Run updated version of black to fix CI

pymare/estimators/combination.py

@@ -112,7 +112,7 @@ def p_value(self, z, w=None):
         """Calculate p-values."""
         if w is None:
             w = np.ones_like(z)
-        cz = (z * w).sum(0) / np.sqrt((w ** 2).sum(0))
+        cz = (z * w).sum(0) / np.sqrt((w**2).sum(0))
         return ss.norm.sf(cz)
 
 

pymare/estimators/estimators.py

@@ -228,7 +228,7 @@ def fit(self, y, v, X):
 
         # Einsum indices: k = studies, p = predictors, i = parallel iterates.
         # q is a dummy for 2nd p when p x p covariance matrix is passed.
-        Xw2 = np.einsum("kp,ki->ipk", X, w ** 2)
+        Xw2 = np.einsum("kp,ki->ipk", X, w**2)
         pXw2 = np.einsum("ipk,qpi->iqk", Xw2, inv_cov)
         A = w_sum - np.trace(pXw2.dot(X), axis1=1, axis2=2)
         tau_dl = (Q - (k - p)) / A

pymare/results.py

@@ -179,7 +179,7 @@ def permutation_test(self, n_perm=1000):
         if has_mods:
             n_exact = np.math.factorial(n_obs)
         else:
-            n_exact = 2 ** n_obs
+            n_exact = 2**n_obs
             if n_exact < n_perm:
                 perms = np.array(list(itertools.product([-1, 1], repeat=n_obs))).T
 
@@ -308,7 +308,7 @@ def permutation_test(self, n_perm=1000):
         p_p = np.zeros_like(self.z)
 
         # Calculate # of permutations and determine whether to use exact test
-        n_exact = 2 ** n_obs
+        n_exact = 2**n_obs
         if n_exact < n_perm:
             perms = np.array(list(itertools.product([-1, 1], repeat=n_obs))).T
             exact = True

pymare/tests/test_effectsize_base.py

@@ -124,7 +124,7 @@ def test_convert_r_to_itself():
         esc.get_v_r()
     esc = OneSampleEffectSizeConverter(r=r, n=n)
     v_r = esc.get("V_R")
-    assert np.allclose(v_r, (1 - r ** 2) / (n - 2))
+    assert np.allclose(v_r, (1 - r**2) / (n - 2))
     ds = esc.to_dataset(measure="R")
     assert np.allclose(ds.y.ravel(), r)
     assert np.allclose(ds.v.ravel(), v_r)