Add new groups to from_numpyro and from_dict (arviz-devs#1125)

* added groups

* update tests and changelog

* lint changes

* update from_dict and tests

* update changelog

* modify io_dict

* minor fixes

* update changelog again

Add warning for log scale default in compare/loo/waic functions (arviz-devs#1150)

* Added scale warning to ELPDData

* Added scale warning to compare function

* Added changes to Changelog

* Moved warning to end of string in loo function

* Removed last test for loo_print

* Ran Black

* Integrated Oriol's comments

add local to docstring, and use lowercase for ess (arviz-devs#1152)

hardcode show=False in plot_posterior subplots (arviz-devs#1151)

* hardcode show=False in plot_posterior subplots

* update changelog

Fix documentation and deprecation warning in pair plot (arviz-devs#1156)

The "kind" argument is not described correctly.

* Fix pairplot warning.

Previously, because of incorrect argument checking, pairplot would
mistakenly warn the caller not to use the "contour" argument when that
argument was NOT supplied.  Changed default value to None and did the
defaulting by hand to fix this issue.

Also added some type declarations.

* Clarified docstring for contour argument.

Co-authored-by: Robert P. Goldman <rpgoldman@sift.net>

add viridis as default cmap (arviz-devs#1160)

add examples to customize 2D KDE (arviz-devs#1158)

* add examples to customize 2D KDE

* blackify gallery example

* update changelog

* briefly explain contour_kwargs and contourf_kwargs

CHANGELOG.md

@@ -11,6 +11,10 @@
 * Revamped the `hpd` function to make it work with mutidimensional arrays, InferenceData and xarray objects (#1117)
 * Skip test for optional/extra dependencies when not installed (#1113)
 * Add option to display rank plots instead of trace (#1134)
+* Add out-of-sample groups (`predictions` and `predictions_constant_data`) to `from_dict` (#1125)
+* Add out-of-sample groups (`predictions` and `predictions_constant_data`) and `constant_data` group to pyro and numpyro translation (#1090, #1125)
+* Add `num_chains` and `pred_dims` arguments to from_pyro and from_numpyro (#1090, #1125)
+* Integrate jointplot into pairplot, add point-estimate and overlay of plot kinds (#1079)
 
 ### Maintenance and fixes
 * Fixed behaviour of `credible_interval=None` in `plot_posterior` (#1115)
@@ -21,14 +25,18 @@
 * Updated benchmarks and moved to asv_benchmarks/benchmarks (#1142)
 * Moved `_fast_kde`, `_fast_kde_2d`, `get_bins` and `_sturges_formula` to `numeric_utils` and `get_coords` to `utils` (#1142)
 * Rank plot: rename `axes` argument to `ax` (#1144)
+* Added a warning specifying log scale is now the default in compare/loo/waic functions ([#1150](https://github.com/arviz-devs/arviz/pull/1150))
+* Fixed bug in `plot_posterior` with rcParam "plot.matplotlib.show" = True (#1151)
+* Set `fill_last` argument of `plot_kde` to False by default (#1158)
 
 ### Deprecation
 
 ### Documentation
 * Add classifier to `setup.py` including Matplotlib framework (#1133)
 * Image thumbs generation updated to be Bokeh 2 compatible (#1116)
 * Add new examples for `plot_pair` (#1110)
-* Add examples for `psislw` and `r2_score` (#1129) 
+* Add examples for `psislw` and `r2_score` (#1129)
+* Add more examples on 2D kde customization (#1158)
 
 ## v0.7.0 (2020 Mar 2)
 

arviz/data/io_dict.py

@@ -16,25 +16,29 @@ def __init__(
         *,
         posterior=None,
         posterior_predictive=None,
+        predictions=None,
         sample_stats=None,
         log_likelihood=None,
         prior=None,
         prior_predictive=None,
         sample_stats_prior=None,
         observed_data=None,
         constant_data=None,
+        predictions_constant_data=None,
         coords=None,
         dims=None
     ):
         self.posterior = posterior
         self.posterior_predictive = posterior_predictive
+        self.predictions = predictions
         self.sample_stats = sample_stats
         self.log_likelihood = log_likelihood
         self.prior = prior
         self.prior_predictive = prior_predictive
         self.sample_stats_prior = sample_stats_prior
         self.observed_data = observed_data
         self.constant_data = constant_data
+        self.predictions_constant_data = predictions_constant_data
         self.coords = coords
         self.dims = dims
 
@@ -89,6 +93,15 @@ def posterior_predictive_to_xarray(self):
 
         return dict_to_dataset(data, library=None, coords=self.coords, dims=self.dims)
 
+    @requires("predictions")
+    def predictions_to_xarray(self):
+        """Convert predictions to xarray."""
+        data = self.predictions
+        if not isinstance(data, dict):
+            raise TypeError("DictConverter.predictions is not a dictionary")
+
+        return dict_to_dataset(data, library=None, coords=self.coords, dims=self.dims)
+
     @requires("prior")
     def prior_to_xarray(self):
         """Convert prior samples to xarray."""
@@ -116,45 +129,41 @@ def prior_predictive_to_xarray(self):
 
         return dict_to_dataset(data, library=None, coords=self.coords, dims=self.dims)
 
