Split up TorConnectionAssistViewModel for better readibility and performance.

Summary

Summarize the bug encountered concisely.

Instead of having big view models that do a lot and are always loaded, we should have small viewmodels that do more specific things and are loaded as needed

• torConnectionAssistViewModel.handleConnect
• torConnectionAssistViewModel.urlToLoadAfterConnecting
• torConnectionAssistViewModel.progress() (for if/when we make the progress bar show on all screens instead of just the bootstrap screen)