use passed in model name in chat completion request (#445)

demos/samples_java/weather_forcecast_service/.project

@@ -22,12 +22,12 @@
 	</natures>
 	<filteredResources>
 		<filter>
-			<id>1739479945236</id>
+			<id>1742579142020</id>
 			<name></name>
 			<type>30</type>
 			<matcher>
 				<id>org.eclipse.core.resources.regexFilterMatcher</id>
-				<arguments>node_modules|.git|__CREATED_BY_JAVA_LANGUAGE_SERVER__</arguments>
+				<arguments>node_modules|\.git|__CREATED_BY_JAVA_LANGUAGE_SERVER__</arguments>
 			</matcher>
 		</filter>
 	</filteredResources>

demos/shared/chatbot_ui/common.py

@@ -38,7 +38,7 @@ def chat(
     try:
         response = client.chat.completions.create(
             # we select model from arch_config file
-            model="--",
+            model="None",
             messages=history,
             temperature=1.0,
             stream=True,

demos/shared/chatbot_ui/run_stream.py

@@ -54,13 +54,13 @@ def chat(
         if model_selector and model_selector != "":
             headers["x-arch-llm-provider-hint"] = model_selector
         client = OpenAI(
-            api_key="--",
+            api_key="None",
             base_url=CHAT_COMPLETION_ENDPOINT,
             default_headers=headers,
         )
         response = client.chat.completions.create(
             # we select model from arch_config file
-            model="--",
+            model="None",
             messages=history,
             temperature=1.0,
             stream=True,

demos/use_cases/llm_routing/arch_config.yaml

@@ -14,11 +14,6 @@ llm_providers:
     model: gpt-4o-mini
     default: true
 
-  - name: gpt-3.5-turbo-0125
-    access_key: $OPENAI_API_KEY
-    provider_interface: openai
-    model: gpt-3.5-turbo-0125
-
   - name: gpt-4o
     access_key: $OPENAI_API_KEY
     provider_interface: openai

demos/use_cases/orchestrating_agents/hurl_tests/simple_issues_repairs.hurl

@@ -0,0 +1,19 @@
+POST http://localhost:10000/v1/chat/completions
+Content-Type: application/json
+
+{
+  "messages": [
+    {
+      "role": "user",
+      "content": "I bought a package recently and it not working properly"
+    }
+  ]
+}
+HTTP 200
+[Asserts]
+header "content-type" == "application/json"
+jsonpath "$.model" matches /^gpt-4o-2/
+jsonpath "$.metadata.x-arch-state" != null
+jsonpath "$.usage" != null
+jsonpath "$.choices[0].message.content" != null
+jsonpath "$.choices[0].message.role" == "assistant"

demos/use_cases/orchestrating_agents/main.py

@@ -31,9 +31,10 @@ openai_client = openai.OpenAI(
 )
 
 
-def call_openai(messages: List[Dict[str, str]], stream: bool):
+def call_openai(messages: List[Dict[str, str]], stream: bool, model: str):
+    logger.info(f"llm agent model: {model}")
     completion = openai_client.chat.completions.create(
-        model="None",  # archgw picks the default LLM configured in the config file
+        model=model,
         messages=messages,
         stream=stream,
     )
@@ -53,14 +54,19 @@ def call_openai(messages: List[Dict[str, str]], stream: bool):
 
 
 class Agent:
-    def __init__(self, role: str, instructions: str):
+    def __init__(self, role: str, instructions: str, model: str = ""):
+        self.model = model
         self.system_prompt = f"You are a {role}.\n{instructions}"
 
     def handle(self, req: ChatCompletionsRequest):
         messages = [{"role": "system", "content": self.get_system_prompt()}] + [
             message.model_dump() for message in req.messages
         ]
-        return call_openai(messages, req.stream)
+
+        model = req.model
+        if self.model:
+            model = self.model
+        return call_openai(messages, req.stream, model)
 
     def get_system_prompt(self) -> str:
         return self.system_prompt
@@ -77,13 +83,17 @@ AGENTS = {
             "2. Quote ridiculous price\n"
             "3. Reveal caveat if user agrees."
         ),
+        model="gpt-4o-mini",
     ),
     "issues_and_repairs": Agent(
         role="issues and repairs agent",
         instructions="Propose a solution, offer refund if necessary.",
+        model="gpt-4o",
     ),
     "escalate_to_human": Agent(
-        role="human escalation agent", instructions="Escalate issues to a human."
+        role="human escalation agent",
+        instructions="Escalate issues to a human.",
+        # skipping model name here as arch gateway will pick the default model from the config file
     ),
     "unknown_agent": Agent(
         role="general assistant", instructions="Assist the user in general queries."