Fixed inconsistency of advi and opvi include_transformed param (#2117)

* Fixed inconsistency of advi and opvi include_transformed param

* Fixed other inconsistencies on the ADVI OPVI API

Author: denadai2

docs/source/notebooks/GLM-model-selection.ipynb

@@ -313,7 +313,7 @@
     "    _ = plt.legend()\n",
     "    _ = ax1d.set_xlim(xlims)\n",
     "    _ = sns.regplot(x='x', y='y', data=rawdata, fit_reg=False\n",
-    "                   ,scatter_kws={'alpha':0.7,'s':100, 'lw':2,'edgecolor':'w'}, ax=ax1d)\n"
+    "                   ,scatter_kws={'alpha':0.7,'s':100, 'lw':2,'edgecolor':'w'}, ax=ax1d)"
    ]
   },
   {
@@ -1054,7 +1054,7 @@
     }
    ],
    "source": [
-    "dftrc_lin = pm.trace_to_dataframe(traces_lin['k1'], hide_transformed_vars=False)\n",
+    "dftrc_lin = pm.trace_to_dataframe(traces_lin['k1'], include_transformed=True)\n",
     "trc_lin_logp = dftrc_lin.apply(lambda x: models_lin['k1'].logp(x.to_dict()), axis=1)\n",
     "mean_deviance = -2 * trc_lin_logp.mean(0)\n",
     "mean_deviance"
@@ -1406,7 +1406,7 @@
   "language_info": {
    "codemirror_mode": {
     "name": "ipython",
-    "version": 3
+    "version": 3.0
    },
    "file_extension": ".py",
    "mimetype": "text/x-python",
@@ -1420,14 +1420,14 @@
     "87b986ac3e5a43ec859cf10e013f2955": {
      "views": [
       {
-       "cell_index": 9
+       "cell_index": 9.0
       }
      ]
     },
     "f1f05f8da738419e8e2c54ee1809c61c": {
      "views": [
       {
-       "cell_index": 47
+       "cell_index": 47.0
       }
      ]
     }
@@ -1437,4 +1437,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}

pymc3/backends/text.py

@@ -202,5 +202,5 @@ def dump(name, trace, chains=None):
     for chain in chains:
         filename = os.path.join(name, 'chain-{}.csv'.format(chain))
         df = ttab.trace_to_dataframe(
-            trace, chains=chain, hide_transformed_vars=False)
+            trace, chains=chain, include_transformed=True)
         df.to_csv(filename, index=False)

pymc3/backends/tracetab.py

@@ -9,7 +9,7 @@
 __all__ = ['trace_to_dataframe']
 
 
-def trace_to_dataframe(trace, chains=None, varnames=None, hide_transformed_vars=True):
+def trace_to_dataframe(trace, chains=None, varnames=None, include_transformed=False):
     """Convert trace to Pandas DataFrame.
 
     Parameters
@@ -21,15 +21,15 @@ def trace_to_dataframe(trace, chains=None, varnames=None, hide_transformed_vars=
     varnames : list of variable names
         Variables to be included in the DataFrame, if None all variable are
         included.
-    hide_transformed_vars: boolean
-        If true transformed variables will not be included in the resulting
+    include_transformed: boolean
+        If true transformed variables will be included in the resulting
         DataFrame.
     """
     var_shapes = trace._straces[0].var_shapes
 
     if varnames is None:
         varnames = get_default_varnames(var_shapes.keys(),
-                                        include_transformed=not hide_transformed_vars)
+                                        include_transformed=include_transformed)
 
     flat_names = {v: create_flat_names(v, var_shapes[v]) for v in varnames}
 
@@ -60,7 +60,7 @@ def create_flat_names(varname, shape):
 
 
 def _create_shape(flat_names):
-    "Determine shape from `create_flat_names` output."
+    """Determine shape from `create_flat_names` output."""
     try:
         _, shape_str = flat_names[-1].rsplit('__', 1)
     except ValueError:

pymc3/tests/test_advi.py

@@ -261,7 +261,7 @@ def test_sample_vp(self):
             p = pm.Beta('p', alpha=1, beta=1)
             pm.Binomial('xs', n=1, p=p, observed=xs)
             v_params = advi(n=1000)
-            trace = sample_vp(v_params, draws=1, hide_transformed=True)
+            trace = sample_vp(v_params, draws=1, include_transformed=False)
             assert trace.varnames == ['p']
-            trace = sample_vp(v_params, draws=1, hide_transformed=False)
+            trace = sample_vp(v_params, draws=1, include_transformed=True)
             assert sorted(trace.varnames) == ['p', 'p_logodds__']

pymc3/tests/test_variational_inference.py

@@ -112,10 +112,10 @@ def test_sample(self):
                 p = pm.Beta('p', alpha=1, beta=1)
                 pm.Binomial('xs', n=1, p=p, observed=xs)
                 app = self.inference().approx
-                trace = app.sample(draws=1, hide_transformed=True)
+                trace = app.sample(draws=1, include_transformed=False)
                 assert trace.varnames == ['p']
                 assert len(trace) == 1
-                trace = app.sample(draws=10, hide_transformed=False)
+                trace = app.sample(draws=10, include_transformed=True)
                 assert sorted(trace.varnames) == ['p', 'p_logodds__']
                 assert len(trace) == 10
 

pymc3/variational/advi.py

@@ -340,7 +340,7 @@ def optimizer(loss, param):
 
 def sample_vp(
         vparams, draws=1000, model=None, local_RVs=None, random_seed=None,
-        hide_transformed=True, progressbar=True):
+        include_transformed=False, progressbar=True):
     """Draw samples from variational posterior.
 
