add testing for aporia guardrails

litellm/proxy/example_config_yaml/otel_test_config.yaml

@@ -9,3 +9,16 @@ litellm_settings:
   cache: true
   callbacks: ["otel"]
 
+guardrails:
+  - guardrail_name: "aporia-pre-guard"
+    litellm_params:
+      guardrail: aporia  # supported values: "aporia", "bedrock", "lakera"
+      mode: "post_call"
+      api_key: os.environ/APORIA_API_KEY_1
+      api_base: os.environ/APORIA_API_BASE_1
+  - guardrail_name: "aporia-post-guard"
+    litellm_params:
+      guardrail: aporia  # supported values: "aporia", "bedrock", "lakera"
+      mode: "post_call"
+      api_key: os.environ/APORIA_API_KEY_2
+      api_base: os.environ/APORIA_API_BASE_2

litellm/proxy/proxy_config.yaml

@@ -1,8 +1,9 @@
 model_list:
-  - model_name: gpt-3.5-turbo
+  - model_name: fake-openai-endpoint
     litellm_params:
-      model: openai/gpt-3.5-turbo
-      api_key: os.environ/OPENAI_API_KEY
+      model: openai/fake
+      api_key: fake-key
+      api_base: https://exampleopenaiendpoint-production.up.railway.app/
 
 guardrails:
   - guardrail_name: "aporia-pre-guard"

tests/otel_tests/test_guardrails.py

@@ -0,0 +1,59 @@
+import pytest
+import asyncio
+import aiohttp, openai
+from openai import OpenAI, AsyncOpenAI
+from typing import Optional, List, Union
+import uuid
+
+
+async def chat_completion(session, key, model: Union[str, List] = "gpt-4"):
+    url = "http://0.0.0.0:4000/chat/completions"
+    headers = {
+        "Authorization": f"Bearer {key}",
+        "Content-Type": "application/json",
+    }
+    data = {
+        "model": model,
+        "messages": [
+            {"role": "user", "content": f"Hello! {str(uuid.uuid4())}"},
+        ],
+        "guardrails": ["aporia-post-guard", "aporia-pre-guard"],
+    }
+
+    async with session.post(url, headers=headers, json=data) as response:
+        status = response.status
+        response_text = await response.text()
+
+        print(response_text)
+        print()
+
+        if status != 200:
+            raise Exception(f"Request did not return a 200 status code: {status}")
+
+        # response headers
+        response_headers = response.headers
+        print("response headers=", response_headers)
+
+        return await response.json(), response_headers
+
+
+@pytest.mark.asyncio
+async def test_no_llm_guard_triggered():
+    """
+    - Tests a request where no content mod is triggered
+    - Assert that the guardrails applied are returned in the response headers
+    """
+    async with aiohttp.ClientSession() as session:
+        response, headers = await chat_completion(
+            session, "sk-1234", model="fake-openai-endpoint"
+        )
+        await asyncio.sleep(3)
+
+        print("response=", response, "response headers", headers)
+
+        assert "x-litellm-applied-guardrails" in headers
+
+        assert (
+            headers["x-litellm-applied-guardrails"]
+            == "aporia-pre-guard,aporia-post-guard"
+        )