fix: adapt fixture usage to newer pytest versions (#879)

* Fix fixture usage

* we don’t even need the fixture

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fmt

* mypy

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: flying-sheep

tests/conftest.py

@@ -3,5 +3,5 @@
 
 
 @pytest.fixture
-def rng():
+def rng() -> np.random.Generator:
     return np.random.default_rng()

tests/tools/_distances/test_distance_tests.py

@@ -1,10 +1,12 @@
 import pytest
 import scanpy as sc
+from anndata import AnnData
 from pandas import DataFrame
 
 import pertpy as pt
+from pertpy.tools._distances._distances import Metric
 
-distances = [
+distances: tuple[Metric, ...] = (
     "edistance",
     "euclidean",
     "mse",
@@ -25,21 +27,21 @@
     # "nbll",
     "mahalanobis",
     "mean_var_distribution",
-]
+)
 
 count_distances = ["nb_ll"]
 
 
 @pytest.fixture
-def adata():
+def adata() -> AnnData:
     adata = pt.dt.distance_example()
     adata = sc.pp.subsample(adata, 0.1, copy=True)
 
     return adata
 
 
 @pytest.mark.parametrize("distance", distances)
-def test_distancetest(adata, distance):
+def test_distancetest(adata: AnnData, distance: Metric) -> None:
     etest = pt.tl.DistanceTest(distance, n_perms=10, obsm_key="X_pca", alpha=0.05, correction="holm-sidak")
     tab = etest(adata, groupby="perturbation", contrast="control")
 

tests/tools/_distances/test_distances.py

@@ -1,12 +1,14 @@
 import numpy as np
 import pytest
 import scanpy as sc
+from anndata import AnnData
 from pandas import DataFrame, Series
 from pytest import fixture, mark
 
 import pertpy as pt
+from pertpy.tools._distances._distances import Distance, Metric
 
-actual_distances = [
+actual_distances: tuple[Metric, ...] = (
     # Euclidean distances and related
     "euclidean",
     "mean_absolute_error",
@@ -22,31 +24,34 @@
     "t_test",
     "wasserstein",
     "mahalanobis",
-]
-semi_distances = ["r2_distance", "sym_kldiv", "ks_test"]
-non_distances = ["classifier_proba"]
-onesided_only = ["classifier_cp"]
-pseudo_counts_distances = ["nb_ll"]
-lognorm_counts_distances = ["mean_var_distribution"]
-all_distances = (
-    actual_distances + semi_distances + non_distances + lognorm_counts_distances + pseudo_counts_distances
-)  # + onesided_only
+)
+semi_distances: tuple[Metric, ...] = ("r2_distance", "sym_kldiv", "ks_test")
+non_distances: tuple[Metric, ...] = ("classifier_proba",)
+onesided_only: tuple[Metric, ...] = ("classifier_cp",)
+pseudo_counts_distances: tuple[Metric, ...] = ("nb_ll",)
+lognorm_counts_distances: tuple[Metric, ...] = ("mean_var_distribution",)
+all_distances: tuple[Metric, ...] = (
+    *actual_distances,
+    *semi_distances,
+    *non_distances,
+    *lognorm_counts_distances,
+    *pseudo_counts_distances,
+    # *onesided_only,
+)
 
 
 @fixture
-def adata(request):
-    low_subsample_distances = [
+def adata(distance: Metric, rng: np.random.Generator) -> AnnData:
+    low_subsample_distances = {
         "sym_kldiv",
         "t_test",
         "ks_test",
         "classifier_proba",
         "classifier_cp",
         "mahalanobis",
         "mean_var_distribution",
-    ]
-    no_subsample_distances = ["mahalanobis"]  # mahalanobis only works on the full data without subsampling
-
-    distance = request.node.callspec.params["distance"]
+    }
+    no_subsample_distances = {"mahalanobis"}  # mahalanobis only works on the full data without subsampling
 
     adata = pt.dt.distance_example()
     if distance not in no_subsample_distances:
@@ -55,7 +60,7 @@ def adata(request):
         else:
             adata = sc.pp.subsample(adata, 0.001, copy=True)
 
