Refactor log message handling and improve word extraction in completions benchmark

workers/openai/worker.py

@@ -10,16 +10,6 @@ MODEL_LOG_FILE             = '/var/log/onstart.log'
 MODEL_HEALTHCHECK_ENDPOINT = "/"
 
 # Ollama-specific log messages
-MODEL_LOAD_LOG_MSG = [
-    "llama runner started in "
-]
-
-MODEL_ERROR_LOG_MSGS = [
-]
-
-MODEL_INFO_LOG_MSGS = [
-]
-
 def request_parser(request):
     data = request
     if request.get("input") is not None:
@@ -33,11 +23,14 @@ def completions_benchmark_generator() -> dict:
     WORD_LIST = []
 
     with open(__file__, 'r') as f:
-        for line in f:
+        # Use regex to extract words from the source code
-            WORD_LIST.extend(line.strip().split())
+        import re
+        source_code = f.read()
+        WORD_LIST = re.findall(r'\b\w+\b', source_code)
 
     prompt = " ".join(random.choices(WORD_LIST, k=int(250)))
     model = os.environ.get("MODEL_NAME")
+
     if not model:
         raise ValueError("MODEL_NAME environment variable not set")
 
@@ -77,9 +70,9 @@ worker_config = WorkerConfig(
         )
     ],
     log_action_config=LogActionConfig(
-        on_load=MODEL_LOAD_LOG_MSG,
+        on_load=["llama runner started in "],
-        on_error=MODEL_ERROR_LOG_MSGS,
+        #on_error=MODEL_ERROR_LOG_MSGS,
-        on_info=MODEL_INFO_LOG_MSGS
+        #on_info=MODEL_INFO_LOG_MSGS
     )
 )