langchain[patch]: fix ChatVertexAI streaming (#14369)

libs/langchain/langchain/chat_models/vertexai.py

@@ -242,7 +242,7 @@ class ChatVertexAI(_VertexAICommon, BaseChatModel):
     ) -> Iterator[ChatGenerationChunk]:
         question = _get_question(messages)
         history = _parse_chat_history(messages[:-1])
-        params = self._prepare_params(stop=stop, **kwargs)
+        params = self._prepare_params(stop=stop, stream=True, **kwargs)
         examples = kwargs.get("examples", None)
         if examples:
             params["examples"] = _parse_examples(examples)

libs/langchain/tests/integration_tests/chat_models/test_vertexai.py

@@ -11,7 +11,12 @@ from typing import Optional
 from unittest.mock import MagicMock, Mock, patch
 
 import pytest
-from langchain_core.messages import AIMessage, HumanMessage, SystemMessage
+from langchain_core.messages import (
+    AIMessage,
+    AIMessageChunk,
+    HumanMessage,
+    SystemMessage,
+)
 from langchain_core.outputs import LLMResult
 
 from langchain.chat_models import ChatVertexAI
@@ -41,6 +46,7 @@ def test_vertexai_single_call(model_name: str) -> None:
     assert isinstance(response.content, str)
 
 
+@pytest.mark.scheduled
 def test_candidates() -> None:
     model = ChatVertexAI(model_name="chat-bison@001", temperature=0.3, n=2)
     message = HumanMessage(content="Hello")
@@ -62,6 +68,16 @@ async def test_vertexai_agenerate() -> None:
     assert response.generations[0][0] == sync_response.generations[0][0]
 
 
+@pytest.mark.scheduled
+async def test_vertexai_stream() -> None:
+    model = ChatVertexAI(temperature=0)
+    message = HumanMessage(content="Hello")
+
+    sync_response = model.stream([message])
+    for chunk in sync_response:
+        assert isinstance(chunk, AIMessageChunk)
+
+
 @pytest.mark.scheduled
 def test_vertexai_single_call_with_context() -> None:
     model = ChatVertexAI()