-    adata = adata[:, np.random.default_rng().choice(adata.n_vars, 100, replace=False)].copy()
+    adata = adata[:, rng.choice(adata.n_vars, 100, replace=False)].copy()
 
     adata.layers["lognorm"] = adata.X.copy()
     adata.layers["counts"] = np.round(adata.X.toarray()).astype(int)
@@ -70,25 +75,23 @@ def adata(request):
 
 
 @fixture
-def distance_obj(request):
-    distance = request.node.callspec.params["distance"]
+def distance_obj(distance: Metric) -> pt.tl.Distance:
     if distance in lognorm_counts_distances:
-        Distance = pt.tl.Distance(distance, layer_key="lognorm")
-    elif distance in pseudo_counts_distances:
-        Distance = pt.tl.Distance(distance, layer_key="counts")
-    else:
-        Distance = pt.tl.Distance(distance, obsm_key="X_pca")
-    return Distance
+        return pt.tl.Distance(distance, layer_key="lognorm")
+    if distance in pseudo_counts_distances:
+        return pt.tl.Distance(distance, layer_key="counts")
+    return pt.tl.Distance(distance, obsm_key="X_pca")
 
 
 @fixture
-@mark.parametrize("distance", all_distances)
-def pairwise_distance(adata, distance_obj, distance):
+def pairwise_distance(adata: AnnData, distance_obj: pt.tl.Distance) -> DataFrame:
     return distance_obj.pairwise(adata, groupby="perturbation", show_progressbar=True)
 
 
 @mark.parametrize("distance", actual_distances + semi_distances)
-def test_distance_axioms(pairwise_distance, distance):
+def test_distance_axioms(pairwise_distance: DataFrame, distance: Metric) -> None:
+    del distance
+
     # This is equivalent to testing for a semimetric, defined as fulfilling all axioms except triangle inequality.
     # (M1) Definiteness
     assert all(np.diag(pairwise_distance.values) == 0)  # distance to self is 0
@@ -102,12 +105,12 @@ def test_distance_axioms(pairwise_distance, distance):
 
 
 @mark.parametrize("distance", actual_distances)
-def test_triangle_inequality(pairwise_distance, distance, rng):
-    # Test if distances are well-defined in accordance with metric axioms
-    # (M4) Triangle inequality (we just probe this for a few random triplets)
-    # Some tests are not well defined for the triangle inequality. We skip those.
+def test_triangle_inequality(pairwise_distance: DataFrame, distance: Metric, rng: np.random.Generator) -> None:
+    """Test if distances are well-defined in accordance with metric axioms
+    (M4) Triangle inequality (we just probe this for a few random triplets)
+    """
     if distance in {"mahalanobis", "wasserstein"}:
-        return
+        pytest.skip("Some tests not well defined for triangle inequality")
 
     for _ in range(5):
         triplet = rng.choice(pairwise_distance.index, size=3, replace=False)
@@ -118,30 +121,33 @@ def test_triangle_inequality(pairwise_distance, distance, rng):
 
 
 @mark.parametrize("distance", all_distances)
-def test_distance_layers(pairwise_distance, distance):
+def test_distance_layers(pairwise_distance: DataFrame, distance: Metric) -> None:
+    del distance
+
     assert isinstance(pairwise_distance, DataFrame)
     assert pairwise_distance.columns.equals(pairwise_distance.index)
     assert np.sum(pairwise_distance.values - pairwise_distance.values.T) == 0  # symmetry
 
 
 @mark.parametrize("distance", actual_distances + pseudo_counts_distances)
-def test_distance_counts(adata, distance):
-    if distance != "mahalanobis":  # skip, doesn't work because covariance matrix is a singular matrix, not invertible
-        distance = pt.tl.Distance(distance, layer_key="counts")
-        df = distance.pairwise(adata, groupby="perturbation")
-        assert isinstance(df, DataFrame)
-        assert df.columns.equals(df.index)
-        assert np.sum(df.values - df.values.T) == 0
+def test_distance_counts(adata: AnnData, distance: Metric) -> None:
+    if distance == "mahalanobis":
+        pytest.skip("covariance matrix is a singular matrix, not invertible")
+    distance_obj = pt.tl.Distance(distance, layer_key="counts")
+    df = distance_obj.pairwise(adata, groupby="perturbation")
+    assert isinstance(df, DataFrame)
+    assert df.columns.equals(df.index)
+    assert np.sum(df.values - df.values.T) == 0
 
 
 @mark.parametrize("distance", all_distances)
-def test_mutually_exclusive_keys(distance):
+def test_mutually_exclusive_keys(distance: Metric) -> None:
     with pytest.raises(ValueError):
         _ = pt.tl.Distance(distance, layer_key="counts", obsm_key="X_pca")
 
 
 @mark.parametrize("distance", actual_distances + semi_distances + non_distances)
-def test_distance_output_type(distance, rng):
+def test_distance_output_type(distance: Metric, rng: np.random.Generator) -> None:
     # Test if distances are outputting floats
     Distance = pt.tl.Distance(distance)
     X = rng.normal(size=(50, 10))
@@ -151,15 +157,16 @@ def test_distance_output_type(distance, rng):
 
 
 @mark.parametrize("distance", all_distances + onesided_only)
-def test_distance_onesided(adata, distance_obj, distance):
+def test_distance_onesided(adata: AnnData, distance_obj: Distance, distance: Metric) -> None:
+    del distance
     # Test consistency of one-sided distance results
-    selected_group = adata.obs.perturbation.unique()[0]
+    selected_group = adata.obs["perturbation"].unique()[0]
     df = distance_obj.onesided_distances(adata, groupby="perturbation", selected_group=selected_group)
     assert isinstance(df, Series)
     assert df.loc[selected_group] == 0  # distance to self is 0
 
 
-def test_bootstrap_distance_output_type(rng):
+def test_bootstrap_distance_output_type(rng: np.random.Generator) -> None:
     # Test if distances are outputting floats
     Distance = pt.tl.Distance(metric="edistance")
     X = rng.normal(size=(50, 10))
@@ -170,7 +177,7 @@ def test_bootstrap_distance_output_type(rng):
 
 
 @mark.parametrize("distance", ["edistance"])
-def test_bootstrap_distance_pairwise(adata, distance):
+def test_bootstrap_distance_pairwise(adata: AnnData, distance: Metric) -> None:
     # Test consistency of pairwise distance results
     Distance = pt.tl.Distance(distance, obsm_key="X_pca")
     bootstrap_output = Distance.pairwise(adata, groupby="perturbation", bootstrap=True, n_bootstrap=3)
@@ -186,9 +193,9 @@ def test_bootstrap_distance_pairwise(adata, distance):
 
 
 @mark.parametrize("distance", ["edistance"])
-def test_bootstrap_distance_onesided(adata, distance):
+def test_bootstrap_distance_onesided(adata: AnnData, distance: Metric) -> None:
     # Test consistency of one-sided distance results
-    selected_group = adata.obs.perturbation.unique()[0]
+    selected_group = adata.obs["perturbation"].unique()[0]
     Distance = pt.tl.Distance(distance, obsm_key="X_pca")
     bootstrap_output = Distance.onesided_distances(
         adata,
@@ -201,7 +208,7 @@ def test_bootstrap_distance_onesided(adata, distance):
     assert isinstance(bootstrap_output, tuple)
 
 
-def test_compare_distance(rng):
+def test_compare_distance(rng: np.random.Generator) -> None:
     X = rng.normal(size=(50, 10))
     Y = rng.normal(size=(50, 10))
     C = rng.normal(size=(50, 10))
@@ -211,4 +218,4 @@ def test_compare_distance(rng):
     res_scaled = Distance.compare_distance(X, Y, C, mode="scaled")
     assert isinstance(res_scaled, float)
     with pytest.raises(ValueError):
-        Distance.compare_distance(X, Y, C, mode="new_mode")
+        Distance.compare_distance(X, Y, C, mode="new_mode")  # type: ignore[arg-type]