMixinMeta):
     )
     def __init__(
         self,
-        X: np.ndarray | sparse.spmatrix | pd.DataFrame | None = None,
+        X: ArrayDataStructureType | pd.DataFrame | None = None,
         obs: pd.DataFrame | Mapping[str, Iterable[Any]] | None = None,
         var: pd.DataFrame | Mapping[str, Iterable[Any]] | None = None,
         uns: Mapping[str, Any] | None = None,
         *,
         obsm: np.ndarray | Mapping[str, Sequence[Any]] | None = None,
         varm: np.ndarray | Mapping[str, Sequence[Any]] | None = None,
-        layers: Mapping[str, np.ndarray | sparse.spmatrix] | None = None,
+        layers: Mapping[str, ArrayDataStructureType] | None = None,
         raw: Mapping[str, Any] | None = None,
         dtype: np.dtype | type | str | None = None,
         shape: tuple[int, int] | None = None,
@@ -592,7 +592,7 @@ def X(self) -> ArrayDataStructureType | None:
         #     return X
 
     @X.setter
-    def X(self, value: np.ndarray | sparse.spmatrix | SpArray | None):
+    def X(self, value: ArrayDataStructureType | None):
         if value is None:
             if self.isbacked:
                 msg = "Cannot currently remove data matrix from backed object."
@@ -1189,7 +1189,7 @@ def _inplace_subset_obs(self, index: Index1D):
         self._init_as_actual(adata_subset)
 
     # TODO: Update, possibly remove
-    def __setitem__(self, index: Index, val: float | np.ndarray | sparse.spmatrix):
+    def __setitem__(self, index: Index, val: ArrayDataStructureType):
         if self.is_view:
             msg = "Object is view and cannot be accessed with `[]`."
             raise ValueError(msg)

src/anndata/_core/index.py

@@ -8,9 +8,9 @@
 import h5py
 import numpy as np
 import pandas as pd
-from scipy.sparse import issparse, spmatrix
+from scipy.sparse import issparse
 
-from ..compat import AwkArray, DaskArray, SpArray
+from ..compat import AwkArray, DaskArray, SpArray, SpMatrix
 
 if TYPE_CHECKING:
     from ..compat import Index, Index1D
@@ -69,13 +69,13 @@ def name_idx(i):
     elif isinstance(indexer, str):
         return index.get_loc(indexer)  # int
     elif isinstance(
-        indexer, Sequence | np.ndarray | pd.Index | spmatrix | np.matrix | SpArray
+        indexer, Sequence | np.ndarray | pd.Index | SpMatrix | np.matrix | SpArray
     ):
         if hasattr(indexer, "shape") and (
             (indexer.shape == (index.shape[0], 1))
             or (indexer.shape == (1, index.shape[0]))
         ):
-            if isinstance(indexer, spmatrix | SpArray):
+            if isinstance(indexer, SpMatrix | SpArray):
                 indexer = indexer.toarray()
             indexer = np.ravel(indexer)
         if not isinstance(indexer, np.ndarray | pd.Index):
@@ -180,9 +180,9 @@ def _subset_dask(a: DaskArray, subset_idx: Index):
     return a[subset_idx]
 
 
-@_subset.register(spmatrix)
+@_subset.register(SpMatrix)
 @_subset.register(SpArray)
-def _subset_sparse(a: spmatrix | SpArray, subset_idx: Index):
+def _subset_sparse(a: SpMatrix | SpArray, subset_idx: Index):
     # Correcting for indexing behaviour of sparse.spmatrix
     if len(subset_idx) > 1 and all(isinstance(x, Iterable) for x in subset_idx):
         first_idx = subset_idx[0]

src/anndata/_core/merge.py

@@ -17,7 +17,6 @@
 import pandas as pd
 from natsort import natsorted
 from scipy import sparse
-from scipy.sparse import spmatrix
 
 from anndata._warnings import ExperimentalFeatureWarning
 
@@ -29,6 +28,7 @@
     CupySparseMatrix,
     DaskArray,
     SpArray,
+    SpMatrix,
     _map_cat_to_str,
 )
 from ..utils import asarray, axis_len, warn_once
@@ -135,7 +135,7 @@ def equal_dask_array(a, b) -> bool:
     if isinstance(b, DaskArray):
         if tokenize(a) == tokenize(b):
             return True
-    if isinstance(a._meta, spmatrix):
+    if isinstance(a._meta, SpMatrix):
         # TODO: Maybe also do this in the other case?
         return da.map_blocks(equal, a, b, drop_axis=(0, 1)).all()
     else:
@@ -165,7 +165,7 @@ def equal_series(a, b) -> bool:
     return a.equals(b)
 
 
-@equal.register(sparse.spmatrix)
+@equal.register(SpMatrix)
 @equal.register(SpArray)
 @equal.register(CupySparseMatrix)
 def equal_sparse(a, b) -> bool:
@@ -174,7 +174,7 @@ def equal_sparse(a, b) -> bool:
 
     xp = array_api_compat.array_namespace(a.data)
 
-    if isinstance(b, CupySparseMatrix | sparse.spmatrix | SpArray):
+    if isinstance(b, CupySparseMatrix | SpMatrix | SpArray):
         if isinstance(a, CupySparseMatrix):
             # Comparison broken for CSC matrices
             # https://github.com/cupy/cupy/issues/7757
@@ -205,8 +205,8 @@ def equal_awkward(a, b) -> bool:
     return ak.almost_equal(a, b)
 
