[python] add sparsity support for new version of pandas and check Series for bad dtypes (#2318)

* reworked pandas dtypes mapper

* added tests

* added sparsity support for new version of pandas

* fixed tests for old pandas

* check pd.Series for bad dtypes as well

* enhanced tests

* fixed pylint

python-package/lightgbm/basic.py

@@ -13,7 +13,7 @@ from tempfile import NamedTemporaryFile
 import numpy as np
 import scipy.sparse
 
-from .compat import (PANDAS_INSTALLED, DataFrame, Series,
+from .compat import (PANDAS_INSTALLED, DataFrame, Series, is_dtype_sparse,
                      DataTable,
                      decode_string, string_type,
                      integer_types, numeric_types,
@@ -78,6 +78,11 @@ def list_to_1d_numpy(data, dtype=np.float32, name='list'):
     elif is_1d_list(data):
         return np.array(data, dtype=dtype, copy=False)
     elif isinstance(data, Series):
+        if _get_bad_pandas_dtypes([data.dtypes]):
+            raise ValueError('Series.dtypes must be int, float or bool')
+        if hasattr(data.values, 'values'):  # SparseArray
+            return data.values.values.astype(dtype)
+        else:
             return data.values.astype(dtype)
     else:
         raise TypeError("Wrong type({0}) for {1}.\n"
@@ -194,11 +199,6 @@ FIELD_TYPE_MAPPER = {"label": C_API_DTYPE_FLOAT32,
                      "feature_penalty": C_API_DTYPE_FLOAT64,
                      "monotone_constraints": C_API_DTYPE_INT8}
 
-PANDAS_DTYPE_MAPPER = {'int8': 'int', 'int16': 'int', 'int32': 'int',
-                       'int64': 'int', 'uint8': 'int', 'uint16': 'int',
-                       'uint32': 'int', 'uint64': 'int', 'bool': 'int',
-                       'float16': 'float', 'float32': 'float', 'float64': 'float'}
-
 
 def convert_from_sliced_object(data):
     """Fix the memory of multi-dimensional sliced object."""
@@ -252,6 +252,17 @@ def c_int_array(data):
     return (ptr_data, type_data, data)  # return `data` to avoid the temporary copy is freed
 
 
+def _get_bad_pandas_dtypes(dtypes):
+    pandas_dtype_mapper = {'int8': 'int', 'int16': 'int', 'int32': 'int',
+                           'int64': 'int', 'uint8': 'int', 'uint16': 'int',
+                           'uint32': 'int', 'uint64': 'int', 'bool': 'int',
+                           'float16': 'float', 'float32': 'float', 'float64': 'float'}
+    bad_indices = [i for i, dtype in enumerate(dtypes) if (dtype.name not in pandas_dtype_mapper
+                                                           and (not is_dtype_sparse(dtype)
+                                                                or dtype.subtype.name not in pandas_dtype_mapper))]
+    return bad_indices
+
+
 def _data_from_pandas(data, feature_name, categorical_feature, pandas_categorical):
     if isinstance(data, DataFrame):
         if len(data.shape) != 2 or data.shape[0] < 1:
@@ -280,13 +291,11 @@ def _data_from_pandas(data, feature_name, categorical_feature, pandas_categorica
                 categorical_feature = list(categorical_feature)
         if feature_name == 'auto':
             feature_name = list(data.columns)
-        data_dtypes = data.dtypes
-        if not all(dtype.name in PANDAS_DTYPE_MAPPER for dtype in data_dtypes):
-            bad_fields = [data.columns[i] for i, dtype in
-                          enumerate(data_dtypes) if dtype.name not in PANDAS_DTYPE_MAPPER]
+        bad_indices = _get_bad_pandas_dtypes(data.dtypes)
+        if bad_indices:
             raise ValueError("DataFrame.dtypes for data must be int, float or bool.\n"
                              "Did not expect the data types in the following fields: "
-                             + ', '.join(bad_fields))
+                             + ', '.join(data.columns[bad_indices]))
         data = data.values.astype('float')
     else:
         if feature_name == 'auto':
@@ -300,8 +309,7 @@ def _label_from_pandas(label):
     if isinstance(label, DataFrame):
         if len(label.columns) > 1:
             raise ValueError('DataFrame for label cannot have multiple columns')
-        label_dtypes = label.dtypes
-        if not all(dtype.name in PANDAS_DTYPE_MAPPER for dtype in label_dtypes):
+        if _get_bad_pandas_dtypes(label.dtypes):
             raise ValueError('DataFrame.dtypes for label must be int, float or bool')
         label = label.values.astype('float').flatten()
     return label

