breaking: Move _CsvDeserializer to sagemaker.deserializers and rename to CSVDeserializer

src/sagemaker/deserializers.py

@@ -13,6 +13,8 @@
 """Implements methods for deserializing data returned from an inference endpoint."""
 from __future__ import absolute_import
 
+import csv
+
 import abc
 
 
@@ -60,3 +62,34 @@ def deserialize(self, data, content_type):
             return data.read()
         finally:
             data.close()
+
+
+class CSVDeserializer(BaseDeserializer):
+    """Deserialize a stream of bytes into a list of lists."""
+
+    ACCEPT = "test/csv"
+
+    def __init__(self, encoding="UTF-8"):
+        """Initialize the string encoding.
+
+        Args:
+            encoding (str): The string encoding to use (default: "UTF-8").
+        """
+        self.encoding = encoding
+
+    def deserialize(self, data, content_type):
+        """Deserialize data from an inference endpoint into a list of lists.
+
+        Args:
+            data (botocore.response.StreamingBody): Data to be deserialized.
+            content_type (str): The MIME type of the data.
+
+        Returns:
+            list: The data deserialized into a list of lists representing the
+                contents of a CSV file.
+        """
+        try:
+            decoded_string = data.read().decode(self.encoding)
+            return list(csv.reader(decoded_string.splitlines()))
+        finally:
+            data.close()

src/sagemaker/predictor.py

@@ -597,32 +597,6 @@ def _row_to_csv(obj):
     return ",".join(obj)
 
 
-class _CsvDeserializer(object):
-    """Placeholder docstring"""
-
-    def __init__(self, encoding="utf-8"):
-        """
-        Args:
-            encoding:
-        """
-        self.accept = CONTENT_TYPE_CSV
-        self.encoding = encoding
-
-    def __call__(self, stream, content_type):
-        """
-        Args:
-            stream:
-            content_type:
-        """
-        try:
-            return list(csv.reader(stream.read().decode(self.encoding).splitlines()))
-        finally:
-            stream.close()
-
-
-csv_deserializer = _CsvDeserializer()
-
-
 class StringDeserializer(object):
     """Return the response as a decoded string.
 

src/sagemaker/xgboost/model.py

@@ -16,10 +16,11 @@
 import logging
 
 import sagemaker
+from sagemaker.deserializers import CSVDeserializer
 from sagemaker.fw_utils import model_code_key_prefix
 from sagemaker.fw_registry import default_framework_uri
 from sagemaker.model import FrameworkModel, MODEL_SERVER_WORKERS_PARAM_NAME
-from sagemaker.predictor import Predictor, npy_serializer, csv_deserializer
+from sagemaker.predictor import Predictor, npy_serializer
 from sagemaker.xgboost.defaults import XGBOOST_NAME
 
 logger = logging.getLogger("sagemaker")
@@ -42,7 +43,7 @@ def __init__(self, endpoint_name, sagemaker_session=None):
                 chain.
         """
         super(XGBoostPredictor, self).__init__(
-            endpoint_name, sagemaker_session, npy_serializer, csv_deserializer
+            endpoint_name, sagemaker_session, npy_serializer, CSVDeserializer()
         )
 
 

tests/unit/sagemaker/test_deserializers.py

@@ -14,7 +14,9 @@
 
 import io
 
-from sagemaker.deserializers import BytesDeserializer
+import pytest
+
+from sagemaker.deserializers import BytesDeserializer, CSVDeserializer
 
 
 def test_bytes_deserializer():
@@ -23,3 +25,28 @@ def test_bytes_deserializer():
     result = deserializer.deserialize(io.BytesIO(b"[1, 2, 3]"), "application/json")
 
     assert result == b"[1, 2, 3]"
+
+
+@pytest.fixture
+def csv_deserializer():
+    return CSVDeserializer()
+
+
+def test_csv_deserializer_single_element(csv_deserializer):
+    result = csv_deserializer.deserialize(io.BytesIO(b"1"), "text/csv")
+    assert result == [["1"]]
+
+
+def test_csv_deserializer_array(csv_deserializer):
+    result = csv_deserializer.deserialize(io.BytesIO(b"1,2,3"), "text/csv")
+    assert result == [["1", "2", "3"]]
+
+
+def test_csv_deserializer_2dimensional(csv_deserializer):
+    result = csv_deserializer.deserialize(io.BytesIO(b"1,2,3\n3,4,5"), "text/csv")
+    assert result == [["1", "2", "3"], ["3", "4", "5"]]
+
+
+def test_csv_deserializer_posix_compliant(csv_deserializer):
+    result = csv_deserializer.deserialize(io.BytesIO(b"1,2,3\n3,4,5\n"), "text/csv")
+    assert result == [["1", "2", "3"], ["3", "4", "5"]]

tests/unit/test_predictor.py

@@ -25,7 +25,6 @@
     json_serializer,
     json_deserializer,
     csv_serializer,
-    csv_deserializer,
     StringDeserializer,
     StreamDeserializer,
     numpy_deserializer,
@@ -150,21 +149,6 @@ def test_csv_serializer_csv_reader():
         assert result == validation_data
 
 
-def test_csv_deserializer_single_element():
-    result = csv_deserializer(io.BytesIO(b"1"), "text/csv")
-    assert result == [["1"]]
-
-
-def test_csv_deserializer_array():
-    result = csv_deserializer(io.BytesIO(b"1,2,3"), "text/csv")
-    assert result == [["1", "2", "3"]]
-
-
-def test_csv_deserializer_2dimensional():
-    result = csv_deserializer(io.BytesIO(b"1,2,3\n3,4,5"), "text/csv")
-    assert result == [["1", "2", "3"], ["3", "4", "5"]]
-
-
 def test_json_deserializer_array():
     result = json_deserializer(io.BytesIO(b"[1, 2, 3]"), "application/json")