-    @requires("observed_data")
-    def observed_data_to_xarray(self):
-        """Convert observed_data to xarray."""
-        data = self.observed_data
+    def data_to_xarray(self, dct, group):
+        """Convert data to xarray."""
+        data = dct
         if not isinstance(data, dict):
-            raise TypeError("DictConverter.observed_data is not a dictionary")
+            raise TypeError("DictConverter.{} is not a dictionary".format(group))
         if self.dims is None:
             dims = {}
         else:
             dims = self.dims
-        observed_data = dict()
+        new_data = dict()
         for key, vals in data.items():
             vals = utils.one_de(vals)
             val_dims = dims.get(key)
             val_dims, coords = generate_dims_coords(
                 vals.shape, key, dims=val_dims, coords=self.coords
             )
-            observed_data[key] = xr.DataArray(vals, dims=val_dims, coords=coords)
-        return xr.Dataset(data_vars=observed_data, attrs=make_attrs(library=None))
+            new_data[key] = xr.DataArray(vals, dims=val_dims, coords=coords)
+        return xr.Dataset(data_vars=new_data, attrs=make_attrs(library=None))
+
+    @requires("observed_data")
+    def observed_data_to_xarray(self):
+        """Convert observed_data to xarray."""
+        return self.data_to_xarray(self.observed_data, group="observed_data")
 
     @requires("constant_data")
     def constant_data_to_xarray(self):
         """Convert constant_data to xarray."""
-        data = self.constant_data
-        if not isinstance(data, dict):
-            raise TypeError("DictConverter.constant_data is not a dictionary")
-        if self.dims is None:
-            dims = {}
-        else:
-            dims = self.dims
-        constant_data = dict()
-        for key, vals in data.items():
-            vals = utils.one_de(vals)
-            val_dims = dims.get(key)
-            val_dims, coords = generate_dims_coords(
-                vals.shape, key, dims=val_dims, coords=self.coords
-            )
-            constant_data[key] = xr.DataArray(vals, dims=val_dims, coords=coords)
-        return xr.Dataset(data_vars=constant_data, attrs=make_attrs(library=None))
+        return self.data_to_xarray(self.constant_data, group="constant_data")
+
+    @requires("predictions_constant_data")
+    def predictions_constant_data_to_xarray(self):
+        """Convert predictions_constant_data to xarray."""
+        return self.data_to_xarray(
+            self.predictions_constant_data, group="predictions_constant_data"
+        )
 
     def to_inference_data(self):
         """Convert all available data to an InferenceData object.
@@ -168,11 +177,13 @@ def to_inference_data(self):
                 "sample_stats": self.sample_stats_to_xarray(),
                 "log_likelihood": self.log_likelihood_to_xarray(),
                 "posterior_predictive": self.posterior_predictive_to_xarray(),
+                "predictions": self.predictions_to_xarray(),
                 "prior": self.prior_to_xarray(),
                 "sample_stats_prior": self.sample_stats_prior_to_xarray(),
                 "prior_predictive": self.prior_predictive_to_xarray(),
                 "observed_data": self.observed_data_to_xarray(),
                 "constant_data": self.constant_data_to_xarray(),
+                "predictions_constant_data": self.predictions_constant_data_to_xarray(),
             }
         )
 
@@ -182,13 +193,15 @@ def from_dict(
     posterior=None,
     *,
     posterior_predictive=None,
+    predictions=None,
     sample_stats=None,
     log_likelihood=None,
     prior=None,
     prior_predictive=None,
     sample_stats_prior=None,
     observed_data=None,
     constant_data=None,
+    predictions_constant_data=None,
     coords=None,
     dims=None
 ):
@@ -198,13 +211,15 @@ def from_dict(
     ----------
     posterior : dict
     posterior_predictive : dict
+    predictions: dict
     sample_stats : dict
     log_likelihood : dict
         For stats functions, log likelihood data should be stored here.
     prior : dict
     prior_predictive : dict
     observed_data : dict
     constant_data : dict
+    predictions_constant_data: dict
     coords : dict[str, iterable]
         A dictionary containing the values that are used as index. The key
         is the name of the dimension, the values are the index values.
@@ -218,13 +233,15 @@ def from_dict(
     return DictConverter(
         posterior=posterior,
         posterior_predictive=posterior_predictive,
+        predictions=predictions,
         sample_stats=sample_stats,
         log_likelihood=log_likelihood,
         prior=prior,
         prior_predictive=prior_predictive,
         sample_stats_prior=sample_stats_prior,
         observed_data=observed_data,
         constant_data=constant_data,
+        predictions_constant_data=predictions_constant_data,
         coords=coords,
         dims=dims,
     ).to_inference_data()