I've noticed that during HASS startup, __init__.py in a subdirectory of pyscript/apps/ can start running before other integrations have finished initializing. This makes sense, since I assume HASS dramatically improves its startup time by initializing integrations in parallel. I assume this is expected behavior.

In my __init__.py, I would like to sanity-check my app's configuration (whether that be in its YAML section or otherwise) by checking whether certain entity IDs referenced in the configuration actually exist, using, for example, state.exist(). This works if pyscript is reloading the app because the code has changed, but it won't work during HASS startup, because a given entity's integration might not have initialized the entity yet.

I can mitigate this issue for HASS startup specifically by putting the initialization code in a function decorated with @time_trigger("startup"). But then that function only runs when HASS is restarted. It won't run if pyscript merely reloads __init__.py because the Python code has changed.

Is there a way to get the best of both worlds--i.e., if HASS is starting up, have the function hold until everything has initialized, but otherwise to just run the function if the app has been reloaded?

I don't think I can wait-with-timeout for a start event from HASS, because the event probably won't be sent until pyscript finishes initializing, and that won't happen until my __init__.py finishes running. And that would be a lame solution anyway since it would impose a significant delay with each app reload. I've so happily ditched needing to reload YAML a million times when developing new code...

My fallback will be to delay handling nonexistent entities until they are actually used, but I was hoping to do a little better so that I can catch typos made when editing configuration when the typos are made, rather than some time possibly much later.

Thanks in advance for any ideas you consider sharing.