Fair enough. I understand that. Different needs - different solutions.
Well my lights don’t come on during the day because they only come on if the light level in that part of the house is low enough to warrant it, using lux sensors.
…but all this is off topic.
People calm down. This is a thread that has become about preferences. Those preferences are personal. Personal choices don’t need to be criticised or even entirely justified. If people like NR, let them. If people prefer HA, let them. Just stop bloody bickering.
2 Likes
That’s not complex…
This has all become so off topic that I make no apologies…
What sensors do you use? I’m hoping based on previous contributions that you use something cheap and non-proprietary (which I can copy 
)
I’m not the one you asked, but I can highly recommend the Xiaomi Lux sensors together with a zigbee stick.
1 Like
No…
One function node and one call service node, that is all.
One function node can get all the states in HA at once
Show me the template to read the .storage folder and output the learned remote commands.
What is not possible with Node red?
Reload xy after you changed something in configuration.yaml, what this topic originally was about before it became about node-red being the best thing since sliced bread…
Why would you need to reload configuration?
That was the whole point of Node red, you don’t need to reload anything, just deploy on the same screen view as you were just working on and everything updates.
You don’t need to open configuration and server controls and so on.
So I don’t get it…
I can do the same with VSCode and a simple key combination, but that’s not the topic.
Can you customize an entity with Node-Red?
Can you create/modify a template switch, light, cover?
Can you create/modify input boolean/select/text/number?
Can you create/modify scenes?
Can you create/modify groups?
Can you create/modify notify services?
Can you create/modify zones
Most of that, yes.
The rest, not sure, but if there is a call service in HA for it then it can be done since they share call service.
But then again I have a hard time believing you can create a template switch with an automation.
Just as you said you don’t like when people suggest node red all the time, the same goes the other way, you understand that right?
When someone claims you can’t create a sensor, or that you need multiple nodes to do a simple automation. I mean it’s not like the yaml writes itself is it?
You still need to write - delay: 00:00:05.
How is that different from dragging out a node?
Or clicking the add action in the GUI.
All I’m saying is, if it bothers you that people suggest Node red then perhaps you can imagine how it bothers some that people claim it’s impossible to create sensors in node red.
And now you say you use VScode, isn’t that kind of the same as using Node red then?
You add an integration and use something not core?
You just argued that you don’t need anything but yaml.
I never said that this is possible with HA automations, this whole topic is not about automations, it’s about why HA needs to be restarted for lots of things. And my problem is that Node-Red was suggested as a solution to this issue, while it is not capable of solving the issue at hand.
The things I listed are things that you configure in YAML (or some of them also from the UI) and can be reloaded now, meaning they don’t need a restart anymore.
VSCode is my YAML editor, so yes I don’t need anything but YAML.
Anyway, this discussion is pointless and I don’t feel like arguing anymore.
This is the last thing I will say about this - but PLEASE anyone reading this topic -
NEVER do this, especially not if you are on something like a Pi. It might seem tempting, but asking Home Assistant for the state of EVERY entity it knows about even though you only plan to use a few of them - is incredibly wasteful on resources, and once that information is in NodeRed - it’s a RAM hog, until you discard them, or the flow has completed. If you do this on several flows, you are likely to start encountering problems.
This is the equivalent of throwing a bucket of paint at a wall, when you only want to paint 3 bricks.
Show me one reason why I would ever want to do that much less do it regularly.
To find out what you named the commands.
This is in my opinion a big flaw when moving from switch remotes to the “new” learn type.
You can’t find the commands you learned, the only way is to open the file in the storage folder.
Or you create a sensor that has the names of all the commands that have learned.
I used to have to open that file every time I wanted to use any of the commands because I never remember if it’s TV_livingroom_power or power_on_livingroom_TV or what I named it.
So that is a reason, now show me the template.
If that is how it works then I agree.
But since I use a raspberry and don’t even notice it I believe the states is already in there.
If it’s not then I have to say that it’s incredible fast.
But those states is not much more than an average json string which is used in many other integrations.
Just look at the Waze API.
It returns an object with every segment along the way you request a driving time for.
Each segment array contains name of street, distance and driving time of the segment.
That is an large object, and HA just sums up the times and discsrds the rest.
I’m fairly sure that is much larger than the states.
So create a sensor. There is never a good reason to be screwing with .storage files particularly when there is a solution.
I actually use Aeotec multisensors (ZWave), so unfortunately not
However you could either use Xiaomi as Burningstone suggested or make your own with ESPhome.
1 Like