Add anthropic thinking + reasoning content support (#8778)

* feat(anthropic/chat/transformation.py): add anthropic thinking param support * feat(anthropic/chat/transformation.py): support returning thinking content for anthropic on streaming responses * feat(anthropic/chat/transformation.py): return list of thinking blocks (include block signature) allows usage in tool call responses * fix(types/utils.py): extract and map reasoning_content from anthropic as content str * test: add testing to ensure thinking_blocks are returned at the root * fix(anthropic/chat/handler.py): return thinking blocks on streaming - include signature * feat(factory.py): handle anthropic thinking blocks translation if in assistant response * test: handle openai internal instability * test: handle openai audio instability * ci: pin anthropic dep * test: handle openai audio instability * fix: fix linting error * refactor(anthropic/chat/transformation.py): refactor function to remain <50 LOC * fix: fix linting error * fix: fix linting error * fix: fix linting error * fix: fix linting error
2025-04-25 18:54:30 +00:00 · 2025-02-24 21:54:30 -08:00 · 2025-02-24 21:54:30 -08:00 · 142b195784
commit 142b195784
parent 9914c166b7
16 changed files with 332 additions and 62 deletions
--- a/tests/local_testing/test_function_calling.py
+++ b/tests/local_testing/test_function_calling.py
@ -157,6 +157,113 @@ def test_aaparallel_function_call(model):
 # test_parallel_function_call()


+@pytest.mark.parametrize(
+    "model",
+    [
+        "anthropic/claude-3-7-sonnet-20250219",
+    ],
+)
+@pytest.mark.flaky(retries=3, delay=1)
+def test_aaparallel_function_call_with_anthropic_thinking(model):
+    try:
+        litellm._turn_on_debug()
+        litellm.modify_params = True
+        # Step 1: send the conversation and available functions to the model
+        messages = [
+            {
+                "role": "user",
+                "content": "What's the weather like in San Francisco, Tokyo, and Paris? - give me 3 responses",
+            }
+        ]
+        tools = [
+            {
+                "type": "function",
+                "function": {
+                    "name": "get_current_weather",
+                    "description": "Get the current weather in a given location",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {
+                            "location": {
+                                "type": "string",
+                                "description": "The city and state",
+                            },
+                            "unit": {
+                                "type": "string",
+                                "enum": ["celsius", "fahrenheit"],
+                            },
+                        },
+                        "required": ["location"],
+                    },
+                },
+            }
+        ]
+        response = litellm.completion(
+            model=model,
+            messages=messages,
+            tools=tools,
+            tool_choice="auto",  # auto is default, but we'll be explicit
+            thinking={"type": "enabled", "budget_tokens": 1024},
+        )
+        print("Response\n", response)
+        response_message = response.choices[0].message
+        tool_calls = response_message.tool_calls
+
+        print("Expecting there to be 3 tool calls")
+        assert (
+            len(tool_calls) > 0
+        )  # this has to call the function for SF, Tokyo and paris
+
+        # Step 2: check if the model wanted to call a function
+        print(f"tool_calls: {tool_calls}")
+        if tool_calls:
+            # Step 3: call the function
+            # Note: the JSON response may not always be valid; be sure to handle errors
+            available_functions = {
+                "get_current_weather": get_current_weather,
+            }  # only one function in this example, but you can have multiple
+            messages.append(
+                response_message
+            )  # extend conversation with assistant's reply
+            print("Response message\n", response_message)
+            # Step 4: send the info for each function call and function response to the model
+            for tool_call in tool_calls:
+                function_name = tool_call.function.name
+                if function_name not in available_functions:
+                    # the model called a function that does not exist in available_functions - don't try calling anything
+                    return
+                function_to_call = available_functions[function_name]
+                function_args = json.loads(tool_call.function.arguments)
+                function_response = function_to_call(
+                    location=function_args.get("location"),
+                    unit=function_args.get("unit"),
+                )
+                messages.append(
+                    {
+                        "tool_call_id": tool_call.id,
+                        "role": "tool",
+                        "name": function_name,
+                        "content": function_response,
+                    }
+                )  # extend conversation with function response
+            print(f"messages: {messages}")
+            second_response = litellm.completion(
+                model=model,
+                messages=messages,
+                seed=22,
+                # tools=tools,
+                drop_params=True,
+                thinking={"type": "enabled", "budget_tokens": 1024},
+            )  # get a new response from the model where it can see the function response
+            print("second response\n", second_response)
+    except litellm.InternalServerError as e:
+        print(e)
+    except litellm.RateLimitError as e:
+        print(e)
+    except Exception as e:
+        pytest.fail(f"Error occurred: {e}")
+
+
 from litellm.types.utils import ChatCompletionMessageToolCall, Function, Message