fix(openai_like): forward acompletion to OpenAILikeChatHandler to fix mock interception

litellm/main.py

@@ -2850,6 +2850,7 @@ def _complete_cohere_chat(ctx: _CompletionDispatchContext) -> _CompletionDispatc
 
 
 def _complete_maritalk(ctx: _CompletionDispatchContext) -> _CompletionDispatchResult:
+    acompletion = ctx.acompletion
     api_base = ctx.api_base
     api_key = ctx.api_key
     custom_prompt_dict = ctx.custom_prompt_dict
@@ -2879,10 +2880,12 @@ def _complete_maritalk(ctx: _CompletionDispatchContext) -> _CompletionDispatchRe
         logging_obj=logging,
         custom_llm_provider="maritalk",
         custom_prompt_dict=custom_prompt_dict,
+        acompletion=acompletion,
     )
 
 
 def _complete_amazon_nova(ctx: _CompletionDispatchContext) -> _CompletionDispatchResult:
+    acompletion = ctx.acompletion
     api_base = ctx.api_base
     api_key = ctx.api_key
     custom_llm_provider = ctx.custom_llm_provider
@@ -2915,6 +2918,7 @@ def _complete_amazon_nova(ctx: _CompletionDispatchContext) -> _CompletionDispatc
         timeout=timeout,
         custom_llm_provider=custom_llm_provider,
         custom_prompt_dict=custom_prompt_dict,
+        acompletion=acompletion,
     )
 
 

