didnt test anything

llm_server/database.py

@@ -74,7 +74,7 @@ def log_prompt(ip, token, prompt, response, gen_time, parameters, headers, backe
     timestamp = int(time.time())
     conn = sqlite3.connect(opts.database_path)
     c = conn.cursor()
-    c.execute("INSERT INTO prompts VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
+    c.execute("INSERT INTO prompts VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)",
               (ip, token, opts.running_model, opts.mode, opts.backend_url, request_url, gen_time, prompt, prompt_tokens, response, response_tokens, backend_response_code, json.dumps(parameters), json.dumps(headers), timestamp))
     conn.commit()
     conn.close()

llm_server/llm/llm_backend.py

@@ -6,7 +6,7 @@ import flask
 class LLMBackend:
     default_params: dict
 
-    def handle_response(self, request: flask.Request, success, response: flask.Response, error_msg, client_ip, token, prompt, elapsed_time, parameters, headers):
+    def handle_response(self, success, request: flask.Request, response: flask.Response, error_msg, client_ip, token, prompt, elapsed_time, parameters, headers):
         raise NotImplementedError
 
     def validate_params(self, params_dict: dict) -> Tuple[bool, str | None]:

llm_server/llm/oobabooga/ooba_backend.py

@@ -9,7 +9,7 @@ from ...routes.helpers.http import validate_json
 
 
 class OobaboogaBackend(LLMBackend):
-    def handle_response(self, request, success, response, error_msg, client_ip, token, prompt, elapsed_time, parameters, headers):
+    def handle_response(self, success, request, response, error_msg, client_ip, token, prompt, elapsed_time, parameters, headers):
         backend_err = False
         response_valid_json, response_json_body = validate_json(response)
         try:

llm_server/llm/vllm/vllm_backend.py

@@ -17,7 +17,7 @@ from llm_server.routes.helpers.http import validate_json
 class VLLMBackend(LLMBackend):
     default_params = vars(SamplingParams())
 
-    def handle_response(self, request, success, response, error_msg, client_ip, token, prompt: str, elapsed_time, parameters, headers):
+    def handle_response(self, success, request, response, error_msg, client_ip, token, prompt: str, elapsed_time, parameters, headers):
         response_valid_json, response_json_body = validate_json(response)
         backend_err = False
         try:
@@ -41,7 +41,7 @@ class VLLMBackend(LLMBackend):
             #         f'HTTP CODE {response_status_code}'
             #     )
 
-            log_prompt(client_ip, token, prompt, backend_response, elapsed_time if not backend_err else None, parameters, headers, response_status_code, request.url, response_tokens=response_json_body.get('details', {}).get('generated_tokens'), is_error=backend_err)
+            log_prompt(ip=client_ip, token=token, prompt=prompt, response=backend_response, gen_time=elapsed_time if not backend_err else None, parameters=parameters, headers=headers, backend_response_code=response_status_code, request_url=request.url, response_tokens=response_json_body.get('details', {}).get('generated_tokens'), is_error=backend_err)
             return jsonify({'results': [{'text': backend_response}]}), 200
         else:
             backend_response = format_sillytavern_err(f'The backend did not return valid JSON.', 'error')

llm_server/routes/openai_request_handler.py

@@ -62,12 +62,12 @@ class OpenAIRequestHandler(RequestHandler):
         elapsed_time = end_time - self.start_time
 
         self.used = True
-        response, response_status_code = self.backend.handle_response(success, backend_response, error_msg, self.client_ip, self.token, self.prompt, elapsed_time, self.parameters, dict(self.request.headers))
+        response, response_status_code = self.backend.handle_response(success=success, request=self.request, response=backend_response, error_msg=error_msg, client_ip=self.client_ip, token=self.token, prompt=self.prompt, elapsed_time=elapsed_time, parameters=self.parameters, headers=dict(self.request.headers))
         return build_openai_response(self.prompt, response.json['results'][0]['text']), 200
 
     def handle_ratelimited(self):
         backend_response = format_sillytavern_err(f'Ratelimited: you are only allowed to have {opts.simultaneous_requests_per_ip} simultaneous requests at a time. Please complete your other requests before sending another.', 'error')
-        log_prompt(self.client_ip, self.token, self.request_json_body.get('prompt', ''), backend_response, None, self.parameters, dict(self.request.headers), 429, self.request.url, is_error=True)
+        log_prompt(ip=self.client_ip, token=self.token, prompt=self.request_json_body.get('prompt', ''), response=backend_response, gen_time=None, parameters=self.parameters, headers=dict(self.request.headers), backend_response_code=429, request_url=self.request.url, is_error=True)
         return build_openai_response(self.prompt, backend_response), 200
 
     def transform_messages_to_prompt(self):