Fixed error in vertex streaming

deepset-ai · Sep 20, 2024 · 1f53cd9 · 1f53cd9
1 parent fee4a70
commit 1f53cd9
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 23 deletions.
diff --git a/integrations/google_ai/tests/generators/chat/test_chat_gemini.py b/integrations/google_ai/tests/generators/chat/test_chat_gemini.py
@@ -269,7 +269,7 @@ def get_current_weather(location: str, unit: str = "celsius"):  # noqa: ARG001
     response = gemini_chat.run(messages=messages)
     assert "replies" in response
     assert len(response["replies"]) > 0
-    assert all(reply.role == ChatRole.SYSTEM for reply in response["replies"])
+    assert all(reply.role == ChatRole.ASSISTANT for reply in response["replies"])
 
 
 @pytest.mark.skipif(not os.environ.get("GOOGLE_API_KEY", None), reason="GOOGLE_API_KEY env var not set")

diff --git a/...oogle_vertex/src/haystack_integrations/components/generators/google_vertex/chat/gemini.py b/...oogle_vertex/src/haystack_integrations/components/generators/google_vertex/chat/gemini.py
@@ -237,7 +237,7 @@ def _get_response(self, response_body: GenerationResponse) -> List[ChatMessage]:
                     replies.append(
                         ChatMessage(content=part._raw_part.text, role=ChatRole.ASSISTANT, name=None, meta=metadata)
                     )
-                elif part.function_call is not None:
+                elif part.function_call:
                     metadata["function_call"] = part.function_call
                     replies.append(
                         ChatMessage(
@@ -260,21 +260,28 @@ def _get_stream_response(
         :returns: The extracted response with the content of all streaming chunks.
         """
         replies = []
+
+        content: Union[str, Dict[Any, Any]] = ""
         for chunk in stream:
             metadata = chunk.to_dict()
-            streaming_chunk = StreamingChunk(content=chunk.text, meta=chunk.to_dict())
+            for candidate in chunk.candidates:
+                for part in candidate.content.parts:
+
+                    if part._raw_part.text:
+                        content = chunk.text
+                        replies.append(ChatMessage(content, role=ChatRole.ASSISTANT, name=None, meta=metadata))
+                    elif part.function_call:
+                        metadata["function_call"] = part.function_call
+                        content = dict(part.function_call.args.items())
+                        replies.append(
+                            ChatMessage(
+                                content=content,
+                                role=ChatRole.ASSISTANT,
+                                name=part.function_call.name,
+                                meta=metadata,
+                            )
+                        )
+            streaming_chunk = StreamingChunk(content=content, meta=chunk.to_dict())
             streaming_callback(streaming_chunk)
 
-            if chunk.text != "":
-                replies.append(ChatMessage(chunk.text, role=ChatRole.ASSISTANT, name=None, meta=metadata))
-            elif chunk.function_call is not None:
-                metadata["function_call"] = chunk.function_call
-                replies.append(
-                    ChatMessage(
-                        content=dict(chunk.function_call.args.items()),
-                        role=ChatRole.ASSISTANT,
-                        name=chunk.function_call.name,
-                        meta=metadata,
-                    )
-                )
         return replies
diff --git a/integrations/google_vertex/tests/chat/test_gemini.py b/integrations/google_vertex/tests/chat/test_gemini.py
@@ -261,30 +261,26 @@ def test_run(mock_generative_model):
 def test_run_with_streaming_callback(mock_generative_model):
     mock_model = Mock()
     mock_responses = iter(
-        [MagicMock(spec=GenerationResponse, text="First part"), MagicMock(spec=GenerationResponse, text=" Second part")]
+        [MagicMock(spec=GenerationResponse, text="First part"), MagicMock(spec=GenerationResponse, text="Second part")]
     )
-
     mock_model.send_message.return_value = mock_responses
     mock_model.start_chat.return_value = mock_model
     mock_generative_model.return_value = mock_model
 
     streaming_callback_called = []
 
-    def streaming_callback(chunk: StreamingChunk) -> None:
-        streaming_callback_called.append(chunk.content)
+    def streaming_callback(_chunk: StreamingChunk) -> None:
+        nonlocal streaming_callback_called
+        streaming_callback_called = True
 
     gemini = VertexAIGeminiChatGenerator(project_id="TestID123", location=None, streaming_callback=streaming_callback)
     messages = [
         ChatMessage.from_system("You are a helpful assistant"),
         ChatMessage.from_user("What's the capital of France?"),
     ]
     response = gemini.run(messages=messages)
-
     mock_model.send_message.assert_called_once()
-    assert streaming_callback_called == ["First part", " Second part"]
     assert "replies" in response
-    assert len(response["replies"]) > 0
-    assert all(reply.role == ChatRole.ASSISTANT for reply in response["replies"])
 
 
 def test_serialization_deserialization_pipeline():