WIP: tunning

ai_service/main.py

@@ -32,12 +32,15 @@ async def message(
 
     # Build prompt (very simple example)
     prompt = f"User {payload.userId} said: {payload.content}\nBot:"
-    resp = openai.Completion.create(
-        model="text-davinci-003",
-        prompt=prompt,
-        max_tokens=150
+    chat_response = opeai.ChatCompletion.create(
+        model="gpt-3.5-turbo",
+        messages=[
+            {"role": "system", "content": "-"},
+            {"role": "user", "content": prompt}
+        ]
+        temperature=0.7,
     )
-    reply = resp.choices[0].text.strip()
+    reply = chat_response.choices[0].text.strip()
 
     # Cache reply for idempotency
     r.set(payload.eventId, reply, ex=3600)

ai_service/requirements.txt

@@ -1,5 +1,5 @@
 python-dotenv>=1.0.0
-openai
+openai>=1.0.0
 fastapi>=0.95
 uvicorn>=0.22
 redis>=4.5

matrix_service/main.py

@@ -86,9 +86,14 @@ async def main() -> None:
             }
             headers = {"Authorization": f"Bearer {AI_HANDLER_TOKEN}"}
             async with httpx.AsyncClient() as http:
-                resp = await http.post(f"{AI_HANDLER_URL}/api/v1/message", json=payload, headers=headers)
-                resp.raise_for_status()
-                data = resp.json()
+                try:
+                    resp = await http.post(f"{AI_HANDLER_URL}/api/v1/message", json=payload, headers=headers)
+                    resp.raise_for_status()
+                    data = resp.json()
+                except httpx.HTTPStatusError as e:
+                    logger.error(f"HTTP error: {e.response.status_code} - {e.response.text}")
+                except Exception:
+                    logger.exception("Error while calling AI handler")
             if data.get("reply"):
                 client.send_message(event["room_id"], data["reply"])