Updates for release 0.2.0

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,4 +1,4 @@
 Make sure you have:
-* added the necessary updates to `doc/content/whatsnew.rst`
-* added documentation for new features
-* updated actibelt/version.py if you are preparing a new release
+- [ ] added the necessary updates to `doc/content/whatsnew.rst`
+- [ ] added documentation for new features
+- [ ] updated sklearn_xarray/version.py if you are preparing a new release

doc/content/api/common.rst

@@ -2,4 +2,4 @@ Top-level functions and classes
 ===============================
 
 .. automodule:: sklearn_xarray
-   :members:
+   :members:

doc/content/whatsnew.rst

@@ -1,19 +1,36 @@
 What's New
 ==========
 
-.. v0.2.0
-.. Breaking changes
-.. The ``dataset`` and ``dataarray`` modules have been removed. Wrappers have
-   to be directly imported from ``sklearn_xarray``.
+
+v0.2.0 (April 9, 2018)
+----------------------
+
+Breaking changes
+~~~~~~~~~~~~~~~~
+- ``wrap`` now returns a decorated ``EstimatorWrapper`` instead of an
+  estimator-specific wrapper class.
+- Removed the ``common.decorators`` module, because the decorated
+  estimators could not be pickled and therefore didn't pass the usual sklearn
+  estimator checks.
+- Removed the ``dataset`` and ``dataarray`` modules. Wrappers have
+  to be directly imported from ``sklearn_xarray``.
+- Removed the ``data`` module (now called ``datasets``).
+
+
+Enhancements
+~~~~~~~~~~~~
+
+- Added wrappers for ``fit_transform``, ``partial_fit``, ``predict_proba``,
+  ``predict_log_proba`` and ``decision_function``.
 
 
 v0.1.4 (March 15, 2018)
---------------------
+-----------------------
 
 Enhancements
 ~~~~~~~~~~~~
 
-- ``preprocessing.Tranposer`` now also accepts a subset of ``X.dims`` for the
+- ``preprocessing.Transposer`` now also accepts a subset of ``X.dims`` for the
   ``order`` parameter.
 - ``preprocessing.Splitter`` and ``preprocessing.Segmenter`` now accept an
   ``axis`` argument that specifies where to insert the new dimension.

doc/content/wrappers.rst

@@ -27,9 +27,9 @@ sklearn for use with a ``DataArray``::
     X = load_dummy_dataarray()
     Xt = wrap(StandardScaler()).fit_transform(X)
 
-The :py:func:`wrap` function will try to guess the type of the estimator it
-wraps and return an object with the corresponding methods for each type of
-estimator (e.g. ``predict`` for classifiers and regressors).
+The :py:func:`wrap` function will return an object with the corresponding
+methods for each type of estimator (e.g. ``predict`` for classifiers and
+regressors).
 
 .. note::
 

examples/plot_activity_recognition.py

@@ -27,7 +27,7 @@
 from sklearn_xarray import wrap, Target
 from sklearn_xarray.preprocessing import (Splitter, Sanitizer, Featurizer)
 from sklearn_xarray.model_selection import CrossValidatorWrapper
-from sklearn_xarray.data import load_wisdm_dataarray
+from sklearn_xarray.datasets import load_wisdm_dataarray
 
 from sklearn.preprocessing import StandardScaler, LabelEncoder
 from sklearn.decomposition import PCA

sklearn_xarray/__init__.py

@@ -1,9 +1,8 @@
 """ ``sklearn_xarray`` """
 
 from sklearn_xarray.common.wrappers import (
-    wrap, ClassifierWrapper, RegressorWrapper, TransformerWrapper)
-from sklearn_xarray.common.decorators import (
-    classifier, regressor, transformer)
+    wrap, EstimatorWrapper, ClassifierWrapper, RegressorWrapper,
+    TransformerWrapper)
 from sklearn_xarray.target import Target
 
 from sklearn_xarray.version import __version__, __release__
@@ -13,9 +12,7 @@
 
 __all__ = [
     'wrap',
-    'classifier',
-    'regressor',
-    'transformer',
+    'EstimatorWrapper',
     'ClassifierWrapper',
     'RegressorWrapper',
     'TransformerWrapper',