Integrate stats.ic_scale rcParam (#993)

arviz/plots/elpdplot.py

@@ -21,7 +21,7 @@ def plot_elpd(
     threshold=None,
     ax=None,
     ic=None,
-    scale="deviance",
+    scale=None,
     plot_kwargs=None,
     backend=None,
     backend_kwargs=None,
@@ -109,6 +109,7 @@ def plot_elpd(
     """
     valid_ics = ["waic", "loo"]
     ic = rcParams["stats.information_criterion"] if ic is None else ic.lower()
+    scale = rcParams["stats.ic_scale"] if scale is None else scale.lower()
     if ic not in valid_ics:
         raise ValueError(
             ("Information Criteria type {} not recognized." "IC must be in {}").format(

arviz/stats/stats.py

@@ -42,13 +42,7 @@
 
 
 def compare(
-    dataset_dict,
-    ic=None,
-    method="BB-pseudo-BMA",
-    b_samples=1000,
-    alpha=1,
-    seed=None,
-    scale="deviance",
+    dataset_dict, ic=None, method="BB-pseudo-BMA", b_samples=1000, alpha=1, seed=None, scale=None
 ):
     r"""Compare models based on WAIC or LOO cross-validation.
 
@@ -136,7 +130,7 @@ def compare(
 
     """
     names = list(dataset_dict.keys())
-    scale = scale.lower()
+    scale = rcParams["stats.ic_scale"] if scale is None else scale.lower()
     if scale == "log":
         scale_value = 1
         ascending = False
@@ -421,7 +415,7 @@ def hpd(ary, credible_interval=0.94, circular=False, multimodal=False):
     return hpd_intervals
 
 
-def loo(data, pointwise=False, reff=None, scale="deviance"):
+def loo(data, pointwise=False, reff=None, scale=None):
     """Pareto-smoothed importance sampling leave-one-out cross-validation.
 
     Calculates leave-one-out (LOO) cross-validation for out of sample predictive model fit,
@@ -493,12 +487,13 @@ def loo(data, pointwise=False, reff=None, scale="deviance"):
     shape = log_likelihood.shape
     n_samples = shape[-1]
     n_data_points = np.product(shape[:-1])
+    scale = rcParams["stats.ic_scale"] if scale is None else scale.lower()
 
-    if scale.lower() == "deviance":
+    if scale == "deviance":
         scale_value = -2
-    elif scale.lower() == "log":
+    elif scale == "log":
         scale_value = 1
-    elif scale.lower() == "negative_log":
+    elif scale == "negative_log":
         scale_value = -1
     else:
         raise TypeError('Valid scale values are "deviance", "log", "negative_log"')
@@ -1101,7 +1096,7 @@ def summary(
     return summary_df
 
 
-def waic(data, pointwise=False, scale="deviance"):
+def waic(data, pointwise=False, scale=None):
     """Calculate the widely available information criterion.
 
     Also calculates the WAIC's standard error and the effective number of
@@ -1165,12 +1160,13 @@ def waic(data, pointwise=False, scale="deviance"):
     if "log_likelihood" not in inference_data.sample_stats:
         raise TypeError("Data must include log_likelihood in sample_stats")
     log_likelihood = inference_data.sample_stats.log_likelihood
+    scale = rcParams["stats.ic_scale"] if scale is None else scale.lower()
 
-    if scale.lower() == "deviance":
+    if scale == "deviance":
         scale_value = -2
-    elif scale.lower() == "log":
+    elif scale == "log":
         scale_value = 1
-    elif scale.lower() == "negative_log":
+    elif scale == "negative_log":
         scale_value = -1
     else:
         raise TypeError('Valid scale values are "deviance", "log", "negative_log"')

arviz/tests/test_plots_bokeh.py

@@ -189,7 +189,7 @@ def test_plot_kde_1d(continuous_model):
     "kwargs",
     [
         {"contour": True, "fill_last": False},
-        {"contour": True, "contourf_kwargs": {"cmap": "plasma"},},
+        {"contour": True, "contourf_kwargs": {"cmap": "plasma"}},
         {"contour": False},
         {"contour": False, "pcolormesh_kwargs": {"cmap": "plasma"}},
     ],
@@ -275,9 +275,7 @@ def test_plot_compare_no_ic(models):
     assert "['waic', 'loo']" in str(err.value)
 
 
-@pytest.mark.parametrize(
-    "kwargs", [{}, {"ic": "loo"}, {"xlabels": True, "scale": "log"},],
-)
+@pytest.mark.parametrize("kwargs", [{}, {"ic": "loo"}, {"xlabels": True, "scale": "log"}])
 @pytest.mark.parametrize("add_model", [False, True])
 @pytest.mark.parametrize("use_elpddata", [False, True])
 def test_plot_elpd(models, add_model, use_elpddata, kwargs):
@@ -300,9 +298,7 @@ def test_plot_elpd(models, add_model, use_elpddata, kwargs):
         assert axes.shape[0] == len(model_dict) - 1
 
 
-@pytest.mark.parametrize(
-    "kwargs", [{}, {"ic": "loo"}, {"xlabels": True, "scale": "log"},],
-)
+@pytest.mark.parametrize("kwargs", [{}, {"ic": "loo"}, {"xlabels": True, "scale": "log"}])
 @pytest.mark.parametrize("add_model", [False, True])
 @pytest.mark.parametrize("use_elpddata", [False, True])
 def test_plot_elpd_multidim(multidim_models, add_model, use_elpddata, kwargs):
@@ -443,9 +439,7 @@ def test_plot_forest(models, model_fits, args_expected):
 
 def test_plot_forest_rope_exception():
     with pytest.raises(ValueError) as err:
-        plot_forest(
-            {"x": [1]}, rope="not_correct_format", backend="bokeh", show=False,
-        )
+        plot_forest({"x": [1]}, rope="not_correct_format", backend="bokeh", show=False)
     assert "Argument `rope` must be None, a dictionary like" in str(err.value)
 
