[python] add return_cvbooster flag to cv func and publish _CVBooster (#…

…283,#2105,#1445) (#3204)

* [python] add return_cvbooster flag to cv function and rename _CVBooster to make public (#283,#2105)

* [python] Reduce expected metric of unit testing

* [docs] add the CVBooster to the documentation

* [python] reflect the review comments

- Add some clarifications to the documentation
- Rename CVBooster.append to make private
- Decrease iteration rounds of testing to save CI time
- Use CVBooster as root member of lgb

* [python] add more checks in testing for cv

Co-authored-by: Nikita Titov <[email protected]>

* [python] add docstring for instance attributes of CVBooster

Co-authored-by: Nikita Titov <[email protected]>

* [python] fix docstring

Co-authored-by: Nikita Titov <[email protected]>

Co-authored-by: Nikita Titov <[email protected]>

docs/Python-API.rst

@@ -11,6 +11,7 @@ Data Structure API
 
     Dataset
     Booster
+    CVBooster
 
 Training API
 ------------

python-package/lightgbm/__init__.py

@@ -8,7 +8,7 @@
 from .basic import Booster, Dataset
 from .callback import (early_stopping, print_evaluation, record_evaluation,
                        reset_parameter)
-from .engine import cv, train
+from .engine import cv, train, CVBooster
 
 import os
 
@@ -29,7 +29,7 @@
     with open(os.path.join(dir_path, 'VERSION.txt')) as version_file:
         __version__ = version_file.read().strip()
 
-__all__ = ['Dataset', 'Booster',
+__all__ = ['Dataset', 'Booster', 'CVBooster',
            'train', 'cv',
            'LGBMModel', 'LGBMRegressor', 'LGBMClassifier', 'LGBMRanker',
            'print_evaluation', 'record_evaluation', 'reset_parameter', 'early_stopping',

python-package/lightgbm/engine.py

@@ -276,19 +276,35 @@ def train(params, train_set, num_boost_round=100,
     return booster
 
 
-class _CVBooster(object):
-    """Auxiliary data struct to hold all boosters of CV."""
+class CVBooster(object):
+    """CVBooster in LightGBM.
+
+    Auxiliary data structure to hold and redirect all boosters of ``cv`` function.
+    This class has the same methods as Booster class.
+    All method calls are actually performed for underlying Boosters and then all returned results are returned in a list.
+
+    Attributes
+    ----------
+    boosters : list of Booster
+        The list of underlying fitted models.
+    best_iteration : int
+        The best iteration of fitted model.
+    """
 
     def __init__(self):
+        """Initialize the CVBooster.
+
+        Generally, no need to instantiate manually.
+        """
         self.boosters = []
         self.best_iteration = -1
 
-    def append(self, booster):
-        """Add a booster to _CVBooster."""
+    def _append(self, booster):
+        """Add a booster to CVBooster."""
         self.boosters.append(booster)
 
     def __getattr__(self, name):
-        """Redirect methods call of _CVBooster."""
+        """Redirect methods call of CVBooster."""
         def handler_function(*args, **kwargs):
             """Call methods with each booster, and concatenate their results."""
             ret = []
@@ -341,7 +357,7 @@ def _make_n_folds(full_data, folds, nfold, params, seed, fpreproc=None, stratifi
             train_id = [np.concatenate([test_id[i] for i in range_(nfold) if k != i]) for k in range_(nfold)]
             folds = zip_(train_id, test_id)
 
-    ret = _CVBooster()
+    ret = CVBooster()
     for train_idx, test_idx in folds:
         train_set = full_data.subset(sorted(train_idx))
         valid_set = full_data.subset(sorted(test_idx))
@@ -354,7 +370,7 @@ def _make_n_folds(full_data, folds, nfold, params, seed, fpreproc=None, stratifi
         if eval_train_metric:
             cvbooster.add_valid(train_set, 'train')
         cvbooster.add_valid(valid_set, 'valid')
-        ret.append(cvbooster)
+        ret._append(cvbooster)
     return ret
 
 
@@ -380,7 +396,8 @@ def cv(params, train_set, num_boost_round=100,
        feature_name='auto', categorical_feature='auto',
        early_stopping_rounds=None, fpreproc=None,
        verbose_eval=None, show_stdv=True, seed=0,
-       callbacks=None, eval_train_metric=False):
+       callbacks=None, eval_train_metric=False,
+       return_cvbooster=False):
     """Perform the cross-validation with given paramaters.
 
