rename “target_order” argument to “targets”

eli5/explain.py

@@ -36,7 +36,7 @@ def explain_weights(estimator, **kwargs):
 
         This argument may be supported or not, depending on estimator type.
 
-    target_order : list, optional
+    targets : list, optional
         Order of class/target names to show. This argument can be also used
         to show information only for a subset of classes. It should be a list
         of class / target names which match either names provided by
@@ -102,7 +102,7 @@ def explain_prediction(estimator, doc, **kwargs):
 
         This argument may be supported or not, depending on estimator type.
 
-    target_order : list, optional
+    targets : list, optional
         Order of class/target names to show. This argument can be also used
         to show information only for a subset of classes. It should be a list
         of class / target names which match either names provided by

eli5/lightning.py

@@ -18,7 +18,7 @@
 @explain_weights.register(BaseEstimator)
 @singledispatch
 def explain_weights_lightning(estimator, vec=None, top=20, target_names=None,
-                              target_order=None, feature_names=None,
+                              targets=None, feature_names=None,
                               coef_scale=None):
     """ Return an explanation of a lightning estimator weights """
     return {
@@ -30,7 +30,7 @@ def explain_weights_lightning(estimator, vec=None, top=20, target_names=None,
 @explain_prediction.register(BaseEstimator)
 @singledispatch
 def explain_prediction_lightning(estimator, doc, vec=None, top=None,
-                                 target_names=None, target_order=None,
+                                 target_names=None, targets=None,
                                  feature_names=None, vectorized=False,
                                  coef_scale=None):
     """ Return an explanation of a lightning estimator predictions """

eli5/sklearn/explain_prediction.py

@@ -46,7 +46,7 @@ def explain_prediction_sklearn(estimator, doc,
                                vec=None,
                                top=None,
                                target_names=None,
-                               target_order=None,
+                               targets=None,
                                feature_names=None,
                                vectorized=False):
     """ Return an explanation of a scikit-learn estimator """
@@ -82,7 +82,7 @@ def explain_prediction_linear_classifier(clf, doc,
                                          vec=None,
                                          top=None,
                                          target_names=None,
-                                         target_order=None,
+                                         targets=None,
                                          feature_names=None,
                                          vectorized=False):
     """ Explain prediction of a linear classifier. """
@@ -113,7 +113,7 @@ def _weights(label_id):
         scores = _multiply(x, coef)
         return get_top_features(feature_names, scores, top)
 
-    display_names = get_display_names(clf.classes_, target_names, target_order)
+    display_names = get_display_names(clf.classes_, target_names, targets)
 
     if is_multiclass_classifier(clf):
         for label_id, label in display_names:
@@ -195,7 +195,7 @@ def explain_prediction_linear_regressor(reg, doc,
                                         vec=None,
                                         top=None,
                                         target_names=None,
-                                        target_order=None,
+                                        targets=None,
                                         feature_names=None,
                                         vectorized=False):
     """ Explain prediction of a linear regressor. """
@@ -221,7 +221,7 @@ def _weights(label_id):
         return get_top_features(feature_names, scores, top)
 
     names = get_default_target_names(reg)
-    display_names = get_display_names(names, target_names, target_order)
+    display_names = get_display_names(names, target_names, targets)
 
     if is_multitarget_regressor(reg):
         for label_id, label in display_names:

eli5/sklearn/explain_weights.py

@@ -97,7 +97,7 @@
 @singledispatch
 def explain_weights_sklearn(estimator, vec=None, top=_TOP,
                             target_names=None,
-                            target_order=None,
+                            targets=None,
                             feature_names=None, coef_scale=None,
                             feature_re=None):
     """ Return an explanation of an estimator """
@@ -117,7 +117,7 @@ def explain_linear_classifier_weights(clf,
                                       vec=None,
                                       top=_TOP,
                                       target_names=None,
-                                      target_order=None,
+                                      targets=None,
                                       feature_names=None,
                                       coef_scale=None,
                                       feature_re=None):
@@ -174,7 +174,7 @@ def _features(label_id):
             coef = coef[flt_indices]
         return get_top_features(feature_names, coef, top)
 
-    display_names = get_display_names(clf.classes_, target_names, target_order)
+    display_names = get_display_names(clf.classes_, target_names, targets)
     if is_multiclass_classifier(clf):
         return Explanation(
             targets=[
@@ -212,7 +212,7 @@ def explain_rf_feature_importance(clf,
                                   vec=None,
                                   top=_TOP,
                                   target_names=None,  # ignored
-                                  target_order=None,  # ignored
+                                  targets=None,       # ignored
                                   feature_names=None,
                                   feature_re=None):
     """
@@ -254,7 +254,7 @@ def explain_decision_tree(clf,
                           vec=None,
                           top=_TOP,
                           target_names=None,
-                          target_order=None,  # ignored
+                          targets=None,  # ignored
                           feature_names=None,
                           feature_re=None,
                           **export_graphviz_kwargs):
@@ -311,7 +311,7 @@ def explain_linear_regressor_weights(reg,
                                      vec=None,
                                      top=_TOP,
                                      target_names=None,
-                                     target_order=None,
+                                     targets=None,
                                      feature_names=None,
                                      coef_scale=None,
                                      feature_re=None):
@@ -368,7 +368,7 @@ def _features(target_id):
         return get_top_features(feature_names, coef, top)
 
     display_names = get_display_names(get_default_target_names(reg),
-                                      target_names, target_order)
+                                      target_names, targets)
     if is_multitarget_regressor(reg):
         return Explanation(
             targets=[

eli5/sklearn_crfsuite/explain_weights.py

@@ -16,7 +16,7 @@
 def explain_weights_sklearn_crfsuite(crf,
                                      top=20,
                                      target_names=None,
-                                     target_order=None,
+                                     targets=None,
                                      feature_re=None):
     """ Explain sklearn_crfsuite.CRF weights """
     feature_names = np.array(crf.attributes_)
@@ -34,10 +34,10 @@ def explain_weights_sklearn_crfsuite(crf,
     def _features(label_id):
         return get_top_features(state_feature_names, state_coef[label_id], top)
 