python-package/lightgbm/compat.py

@@ -62,6 +62,7 @@ def json_default_with_numpy(obj):
 """pandas"""
 try:
     from pandas import Series, DataFrame
+    from pandas.api.types import is_sparse as is_dtype_sparse
     PANDAS_INSTALLED = True
 except ImportError:
     PANDAS_INSTALLED = False
@@ -76,6 +77,8 @@ except ImportError:
 
         pass
 
+    is_dtype_sparse = None
+
 """matplotlib"""
 try:
     import matplotlib

tests/python_package_test/test_engine.py

@@ -719,6 +719,32 @@ class TestEngine(unittest.TestCase):
         self.assertListEqual(gbm6.pandas_categorical, cat_values)
         self.assertListEqual(gbm7.pandas_categorical, cat_values)
 
+    @unittest.skipIf(not lgb.compat.PANDAS_INSTALLED, 'pandas is not installed')
+    def test_pandas_sparse(self):
+        import pandas as pd
+        X = pd.DataFrame({"A": pd.SparseArray(np.random.permutation([0, 1, 2] * 100)),
+                          "B": pd.SparseArray(np.random.permutation([0.0, 0.1, 0.2, -0.1, 0.2] * 60)),
+                          "C": pd.SparseArray(np.random.permutation([True, False] * 150))})
+        y = pd.Series(pd.SparseArray(np.random.permutation([0, 1] * 150)))
+        X_test = pd.DataFrame({"A": pd.SparseArray(np.random.permutation([0, 2] * 30)),
+                               "B": pd.SparseArray(np.random.permutation([0.0, 0.1, 0.2, -0.1] * 15)),
+                               "C": pd.SparseArray(np.random.permutation([True, False] * 30))})
+        if pd.__version__ >= '0.24.0':
+            for dtype in pd.concat([X.dtypes, X_test.dtypes, pd.Series(y.dtypes)]):
+                self.assertTrue(pd.api.types.is_sparse(dtype))
+        params = {
+            'objective': 'binary',
+            'verbose': -1
+        }
+        lgb_train = lgb.Dataset(X, y)
+        gbm = lgb.train(params, lgb_train, num_boost_round=10)
+        pred_sparse = gbm.predict(X_test, raw_score=True)
+        if hasattr(X_test, 'sparse'):
+            pred_dense = gbm.predict(X_test.sparse.to_dense(), raw_score=True)
+        else:
+            pred_dense = gbm.predict(X_test.to_dense(), raw_score=True)
+        np.testing.assert_allclose(pred_sparse, pred_dense)
+
     def test_reference_chain(self):
         X = np.random.normal(size=(100, 2))
         y = np.random.normal(size=100)

tests/python_package_test/test_sklearn.py

@@ -277,6 +277,27 @@ class TestSklearn(unittest.TestCase):
         self.assertListEqual(gbm5.booster_.pandas_categorical, cat_values)
         self.assertListEqual(gbm6.booster_.pandas_categorical, cat_values)
 
+    @unittest.skipIf(not lgb.compat.PANDAS_INSTALLED, 'pandas is not installed')
+    def test_pandas_sparse(self):
+        import pandas as pd
+        X = pd.DataFrame({"A": pd.SparseArray(np.random.permutation([0, 1, 2] * 100)),
+                          "B": pd.SparseArray(np.random.permutation([0.0, 0.1, 0.2, -0.1, 0.2] * 60)),
+                          "C": pd.SparseArray(np.random.permutation([True, False] * 150))})
+        y = pd.Series(pd.SparseArray(np.random.permutation([0, 1] * 150)))
+        X_test = pd.DataFrame({"A": pd.SparseArray(np.random.permutation([0, 2] * 30)),
+                               "B": pd.SparseArray(np.random.permutation([0.0, 0.1, 0.2, -0.1] * 15)),
+                               "C": pd.SparseArray(np.random.permutation([True, False] * 30))})
+        if pd.__version__ >= '0.24.0':
+            for dtype in pd.concat([X.dtypes, X_test.dtypes, pd.Series(y.dtypes)]):
+                self.assertTrue(pd.api.types.is_sparse(dtype))
+        gbm = lgb.sklearn.LGBMClassifier().fit(X, y)
+        pred_sparse = gbm.predict(X_test, raw_score=True)
+        if hasattr(X_test, 'sparse'):
+            pred_dense = gbm.predict(X_test.sparse.to_dense(), raw_score=True)
+        else:
+            pred_dense = gbm.predict(X_test.to_dense(), raw_score=True)
+        np.testing.assert_allclose(pred_sparse, pred_dense)
+
     def test_predict(self):
         iris = load_iris()
         X_train, X_test, y_train, y_test = train_test_split(iris.data, iris.target,