Merge branch 'main' into no-positional-bools

.azure-pipelines.yml

@@ -49,11 +49,7 @@ jobs:
           path: $(uv_cache_dir)
         displayName: Cache pip packages
 
-      # This prevents broken and slow back-tracking in dependency resolution
-      - script: printf "llvmlite>=0.43\nscanpy>=1.10.0rc1" | tee /tmp/constraints.txt
-        displayName: "Create constraints file for `pre-release` and `latest` jobs"
-
-      - script: uv pip install --system --compile "anndata[dev,test] @ ." -c /tmp/constraints.txt
+      - script: uv pip install --system --compile "anndata[dev,test] @ ." -c ci/constraints.txt
         displayName: "Install dependencies"
         condition: eq(variables['DEPENDENCIES_VERSION'], 'latest')
 
@@ -65,7 +61,7 @@ jobs:
         displayName: "Install minimum dependencies"
         condition: eq(variables['DEPENDENCIES_VERSION'], 'minimum')
 
-      - script: uv pip install -v --system --compile --pre "anndata[dev,test] @ ." -c /tmp/constraints.txt
+      - script: uv pip install -v --system --compile --pre "anndata[dev,test] @ ." -c ci/constraints.txt
         displayName: "Install dependencies release candidates"
         condition: eq(variables['DEPENDENCIES_VERSION'], 'pre-release')
 

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -53,10 +53,18 @@ body:
       description: |
         Which version of anndata and other related software you used.
 
-        Please install `session-info`, run the following command, then paste the results into the text box below.
+        Please install `session-info2`, run the following command in a notebook,
+        click the “Copy as Markdown” button,
+        then paste the results into the text box below.
 
         ```python
-        >>> import anndata, session_info; session_info.show(html=False, dependencies=True)
+        In[1]: import anndata, session_info2; session_info2.session_info(dependencies=True)
+        ```
+
+        Alternatively, run this in a console:
+
+        ```python
+        >>> import session_info2; print(session_info2.session_info(dependencies=True)._repr_mimebundle_()["text/markdown"])
         ```
       render: python
     validations:

.github/workflows/test-gpu.yml

@@ -72,9 +72,7 @@ jobs:
           cache-dependency-path: pyproject.toml
 
       - name: Install AnnData
-        run: |
-          printf "llvmlite>=0.43\nscanpy>=1.10.0rc1" | tee /tmp/constraints.txt
-          uv pip install --system -e ".[dev,test,cu12]" -c /tmp/constraints.txt
+        run: uv pip install --system -e ".[dev,test,cu12]" -c ci/constraints.txt
 
       - name: Env list
         run: pip list

.gitignore

@@ -10,6 +10,7 @@ __pycache__/
 
 # Distribution / packaging
 /dist/
+/ci/min-deps.txt
 /src/anndata/_version.py
 /requirements*.lock
 /.python-version

.pre-commit-config.yaml

@@ -1,6 +1,6 @@
 repos:
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.8.2
+    rev: v0.8.6
     hooks:
       - id: ruff
         types_or: [python, pyi, jupyter]

ci/constraints.txt

@@ -0,0 +1 @@
+numba>=0.56

ci/scripts/min-deps.py

@@ -11,6 +11,8 @@
 import argparse
 import sys
 from collections import deque
+from contextlib import ExitStack
+from functools import cached_property
 from pathlib import Path
 from typing import TYPE_CHECKING
 
@@ -23,7 +25,9 @@
 from packaging.version import Version
 
 if TYPE_CHECKING:
-    from collections.abc import Generator, Iterable
+    from collections.abc import Generator, Iterable, Sequence
+    from collections.abc import Set as AbstractSet
+    from typing import Any, Self
 
 
 def min_dep(req: Requirement) -> Requirement:
@@ -75,34 +79,92 @@ def extract_min_deps(
             yield min_dep(req)
 
 
-def main():
-    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`.""",
-        usage="pip install `python min-deps.py pyproject.toml`",
-    )
-    parser.add_argument(
-        "path", type=Path, help="pyproject.toml to parse minimum dependencies from"
-    )
-    parser.add_argument(
-        "--extras", type=str, nargs="*", default=(), help="extras to install"
-    )
-
-    args = parser.parse_args()
-
-    pyproject = tomllib.loads(args.path.read_text())
+class Args(argparse.Namespace):
+    """\
+    Parse a pyproject.toml file and output a list of minimum dependencies.
+    Output is optimized for `[uv] pip install` (see `-o`/`--output` for details).
+    """
 
