[python] remove `verbose` argument of `model_from_string()` method of Booster class (#4877)

python-package/lightgbm/basic.py

@@ -2619,7 +2619,7 @@ class Booster:
             self.__num_class = out_num_class.value
             self.pandas_categorical = _load_pandas_categorical(file_name=model_file)
         elif model_str is not None:
-            self.model_from_string(model_str, verbose="_silent_false")
+            self.model_from_string(model_str)
         else:
             raise TypeError('Need at least one training dataset or model file or model string '
                             'to create Booster instance')
@@ -3300,15 +3300,13 @@ class Booster:
             ctypes.c_int(end_iteration)))
         return self
 
-    def model_from_string(self, model_str, verbose='warn'):
+    def model_from_string(self, model_str):
         """Load Booster from a string.
 
         Parameters
         ----------
         model_str : str
             Model will be loaded from this string.
-        verbose : bool, optional (default=True)
-            Whether to print messages while loading model.
 
         Returns
         -------

python-package/lightgbm/engine.py

@@ -303,7 +303,7 @@ def train(
     for dataset_name, eval_name, score, _ in evaluation_result_list:
         booster.best_score[dataset_name][eval_name] = score
     if not keep_training_booster:
-        booster.model_from_string(booster.model_to_string(), verbose='_silent_false').free_dataset()
+        booster.model_from_string(booster.model_to_string()).free_dataset()
     return booster
 
 

tests/python_package_test/test_engine.py

@@ -1023,7 +1023,7 @@ def test_pandas_categorical():
     gbm4 = lgb.Booster(model_file='categorical.model')
     pred4 = gbm4.predict(X_test)
     model_str = gbm4.model_to_string()
-    gbm4.model_from_string(model_str, False)
+    gbm4.model_from_string(model_str)
     pred5 = gbm4.predict(X_test)
     gbm5 = lgb.Booster(model_str=model_str)
     pred6 = gbm5.predict(X_test)
@@ -2146,7 +2146,7 @@ def test_model_size():
         num_end_spaces = 2**31 - one_tree_size * total_trees
         new_model_str = f"{before_tree_sizes}\n\n{trees}{more_trees}{after_trees}{'':{num_end_spaces}}"
         assert len(new_model_str) > 2**31
-        bst.model_from_string(new_model_str, verbose=False)
+        bst.model_from_string(new_model_str)
         assert bst.num_trees() == total_trees
         y_pred_new = bst.predict(X, num_iteration=2)
         np.testing.assert_allclose(y_pred, y_pred_new)