Merge branch 'main' into harmony_doc_fix

.gitignore

@@ -28,6 +28,7 @@
 # Python build files
 __pycache__/
 /src/scanpy/_version.py
+/ci/scanpy-min-deps.txt
 /dist/
 /*-env/
 /env-*/

.pre-commit-config.yaml

@@ -1,6 +1,6 @@
 repos:
 -   repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.7.3
+    rev: v0.7.4
     hooks:
     -   id: ruff
         types_or: [python, pyi, jupyter]

ci/scripts/min-deps.py

@@ -11,6 +11,7 @@
 import argparse
 import sys
 from collections import deque
+from contextlib import ExitStack
 from pathlib import Path
 from typing import TYPE_CHECKING
 
@@ -23,7 +24,7 @@
 from packaging.version import Version
 
 if TYPE_CHECKING:
-    from collections.abc import Generator, Iterable
+    from collections.abc import Generator, Iterable, Sequence
 
 
 def min_dep(req: Requirement) -> Requirement:
@@ -34,7 +35,7 @@ def min_dep(req: Requirement) -> Requirement:
     -------
 
     >>> min_dep(Requirement("numpy>=1.0"))
-    "numpy==1.0"
+    <Requirement('numpy==1.0.*')>
     """
     req_name = req.name
     if req.extras:
@@ -75,12 +76,19 @@ def extract_min_deps(
             yield min_dep(req)
 
 
-def main():
+class Args(argparse.Namespace):
+    path: Path
+    output: Path | None
+    extras: list[str]
+
+
+def main(argv: Sequence[str] | None = None) -> None:
     parser = argparse.ArgumentParser(
         prog="min-deps",
-        description="""Parse a pyproject.toml file and output a list of minimum dependencies.
-
-        Output is directly passable to `pip install`.""",
+        description=(
+            "Parse a pyproject.toml file and output a list of minimum dependencies. "
+            "Output is optimized for `[uv] pip install` (see `-o`/`--output` for details)."
+        ),
         usage="pip install `python min-deps.py pyproject.toml`",
     )
     parser.add_argument(
@@ -89,8 +97,18 @@ def main():
     parser.add_argument(
         "--extras", type=str, nargs="*", default=(), help="extras to install"
     )
+    parser.add_argument(
+        *("--output", "-o"),
+        type=Path,
+        default=None,
+        help=(
+            "output file (default: stdout). "
+            "Without this option, output is space-separated for direct passing to `pip install`. "
+            "With this option, output written to a file newline-separated file usable as `requirements.txt` or `constraints.txt`."
+        ),
+    )
 
-    args = parser.parse_args()
+    args = parser.parse_args(argv, Args())
 
     pyproject = tomllib.loads(args.path.read_text())
 
@@ -102,7 +120,10 @@ def main():
 
     min_deps = extract_min_deps(deps, pyproject=pyproject)
 
-    print(" ".join(map(str, min_deps)))
+    sep = "\n" if args.output else " "
+    with ExitStack() as stack:
+        f = stack.enter_context(args.output.open("w")) if args.output else sys.stdout
+        print(sep.join(map(str, min_deps)), file=f)
 
 
 if __name__ == "__main__":

ci/scripts/towncrier_automation.py

@@ -1,4 +1,8 @@
 #!/usr/bin/env python3
+# /// script
+# dependencies = [ "towncrier", "packaging" ]
+# ///
+
 from __future__ import annotations
 
 import argparse
@@ -62,9 +66,7 @@ def main(argv: Sequence[str] | None = None) -> None:
         text=True,
         check=True,
     ).stdout.strip()
-    pr_description = (
-        "" if base_branch == "main" else "@meeseeksmachine backport to main"
-    )
+    pr_description = "" if base_branch == "main" else "@meeseeksdev backport to main"
     branch_name = f"release_notes_{args.version}"
 
     # Create a new branch + commit

hatch.toml

@@ -19,14 +19,17 @@ features = ["test", "dask-ml"]
 extra-dependencies = ["ipykernel"]
 overrides.matrix.deps.env-vars = [
     { if = ["pre"], key = "UV_PRERELEASE", value = "allow" },
-    { if = ["min"], key = "UV_RESOLUTION", value = "lowest-direct" },
+    { if = ["min"], key = "UV_CONSTRAINT", value = "ci/scanpy-min-deps.txt" },
+]
+overrides.matrix.deps.pre-install-commands = [
+    { if = ["min"], value = "uv run ci/scripts/min-deps.py pyproject.toml -o ci/scanpy-min-deps.txt" },
 ]
 overrides.matrix.deps.python = [
-    { if = ["min"] , value = "3.10" },
+    { if = ["min"], value = "3.10" },
     { if = ["stable", "full", "pre"], value = "3.12" },
 ]
 overrides.matrix.deps.features = [
-    { if = ["full"] , value = "test-full" },
+    { if = ["full"], value = "test-full" },
 ]
 
