From e2f552a1c504c4f63bd879793633b9e47a499521 Mon Sep 17 00:00:00 2001
From: Adam Wilson <lightbrok3r@gmail.com>
Date: Fri, 30 May 2025 08:40:38 -0600
Subject: [PATCH] initial pass at logging service

---
 src/text_generation/entrypoints/http_api.py   | 35 -------------------
 .../entrypoints/http_api_controller.py        | 10 ++++--
 src/text_generation/entrypoints/server.py     | 13 ++-----
 .../service/logging_service.py                | 10 ++++++
 4 files changed, 21 insertions(+), 47 deletions(-)
 delete mode 100644 src/text_generation/entrypoints/http_api.py
 create mode 100644 src/text_generation/service/logging_service.py

diff --git a/src/text_generation/entrypoints/http_api.py b/src/text_generation/entrypoints/http_api.py
deleted file mode 100644
index 5f0225bc3..000000000
--- a/src/text_generation/entrypoints/http_api.py
+++ /dev/null
@@ -1,35 +0,0 @@
-# """
-#     Usage:
-#         $ uvicorn src.api.http_api:app --host 0.0.0.0 --port 9999
-# """
-
-# from fastapi import FastAPI
-# from pathlib import Path
-# from pydantic import BaseModel
-# from src.llm.llm import Phi3LanguageModel
-
-
-# STATIC_PATH = Path(__file__).parent.absolute() / 'static'
-
-# app = FastAPI(
-#     title='Phi-3 Language Model API',
-#     description='HTTP API for interacting with Phi-3 Mini 4K language model'
-# )
-
-# class LanguageModelPrompt(BaseModel):
-#     prompt: str
-
-# class LanguageModelResponse(BaseModel):
-#     response: str
-
-
-# @app.get('/', response_model=str)
-# async def health_check():
-#     return 'success'
-
-
-# @app.post('/api/conversations', response_model=LanguageModelResponse)
-# async def get_llm_conversation_response(request: LanguageModelPrompt):
-#     service = Phi3LanguageModel()
-#     response = service.invoke(user_input=request.prompt)
-#     return LanguageModelResponse(response=response)
diff --git a/src/text_generation/entrypoints/http_api_controller.py b/src/text_generation/entrypoints/http_api_controller.py
index 420255a77..d883979cc 100644
--- a/src/text_generation/entrypoints/http_api_controller.py
+++ b/src/text_generation/entrypoints/http_api_controller.py
@@ -3,9 +3,11 @@ import traceback
 
 from src.text_generation.adapters.llm.llm import Phi3LanguageModel
 from src.text_generation.adapters.llm.llm_rag import Phi3LanguageModelWithRag
+from src.text_generation.service.logging_service import LoggingService
 
 class HttpApiController:
     def __init__(self):
+        self.logger = LoggingService(filename='text_generation.controller.log').logger
         self.routes = {}
         # Register routes
         self.register_routes()
@@ -67,8 +69,10 @@ class HttpApiController:
         data = self.get_service_response(prompt)
         response_body = self.format_response(data)
         
+        http_status_code = 200 # make enum
         response_headers = [('Content-Type', 'application/json'), ('Content-Length', str(len(response_body)))]
-        start_response('200 OK', response_headers)    
+        start_response(f'{http_status_code} OK', response_headers)
+        self.logger.info('non-RAG response', request_body, http_status_code, response_body)
         return [response_body]
 
     def handle_conversations_with_rag(self, env, start_response):
@@ -91,8 +95,10 @@ class HttpApiController:
         data = self.get_service_response_with_rag(prompt)
         response_body = self.format_response(data)
         
+        http_status_code = 200 # make enum
         response_headers = [('Content-Type', 'application/json'), ('Content-Length', str(len(response_body)))]
-        start_response('200 OK', response_headers)    
+        start_response(f'{http_status_code} OK', response_headers)
+        self.logger.info('RAG response', request_body, http_status_code, response_body)
         return [response_body]
 
     def __http_200_ok(self, env, start_response):
diff --git a/src/text_generation/entrypoints/server.py b/src/text_generation/entrypoints/server.py
index 665be9f9b..f8081944b 100644
--- a/src/text_generation/entrypoints/server.py
+++ b/src/text_generation/entrypoints/server.py
@@ -1,17 +1,11 @@
-import json
-import logging
-
 from src.text_generation.entrypoints.http_api_controller import HttpApiController
+from src.text_generation.service.logging_service import LoggingService
 from wsgiref.simple_server import make_server
 
 
 class RestApiServer:
     def __init__(self):
-        pass
-
-    def post_response(self, env, start_response):
-        start_response('200 OK', [('Content-Type', 'application/json')])
-        yield [json.dumps({'received': 'data'}).encode('utf-8')]
+        self.logger = LoggingService(filename='text_generation.server.log').logger
 
     def listen(self):
         try:
@@ -21,8 +15,7 @@ class RestApiServer:
                 print(f'listening on port {port}...')
                 wsgi_srv.serve_forever()
         except Exception as e:
-            logging.warning(e)
-
+            self.logger.debug(e)
 
 if __name__ == '__main__':
     srv = RestApiServer()
diff --git a/src/text_generation/service/logging_service.py b/src/text_generation/service/logging_service.py
new file mode 100644
index 000000000..a8a79a230
--- /dev/null
+++ b/src/text_generation/service/logging_service.py
@@ -0,0 +1,10 @@
+import logging
+
+class LoggingService:
+    # TODO use base class and make this file logging service
+
+    def __init__(self, filename):
+        logger = logging.getLogger()
+        logger.setLevel(logging.DEBUG)
+        logger.addHandler(logging.FileHandler(filename))
+        self.logger = logger