test - wildcard openai models on proxy

litellm/proxy/proxy_config.yaml

@@ -4,6 +4,10 @@ model_list:
       model: openai/fake
       api_key: fake-key
       api_base: https://exampleopenaiendpoint-production.up.railway.app/
+  - model_name: "*"
+    litellm_params:
+      model: openai/*
+      api_key: os.environ/OPENAI_API_KEY
 general_settings:
   store_model_in_db: true
   master_key: sk-1234

proxy_server_config.yaml

@@ -55,6 +55,10 @@ model_list:
       api_base: https://openai-function-calling-workers.tasslexyz.workers.dev/
       stream_timeout: 0.001
       rpm: 1
+  - model_name: "*"
+    litellm_params:
+      model: openai/*
+      api_key: os.environ/OPENAI_API_KEY
   - model_name: gpt-instruct # [PROD TEST] - tests if `/health` automatically infers this to be a text completion model
     litellm_params:
       model: text-completion-openai/gpt-3.5-turbo-instruct

tests/test_openai_endpoints.py

@@ -14,16 +14,19 @@ def response_header_check(response):
     assert headers_size < 4096, "Response headers exceed the 4kb limit"
 
 
-async def generate_key(session):
+async def generate_key(
+    session,
+    models=[
+        "gpt-4",
+        "text-embedding-ada-002",
+        "dall-e-2",
+        "fake-openai-endpoint-2",
+    ],
+):
     url = "http://0.0.0.0:4000/key/generate"
     headers = {"Authorization": "Bearer sk-1234", "Content-Type": "application/json"}
     data = {
-        "models": [
-            "gpt-4",
-            "text-embedding-ada-002",
-            "dall-e-2",
-            "fake-openai-endpoint-2",
-        ],
+        "models": models,
         "duration": None,
     }
 
@@ -294,3 +297,19 @@ async def test_image_generation():
         key_gen = await new_user(session=session)
         key_2 = key_gen["key"]
         await image_generation(session=session, key=key_2)
+
+
+@pytest.mark.asyncio
+async def test_openai_wildcard_chat_completion():
+    """
+    - Create key for model = "*" -> this has access to all models
+    - proxy_server_config.yaml has model = *
+    - Make chat completion call
+
+    """
+    async with aiohttp.ClientSession() as session:
+        key_gen = await generate_key(session=session, models=["*"])
+        key = key_gen["key"]
+
+        # call chat/completions with a model that the key was not created for + the model is not on the config.yaml
+        await chat_completion(session=session, key=key, model="gpt-3.5-turbo-0125")