Merge branch 'master' into fix-clarify-mm

Author: zicanl-amazon

pyproject.toml

@@ -40,7 +40,7 @@ dependencies = [
   "importlib-metadata>=1.4.0,<7.0",
   "jsonschema",
   "numpy==1.26.4",
-  "omegaconf>=2.2,<=2.3",
+  "omegaconf>=2.2,<3",
   "packaging>=23.0,<25",
   "pandas",
   "pathos",

src/sagemaker/feature_store/feature_processor/_input_offset_parser.py

@@ -72,14 +72,16 @@ def get_offset_datetime(self, offset: Optional[str]) -> datetime:
 
         return self.now + offset_td
 
-    def get_offset_date_year_month_day_hour(self, offset: Optional[str]) -> Tuple[str]:
+    def get_offset_date_year_month_day_hour(
+        self, offset: Optional[str]
+    ) -> Tuple[str, str, str, str]:
         """Get the year, month, day and hour based on offset diff.
 
         Args:
             offset (Optional[str]): Offset that is used for target date calcluation.
 
         Returns:
-            Tuple[str]: A tuple that consists of extracted year, month, day, hour from offset date.
+            Tuple[str, str, str, str]: A tuple that consists of extracted year, month, day, hour from offset date.
         """
         if offset is None:
             return (None, None, None, None)

src/sagemaker/modules/train/container_drivers/common/utils.py

@@ -124,8 +124,6 @@ def safe_deserialize(data: Any) -> Any:
 
     This function handles the following cases:
     1. If `data` is not a string, it returns the input as-is.
-    2. If `data` is a string and matches common boolean values ("true" or "false"),
-    it returns the corresponding boolean value (True or False).
     3. If `data` is a JSON-encoded string, it attempts to deserialize it using `json.loads()`.
     4. If `data` is a string but cannot be decoded as JSON, it returns the original string.
 
@@ -134,13 +132,6 @@ def safe_deserialize(data: Any) -> Any:
     """
     if not isinstance(data, str):
         return data
-
-    lower_data = data.lower()
-    if lower_data in ["true"]:
-        return True
-    if lower_data in ["false"]:
-        return False
-
     try:
         return json.loads(data)
     except json.JSONDecodeError:

src/sagemaker/modules/train/model_trainer.py

@@ -865,7 +865,7 @@ def _prepare_train_script(
             working_dir = f"cd {SM_CODE_CONTAINER_PATH} \n"
             if source_code.source_dir.endswith(".tar.gz"):
                 tarfile_name = os.path.basename(source_code.source_dir)
-                working_dir += f"tar --strip-components=1 -xzf {tarfile_name} \n"
+                working_dir += f"tar -xzf {tarfile_name} \n"
 
         if base_command:
             execute_driver = EXECUTE_BASE_COMMANDS.format(base_command=base_command)

tests/data/modules/script_mode/code.tar.gz

@@ -59,8 +59,14 @@ def test_safe_deserialize_not_a_string():
 def test_safe_deserialize_boolean_strings():
     assert safe_deserialize("true") is True
     assert safe_deserialize("false") is False
-    assert safe_deserialize("True") is True
-    assert safe_deserialize("False") is False
+
+    # The below are not valid JSON booleans
+    assert safe_deserialize("True") == "True"
+    assert safe_deserialize("False") == "False"
+    assert safe_deserialize("TRUE") == "TRUE"
+    assert safe_deserialize("FALSE") == "FALSE"
+    assert safe_deserialize("tRuE") == "tRuE"
+    assert safe_deserialize("fAlSe") == "fAlSe"
 
 
 def test_safe_deserialize_valid_json_string():