tests/test_litellm/test_openai_params_strip.py

@@ -229,3 +229,285 @@ async def test_openai_metadata_preview_feature():
         mock_create_no_preview.assert_called_once()
         call_kwargs = mock_create_no_preview.call_args[1]
         assert "metadata" not in call_kwargs
+
+
+@pytest.mark.asyncio
+async def test_strip_litellm_internal_params_non_dict():
+    """
+    Test that strip_litellm_internal_params returns the input directly
+    if it is not a dictionary.
+    """
+    from litellm.litellm_core_utils.param_utils import strip_litellm_internal_params
+
+    assert strip_litellm_internal_params("not a dict") == "not a dict"
+    assert strip_litellm_internal_params(None) is None
+    assert strip_litellm_internal_params(123) == 123
+
+
+@pytest.mark.asyncio
+async def test_azure_chat_completion_params_strip():
+    """
+    Test that litellm_params and _litellm_* prefixed params are stripped
+    from Azure OpenAI completion calls.
+    """
+    mock_choice = MagicMock()
+    mock_choice.finish_reason = "stop"
+    mock_choice.index = 0
+    mock_choice.message = MagicMock(content="Mock response", role="assistant")
+    mock_choice.message.tool_calls = None
+    mock_choice.message.function_call = None
+    mock_choice.message.provider_specific_fields = {}
+
+    mock_response_data = MagicMock()
+    mock_response_data.choices = [mock_choice]
+    mock_response_data.id = "chatcmpl-123"
+    mock_response_data.created = 1677858242
+    mock_response_data.model = "gpt-4"
+    mock_response_data.object = "chat.completion"
+    mock_response_data.usage = MagicMock(completion_tokens=10, prompt_tokens=5, total_tokens=15)
+
+    mock_create = MagicMock()
+    mock_raw_resp = MagicMock()
+    mock_raw_resp.headers = {"x-test-header": "test"}
+    mock_raw_resp.parse.return_value = mock_response_data
+    mock_create.return_value = mock_raw_resp
+
+    with patch("openai.resources.chat.completions.Completions.create", mock_create):
+        completion(
+            model="azure/gpt-4",
+            messages=[{"role": "user", "content": "hi"}],
+            api_key="mock-key",
+            api_base="https://mock.openai.azure.com",
+            api_version="2023-05-15",
+            # internal params that should be stripped
+            litellm_params={"metadata": {"some_internal_key": "some_value"}},
+            _litellm_test_param="test_value",
+        )
+
+        mock_create.assert_called_once()
+        call_kwargs = mock_create.call_args[1]
+
+        assert "litellm_params" not in call_kwargs
+        assert "_litellm_test_param" not in call_kwargs
+
+
+@pytest.mark.asyncio
+async def test_azure_chat_acompletion_params_strip():
+    """
+    Test that litellm_params and _litellm_* prefixed params are stripped
+    from Azure OpenAI async completion calls.
+    """
+    mock_choice = MagicMock()
+    mock_choice.finish_reason = "stop"
+    mock_choice.index = 0
+    mock_choice.message = MagicMock(content="Mock response", role="assistant")
+    mock_choice.message.tool_calls = None
+    mock_choice.message.function_call = None
+    mock_choice.message.provider_specific_fields = {}
+
+    mock_response_data = MagicMock()
+    mock_response_data.choices = [mock_choice]
+    mock_response_data.id = "chatcmpl-123"
+    mock_response_data.created = 1677858242
+    mock_response_data.model = "gpt-4"
+    mock_response_data.object = "chat.completion"
+    mock_response_data.usage = MagicMock(completion_tokens=10, prompt_tokens=5, total_tokens=15)
+
+    mock_raw_resp = MagicMock()
+    mock_raw_resp.headers = {"x-test-header": "test"}
+    mock_raw_resp.parse.return_value = mock_response_data
+
+    mock_acreate = AsyncMock(return_value=mock_raw_resp)
+
+    with patch("openai.resources.chat.completions.AsyncCompletions.create", mock_acreate):
+        await acompletion(
+            model="azure/gpt-4",
+            messages=[{"role": "user", "content": "hi"}],
+            api_key="mock-key",
+            api_base="https://mock.openai.azure.com",
+            api_version="2023-05-15",
+            litellm_params={"metadata": {"some_internal_key": "some_value"}},
+            _litellm_test_param="test_value",
+        )
+
+        mock_acreate.assert_called_once()
+        call_kwargs = mock_acreate.call_args[1]
+
+        assert "litellm_params" not in call_kwargs
+        assert "_litellm_test_param" not in call_kwargs
+
+
+@pytest.mark.asyncio
+async def test_azure_embedding_params_strip():
+    """
+    Test that litellm_params and _litellm_* prefixed params are stripped
+    from Azure OpenAI embedding calls.
+    """
+    mock_response_data = MagicMock()
+    mock_response_data.model = "text-embedding-ada-002"
+    mock_response_data.object = "list"
+    mock_response_data.data = [MagicMock(embedding=[0.1, 0.2])]
+    mock_response_data.usage = MagicMock(prompt_tokens=5, total_tokens=5)
+
+    mock_create = MagicMock()
+    mock_raw_resp = MagicMock()
+    mock_raw_resp.headers = {"x-test-header": "test"}
+    mock_raw_resp.parse.return_value = mock_response_data
+    mock_create.return_value = mock_raw_resp
+
+    with patch("openai.resources.embeddings.Embeddings.create", mock_create):
+        embedding(
+            model="azure/text-embedding-ada-002",
+            input=["hello"],
+            api_key="mock-key",
+            api_base="https://mock.openai.azure.com",
+            api_version="2023-05-15",
