I write a blueprint about once every 2 years. While this may make me one of the least capable blueprint authors, it does provide perspective.
Blueprints do not seem to be advancing in usability anywhere near as fast as most of home assistant. They were hot stuff when they came out, but they are roughly the same stuff now (though actions/scripts have indirectly added a fair amount).
They are still very confusing (triggers can’t use full templates, if any templates, or sometimes templates), where and when inputs can be used vs variables vs trigger variables. Lots of weaknesses (and at best confusions) that could be addressed, but haven’t been. That’s not a complaint so much as begs the question:
Is there a HA roadmap for Blueprints? Or are they more like Zune - nothing exactly wrong with it, just left behind and forgotten?
Won’t stop me from, in 2028 or so, writing another one, but I am curious and can’t find a lot (of course “blueprint” is used a lot in the not-that-kind-of-blueprint way so it’s hard to search).
Or maybe the more interesting question: what are people using in place of them, if you are not using blueprints? I just made one for matter smoke detectors so I wouldn’t have to replicate so much code. Maybe I should have tried using scripts withing automations and variables?
Among who ? , i still haven’t got any, read through a few, but i stick to automation and templates
For various reason as you mention your self, and i have hard to believe they should be more “efficient”
And Many examples i’ve seen is humongous and Prone
ahh that was before my time, but a few year back i did saw a release with it mentioned , And seen a few "releases from people/users, i doubt HA Dev team will get to involved in this, as it can be a highly “individual” topic , depending upon devices, use case, etc
Edit: beside the Category to the left in here, there are a few people which have a dedicated github repos because they like to write Blueprints
When they first came out I found them to just be strange… they were often less convenient to set up than using the UI automation editor. Then they got added to the UI in late 2020, but you were still locked into whatever limitations the author built in.
Selectors have been expanded.
The ability to take them over has been added. So you can make modifications to the end script or automation to better fit your needs.
I haven’t seen anything official, but my feeling is that they have always been seen as more of an aid for new or low-skill users.
With the (currently experimental) addition of “purpose-specific” triggers and conditions and other additions to the Automation Editor, it seems the current focus is to make that easier to use and more powerful, which would naturally decrease the need for blueprints.
I use a few from the community, because they were well made, did what I needed… and I was feeling lazy. I’ve written a few too, but I rarely use those myself. For the most part the one’s I’ve written were for things I’ve seen people struggling with here on the forums. I haven’t developed the habit of checking the Exchange when I come up with an idea or find a need for an automation.
Creating an automation or script is so much easier than abstracting it to a blueprint that it seems kind of a waste of time to do it for something just for my own use. There doesn’t really seem to be much of an advantage to using them, in terms of code repetition, compared to using scripts or other tools like YAML anchors.
The start of template blueprints brought out the complete mess that templates had become. With years of multiple people touching the code, inconsistent ideas, and patched code it became very clear that the code had to be fixed before template blueprints could be generalized.
One of the community members has taken on that task and is currently working on it. Several of the more visible deprecating changes to HA in the last year have happened as a result so it is a slow process.
Blueprints themselves were for people (like me) that are members of the community that want to help others do some of the more complex scripting using scripts, automations.Templates are on the way. Now I would like to have a better system to distribute, and there is talk again to use HACS for this which would be awesome, I don’t see a lot else missing.
Thank you. That’s what I was looking for, some kind of “what are the developers planning” thoughts.
Template completeness is probably the worst, better selectors (e.g. wildcards) would be nice. The template issue may address the other big problem I have which is visibility into template expansion when I get an error. I frequently get something like a malformed entity ID on substitution but have no idea what it looked like. Just passing that out into the error message would be a huge help. It’s easier if they are actual runtime errors as they appear in the trace, but I spent an hour or more getting errors on the save with no clue until I saw the notes on templates not being valid in triggers (and similar for variables).
Just keep in mind the template blueprints to date have been disappointing to me. You set them up and they are kind of static.
In other words You load an input or something, and it builds the template, then that’s it. Changing the variable value doesn’t work the way I expect and won’t change again until restart when the template is created again from the template., or something. No dynamicing the template values after. I played a couple of times and decided either I’m not understanding or it’s not for me.
Me hoping it’s just a WIP or PEBKAC problem.
I decided to move my blueprint automation which was working to an automation. With templating I could then list all the smoke detectors (as opposed to one automation for each from a blueprint), and it was much more straightforward to manipulate the entity names into voice announcements and such.
Yes, I can see them for distribution of code. But if I’m writing the code, not better, harder and less flexible.
, i still haven’t got any, read through a few, but i stick to automation and templates