 
-def as_sparse(x, *, use_sparse_array: bool = False):
-    if not isinstance(x, sparse.spmatrix | SpArray):
+def as_sparse(x, *, use_sparse_array=False):
+    if not isinstance(x, SpMatrix | SpArray):
         if CAN_USE_SPARSE_ARRAY and use_sparse_array:
             return sparse.csr_array(x)
         return sparse.csr_matrix(x)
@@ -537,7 +537,7 @@ def apply(self, el, *, axis, fill_value=None):
             return el
         if isinstance(el, pd.DataFrame):
             return self._apply_to_df(el, axis=axis, fill_value=fill_value)
-        elif isinstance(el, sparse.spmatrix | SpArray | CupySparseMatrix):
+        elif isinstance(el, SpMatrix | SpArray | CupySparseMatrix):
             return self._apply_to_sparse(el, axis=axis, fill_value=fill_value)
         elif isinstance(el, AwkArray):
             return self._apply_to_awkward(el, axis=axis, fill_value=fill_value)
@@ -615,8 +615,8 @@ def _apply_to_array(self, el, *, axis, fill_value=None):
         )
 
     def _apply_to_sparse(
-        self, el: sparse.spmatrix | SpArray, *, axis, fill_value=None
-    ) -> spmatrix:
+        self, el: SpMatrix | SpArray, *, axis, fill_value=None
+    ) -> SpMatrix:
         if isinstance(el, CupySparseMatrix):
             from cupyx.scipy import sparse
         else:
@@ -726,11 +726,8 @@ def default_fill_value(els):
     This is largely due to backwards compat, and might not be the ideal solution.
     """
     if any(
-        isinstance(el, sparse.spmatrix | SpArray)
-        or (
-            isinstance(el, DaskArray)
-            and isinstance(el._meta, sparse.spmatrix | SpArray)
-        )
+        isinstance(el, SpMatrix | SpArray)
+        or (isinstance(el, DaskArray) and isinstance(el._meta, SpMatrix | SpArray))
         for el in els
     ):
         return 0
@@ -826,7 +823,7 @@ def concat_arrays(arrays, reindexers, axis=0, index=None, fill_value=None):
             ],
             axis=axis,
         )
-    elif any(isinstance(a, sparse.spmatrix | SpArray) for a in arrays):
+    elif any(isinstance(a, SpMatrix | SpArray) for a in arrays):
         sparse_stack = (sparse.vstack, sparse.hstack)[axis]
         use_sparse_array = any(issubclass(type(a), SpArray) for a in arrays)
         return sparse_stack(

src/anndata/_core/raw.py

@@ -17,8 +17,7 @@
     from collections.abc import Mapping, Sequence
     from typing import ClassVar
 
-    from scipy import sparse
-
+    from ..compat import SpMatrix
     from .aligned_mapping import AxisArraysView
     from .anndata import AnnData
     from .sparse_dataset import BaseCompressedSparseDataset
@@ -31,7 +30,7 @@ class Raw:
     def __init__(
         self,
         adata: AnnData,
-        X: np.ndarray | sparse.spmatrix | None = None,
+        X: np.ndarray | SpMatrix | None = None,
         var: pd.DataFrame | Mapping[str, Sequence] | None = None,
         varm: AxisArrays | Mapping[str, np.ndarray] | None = None,
     ):
@@ -67,7 +66,7 @@ def _get_X(self, layer=None):
         return self.X
 
     @property
-    def X(self) -> BaseCompressedSparseDataset | np.ndarray | sparse.spmatrix:
+    def X(self) -> BaseCompressedSparseDataset | np.ndarray | SpMatrix:
         # TODO: Handle unsorted array of integer indices for h5py.Datasets
         if not self._adata.isbacked:
             return self._X

src/anndata/_core/sparse_dataset.py

@@ -30,7 +30,7 @@
 
 from .. import abc
 from .._settings import settings
-from ..compat import H5Group, SpArray, ZarrArray, ZarrGroup, _read_attr
+from ..compat import H5Group, SpArray, SpMatrix, ZarrArray, ZarrGroup, _read_attr
 from .index import _fix_slice_bounds, _subset, unpack_index
 
 if TYPE_CHECKING:
@@ -329,7 +329,7 @@ def get_memory_class(
         if format == fmt:
             if use_sparray_in_io and issubclass(memory_class, SpArray):
                 return memory_class
-            elif not use_sparray_in_io and issubclass(memory_class, ss.spmatrix):
+            elif not use_sparray_in_io and issubclass(memory_class, SpMatrix):
                 return memory_class
     msg = f"Format string {format} is not supported."
     raise ValueError(msg)
@@ -342,7 +342,7 @@ def get_backed_class(
         if format == fmt:
             if use_sparray_in_io and issubclass(backed_class, SpArray):
                 return backed_class
-            elif not use_sparray_in_io and issubclass(backed_class, ss.spmatrix):
+            elif not use_sparray_in_io and issubclass(backed_class, SpMatrix):
                 return backed_class
     msg = f"Format string {format} is not supported."
     raise ValueError(msg)