Merge pull request #66 from eric-czech/regenie_wgr

REGENIE Implementation

Author: eric-czech

.gitignore

@@ -131,6 +131,7 @@ dmypy.json
 # IDE
 .vscode
 .idea
+.DS_Store
 
 # sgkit
 docs/generated

requirements-dev.txt

@@ -2,7 +2,9 @@ codecov
 pre-commit
 pytest
 pytest-cov
+pytest-datadir
 hypothesis
+statsmodels
+zarr
 sphinx
 sphinx_rtd_theme
-statsmodels

requirements.txt

@@ -1,4 +1,4 @@
 numpy
 xarray
 dask[array]
-scipy
+scipy

setup.cfg

@@ -27,6 +27,8 @@ python_requires = >=3.6
 install_requires =
     numpy
     xarray
+    dask[array]
+    scipy
     setuptools >= 41.2  # For pkg_resources
 setup_requires =
     setuptools >= 41.2
@@ -36,7 +38,7 @@ setup_requires =
 fail_under = 100
 
 [tool:pytest]
-addopts = --doctest-modules
+addopts = --doctest-modules --ignore=validation
 norecursedirs = .eggs docs
 
 [flake8]
@@ -55,7 +57,7 @@ ignore =
 profile = black
 default_section = THIRDPARTY
 known_first_party = sgkit
-known_third_party = dask,hypothesis,numpy,pandas,pytest,setuptools,xarray
+known_third_party = dask,fire,glow,hail,hypothesis,invoke,numpy,pandas,pkg_resources,pyspark,pytest,setuptools,sgkit_plink,xarray,yaml,zarr
 multi_line_output = 3
 include_trailing_comma = True
 force_grid_wrap = 0
@@ -74,7 +76,13 @@ ignore_missing_imports = True
 ignore_missing_imports = True
 [mypy-hypothesis.*]
 ignore_missing_imports = True
+[mypy-zarr.*]
+ignore_missing_imports = True
 [mypy-setuptools]
 ignore_missing_imports = True
+[mypy-sgkit.*]
+allow_redefinition = True
 [mypy-sgkit.tests.*]
 disallow_untyped_defs = False
+[mypy-validation.*]
+ignore_errors = True

sgkit/__init__.py

@@ -9,6 +9,7 @@
 from .display import display_genotypes
 from .stats.aggregation import count_alleles
 from .stats.association import gwas_linear_regression
+from .stats.regenie import regenie
 
 __all__ = [
     "DIM_ALLELE",
@@ -20,4 +21,5 @@
     "create_genotype_dosage_dataset",
     "display_genotypes",
     "gwas_linear_regression",
+    "regenie",
 ]

sgkit/stats/aggregation.py

@@ -55,7 +55,7 @@ def count_alleles(ds: Dataset) -> DataArray:
     CT = da.map_blocks(
         lambda x: np.apply_along_axis(np.bincount, 1, x, minlength=max_allele),
         G,
-        chunks=(G.chunks[0], max_allele),  # type: ignore[index]
+        chunks=(G.chunks[0], max_allele),
     )
     assert CT.shape == (n_variant, G.numblocks[1] * max_allele)