     Parameters
@@ -486,6 +503,8 @@ def cv(params, train_set, num_boost_round=100,
     eval_train_metric : bool, optional (default=False)
         Whether to display the train metric in progress.
         The score of the metric is calculated again after each training step, so there is some impact on performance.
+    return_cvbooster : bool, optional (default=False)
+        Whether to return Booster models trained on each fold through ``CVBooster``.
 
     Returns
     -------
@@ -495,6 +514,7 @@ def cv(params, train_set, num_boost_round=100,
         {'metric1-mean': [values], 'metric1-stdv': [values],
         'metric2-mean': [values], 'metric2-stdv': [values],
         ...}.
+        If ``return_cvbooster=True``, also returns trained boosters via ``cvbooster`` key.
     """
     if not isinstance(train_set, Dataset):
         raise TypeError("Training only accepts Dataset object")
@@ -586,4 +606,8 @@ def cv(params, train_set, num_boost_round=100,
             for k in results:
                 results[k] = results[k][:cvfolds.best_iteration]
             break
+
+    if return_cvbooster:
+        results['cvbooster'] = cvfolds
+
     return dict(results)

tests/python_package_test/test_engine.py

@@ -735,6 +735,50 @@ def test_cv(self):
                                    verbose_eval=False)
         np.testing.assert_allclose(cv_res_lambda['ndcg@3-mean'], cv_res_lambda_obj['ndcg@3-mean'])
 
+    def test_cvbooster(self):
+        X, y = load_breast_cancer(True)
+        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=42)
+        params = {
+            'objective': 'binary',
+            'metric': 'binary_logloss',
+            'verbose': -1,
+        }
+        lgb_train = lgb.Dataset(X_train, y_train)
+        # with early stopping
+        cv_res = lgb.cv(params, lgb_train,
+                        num_boost_round=25,
+                        early_stopping_rounds=5,
+                        verbose_eval=False,
+                        nfold=3,
+                        return_cvbooster=True)
+        self.assertIn('cvbooster', cv_res)
+        cvb = cv_res['cvbooster']
+        self.assertIsInstance(cvb, lgb.CVBooster)
+        self.assertIsInstance(cvb.boosters, list)
+        self.assertEqual(len(cvb.boosters), 3)
+        self.assertTrue(all(isinstance(bst, lgb.Booster) for bst in cvb.boosters))
+        self.assertGreater(cvb.best_iteration, 0)
+        # predict by each fold booster
+        preds = cvb.predict(X_test, num_iteration=cvb.best_iteration)
+        self.assertIsInstance(preds, list)
+        self.assertEqual(len(preds), 3)
+        # fold averaging
+        avg_pred = np.mean(preds, axis=0)
+        ret = log_loss(y_test, avg_pred)
+        self.assertLess(ret, 0.13)
+        # without early stopping
+        cv_res = lgb.cv(params, lgb_train,
+                        num_boost_round=20,
+                        verbose_eval=False,
+                        nfold=3,
+                        return_cvbooster=True)
+        cvb = cv_res['cvbooster']
+        self.assertEqual(cvb.best_iteration, -1)
+        preds = cvb.predict(X_test)
+        avg_pred = np.mean(preds, axis=0)
+        ret = log_loss(y_test, avg_pred)
+        self.assertLess(ret, 0.15)
+
     def test_feature_name(self):
         X_train, y_train = load_boston(True)
         params = {'verbose': -1}