Use pyupgrade *.py --py310-plus (#364)

pingouin/bayesian.py

@@ -300,7 +300,7 @@ def fun(g, r, n):
             )
 
         integr = quad(fun, 0, np.inf, args=(r, n))[0]
-        bf10 = np.sqrt((n / 2)) / gamma(1 / 2) * integr
+        bf10 = np.sqrt(n / 2) / gamma(1 / 2) * integr
 
     else:
         # Ly et al, 2016. Analytical solution.

pingouin/contingency.py

@@ -147,7 +147,7 @@ def chi2_independence(data, x, y, correction=True):
     # All count frequencies should be at least 5
     for df, name in zip([observed, expected], ["observed", "expected"]):
         if (df < 5).any(axis=None):
-            warnings.warn("Low count on {} frequencies.".format(name))
+            warnings.warn(f"Low count on {name} frequencies.")
 
     dof = float(expected.size - sum(expected.shape) + expected.ndim - 1)
 

pingouin/effsize.py

@@ -598,7 +598,7 @@ def convert_effsize(ef, input_type, output_type, nx=None, ny=None):
     # Check input and output type
     for inp in [it, ot]:
         if not _check_eftype(inp):
-            err = "Could not interpret input '{}'".format(inp)
+            err = f"Could not interpret input '{inp}'"
             raise ValueError(err)
     if it not in ["pointbiserialr", "cohen"]:
         raise ValueError("Input type must be 'cohen' or 'pointbiserialr'")
@@ -768,7 +768,7 @@ def compute_effsize(x, y, paired=False, eftype="cohen"):
     """
     # Check arguments
     if not _check_eftype(eftype):
-        err = "Could not interpret input '{}'".format(eftype)
+        err = f"Could not interpret input '{eftype}'"
         raise ValueError(err)
 
     x = np.asarray(x)
@@ -862,7 +862,7 @@ def compute_effsize_from_t(tval, nx=None, ny=None, N=None, eftype="cohen"):
     0.7487767802667672
     """
     if not _check_eftype(eftype):
-        err = "Could not interpret input '{}'".format(eftype)
+        err = f"Could not interpret input '{eftype}'"
         raise ValueError(err)
 
     if not isinstance(tval, float):

pingouin/pairwise.py

@@ -1324,8 +1324,7 @@ def traverse(o, tree_types=(list, tuple)):
         """
         if isinstance(o, tree_types):
             for value in o:
-                for subvalue in traverse(value, tree_types):
-                    yield subvalue
+                yield from traverse(value, tree_types)
         else:
             yield o
 

pingouin/parametric.py

@@ -594,7 +594,7 @@ def rm_anova(
 
     # If required, apply Greenhouse-Geisser correction for sphericity
     if correction:
-        corr_ddof1, corr_ddof2 = [np.maximum(d * eps, 1.0) for d in (ddof1, ddof2)]
+        corr_ddof1, corr_ddof2 = (np.maximum(d * eps, 1.0) for d in (ddof1, ddof2))
         p_corr = f(corr_ddof1, corr_ddof2).sf(fval)
 
     # Create output dataframe
@@ -781,9 +781,9 @@ def rm_anova2(data=None, dv=None, within=None, subject=None, effsize="ng2"):
     eps_ab = epsilon(data_piv, correction="gg")
 
     # Greenhouse-Geisser correction
-    df_a_c, df_as_c = [np.maximum(d * eps_a, 1.0) for d in (df_a, df_as)]
-    df_b_c, df_bs_c = [np.maximum(d * eps_b, 1.0) for d in (df_b, df_bs)]
-    df_ab_c, df_abs_c = [np.maximum(d * eps_ab, 1.0) for d in (df_ab, df_abs)]
+    df_a_c, df_as_c = (np.maximum(d * eps_a, 1.0) for d in (df_a, df_as))
+    df_b_c, df_bs_c = (np.maximum(d * eps_b, 1.0) for d in (df_b, df_bs))
+    df_ab_c, df_abs_c = (np.maximum(d * eps_ab, 1.0) for d in (df_ab, df_abs))
     p_a_corr = f(df_a_c, df_as_c).sf(f_a)
     p_b_corr = f(df_b_c, df_bs_c).sf(f_b)
     p_ab_corr = f(df_ab_c, df_abs_c).sf(f_ab)
@@ -1714,7 +1714,7 @@ def ancova(data=None, dv=None, between=None, covar=None, effsize="np2"):
     assert dv not in ["C", "Q"], "`dv` must not be 'C' or 'Q'."
     assert between not in ["C", "Q"], "`between` must not be 'C' or 'Q'."
     assert all(c not in ["C", "Q"] for c in covar), "`covar` must not contain 'C' or 'Q'."
-    formula = "Q('%s') ~ C(Q('%s'))" % (dv, between)
+    formula = f"Q('{dv}') ~ C(Q('{between}'))"
     for c in covar:
         formula += " + Q('%s')" % (c)
     model = ols(formula, data=data).fit()

