feat(litellm): Added Support for responses API in litellm instrumentor

python/instrumentation/openinference-instrumentation-litellm/examples/litellm_responses.py

@@ -0,0 +1,207 @@
+import asyncio
+
+import litellm
+from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
+from opentelemetry.sdk.trace import TracerProvider
+from opentelemetry.sdk.trace.export import SimpleSpanProcessor
+
+from openinference.instrumentation.litellm import LiteLLMInstrumentor
+
+endpoint = "http://127.0.0.1:6006/v1/traces"
+tracer_provider = TracerProvider()
+tracer_provider.add_span_processor(SimpleSpanProcessor(OTLPSpanExporter(endpoint)))
+LiteLLMInstrumentor().instrument(tracer_provider=tracer_provider)
+
+
+stream = True
+
+
+def responses_image_input():
+    response = litellm.responses(
+        model="openai/gpt-4.1",
+        input=[
+            {
+                "role": "user",
+                "content": [
+                    {"type": "input_text", "text": "what is in this image?"},
+                    {
+                        "type": "input_image",
+                        "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg",
+                    },
+                ],
+            }
+        ],
+        stream=stream,
+    )
+    print(list(response))
+
+
+def responses_file_input():
+    response = litellm.responses(
+        model="openai/gpt-4.1",
+        input=[
+            {
+                "role": "user",
+                "content": [
+                    {"type": "input_text", "text": "what is in this file?"},
+                    {
+                        "type": "input_file",
+                        "file_url": "https://www.berkshirehathaway.com/letters/2024ltr.pdf",
+                    },
+                ],
+            }
+        ],
+        stream=stream,
+    )
+    print(list(response))
+
+
+def websearch_input():
+    response = litellm.responses(
+        model="openai/gpt-4.1",
+        tools=[{"type": "web_search_preview"}],
+        input="What was a positive news story from today?",
+        stream=stream,
+    )
+    print(list(response))
+
+
+def function_call():
+    tools = [
+        {
+            "type": "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, e.g. San Francisco, CA",
+                    },
+                    "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
+                },
+                "required": ["location", "unit"],
+            },
+        }
+    ]
+    response = litellm.responses(
+        model="openai/gpt-4.1",
+        tools=tools,
+        input="What is the weather like in Boston today?",
+        tool_choice="auto",
+        stream=stream,
+    )
+    print(list(response))
+
+
+def reasoning_input():
+    response = litellm.responses(
+        model="openai/o3-mini",
+        input="How much wood would a woodchuck chuck?",
+        reasoning={"effort": "high"},
+        stream=stream,
+    )
+    print(list(response))
+
+
+async def async_responses_image_input():
+    response = await litellm.aresponses(
+        model="openai/gpt-4.1",
+        input=[
+            {
+                "role": "user",
+                "content": [
+                    {"type": "input_text", "text": "what is in this image?"},
+                    {
+                        "type": "input_image",
+                        "image_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg",
+                    },
+                ],
+            }
+        ],
+        stream=stream,
+    )
+    print(list(response))
+
+
+async def async_responses_file_input():
+    response = await litellm.aresponses(
+        model="openai/gpt-4.1",
+        input=[
+            {
+                "role": "user",
+                "content": [
+                    {"type": "input_text", "text": "what is in this file?"},
+                    {
+                        "type": "input_file",
+                        "file_url": "https://www.berkshirehathaway.com/letters/2024ltr.pdf",
+                    },
+                ],
+            }
+        ],
+        stream=stream,
+    )
+    print(list(response))
+
+
+async def async_websearch_input():
+    response = await litellm.aresponses(
+        model="openai/gpt-4.1",
+        tools=[{"type": "web_search_preview"}],
+        input="What was a positive news story from today?",
+        stream=stream,
+    )
+    print(list(response))
+
+
+async def async_function_call():
+    tools = [
+        {
+            "type": "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, e.g. San Francisco, CA",
+                    },
+                    "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
+                },
+                "required": ["location", "unit"],
+            },
+        }
+    ]
+    response = await litellm.aresponses(
+        model="openai/gpt-4.1",
+        tools=tools,
+        input="What is the weather like in Boston today?",
+        tool_choice="auto",
+        stream=stream,
+    )
+    print(list(response))
+
+
+async def async_reasoning_input():
+    response = await litellm.aresponses(
+        model="openai/o3-mini",
+        input="How much wood would a woodchuck chuck?",
+        reasoning={"effort": "high"},
+        stream=stream,
+    )
+    print(list(response))
+
+
+if __name__ == "__main__":
+    responses_image_input()
+    responses_file_input()
+    websearch_input()
+    function_call()
+    reasoning_input()
+    asyncio.run(async_responses_image_input())
+    asyncio.run(async_responses_file_input())
+    asyncio.run(async_websearch_input())
+    asyncio.run(async_function_call())
+    asyncio.run(async_reasoning_input())