 [[envs.hatch-test.matrix]]

pyproject.toml

@@ -52,7 +52,7 @@ dependencies = [
     "pandas >=1.5",
     "scipy>=1.8",
     "seaborn>=0.13",
-    "h5py>=3.6",
+    "h5py>=3.7",
     "tqdm",
     "scikit-learn>=1.1",
     "statsmodels>=0.13",
@@ -166,7 +166,7 @@ addopts = [
     "-ptesting.scanpy._pytest",
     "--pyargs",
 ]
-testpaths = ["./tests", "scanpy"]
+testpaths = ["./tests", "./ci", "scanpy"]
 norecursedirs = ["tests/_images"]
 xfail_strict = true
 nunit_attach_on = "fail"
@@ -211,7 +211,7 @@ exclude_also = [
     "if __name__ == .__main__.:",
     "if TYPE_CHECKING:",
     # https://github.com/numba/numba/issues/4268
-    "@numba.njit.*",
+    '@(numba\.|nb\.)njit.*',
 ]
 
 [tool.ruff]

src/scanpy/_compat.py

@@ -9,15 +9,19 @@
 from pathlib import Path
 from typing import TYPE_CHECKING, Literal, ParamSpec, TypeVar, cast, overload
 
+import numpy as np
 from packaging.version import Version
 
 if TYPE_CHECKING:
     from collections.abc import Callable
     from importlib.metadata import PackageMetadata
 
+
 P = ParamSpec("P")
 R = TypeVar("R")
 
+_LegacyRandom = int | np.random.RandomState | None
+
 
 if TYPE_CHECKING:
     # type checkers are confused and can only see …core.Array

src/scanpy/_utils/__init__.py

@@ -12,6 +12,7 @@
 import re
 import sys
 import warnings
+from collections.abc import Sequence
 from contextlib import contextmanager, suppress
 from enum import Enum
 from functools import partial, reduce, singledispatch, wraps
@@ -56,12 +57,13 @@
     from anndata import AnnData
     from numpy.typing import ArrayLike, DTypeLike, NDArray
 
+    from .._compat import _LegacyRandom
     from ..neighbors import NeighborsParams, RPForestDict
 
 
-# e.g. https://scikit-learn.org/stable/modules/generated/sklearn.decomposition.PCA.html
-# maybe in the future random.Generator
-AnyRandom = int | np.random.RandomState | None
+SeedLike = int | np.integer | Sequence[int] | np.random.SeedSequence
+RNGLike = np.random.Generator | np.random.BitGenerator
+
 LegacyUnionType = type(Union[int, str])  # noqa: UP007
 
 
@@ -493,7 +495,7 @@ def moving_average(a: np.ndarray, n: int):
     return ret[n - 1 :] / n
 
 
-def get_random_state(seed: AnyRandom) -> np.random.RandomState:
+def _get_legacy_random(seed: _LegacyRandom) -> np.random.RandomState:
     if isinstance(seed, np.random.RandomState):
         return seed
     return np.random.RandomState(seed)

src/scanpy/datasets/_datasets.py

@@ -18,7 +18,7 @@
 if TYPE_CHECKING:
     from typing import Literal
 
-    from .._utils import AnyRandom
+    from .._compat import _LegacyRandom
 
     VisiumSampleID = Literal[
         "V1_Breast_Cancer_Block_A_Section_1",
@@ -63,7 +63,7 @@ def blobs(
     n_centers: int = 5,
     cluster_std: float = 1.0,
     n_observations: int = 640,
-    random_state: AnyRandom = 0,
+    random_state: _LegacyRandom = 0,
 ) -> AnnData:
     """\
     Gaussian Blobs.

src/scanpy/external/pp/_dca.py

@@ -11,7 +11,7 @@
 
     from anndata import AnnData
 
-    from ..._utils import AnyRandom
+    from ..._compat import _LegacyRandom
 
     _AEType = Literal["zinb-conddisp", "zinb", "nb-conddisp", "nb"]
 
@@ -62,7 +62,7 @@ def dca(
     early_stop: int = 15,
     batch_size: int = 32,
     optimizer: str = "RMSprop",
-    random_state: AnyRandom = 0,
+    random_state: _LegacyRandom = 0,
     threads: int | None = None,
     learning_rate: float | None = None,
     verbose: bool = False,

src/scanpy/external/pp/_magic.py

@@ -19,7 +19,7 @@
 
     from anndata import AnnData
 
-    from ..._utils import AnyRandom
+    from ..._compat import _LegacyRandom
 
 MIN_VERSION = "2.0"
 