pingouin/plotting.py

@@ -1004,7 +1004,7 @@ def plot_rm_corr(
     assert x not in ["C", "Q"], "`x` must not be 'C' or 'Q'."
     assert y not in ["C", "Q"], "`y` must not be 'C' or 'Q'."
     assert subject not in ["C", "Q"], "`subject` must not be 'C' or 'Q'."
-    formula = "Q('%s') ~ C(Q('%s')) + Q('%s')" % (y, subject, x)
+    formula = f"Q('{y}') ~ C(Q('{subject}')) + Q('{x}')"
     model = ols(formula, data=data).fit()
 
     # Fitted values

pingouin/reliability.py

@@ -359,7 +359,7 @@ def intraclass_corr(data=None, targets=None, raters=None, ratings=None, nan_poli
     u3 = (f3u - 1) / (f3u + (k - 1))
     # Case 2
     fj = msj / mse
-    vn = df2kd * ((k * icc2 * fj + n * (1 + (k - 1) * icc2) - k * icc2)) ** 2
+    vn = df2kd * (k * icc2 * fj + n * (1 + (k - 1) * icc2) - k * icc2) ** 2
     vd = df1 * k**2 * icc2**2 * fj**2 + (n * (1 + (k - 1) * icc2) - k * icc2) ** 2
     v = vn / vd
     f2u = f.ppf(1 - alpha / 2, n - 1, v)

pingouin/tests/__init__.py

@@ -1 +0,0 @@
-

pingouin/utils.py

@@ -145,9 +145,9 @@ def _postprocess_dataframe(df):
 
 def _get_round_setting_for(row, col):
     keys_to_check = (
-        "round.cell.[{}]x[{}]".format(row, col),
-        "round.column.{}".format(col),
-        "round.row.{}".format(row),
+        f"round.cell.[{row}]x[{col}]",
+        f"round.column.{col}",
+        f"round.row.{row}",
     )
     for key in keys_to_check:
         try:
@@ -397,11 +397,11 @@ def _is_statsmodels_installed(raise_error=False):
         import statsmodels  # noqa
 
         is_installed = True
-    except IOError:  # pragma: no cover
+    except OSError:  # pragma: no cover
         is_installed = False
     # Raise error (if needed) :
     if raise_error and not is_installed:  # pragma: no cover
-        raise IOError("statsmodels needs to be installed. Please use `pip " "install statsmodels`.")
+        raise OSError("statsmodels needs to be installed. Please use `pip " "install statsmodels`.")
     return is_installed
 
 
@@ -411,11 +411,11 @@ def _is_sklearn_installed(raise_error=False):
         import sklearn  # noqa
 
         is_installed = True
-    except IOError:  # pragma: no cover
+    except OSError:  # pragma: no cover
         is_installed = False
     # Raise error (if needed) :
     if raise_error and not is_installed:  # pragma: no cover
-        raise IOError("sklearn needs to be installed. Please use `pip " "install scikit-learn`.")
+        raise OSError("sklearn needs to be installed. Please use `pip " "install scikit-learn`.")
     return is_installed
 
 
@@ -425,9 +425,9 @@ def _is_mpmath_installed(raise_error=False):
         import mpmath  # noqa
 
         is_installed = True
-    except IOError:  # pragma: no cover
+    except OSError:  # pragma: no cover
         is_installed = False
     # Raise error (if needed) :
     if raise_error and not is_installed:  # pragma: no cover
-        raise IOError("mpmath needs to be installed. Please use `pip " "install mpmath`.")
+        raise OSError("mpmath needs to be installed. Please use `pip " "install mpmath`.")
     return is_installed