+            litellm_params={"metadata": {"some_internal_key": "some_value"}},
+            _litellm_test_param="test_value",
+        )
+
+        mock_create.assert_called_once()
+        call_kwargs = mock_create.call_args[1]
+
+        assert "litellm_params" not in call_kwargs
+        assert "_litellm_test_param" not in call_kwargs
+
+
+@pytest.mark.asyncio
+async def test_azure_aembedding_params_strip():
+    """
+    Test that litellm_params and _litellm_* prefixed params are stripped
+    from Azure OpenAI async embedding calls.
+    """
+    from litellm import aembedding
+
+    mock_response_data = MagicMock()
+    mock_response_data.model = "text-embedding-ada-002"
+    mock_response_data.object = "list"
+    mock_response_data.data = [MagicMock(embedding=[0.1, 0.2])]
+    mock_response_data.usage = MagicMock(prompt_tokens=5, total_tokens=5)
+
+    mock_raw_resp = MagicMock()
+    mock_raw_resp.headers = {"x-test-header": "test"}
+    mock_raw_resp.parse.return_value = mock_response_data
+
+    mock_acreate = AsyncMock(return_value=mock_raw_resp)
+
+    with patch("openai.resources.embeddings.AsyncEmbeddings.create", mock_acreate):
+        await aembedding(
+            model="azure/text-embedding-ada-002",
+            input=["hello"],
+            api_key="mock-key",
+            api_base="https://mock.openai.azure.com",
+            api_version="2023-05-15",
+            litellm_params={"metadata": {"some_internal_key": "some_value"}},
+            _litellm_test_param="test_value",
+        )
+
+        mock_acreate.assert_called_once()
+        call_kwargs = mock_acreate.call_args[1]
+
+        assert "litellm_params" not in call_kwargs
+        assert "_litellm_test_param" not in call_kwargs
+
+
+@pytest.mark.asyncio
+async def test_openai_like_chat_completion_params_strip():
+    """
+    Test that litellm_params and _litellm_* prefixed params are stripped
+    from OpenAI-like completion calls.
+    """
+    import json
+    import httpx
+
+    # Mock Response object
+    mock_response = MagicMock(spec=httpx.Response)
+    mock_response.status_code = 200
+    mock_response.headers = httpx.Headers({"content-type": "application/json"})
+
+    mock_body = {
+        "choices": [
+            {"finish_reason": "stop", "index": 0, "message": {"content": "Mock response", "role": "assistant"}}
+        ],
+        "id": "chatcmpl-123",
+        "created": 1677858242,
+        "model": "maritalk",
+        "object": "chat.completion",
+        "usage": {"completion_tokens": 10, "prompt_tokens": 5, "total_tokens": 15},
+    }
+
+    mock_response.text = json.dumps(mock_body)
+    mock_response.json.return_value = mock_body
+    mock_response.content = json.dumps(mock_body).encode("utf-8")
+
+    mock_post = MagicMock(return_value=mock_response)
+
+    with patch("litellm.llms.custom_httpx.http_handler.HTTPHandler.post", mock_post):
+        completion(
+            model="maritalk/maritalk",
+            messages=[{"role": "user", "content": "hi"}],
+            api_key="mock-key",
+            api_base="https://example.com",
+            # internal params that should be stripped
+            litellm_params={"metadata": {"some_internal_key": "some_value"}},
+            _litellm_test_param="test_value",
+        )
+
+        mock_post.assert_called_once()
+        call_kwargs = mock_post.call_args[1]
+
+        # Verify that data parameter passed to post has been stripped
+        sent_data = json.loads(call_kwargs.get("data", "{}"))
+        assert "litellm_params" not in sent_data
+        assert "_litellm_test_param" not in sent_data
+
+
+@pytest.mark.asyncio
+async def test_openai_like_chat_acompletion_params_strip():
+    """
+    Test that litellm_params and _litellm_* prefixed params are stripped
+    from OpenAI-like async completion calls.
+    """
+    import json
+    import httpx
+
+    # Mock Response object
+    mock_response = MagicMock(spec=httpx.Response)
+    mock_response.status_code = 200
+    mock_response.headers = httpx.Headers({"content-type": "application/json"})
+
+    mock_body = {
+        "choices": [
+            {"finish_reason": "stop", "index": 0, "message": {"content": "Mock response", "role": "assistant"}}
+        ],
+        "id": "chatcmpl-123",
+        "created": 1677858242,
+        "model": "maritalk",
+        "object": "chat.completion",
+        "usage": {"completion_tokens": 10, "prompt_tokens": 5, "total_tokens": 15},
+    }
+
+    mock_response.text = json.dumps(mock_body)
+    mock_response.json.return_value = mock_body
+    mock_response.content = json.dumps(mock_body).encode("utf-8")
+
+    mock_apost = AsyncMock(return_value=mock_response)
+
+    with patch.object(litellm.module_level_aclient, "post", mock_apost):
+        await acompletion(
+            model="maritalk/maritalk",
+            messages=[{"role": "user", "content": "hi"}],
+            api_key="mock-key",
+            api_base="https://example.com",
+            # internal params that should be stripped
+            litellm_params={"metadata": {"some_internal_key": "some_value"}},
+            _litellm_test_param="test_value",
+        )
+
+        mock_apost.assert_called_once()
+        call_kwargs = mock_apost.call_args[1]
+
+        # Verify that data parameter passed to post has been stripped
+        sent_data = json.loads(call_kwargs.get("data", "{}"))
+        assert "litellm_params" not in sent_data
+        assert "_litellm_test_param" not in sent_data