@@ -36,7 +36,7 @@ def magic(
     n_pca: int | None = 100,
     solver: Literal["exact", "approximate"] = "exact",
     knn_dist: str = "euclidean",
-    random_state: AnyRandom = None,
+    random_state: _LegacyRandom = None,
     n_jobs: int | None = None,
     verbose: bool = False,
     copy: bool | None = None,

src/scanpy/external/tl/_phate.py

@@ -16,7 +16,7 @@
 
     from anndata import AnnData
 
-    from ..._utils import AnyRandom
+    from ..._compat import _LegacyRandom
 
 
 @old_positionals(
@@ -49,7 +49,7 @@ def phate(
     mds_dist: str = "euclidean",
     mds: Literal["classic", "metric", "nonmetric"] = "metric",
     n_jobs: int | None = None,
-    random_state: AnyRandom = None,
+    random_state: _LegacyRandom = None,
     verbose: bool | int | None = None,
     copy: bool = False,
     **kwargs,

src/scanpy/neighbors/__init__.py

@@ -33,7 +33,7 @@
     from igraph import Graph
     from scipy.sparse import csr_matrix
 
-    from .._utils import AnyRandom
+    from .._compat import _LegacyRandom
     from ._types import KnnTransformerLike, _Metric, _MetricFn
 
 
@@ -54,13 +54,13 @@ class KwdsForTransformer(TypedDict):
     n_neighbors: int
     metric: _Metric | _MetricFn
     metric_params: Mapping[str, Any]
-    random_state: AnyRandom
+    random_state: _LegacyRandom
 
 
 class NeighborsParams(TypedDict):
     n_neighbors: int
     method: _Method
-    random_state: AnyRandom
+    random_state: _LegacyRandom
     metric: _Metric | _MetricFn
     metric_kwds: NotRequired[Mapping[str, Any]]
     use_rep: NotRequired[str]
@@ -79,7 +79,7 @@ def neighbors(
     transformer: KnnTransformerLike | _KnownTransformer | None = None,
     metric: _Metric | _MetricFn = "euclidean",
     metric_kwds: Mapping[str, Any] = MappingProxyType({}),
-    random_state: AnyRandom = 0,
+    random_state: _LegacyRandom = 0,
     key_added: str | None = None,
     copy: bool = False,
 ) -> AnnData | None:
@@ -521,7 +521,7 @@ def compute_neighbors(
         transformer: KnnTransformerLike | _KnownTransformer | None = None,
         metric: _Metric | _MetricFn = "euclidean",
         metric_kwds: Mapping[str, Any] = MappingProxyType({}),
-        random_state: AnyRandom = 0,
+        random_state: _LegacyRandom = 0,
     ) -> None:
         """\
         Compute distances and connectivities of neighbors.
@@ -757,7 +757,7 @@ def compute_eigen(
         n_comps: int = 15,
         sym: bool | None = None,
         sort: Literal["decrease", "increase"] = "decrease",
-        random_state: AnyRandom = 0,
+        random_state: _LegacyRandom = 0,
     ):
         """\
         Compute eigen decomposition of transition matrix.

src/scanpy/plotting/_tools/paga.py

@@ -33,6 +33,7 @@
     from matplotlib.colors import Colormap
     from scipy.sparse import spmatrix
 
+    from ..._compat import _LegacyRandom
     from ...tools._draw_graph import _Layout as _LayoutWithoutEqTree
     from .._utils import _FontSize, _FontWeight, _LegendLoc
 
@@ -210,7 +211,7 @@ def _compute_pos(
     adjacency_solid: spmatrix | np.ndarray,
     *,
     layout: _Layout | None = None,
-    random_state: _sc_utils.AnyRandom = 0,
+    random_state: _LegacyRandom = 0,
     init_pos: np.ndarray | None = None,
     adj_tree=None,
     root: int = 0,

src/scanpy/preprocessing/_pca/__init__.py

@@ -30,7 +30,8 @@
     from scipy import sparse
     from scipy.sparse import spmatrix
 
-    from ..._utils import AnyRandom, Empty
+    from ..._compat import _LegacyRandom
+    from ..._utils import Empty
 
     CSMatrix = sparse.csr_matrix | sparse.csc_matrix
 
@@ -70,7 +71,7 @@ def pca(
     layer: str | None = None,
     zero_center: bool | None = True,
     svd_solver: SvdSolver | None = None,
-    random_state: AnyRandom = 0,
+    random_state: _LegacyRandom = 0,
     return_info: bool = False,
     mask_var: NDArray[np.bool_] | str | None | Empty = _empty,
     use_highly_variable: bool | None = None,

src/scanpy/preprocessing/_pca/_compat.py

@@ -18,7 +18,7 @@
     from scipy import sparse
     from sklearn.decomposition import PCA
 
-    from .._utils import AnyRandom
+    from ..._compat import _LegacyRandom
 
     CSMatrix = sparse.csr_matrix | sparse.csc_matrix
 
@@ -29,7 +29,7 @@ def _pca_compat_sparse(
     *,
     solver: Literal["arpack", "lobpcg"],
     mu: NDArray[np.floating] | None = None,
-    random_state: AnyRandom = None,
+    random_state: _LegacyRandom = None,
 ) -> tuple[NDArray[np.floating], PCA]:
     """Sparse PCA for scikit-learn <1.4"""
     random_state = check_random_state(random_state)