Test on Python 3.12 - but only on NumPy 2 due to cbgen version constraints

Update flake8

Fix flake8 errors

Don't run pre-commit for NumPy 2 tests

Try installing latest bio2zarr at end

Replace pkg_resources with importlib.metadata

Author: tomwhite

.github/workflows/build-numpy-2.yml

@@ -11,7 +11,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ["3.10", "3.11"]
+        python-version: ["3.10", "3.11", "3.12"]
 
     steps:
     - uses: actions/checkout@v2
@@ -22,11 +22,11 @@ jobs:
     - name: Install dependencies
       run: |
         python -m pip install --upgrade pip
-        pip install -r requirements.txt -r requirements-dev.txt
+        pip install -r requirements-numpy2.txt -r requirements-numpy2-dev.txt
         # update bio2zarr for NumPy 2, see https://github.com/sgkit-dev/bio2zarr/issues/256
         pip install -U 'numpy<2.1' -U git+https://github.com/sgkit-dev/bio2zarr.git
-    - name: Run pre-commit
-      uses: pre-commit/action@v2.0.0
+    # - name: Run pre-commit
+    #   uses: pre-commit/action@v2.0.0
     - name: Test with pytest (numba jit disabled)
       env:
         NUMBA_DISABLE_JIT: 1

.pre-commit-config.yaml

@@ -17,7 +17,7 @@ repos:
       - id: black
         language_version: python3
   - repo: https://github.com/pycqa/flake8
-    rev: 6.0.0
+    rev: 6.1.0
     hooks:
       - id: flake8
         language_version: python3

requirements-numpy2-dev.txt

@@ -0,0 +1,23 @@
+callee
+pre-commit
+pytest
+pytest-cov
+pytest-datadir
+pytest-mock
+hypothesis
+scikit-allel
+statsmodels
+msprime>=1.0
+scikit-learn
+partd
+bed-reader
+rechunker
+cbgen > 1.0.5; platform_system != "Windows"
+cbgen == 1.0.1; platform_system == "Windows"
+yarl
+matplotlib
+asv
+networkx
+aiohttp
+requests
+graphviz

requirements-numpy2.txt

@@ -0,0 +1,12 @@
+numpy < 2.1
+xarray
+dask[array] >= 2023.01.0, <= 2024.8.0
+distributed >= 2023.01.0, <= 2024.8.0
+dask-ml
+scipy
+typing-extensions
+numba
+zarr >= 2.10.0, != 2.11.0, != 2.11.1, != 2.11.2, < 3
+fsspec != 2021.6.*
+scikit-learn
+pandas

sgkit/__init__.py

@@ -1,4 +1,4 @@
-from pkg_resources import DistributionNotFound, get_distribution  # type: ignore[import]
+from importlib.metadata import PackageNotFoundError, version  # type: ignore[import]
 
 from .display import display_genotypes, display_pedigree
 from .distance.api import pairwise_distance
@@ -68,8 +68,8 @@
 )
 
 try:
-    __version__ = get_distribution(__name__).version
-except DistributionNotFound:  # pragma: no cover
+    __version__ = version(__name__)
+except PackageNotFoundError:  # pragma: no cover
     __version__ = "unknown"
 
 __all__ = [

sgkit/tests/test_stats_utils.py

@@ -107,7 +107,7 @@ def sample_dataset(draw):
     # Split first dimension assignment into two groups and assign second
     # dimension with unique names since n cols will differ
     dims = [
-        (f"dim-0-{i%2}",) if i < n_1d else (f"dim-0-{i%2}", f"dim-1-{i}")
+        (f"dim-0-{i % 2}",) if i < n_1d else (f"dim-0-{i % 2}", f"dim-1-{i}")
         for i in range(n_arrs)
     ]
 

validation/gwas/method/regenie/hail_sim.py

@@ -49,8 +49,8 @@ def get_plink_sim_dataset(n_variants=16, n_samples=4, n_contigs=2, seed=0):
         for s in range(n_samples):
             data.append(
                 {
-                    "v": f"{c+1}:{v+1}:A:C",
-                    "s": f"S{s+1:07d}",
+                    "v": f"{c + 1}:{v + 1}:A:C",
+                    "s": f"S{s + 1:07d}",
                     "cm": 0.1,
                     "GT": hl.Call([rs.randint(0, 2), rs.randint(0, 2)]),
                 }