-    project_name = pyproject["project"]["name"]
+    _path: Path
+    output: Path | None
+    _extras: list[str]
+    _all_extras: bool
+
+    @classmethod
+    def parse(cls, argv: Sequence[str] | None = None) -> Self:
+        return cls.parser().parse_args(argv, cls())
+
+    @classmethod
+    def parser(cls) -> argparse.ArgumentParser:
+        parser = argparse.ArgumentParser(
+            prog="min-deps",
+            description=cls.__doc__,
+            usage="pip install `python min-deps.py pyproject.toml`",
+        )
+        parser.add_argument(
+            "_path",
+            metavar="pyproject.toml",
+            type=Path,
+            help="Path to pyproject.toml to parse minimum dependencies from",
+        )
+        parser.add_argument(
+            "--extras",
+            dest="_extras",
+            metavar="EXTRA",
+            type=str,
+            nargs="*",
+            default=(),
+            help="extras to install",
+        )
+        parser.add_argument(
+            "--all-extras",
+            dest="_all_extras",
+            action="store_true",
+            help="get all extras",
+        )
+        parser.add_argument(
+            *("--output", "-o"),
+            metavar="FILE",
+            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`."
+            ),
+        )
+        return parser
+
+    @cached_property
+    def pyproject(self) -> dict[str, Any]:
+        return tomllib.loads(self._path.read_text())
+
+    @cached_property
+    def extras(self) -> AbstractSet[str]:
+        if self._extras:
+            if self._all_extras:
+                sys.exit("Cannot specify both --extras and --all-extras")
+            return dict.fromkeys(self._extras).keys()
+        if not self._all_extras:
+            return set()
+        return self.pyproject["project"]["optional-dependencies"].keys()
+
+
+def main(argv: Sequence[str] | None = None) -> None:
+    args = Args.parse(argv)
+
+    project_name = args.pyproject["project"]["name"]
     deps = [
-        *map(Requirement, pyproject["project"]["dependencies"]),
+        *map(Requirement, args.pyproject["project"]["dependencies"]),
         *(Requirement(f"{project_name}[{extra}]") for extra in args.extras),
     ]
 
-    min_deps = extract_min_deps(deps, pyproject=pyproject)
+    min_deps = extract_min_deps(deps, pyproject=args.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 +1,7 @@
 #!/usr/bin/env python3
+# /// script
+# dependencies = [ "towncrier", "packaging" ]
+# ///
 from __future__ import annotations
 
 import argparse

docs/release-notes/1806.bugfix.md

@@ -0,0 +1 @@
+Add {mod}`scipy` 1.5 compatibility {user}`flying-sheep`

hatch.toml

@@ -4,21 +4,26 @@ features = ["dev"]
 
 [envs.docs]
 features = ["doc"]
-extra-dependencies = ["setuptools"]  # https://bitbucket.org/pybtex-devs/pybtex/issues/169
 scripts.build = "sphinx-build -M html docs docs/_build -W --keep-going {args}"
+scripts.open = "python3 -m webbrowser -t docs/_build/html/index.html"
 scripts.clean = "git clean -fdX -- {args:docs}"
 
 [envs.towncrier]
+scripts.create = "towncrier create {args}"
 scripts.build = "python3 ci/scripts/towncrier_automation.py {args}"
 scripts.clean = "git restore --source=HEAD --staged --worktree -- docs/release-notes"
 
 [envs.hatch-test]
 default-args = []
-extra-dependencies = ["ipykernel"]
 features = ["dev", "test"]
+extra-dependencies = ["ipykernel"]
+env-vars.UV_CONSTRAINT = "ci/constraints.txt"
 overrides.matrix.deps.env-vars = [
-    { key = "UV_PRERELEASE", value = "allow", if = ["pre"] },
-    { key = "UV_RESOLUTION", value = "lowest-direct", if = ["min"] },
+    { if = ["pre"], key = "UV_PRERELEASE", value = "allow" },
+    { if = ["min"], key = "UV_CONSTRAINT", value = "ci/constraints.txt ci/min-deps.txt" },
+]
+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" },

pyproject.toml

@@ -42,7 +42,7 @@ dependencies = [
     "numpy>=1.23",
     # https://github.com/scverse/anndata/issues/1434
     "scipy >1.8",
-    "h5py>=3.6",
+    "h5py>=3.7",
     "exceptiongroup; python_version<'3.11'",
     "natsort",
     "packaging>=20.0",
@@ -75,7 +75,7 @@ doc = [
     "nbsphinx",
     "scanpydoc[theme,typehints] >=0.14.1",
     "zarr",
-    "awkward>=2.0.7",
+    "awkward>=2.3",
     "IPython",                          # For syntax highlighting in notebooks
     "myst_parser",
     "sphinx_design>=0.5.0",
@@ -94,7 +94,7 @@ test = [
     "openpyxl",
     "joblib",
     "boltons",
-    "scanpy",
+    "scanpy>=1.9.8",
     "httpx", # For data downloading
     "dask[distributed]",
     "awkward>=2.3",

