Add stream option to _agenerate

libs/ibm/langchain_ibm/llms.py

@@ -471,17 +471,25 @@ async def _agenerate(
         prompts: List[str],
         stop: Optional[List[str]] = None,
         run_manager: Optional[AsyncCallbackManagerForLLMRun] = None,
+        stream: Optional[bool] = None,
         **kwargs: Any,
     ) -> LLMResult:
         """Async run the LLM on the given prompt and input."""
         params, kwargs = self._get_chat_params(stop=stop, **kwargs)
         params = self._validate_chat_params(params)
-        responses = [
-            await self.watsonx_model.agenerate(prompt=prompt, params=params, **kwargs)
-            for prompt in prompts
-        ]
+        if stream:
+            return await super()._agenerate(
+                prompts=prompts, stop=stop, run_manager=run_manager, **kwargs
+            )
+        else:
+            responses = [
+                await self.watsonx_model.agenerate(
+                    prompt=prompt, params=params, **kwargs
+                )
+                for prompt in prompts
+            ]
 
-        return self._create_llm_result(responses)
+            return self._create_llm_result(responses)
 
     def _stream(
         self,

libs/ibm/tests/integration_tests/test_llms.py

@@ -445,6 +445,16 @@ async def test_watsonx_agenerate() -> None:
     assert response.llm_output["token_usage"]["generated_token_count"] != 0  # type: ignore
 
 
+async def test_watsonx_agenerate_with_stream() -> None:
+    watsonxllm = WatsonxLLM(
+        model_id=MODEL_ID,
+        url="https://us-south.ml.cloud.ibm.com",  # type: ignore[arg-type]
+        project_id=WX_PROJECT_ID,
+    )
+    response = await watsonxllm.agenerate(["What color sunflower is?"], stream=True)
+    assert "yellow" in response.generations[0][0].text.lower()
+
+
 def test_get_num_tokens() -> None:
     watsonxllm = WatsonxLLM(
         model_id=MODEL_ID,