Use large github action machine to run e2e tests (#230)

model_server/.vscode/launch.json

@@ -9,7 +9,7 @@
       "type": "debugpy",
       "request": "launch",
       "module": "uvicorn",
-      "args": ["app.main:app","--reload", "--port", "8000"]
+      "args": ["app.main:app","--reload", "--port", "51000"]
     }
   ]
 }

model_server/app/cli.py

@@ -1,3 +1,4 @@
+import importlib
 import sys
 import os
 import time
@@ -7,6 +8,15 @@ import tempfile
 import subprocess
 import logging
 
+
+def get_version():
+    try:
+        version = importlib.metadata.version("archgw_modelserver")
+        return version
+    except importlib.metadata.PackageNotFoundError:
+        return "version not found"
+
+
 logging.basicConfig(
     level=logging.INFO,
     format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
@@ -14,6 +24,7 @@ logging.basicConfig(
 
 log = logging.getLogger("model_server.cli")
 log.setLevel(logging.INFO)
+log.info(f"model server version: {get_version()}")
 
 
 def run_server(port=51000):
@@ -37,8 +48,9 @@ def run_server(port=51000):
 def start_server(port=51000):
     """Start the Uvicorn server"""
     log.info(
-        "Starting model server - loading some awesomeness, this may take some time :)"
+        "starting model server - loading some awesomeness, this may take some time :)"
     )
+
     process = subprocess.Popen(
         [
             "python",
@@ -61,7 +73,7 @@ def start_server(port=51000):
         log.info(f"Model server started with PID {process.pid}")
     else:
         # Add model_server boot-up logs
-        log.info("Model server - Didn't Sart In Time. Shutting Down")
+        log.info("model server - didn't start in time, shutting down")
         process.terminate()
 
 

model_server/app/function_calling/model_utils.py

@@ -67,12 +67,16 @@ async def chat_completion(req: ChatMessage, res: Response):
         f"model_server => arch_function: {client_model_name}, messages: {json.dumps(messages)}"
     )
 
-    resp = const.arch_function_client.chat.completions.create(
-        messages=messages,
-        model=client_model_name,
-        stream=False,
-        extra_body=const.arch_function_generation_params,
-    )
+    try:
+        resp = const.arch_function_client.chat.completions.create(
+            messages=messages,
+            model=client_model_name,
+            stream=False,
+            extra_body=const.arch_function_generation_params,
+        )
+    except Exception as e:
+        logger.error(f"model_server <= arch_function: error: {e}")
+        raise
 
     tool_calls = const.arch_function_hanlder.extract_tool_calls(
         resp.choices[0].message.content

model_server/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "archgw_modelserver"
-version = "0.0.4"
+version = "0.0.5"
 description = "A model server for serving models"
 authors = ["Katanemo Labs, Inc <archgw@katanemo.com>"]
 license = "Apache 2.0"