Update model names and configuration parameters for improved performance

src/chat.py

@@ -8,6 +8,7 @@ from ollama import AsyncClient, ChatResponse, Message
 from .config import (
     MAX_TOOL_CALL_DEPTH,
     MODEL_NAME,
+    EMBEDDING_MODEL_NAME,
     NUM_CTX,
     OLLAMA_HOST,
     SYSTEM_PROMPT,
@@ -28,6 +29,7 @@ class ChatSession:
         session: str = "default",
         host: str = OLLAMA_HOST,
         model: str = MODEL_NAME,
+        embedding_model: str = EMBEDDING_MODEL_NAME,
     ) -> None:
         init_db()
         self._client = AsyncClient(host=host)

src/config.py

@@ -3,10 +3,11 @@ from __future__ import annotations
 import os
 from typing import Final
 
-MODEL_NAME: Final[str] = os.getenv("OLLAMA_MODEL", "qwen3")
+MODEL_NAME: Final[str] = os.getenv("OLLAMA_MODEL", "qwen3:1.7b")
+EMBEDDING_MODEL_NAME: Final[str] = os.getenv("OLLAMA_EMBEDDING_MODEL", "snowflake-arctic-embed:137m")
 OLLAMA_HOST: Final[str] = os.getenv("OLLAMA_HOST", "http://localhost:11434")
 MAX_TOOL_CALL_DEPTH: Final[int] = 5
-NUM_CTX: Final[int] = int(os.getenv("OLLAMA_NUM_CTX", "32000"))
+NUM_CTX: Final[int] = int(os.getenv("OLLAMA_NUM_CTX", "16000"))
 
 SYSTEM_PROMPT: Final[str] = (
     "You are a versatile AI assistant named Starlette able to orchestrate several tools to "

src/vm.py

@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-from typing import Optional
 import subprocess
 import uuid