[python] Fix continued train by reusing the same dataset (#2906)

* fix

* fix return

* fix test

* fix test

* fix predictor is none

* Apply suggestions from code review

* Update basic.py

* Update basic.py

* Apply suggestions from code review
Co-Authored-By: Nikita Titov <nekit94-08@mail.ru>
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>

python-package/lightgbm/basic.py

@@ -465,11 +465,7 @@ class _InnerPredictor(object):
                 self.handle,
                 ctypes.byref(out_num_class)))
             self.num_class = out_num_class.value
-            out_num_iterations = ctypes.c_int(0)
-            _safe_call(_LIB.LGBM_BoosterGetCurrentIteration(
-                self.handle,
-                ctypes.byref(out_num_iterations)))
-            self.num_total_iteration = out_num_iterations.value
+            self.num_total_iteration = self.current_iteration()
             self.pandas_categorical = None
         else:
             raise TypeError('Need model_file or booster_handle to create a predictor')
@@ -726,6 +722,20 @@ class _InnerPredictor(object):
             raise ValueError("Wrong length for predict results")
         return preds, nrow
 
+    def current_iteration(self):
+        """Get the index of the current iteration.
+
+        Returns
+        -------
+        cur_iter : int
+            The index of the current iteration.
+        """
+        out_cur_iter = ctypes.c_int(0)
+        _safe_call(_LIB.LGBM_BoosterGetCurrentIteration(
+            self.handle,
+            ctypes.byref(out_cur_iter)))
+        return out_cur_iter.value
+
 
 class Dataset(object):
     """Dataset in LightGBM."""
@@ -842,27 +852,32 @@ class Dataset(object):
         if isinstance(data, string_type):
             # check data has header or not
             data_has_header = any(self.params.get(alias, False) for alias in _ConfigAliases.get("header"))
-        init_score = predictor.predict(data,
-                                       raw_score=True,
-                                       data_has_header=data_has_header,
-                                       is_reshape=False)
         num_data = self.num_data()
-        if used_indices is not None:
-            assert not self.need_slice
-            if isinstance(data, string_type):
-                sub_init_score = np.zeros(num_data * predictor.num_class, dtype=np.float32)
-                assert num_data == len(used_indices)
-                for i in range_(len(used_indices)):
+        if predictor is not None:
+            init_score = predictor.predict(data,
+                                           raw_score=True,
+                                           data_has_header=data_has_header,
+                                           is_reshape=False)
+            if used_indices is not None:
+                assert not self.need_slice
+                if isinstance(data, string_type):
+                    sub_init_score = np.zeros(num_data * predictor.num_class, dtype=np.float32)
+                    assert num_data == len(used_indices)
+                    for i in range_(len(used_indices)):
+                        for j in range_(predictor.num_class):
+                            sub_init_score[i * predictor.num_class + j] = init_score[used_indices[i] * predictor.num_class + j]
+                    init_score = sub_init_score
+            if predictor.num_class > 1:
+                # need to regroup init_score
+                new_init_score = np.zeros(init_score.size, dtype=np.float32)
+                for i in range_(num_data):
                     for j in range_(predictor.num_class):
-                        sub_init_score[i * predictor.num_class + j] = init_score[used_indices[i] * predictor.num_class + j]
-                init_score = sub_init_score
-        if predictor.num_class > 1:
-            # need to regroup init_score
-            new_init_score = np.zeros(init_score.size, dtype=np.float32)
-            for i in range_(num_data):
-                for j in range_(predictor.num_class):
-                    new_init_score[j * num_data + i] = init_score[i * predictor.num_class + j]
-            init_score = new_init_score
+                        new_init_score[j * num_data + i] = init_score[i * predictor.num_class + j]
+                init_score = new_init_score
+        elif self.init_score is not None:
+            init_score = np.zeros(self.init_score.shape, dtype=np.float32)
+        else:
+            return self
         self.set_init_score(init_score)
 
     def _lazy_init(self, data, label=None, reference=None,
@@ -1381,16 +1396,20 @@ class Dataset(object):
         It is not recommended for user to call this function.
         Please use init_model argument in engine.train() or engine.cv() instead.
         """
-        if predictor is self._predictor:
+        if predictor is self._predictor and (predictor is None or predictor.current_iteration() == self._predictor.current_iteration()):
             return self
-        if self.data is not None or (self.used_indices is not None
-                                     and self.reference is not None
-                                     and self.reference.data is not None):
+        if self.handle is None:
             self._predictor = predictor
-            return self._free_handle()
+        elif self.data is not None:
+            self._predictor = predictor
+            self._set_init_score_by_predictor(self._predictor, self.data)
+        elif self.used_indices is not None and self.reference is not None and self.reference.data is not None:
+            self._predictor = predictor
+            self._set_init_score_by_predictor(self._predictor, self.reference.data, self.used_indices)
         else:
             raise LightGBMError("Cannot set predictor after freed raw data, "
                                 "set free_raw_data=False when construct Dataset to avoid this.")
+        return self
 
     def set_reference(self, reference):
         """Set reference Dataset.

tests/python_package_test/test_engine.py

@@ -613,6 +613,19 @@ class TestEngine(unittest.TestCase):
         np.testing.assert_allclose(evals_result['valid_0']['l1'], evals_result['valid_0']['custom_mae'])
         os.remove(model_name)
 
+    def test_continue_train_reused_dataset(self):
+        X, y = load_boston(True)
+        params = {
+            'objective': 'regression',
+            'verbose': -1
+        }
+        lgb_train = lgb.Dataset(X, y, free_raw_data=False)
+        init_gbm = lgb.train(params, lgb_train, num_boost_round=5)
+        init_gbm_2 = lgb.train(params, lgb_train, num_boost_round=5, init_model=init_gbm)
+        init_gbm_3 = lgb.train(params, lgb_train, num_boost_round=5, init_model=init_gbm_2)
+        gbm = lgb.train(params, lgb_train, num_boost_round=5, init_model=init_gbm_3)
+        self.assertEqual(gbm.current_iteration(), 20)
+
     def test_continue_train_dart(self):
         X, y = load_boston(True)
         X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=42)