ModelResults.summary output is mo concise, mo infomative, mo betta

It is now based on pm.df_summary() rather than pm.summary(). Terms with
multiple levels have the actual level names printed in the output.
Internally transformed variables are suppressed by default. Random
effects are not summarized by default (but their SDs are). Between this
and e02a276, this effectively closes #36 and definitely closes #41.

Also adds test both for this and for the traceplots and priors plots.
The tests are in test_model basically because we have to fit a model
before we can test the fitted model summary functions.

bambi/results.py

@@ -63,8 +63,21 @@ def __init__(self, model, trace):
     def plot(self, burn_in=0, names=None, annotate=True, hide_transformed=True,
         kind='trace', **kwargs):
         '''
-        Plots posterior distributions and sample traces. Code slightly modified from:
+        Plots posterior distributions and sample traces. Basically a wrapper
+        for pm.traceplot() plus some niceties, based partly on code from:
         https://pymc-devs.github.io/pymc3/notebooks/GLM-model-selection.html
+        Args:
+            burn_in (int): Number of initial samples to exclude before
+                summary statistics are computed.
+            names (list): Optional list of variable names to summarize.
+            annotate (bool): If True (default), add lines marking the
+                posterior means, write the posterior means next to the
+                lines, and add factor level names for fixed factors with
+                more than one distribution on the traceplot.
+            hide_transformed (bool): If True (default), do not print
+                summary statistics for internally transformed variables.
+            kind (str): Either 'trace' (default) or 'priors'. If 'priors',
+                this just internally calls Model.plot()
         '''
         if kind == 'priors':
             return self.model.plot()
@@ -117,16 +130,46 @@ def plot(self, burn_in=0, names=None, annotate=True, hide_transformed=True,
 
         return ax
 
-    def summary(self, burn_in=0, fixed=True, random=True, names=None,
+    def summary(self, burn_in=0, exclude_ranefs=True, names=None,
                 hide_transformed=True, **kwargs):
         '''
-        Summarizes all parameter estimates. Currently just a wrapper for
-        pm.summary().
+        Summarizes all parameter estimates. Basically a wrapper for
+        pm.df_summary() plus some niceties.
+        Args:
+            burn_in (int): Number of initial samples to exclude before
+                summary statistics are computed.
+            exclude_ranefs (bool): If True (default), do not print
+                summary statistics for individual random effects.
+            names (list): Optional list of variable names to summarize.
+            hide_transformed (bool): If True (default), do not print
+                summary statistics for internally transformed variables.
         '''
+
+        # if no 'names' specified, filter out unwanted variables
         if names is None:
             names = self.untransformed_vars if hide_transformed else self.trace.varnames
-
-        return pm.summary(self.trace[burn_in:], varnames=names, **kwargs)
+            if exclude_ranefs:
+                names = [x for x in names
+                    if x[2:] not in list(self.model.random_terms.keys())]
+
+        # get the basic DataFrame
+        df = pm.df_summary(self.trace[burn_in:], varnames=names, **kwargs)
+
+        # replace the "__\d" suffixes with an informative factor level name
+        match = [re.match('^(.*)(?:__)(\d+)?$', x) for x in df.index]
+        def replace_with_name(match):
+            term = self.model.terms[match.group(1)[2:]]
+            # handle fixed effects
+            if term in self.model.fixed_terms.values():
+                return term.levels[int(match.group(2))]
+            # handle random effects
+            else:
+                return '{}[{}]'.format(term.name, term.levels[int(match.group(2))])
+        new = [replace_with_name(x) if x is not None else df.index[i]
+            for i,x in enumerate(match)]
+        df.set_index([new], inplace=True)
+
+        return df
 
 
 class PyMC3ADVIResults(ModelResults):

bambi/tests/test_model.py

@@ -351,7 +351,7 @@ def test_cell_means_with_random_intercepts(crossed_data):
     model0 = Model(crossed_data)
     model0.fit('Y ~ 0 + threecats', random=['subj'], run=False)
     model0.build()
-    model0.fit(samples=1)
+    fitted = model0.fit(samples=10)
 
     # using add_term
     model1 = Model(crossed_data, intercept=False)
@@ -380,6 +380,14 @@ def test_cell_means_with_random_intercepts(crossed_data):
     priors1 = {x.name:x.prior.args['sd'].args for x in model1.terms.values() if x.random}
     assert set(priors0) == set(priors1)
 
+    # test summary
+    fitted.summary()
+    fitted.summary(exclude_ranefs=False)
+
+    # test plots
+    fitted.plot(kind='priors')
+    fitted.plot()
+
 
 def test_random_intercepts(crossed_data):
     # using formula and '1|' syntax