feat(assistant): Improve screen context processing with model selection
This commit is contained in:
@@ -315,12 +315,10 @@ class _ChatPageState extends ConsumerState<ChatPage> {
|
||||
// Clear the context so we don't process it again
|
||||
WidgetsBinding.instance.addPostFrameCallback((_) {
|
||||
ref.read(screenContextProvider.notifier).setContext(null);
|
||||
// Pre-fill the input or send a message
|
||||
// For now, let's just pre-fill the input with a prompt
|
||||
// TODO: Ideally we should add this as a system message or attachment
|
||||
final currentModel = ref.read(selectedModelProvider);
|
||||
_handleMessageSend(
|
||||
"Here is the content of my screen:\n\n$screenContext\n\nCan you summarize this?",
|
||||
null,
|
||||
currentModel,
|
||||
);
|
||||
});
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user