src/anndata/_core/sparse_dataset.py

@@ -23,7 +23,9 @@
 
 import h5py
 import numpy as np
+import scipy
 import scipy.sparse as ss
+from packaging.version import Version
 from scipy.sparse import _sparsetools
 
 from .. import abc
@@ -39,11 +41,14 @@
 
     from .._types import GroupStorageType
     from ..compat import H5Array
-    from .index import Index
+    from .index import Index, Index1D
 else:
     from scipy.sparse import spmatrix as _cs_matrix
 
 
+SCIPY_1_15 = Version(scipy.__version__) >= Version("1.15rc0")
+
+
 class BackedFormat(NamedTuple):
     format: Literal["csr", "csc"]
     backed_type: type[BackedSparseMatrix]
@@ -353,7 +358,9 @@ def _get_group_format(group: GroupStorageType) -> str:
 
 
 # Check for the overridden few methods above in our BackedSparseMatrix subclasses
-def is_sparse_indexing_overridden(format: Literal["csr", "csc"], row, col):
+def is_sparse_indexing_overridden(
+    format: Literal["csr", "csc"], row: Index1D, col: Index1D
+):
     major_indexer, minor_indexer = (row, col) if format == "csr" else (col, row)
     return isinstance(minor_indexer, slice) and (
         (isinstance(major_indexer, int | np.integer))
@@ -362,6 +369,13 @@ def is_sparse_indexing_overridden(format: Literal["csr", "csc"], row, col):
     )
 
 
+def validate_indices(
+    mtx: BackedSparseMatrix, indices: tuple[Index1D, Index1D]
+) -> tuple[Index1D, Index1D]:
+    res = mtx._validate_indices(indices)
+    return res[0] if SCIPY_1_15 else res
+
+
 class BaseCompressedSparseDataset(abc._AbstractCSDataset, ABC):
     _group: GroupStorageType
 
@@ -424,8 +438,8 @@ def __getitem__(
         indices = self._normalize_index(index)
         row, col = indices
         mtx = self._to_backed()
-        row_sp_matrix_validated, col_sp_matrix_validated = mtx._validate_indices(
-            (row, col)
+        row_sp_matrix_validated, col_sp_matrix_validated = validate_indices(
+            mtx, indices
         )
 
         # Handle masked indexing along major axis

src/anndata/compat/__init__.py

@@ -47,7 +47,7 @@ class Empty:
     pass
 
 
-Index1D = slice | int | str | np.int64 | np.ndarray
+Index1D = slice | int | str | np.int64 | np.ndarray | pd.Series
 IndexRest = Index1D | EllipsisType
 Index = (
     IndexRest

tests/test_backed_sparse.py

@@ -615,17 +615,28 @@ def test_backed_sizeof(
     assert csr_mem.__sizeof__() > csc_disk.__sizeof__()
 
 
+sparray_scipy_bug_marks = (
+    [pytest.mark.skip(reason="scipy bug causes view to be allocated")]
+    if CAN_USE_SPARSE_ARRAY
+    else []
+)
+
+
 @pytest.mark.parametrize(
     "group_fn",
     [
         pytest.param(lambda _: zarr.group(), id="zarr"),
         pytest.param(lambda p: h5py.File(p / "test.h5", mode="a"), id="h5py"),
     ],
 )
-@pytest.mark.parametrize("sparse_class", [sparse.csr_matrix, sparse.csr_array])
+@pytest.mark.parametrize(
+    "sparse_class",
+    [
+        sparse.csr_matrix,
+        pytest.param(sparse.csr_array, marks=[*sparray_scipy_bug_marks]),
+    ],
+)
 def test_append_overflow_check(group_fn, sparse_class, tmp_path):
-    if CAN_USE_SPARSE_ARRAY and issubclass(sparse_class, SpArray):
-        pytest.skip("scipy bug causes view to be allocated")
     group = group_fn(tmp_path)
     typemax_int32 = np.iinfo(np.int32).max
     orig_mtx = sparse_class(np.ones((1, 1), dtype=bool))

tests/test_concatenate.py

@@ -1446,7 +1446,7 @@ def test_concat_outer_aligned_mapping(elem):
     del b.obsm[elem]
 
     concated = concat({"a": a, "b": b}, join="outer", label="group")
-    result = concated.obsm[elem][concated.obs["group"] == "b"]
+    result = concated[concated.obs["group"] == "b"].obsm[elem]
 
     check_filled_like(result, elem_name=f"obsm/{elem}")