     Parameters
@@ -353,8 +353,8 @@ def sample_vp(
         Probabilistic model.
     random_seed : int or None
         Seed of random number generator.  None to use current seed.
-    hide_transformed : bool
-        If False, transformed variables are also sampled. Default is True.
+    include_transformed : bool
+        If True, transformed variables are also sampled. Default is False.
 
     Returns
     -------
@@ -411,7 +411,7 @@ def rvs(x):
 
     # Random variables which will be sampled
     vars_sampled = pm.util.get_default_varnames(model.unobserved_RVs,
-                                                include_transformed=not hide_transformed)
+                                                include_transformed=include_transformed)
 
     varnames = [str(var) for var in model.unobserved_RVs]
     trace = pm.sampling.NDArray(model=model, vars=vars_sampled)

pymc3/variational/approximations.py

@@ -17,8 +17,7 @@
 
 
 class MeanField(Approximation):
-    """
-    Mean Field approximation to the posterior where spherical Gaussian family
+    """Mean Field approximation to the posterior where spherical Gaussian family
     is fitted to minimize KL divergence from True posterior. It is assumed
     that latent space variables are uncorrelated that is the main drawback
     of the method
@@ -92,8 +91,7 @@ def random_global(self, size=None, no_rand=False):
 
 
 class FullRank(Approximation):
-    """
-    Full Rank approximation to the posterior where Multivariate Gaussian family
+    """Full Rank approximation to the posterior where Multivariate Gaussian family
     is fitted to minimize KL divergence from True posterior. In contrast to
     MeanField approach correlations between variables are taken in account. The
     main drawback of the method is computational cost.
@@ -175,8 +173,7 @@ def create_shared_params(self, **kwargs):
                 }
 
     def log_q_W_global(self, z):
-        """
-        log_q_W samples over q for global vars
+        """log_q_W samples over q for global vars
         """
         mu = self.scale_grad(self.mean)
         L = self.scale_grad(self.L)
@@ -197,8 +194,7 @@ def random_global(self, size=None, no_rand=False):
 
     @classmethod
     def from_mean_field(cls, mean_field, gpu_compat=False):
-        """
-        Construct FullRank from MeanField approximation
+        """Construct FullRank from MeanField approximation
 
         Parameters
         ----------
@@ -233,8 +229,7 @@ def from_mean_field(cls, mean_field, gpu_compat=False):
 
 
 class Empirical(Approximation):
-    """
-    Builds Approximation instance from a given trace,
+    """Builds Approximation instance from a given trace,
     it has the same interface as variational approximation
 
     Parameters
@@ -309,16 +304,14 @@ def random_global(self, size=None, no_rand=False):
 
     @property
     def histogram(self):
-        """
-        Shortcut to flattened Trace
+        """Shortcut to flattened Trace
         """
         return self.shared_params
 
     @property
     @memoize
     def histogram_logp(self):
-        """
-        Symbolic logp for every point in trace
+        """Symbolic logp for every point in trace
         """
         node = self.to_flat_input(self.model.logpt)
 
@@ -341,8 +334,7 @@ def cov(self):
 
     @classmethod
     def from_noise(cls, size, jitter=.01, local_rv=None, start=None, model=None, seed=None):
-        """
-        Initialize Histogram with random noise
+        """Initialize Histogram with random noise
 
         Parameters
         ----------
@@ -371,17 +363,16 @@ def from_noise(cls, size, jitter=.01, local_rv=None, start=None, model=None, see
         return hist
 
 
-def sample_approx(approx, draws=100, hide_transformed=False):
-    """
-    Draw samples from variational posterior.
+def sample_approx(approx, draws=100, include_transformed=True):
+    """Draw samples from variational posterior.
 
     Parameters
     ----------
     approx : Approximation
     draws : int
         Number of random samples.
-    hide_transformed : bool
-        If False, transformed variables are also sampled. Default is True.
+    include_transformed : bool
+        If True, transformed variables are also sampled. Default is True.
 
     Returns
     -------
@@ -390,4 +381,4 @@ def sample_approx(approx, draws=100, hide_transformed=False):
     """
     if not isinstance(approx, Approximation):
         raise TypeError('Need Approximation instance, got %r' % approx)
-    return approx.sample(draws=draws, hide_transformed=hide_transformed)
+    return approx.sample(draws=draws, include_transformed=include_transformed)