 
@@ -508,19 +502,15 @@ def test_plot_joint_discrete(discrete_model):
 def test_plot_joint_bad(models):
     with pytest.raises(ValueError):
         plot_joint(
-            models.model_1, var_names=("mu", "tau"), kind="bad_kind", backend="bokeh", show=False,
+            models.model_1, var_names=("mu", "tau"), kind="bad_kind", backend="bokeh", show=False
         )
 
     with pytest.raises(Exception):
-        plot_joint(
-            models.model_1, var_names=("mu", "tau", "eta"), backend="bokeh", show=False,
-        )
+        plot_joint(models.model_1, var_names=("mu", "tau", "eta"), backend="bokeh", show=False)
 
     with pytest.raises(ValueError):
         _, axes = list(range(5))
-        plot_joint(
-            models.model_1, var_names=("mu", "tau"), ax=axes, backend="bokeh", show=False,
-        )
+        plot_joint(models.model_1, var_names=("mu", "tau"), ax=axes, backend="bokeh", show=False)
 
 
 @pytest.mark.parametrize(
@@ -614,7 +604,7 @@ def test_plot_loo_pit_incompatible_args(models):
     """Test error when both ecdf and use_hpd are True."""
     with pytest.raises(ValueError, match="incompatible"):
         plot_loo_pit(
-            idata=models.model_1, y="y", ecdf=True, use_hpd=True, backend="bokeh", show=False,
+            idata=models.model_1, y="y", ecdf=True, use_hpd=True, backend="bokeh", show=False
         )
 
 
@@ -694,8 +684,8 @@ def test_plot_mcse_no_divergences(models):
 @pytest.mark.parametrize(
     "kwargs",
     [
-        {"var_names": "theta", "divergences": True, "coords": {"theta_dim_0": [0, 1]},},
-        {"divergences": True, "var_names": ["theta", "mu"],},
+        {"var_names": "theta", "divergences": True, "coords": {"theta_dim_0": [0, 1]}},
+        {"divergences": True, "var_names": ["theta", "mu"]},
         {"kind": "kde", "var_names": ["theta"]},
         {"kind": "hexbin", "var_names": ["theta"]},
         {"kind": "hexbin", "var_names": ["theta"]},
@@ -760,7 +750,7 @@ def test_plot_parallel_exception(models, var_names):
     """Ensure that correct exception is raised when one variable is passed."""
     with pytest.raises(ValueError):
         assert plot_parallel(
-            models.model_1, var_names=var_names, norm_method="foo", backend="bokeh", show=False,
+            models.model_1, var_names=var_names, norm_method="foo", backend="bokeh", show=False
         )
 
 
@@ -860,9 +850,7 @@ def test_plot_ppc_bad(models, kind):
     with pytest.raises(TypeError):
         plot_ppc(models.model_1, kind="bad_val", backend="bokeh", show=False)
     with pytest.raises(TypeError):
-        plot_ppc(
-            models.model_1, num_pp_samples="bad_val", backend="bokeh", show=False,
-        )
+        plot_ppc(models.model_1, num_pp_samples="bad_val", backend="bokeh", show=False)
 
 
 @pytest.mark.parametrize("kind", ["kde", "cumulative", "scatter"])
@@ -913,13 +901,9 @@ def test_plot_posterior_bad(models):
     with pytest.raises(ValueError):
         plot_posterior(models.model_1, backend="bokeh", show=False, rope="bad_value")
     with pytest.raises(ValueError):
-        plot_posterior(
-            models.model_1, ref_val="bad_value", backend="bokeh", show=False,
-        )
+        plot_posterior(models.model_1, ref_val="bad_value", backend="bokeh", show=False)
     with pytest.raises(ValueError):
-        plot_posterior(
-            models.model_1, point_estimate="bad_value", backend="bokeh", show=False,
-        )
+        plot_posterior(models.model_1, point_estimate="bad_value", backend="bokeh", show=False)
 
 
 @pytest.mark.parametrize("point_estimate", ("mode", "mean", "median"))

examples/bokeh/bokeh_plot_dist.py

@@ -16,9 +16,7 @@
 ax_poisson = bkp.figure(**figure_kwargs)
 ax_normal = bkp.figure(**figure_kwargs)
 
-az.plot_dist(
-    a, color="black", label="Poisson", ax=ax_poisson, backend="bokeh", show=False,
-)
+az.plot_dist(a, color="black", label="Poisson", ax=ax_poisson, backend="bokeh", show=False)
 az.plot_dist(b, color="red", label="Gaussian", ax=ax_normal, backend="bokeh", show=False)
 
 ax = row(ax_poisson, ax_normal)

examples/bokeh/bokeh_plot_loo_pit_ecdf.py

@@ -11,5 +11,5 @@
 log_weights = az.psislw(-log_like)[0]
 
 ax = az.plot_loo_pit(
-    idata, y="y_like", log_weights=log_weights, ecdf=True, color="orange", backend="bokeh",
+    idata, y="y_like", log_weights=log_weights, ecdf=True, color="orange", backend="bokeh"
 )

examples/bokeh/bokeh_plot_pair.py

@@ -10,5 +10,5 @@
 
 coords = {"school": ["Choate", "Deerfield"]}
 ax = az.plot_pair(
-    centered, var_names=["theta", "mu", "tau"], coords=coords, divergences=True, backend="bokeh",
+    centered, var_names=["theta", "mu", "tau"], coords=coords, divergences=True, backend="bokeh"
 )