Fix preset override issue - always use current UI values for execution

app_with_mcp.py

@@ -191,38 +191,33 @@ async def minion_respond_async(query: str, preset_model: str, api_type: str,
                              use_tools: bool):
     """Respond to query using specified configuration with optional MCP tools"""
     
-    # Build configuration based on preset or custom settings
+    # Always use the current UI values, regardless of preset selection
+    # Preset is only used for initializing UI fields, not for actual execution
+    llm_config_dict = create_custom_llm_config(
+        api_type, api_key, base_url, api_version, model, temperature, max_tokens
+    )
+    
     if preset_model != "Custom":
-        # Use preset as base, but allow UI overrides for temperature and max_tokens
-        config_obj = preset_configs.get(preset_model, default_config)
-        llm_config_dict = {
-            'api_type': config_obj.api_type,
-            'api_key': config_obj.api_key,
-            'base_url': config_obj.base_url,
-            'api_version': config_obj.api_version,
-            'model': config_obj.model,
-            'temperature': temperature,  # Always use UI value for dynamic parameters
-            'max_tokens': max_tokens,    # Always use UI value for dynamic parameters
-            'vision_enabled': config_obj.vision_enabled
-        }
-        print(f"🔧 Using preset '{preset_model}' with overrides:")
-        print(f"   - Temperature: {temperature} (preset: {config_obj.temperature})")
-        print(f"   - Max tokens: {max_tokens} (preset: {config_obj.max_tokens})")
-        print(f"   - Model: {config_obj.model}")
-        print(f"   - API Type: {config_obj.api_type}")
+        print(f"🔧 Using preset '{preset_model}' base with UI overrides:")
+        print(f"   - API Type: {api_type}")
+        print(f"   - Model: {model}")
+        print(f"   - Base URL: {base_url}")
+        print(f"   - API Version: {api_version}")
+        print(f"   - Temperature: {temperature}")
+        print(f"   - Max tokens: {max_tokens}")
     else:
-        # Use completely custom configuration
-        llm_config_dict = create_custom_llm_config(
-            api_type, api_key, base_url, api_version, model, temperature, max_tokens
-        )
         print(f"🔧 Using custom configuration:")
+        print(f"   - API Type: {api_type}")
+        print(f"   - Model: {model}")
+        print(f"   - Base URL: {base_url}")
+        print(f"   - API Version: {api_version}")
         print(f"   - Temperature: {temperature}")
         print(f"   - Max tokens: {max_tokens}")
-        print(f"   - Model: {model}")
-        print(f"   - API Type: {api_type}")
     
-    # Always rebuild brain with current configuration
+    # Always rebuild brain with current UI configuration
     print(f"🧠 Building brain with final config:")
+    print(f"   - Final API type: {llm_config_dict['api_type']}")
+    print(f"   - Final model: {llm_config_dict['model']}")
     print(f"   - Final temperature: {llm_config_dict['temperature']}")
     print(f"   - Final max_tokens: {llm_config_dict['max_tokens']}")
     brain = build_brain_with_config(llm_config_dict)