-    if target_order is None:
-        target_order = ner_default_target_order(crf.classes_)
+    if targets is None:
+        targets = sorted_for_ner(crf.classes_)
 
-    display_names = get_display_names(crf.classes_, target_names, target_order)
+    display_names = get_display_names(crf.classes_, target_names, targets)
     indices, names = zip(*display_names)
     transition_coef = filter_transition_coefs(transition_coef, indices)
 
@@ -105,11 +105,11 @@ def filter_transition_coefs(transition_coef, indices):
     return rows[:,indices]
 
 
-def ner_default_target_order(crf_classes):
+def sorted_for_ner(crf_classes):
     """
-    Return default order of labels for NER tasks:
+    Return labels sorted in a default order suitable for NER tasks:
 
-    >>> ner_default_target_order(['B-ORG', 'B-PER', 'O', 'I-PER'])
+    >>> sorted_for_ner(['B-ORG', 'B-PER', 'O', 'I-PER'])
     ['O', 'B-ORG', 'B-PER', 'I-PER']
     """
     def key(cls):

eli5/utils.py

@@ -42,13 +42,13 @@ def vstack(blocks, format=None, dtype=None):
         return np.vstack(blocks)
 
 
-def get_display_names(original_names=None, target_names=None, target_order=None):
+def get_display_names(original_names=None, target_names=None, targets=None):
     """
     Return a list of (class_id, display_name) tuples.
 
-    target_order can be written using both names from target_names and
-    from original_names:
-    >>> get_display_names(['x', 'y'], target_order=['y', 'X'],
+    ``targets`` can be written using both names from ``target_names` and
+    from ``original_names``:
+    >>> get_display_names(['x', 'y'], targets=['y', 'X'],
     ...                   target_names={'x': 'X'})
     [(1, 'y'), (0, 'X')]
 
@@ -57,15 +57,15 @@ def get_display_names(original_names=None, target_names=None, target_order=None)
     [(0, 'foo'), (1, 'bar')]
 
     Change order of labels:
-    >>> get_display_names(['x', 'y'], target_order=['y', 'x'])
+    >>> get_display_names(['x', 'y'], targets=['y', 'x'])
     [(1, 'y'), (0, 'x')]
 
     Provide display names, choose only a subset of labels:
-    >>> get_display_names([0, 2], target_names=['foo', 'bar'], target_order=[2])
+    >>> get_display_names([0, 2], target_names=['foo', 'bar'], targets=[2])
     [(1, 'bar')]
 
     target_names can be a dictionary with {old_name: new_name} labels:
-    >>> get_display_names(['x', 'y'], target_order=['y', 'x'],
+    >>> get_display_names(['x', 'y'], targets=['y', 'x'],
     ...                   target_names={'x': 'X'})
     [(1, 'y'), (0, 'X')]
 
@@ -89,11 +89,11 @@ def get_display_names(original_names=None, target_names=None, target_order=None)
     else:
         display_names = original_names
 
-    if target_order is None:
-        target_order = original_names
+    if targets is None:
+        targets = original_names
 
     class_indices = _get_value_indices(original_names, display_names,
-                                       target_order)
+                                       targets)
     names = [display_names[i] for i in class_indices]
     return list(zip(class_indices, names))
 

notebooks/sklearn-crfsuite.ipynb

@@ -6454,7 +6454,7 @@
     }
    ],
    "source": [
-    "explain_weights(crf, top=20, target_order=['O', 'B-ORG', 'I-ORG'])"
+    "explain_weights(crf, top=20, targets=['O', 'B-ORG', 'I-ORG'])"
    ]
   },
   {

tests/test_sklearn_crfsuite.py

@@ -59,20 +59,20 @@ def test_sklearn_crfsuite_feature_re(xseq, yseq):
         assert 'walk' not in expl
 
 
-@pytest.mark.parametrize(['target_order'], [
+@pytest.mark.parametrize(['targets'], [
     [['rainy', 'sunny']],
     [['sunny', 'rainy']],
 ])
-def test_sklearn_targets(xseq, yseq, target_order):
+def test_sklearn_targets(xseq, yseq, targets):
     crf = CRF(c1=0.0, c2=0.1, max_iterations=50)
     crf.fit([xseq], [yseq])
 
     res = explain_weights(crf,
                           target_names={'sunny': u'☀'},
-                          target_order=target_order)
+                          targets=targets)
     for expl in format_as_all(res, crf):
         assert u'☀' in expl
-        if target_order[0] == 'rainy':
+        if targets[0] == 'rainy':
             assert expl.index('rainy') < expl.index(u'☀')
         else:
             assert expl.index('rainy') > expl.index(u'☀')