With version 0.108 released we saw plenty more integrations available to set up from the UI.
We also saw the Unifi manual integration being deprecated.
I was wondering if Home Assistant is gradually moving away from manual (yaml) configurations, to UI configurations?
I’m sure that’s fine for the majority of users, but some of you will agree that manual configuration is a much more satisfying learning experience, providing a lot more inside on how the integration’s options work, and perhaps giving a more granular way of configuring the integration.
Then again, I’m guessing the UI integrations also save their configurations in a file (xml?) that can be reviewed and tinkered with?
Anyway, I’m just thinking out loud here.
Two questions in this post, if you know their answers.
Thanks.
Yes and yes.
It’s been known for a while that a lot of things will be moved over to the UI. This way the platform will become more accessible to a broader audience.
The config is saved in yaml JSON files. But… It’s definitely made NOT to edit by hand. It’s located in .storage folder in your config folder.
the important question that I haven’t seen answered: how is this type of configuration supposed to be backed up? HomeAssistant doesn’t offer any of this functionality out of the box… what happens in the event of data storage failure? Start over?
Having YAML based configuration in a version controlled repository is extremely useful. I’m fine with this direction as long as there is ability to back up the settings somewhere, automatically and periodically.
Shit happens. And when it does it is a lot more difficult to fix the .storage files than YAML.
We were promised that YAML would ‘not be going away’. But significantly - not that everything could continue to be done with it. It could come to a point where the only thing you can configure with YAML is your server name and that initial statement would still be true.
But that’s the way it’s going to be and bitching and moaning isn’t going to change anything.
I’m sure you would be welcome to fork the code (though maybe I’m wrong)?
Personally I approve of these changes because I think it makes HA more marketable and easier to use for newer people.
Most of the integrations that can be configured in the GUI are tightly managed and are being actively maintained which is exactly what we need (as the maintainer keeps up with changes in the end product).
You note that “shit happens” and I don’t disagree, but what sort of “fixing” would you be doing to the .storage folder that couldn’t be fixed by tweaking in the GUI or by restoring your latest snapshot? I’m honestly asking, not trying to be an ass. Change my mind.
You are not wrong at all and you know it. You are correct. Anyone is free to fork, this is really open source.
Look closely at the competency of the current code developers. Yah, they smart coders. And I’m sure they have considered the ramifications of alienating their existing userbase vs potential gains.
Duplicate entities for one. If it has not happened to you yet you do not know the pain in the arse it is to rectify.
However, as I said, pointless arguing about it. It’s going to continue to happen without consultation with the usrebase… STFU or fork it yourself. I think we agree. But I am very drunk and don’t really give a shit. I could quite easily pull the whole fucking thing and go back to being a Luddite that has to check an empty mailbox IRL.
HA HA, didn’t mean to fire you up while you’re drunk so sorry for that.
As for duplicate entities, I do think that’s less likely (impossible?) when you configure via the integrations panel as the system is designed to prevent that from happening.
The bottom side of her was allright. The top bit though <shudder>. Think ‘toxic avenger’ in female form.
Right that’s it. I’m done being a drunken exist pig and going to bed and look forward to cursing my over indulgent past self tomorrow. Hopefully I don;t have to delete too many posts.
