Add show results function (a wrapper around show_priors but wit ha be…

…tter name) (#45)

cobaya_utilities/plots.py

@@ -297,6 +297,41 @@ def get_mc_samples(
     return samples, labels, colors
 
 
+def show_results(g, results, with_legend=True, legend_kwargs=None):
+    """Show results values on a given set of axes
+    Parameters
+    ----------
+    g: getdist.plots
+      the getdist plotter instance
+    results: dict
+      dictionary holding result values. The dict can be either a parameter: loc/scale combo or
+      parameter: dict(loc/scale +  and other settings such as color, linestyle)
+    with_legend: bool
+      add legend
+    legend_kwargs: dict
+      set legend kwargs
+    """
+    for name, result in results.items():
+        show_priors(
+            g,
+            result.get("values", result),
+            color=result.get("color"),
+            ls=result.get("ls", "-"),
+            with_legend=False,
+        )
+    kwargs = dict(
+        ax=g.fig.axes[-1],
+        labels=results.keys(),
+        colors=[v.get("color") for v in results.values()],
+        fontsize=10,
+        bbox_to_anchor=(1, 1),
+        loc="upper left",
+    )
+    kwargs |= legend_kwargs or {}
+    if with_legend:
+        add_legend(**kwargs)
+
+
 def show_priors(g, priors, color="gray", ls="--", with_legend=True, legend_kwargs=None):
     """Show prior values on a given set of axes
     Parameters
@@ -528,18 +563,29 @@ def plot_mean_distributions(samples, params, colors="0.7", return_results=False,
         return pd.DataFrame.from_dict(results, orient="index")
 
 
-def add_legend(fig=None, ax=None, labels=None, colors=None, ls=None, **kwargs):
+def add_legend(obj, labels=None, colors=None, ls=None, **kwargs):
+    """Add legend
 
-    if not fig and not ax:
-        raise ValueError("Missing either fig or axis instance!")
+    Parameters
+    ----------
+    obj: figure or axis
+      a matplotlib figure or axis
+    params: dict or list
+      a dict holding the parameter names and its default value or
+      a unique list of parameter names
+    colors: list or str
+      the colors of the different markers. If colors == "chi2" then the markers will be colored
+      relatively to their chi2 values (if default values are given for parameters)
+    return_results: bool
+      return a pandas Dataframe holding the results
+    """
 
     labels = labels or kwargs.get("legend_labels")
 
     colors = colors or [None for label in labels]
     ls = ls or ["-" for label in labels]
     handles = [Line2D([0], [0], color=colors[i], ls=ls[i]) for i, label in enumerate(labels)]
 
-    obj = fig or ax
     leg = obj.legend(
         handles,
         labels,