fix(core): _should_stream respects streaming param

Author: mdrxy

libs/core/langchain_core/language_models/chat_models.py

@@ -451,6 +451,10 @@ def _should_stream(
         if self.disable_streaming == "tool_calling" and kwargs.get("tools"):
             return False
 
+        # Check if streaming has been disabled via the streaming parameter.
+        if hasattr(self, "streaming") and not self.streaming:
+            return False
+
         # Check if a runtime streaming flag has been passed in.
         if "stream" in kwargs:
             return kwargs["stream"]

libs/partners/openai/tests/unit_tests/chat_models/test_base.py

@@ -350,7 +350,7 @@ def mock_glm4_completion() -> list:
 
 async def test_glm4_astream(mock_glm4_completion: list) -> None:
     llm_name = "glm-4"
-    llm = ChatOpenAI(model=llm_name, stream_usage=True)
+    llm = ChatOpenAI(model=llm_name, stream_usage=True, streaming=True)
     mock_client = AsyncMock()
 
     async def mock_create(*args: Any, **kwargs: Any) -> MockAsyncContextManager:
@@ -375,7 +375,7 @@ async def mock_create(*args: Any, **kwargs: Any) -> MockAsyncContextManager:
 
 def test_glm4_stream(mock_glm4_completion: list) -> None:
     llm_name = "glm-4"
-    llm = ChatOpenAI(model=llm_name, stream_usage=True)
+    llm = ChatOpenAI(model=llm_name, stream_usage=True, streaming=True)
     mock_client = MagicMock()
 
     def mock_create(*args: Any, **kwargs: Any) -> MockSyncContextManager:
@@ -431,7 +431,7 @@ def mock_deepseek_completion() -> list[dict]:
 
 async def test_deepseek_astream(mock_deepseek_completion: list) -> None:
     llm_name = "deepseek-chat"
-    llm = ChatOpenAI(model=llm_name, stream_usage=True)
+    llm = ChatOpenAI(model=llm_name, stream_usage=True, streaming=True)
     mock_client = AsyncMock()
 
     async def mock_create(*args: Any, **kwargs: Any) -> MockAsyncContextManager:
@@ -455,7 +455,7 @@ async def mock_create(*args: Any, **kwargs: Any) -> MockAsyncContextManager:
 
 def test_deepseek_stream(mock_deepseek_completion: list) -> None:
     llm_name = "deepseek-chat"
-    llm = ChatOpenAI(model=llm_name, stream_usage=True)
+    llm = ChatOpenAI(model=llm_name, stream_usage=True, streaming=True)
     mock_client = MagicMock()
 
     def mock_create(*args: Any, **kwargs: Any) -> MockSyncContextManager:
@@ -499,7 +499,7 @@ def mock_openai_completion() -> list[dict]:
 
 async def test_openai_astream(mock_openai_completion: list) -> None:
     llm_name = "gpt-4o"
-    llm = ChatOpenAI(model=llm_name, stream_usage=True)
+    llm = ChatOpenAI(model=llm_name, stream_usage=True, streaming=True)
     mock_client = AsyncMock()
 
     async def mock_create(*args: Any, **kwargs: Any) -> MockAsyncContextManager:
@@ -523,7 +523,7 @@ async def mock_create(*args: Any, **kwargs: Any) -> MockAsyncContextManager:
 
 def test_openai_stream(mock_openai_completion: list) -> None:
     llm_name = "gpt-4o"
-    llm = ChatOpenAI(model=llm_name, stream_usage=True)
+    llm = ChatOpenAI(model=llm_name, stream_usage=True, streaming=True)
     mock_client = MagicMock()
 
     def mock_create(*args: Any, **kwargs: Any) -> MockSyncContextManager:

libs/partners/openai/tests/unit_tests/chat_models/test_responses_stream.py

@@ -621,7 +621,7 @@ def _strip_none(obj: Any) -> Any:
 
 
 def test_responses_stream() -> None:
-    llm = ChatOpenAI(model="o4-mini", output_version="responses/v1")
+    llm = ChatOpenAI(model="o4-mini", output_version="responses/v1", streaming=True)
     mock_client = MagicMock()
 
     def mock_create(*args: Any, **kwargs: Any) -> MockSyncContextManager: