New guy doubting if HA is for him

[SgtFlippy]

I thought I’d make this a project post, but it’s a bit of introduction/questions/wondering. Very start of the project as I see it.

Some background: At the moment, almost everything is linked to Smartthings, but tied together with Sharptools. All my rules/automations are configured in Sharptools. But I do like the idea of full local control and I saw a few things that might be interesting to let my smart home grow. But I’m also having doubts, because there seems to be a lot of coding going on. I can follow guides, but all the coding I have done is mostly just that.
So about my current setup:
A short list of my device list, connected to an Aeotec hub:

• Frient sensors
• Sonoff mini switches
• MHcozy switches
• tablet running Sharptools

Then there’s a DELL Optiplex, running 4 camera’s through Blue Iris, this PC also runs AdGuard in a docker and 2 apps that checks the network for our phones and reports to Smartthings.

So on to my idea for the project:
Run HA on the DELL on a VM, for Zigbee antenna’s, I want 2 SLZB-06M, one for the house and one for the garage in my backyard.

I’m a bit scared I can’t fully replicate what I’m running at the moment. I use quite a lot of variables in Sharptools, which seems to be a bit more complicated in HA. I’m not sure if the presence detection will be as good as it is now. Not even sure if I would get the antenna’s running. I would like to add my philips TV (integration), was hoping to be able to add some Eurom heaters (I think I can link them through Tuya, they are not linked to my Smartthings at the moment), add the camera’s to a dashboard and I searched to add my Cielo Breez, but can’t find anything on those.

I must say I will soon try the install and have a look around, but haven’t gotten that far yet.

[tom_l]

Moved. The share your project category is for showing off projects you have completed. Please read the category descriptions.

[SgtFlippy]

Ah yes, missed the finished part. Sorry!

[Edwin_D]

I came from a different home automation system that could do pretty much anything you need with no coding whatsoever. Home assistant can do the same things without any coding using the automation editor, dashboard editor and helper entities. So you won’t run into any problems soon.

But you can do so much extra fun stuff with a little “coding” (yaml configurations, templates, mostly scripting) that it will be hard to resist. You can get along following examples for the forum for a long time, you’ll get to learn how they work in due time, and you’ll come to appreciate and use it.

The thing to consider though, you’ll likely spend a lot of time playing around, setting things up, tweaking, finding there’s so much that you could have done even nicer, … So if you’re willing to put some time in it (it is fun to do) go for it. If you’re only in it for turning on the light when some one moves, you can find less time consuming systems (with more limitations).

[SgtFlippy]

I do know the concept of the rabbit hole it is. It all started with 2 wifi plugs because I hated the classic “clock” type plugs that didn’t sync up when to turn the lights on.

At the moment I have solar timers for my outside lighting, depending on presence inside as well. A variable for Christmas to run the lights differently, an alarm system, a sliding gate and garage doors which open depending on selection at the opening of either the back door or garage door and bluetooth buttons in the car to open them remotely.
The tablet says morning/night, I can send snapshots of the camera’s and alarms to Pushover (my work network is very strict), …

So I guess not super advanced, but more than just turning a light on and off.

[Edwin_D]

Then HA is the best rabbit hole you’ll find
Seriously, if you invest a lot in something else, find out you want more, and then come back to HA, you’ll be spending even more time in vain.

[JulianDH (Julian Hall)]

A rabbit hole that keeps on giving. Perfect for me. But I can see that this ecosystem can require a bit of effort. Lots and lots of very good video tutorials

[Edwin_D]

One caviat though. The good thing is Home Assistant keeps improving on a monthly basis. Especially to avoid having only yaml configuration and improve the UI. The downside is that even pretty recent video tutorials may use an old version of the UI. So things may not look exactly the same, or the function you need might be moved to a more sensical location.

[NathanCu (Nathan Curtis)]

Also biggest hint as you migrate…

… forget everything SmartThings taught you about how smart home hubs work. They do a lot behind the scenes for end users that honestly tends to lead to very bad things of you do the same thing in HA.

For instance never run a ZWave full network heal just because. Yes that was the goto fix in ST but in HA you’ll be waiting an hour for your network to recover because it just works differently. The differences are enough where if you make SmartThings assumptions you get bad results.

So, go in planning the results (what you want done) not how you do it with SmartThings. Good luck!

[alex3m3ti8]

Howdy. I switched from SmartThings to HomeAssistant back in March 2021. I was heavily invested in SmartThings and WebCore and had a ton of automations in there (146 pistons). When SmartThings announced that the IDE was changing and WebCore was slated for shutdown I started my journey over to HomeAssistant.

Since HomeAssistant can connect to SmartThings, I found that I was able to install HomeAssistant, connect it to SmartThings, get to understand things like entities, services, and start to learn how to automate. All without shutting down anything in SmartThings initially. At the time, a more seasoned HomeAssistant friend of mine suggested that for the complexity of automations I already had that I would probably be better served using NodeRed, but certainly could put some automations natively within HomeAssistant. He was 100% right.

I outputed all my pistons in WebCore, and painstakingly evaluated and moved one at a time from WebCore to NodeRed. Did the easiest/simplest ones first so I could learn and debug. Once I got my desired result, I disabled the WebCore piston. It took me a good two months of learning, trial and error, searching forums to get there, but I did it. Once I converted all my pistons and was able to shutdown WebCore, I then started on the process of decoupling from SmartThings by moving things that were in SmartThings over to native control in Home Assistant.

When I started, I would say that debugging in NodeRed was a bit easier and fit my brain better, there have been vast improvements to the native automations engine in HomeAssistant since then. I would also add that learning Templates helped tremendously, but a lot of it was learning YAML.

My personal opinion/breakdown of automation methods would be (your milage may vary):

NodeRed - Great for visual thinkers, excellent debugging, great capabilities for reusing code and especially good at highly complex automations which may have lots of branches, conditions or variable usage. For example, I have a ton of helper flows where I pass variables to it and reuse code and minimize my rewrites. Will add that NodeRed is pretty terrible for editing on a mobile platforms however.
HA Automation - Really well executed, and constantly improving. Would say that very simple to mid-tier complex automations. Much easier to edit and modify on mobile platform.
Templates - Great for simple tasks (i.e. change this to this, or reformat this to this), but really need to learn YAML. About halfway through my WebCore to NodeRed migration I started to figure out Templates and was able to eliminate code in NodeRed using templates. But learning how to do even the simplest in NodeRed first was really helpful to get me started.

You expressed concern that you won’t be able to replicate what you’ve already done in Sharptools. Having been on the HA train for 3 years, I find that nearly implausible. There are so many kind helpful people here who see a problem and take the time to guide people through it or just get cracking on the code and fix it. Just browsing through the forums it seems like every problem there’s five or six ways to approach the solution and a ton of creative folks willing to help. I remember when I started to see the WebCore community whither that I knew I needed to start migration; the vibrancy of the HA community reinforces my bias that I made the right choice.

Lastly I will say the first time I did something natively in HA (without SmartThings) and saw the speed and performance, I was amazed. Everything is so much snappier when it doesn’t have to go out SmartThings cloud and back. Native Local control FTW.

[PecosKidd]

Post of the day! Brilliant!

[FriedCheese (Ryan)]

@SgtFlippy

lot of coding going on

I was under this same impression coming in and quickly realized, it a lot of cases, I was wrong. You might see things like someone asking for help with automation. You’ll end up seeing a lot of code blocks like this:

alias: Pantry Light with Motion
description: ""
trigger:
  - platform: state
    entity_id:
      - binary_sensor.pantry_motion_sensor_occupancy
    from: "off"
    to: "on"
    id: motionActive
  - platform: state
    entity_id:
      - binary_sensor.pantry_motion_sensor_occupancy
    from: "on"
    to: "off"
    id: motionInactive
    for:
      hours: 0
      minutes: 1
      seconds: 0
condition: []
action:
  - choose:
      - conditions:
          - condition: trigger
            id:
              - motionActive
        sequence:
          - service: switch.turn_on
            target:
              entity_id: switch.pantry_light_switch
            data: {}
      - conditions:
          - condition: trigger
            id:
              - motionInactive
        sequence:
          - service: switch.turn_off
            metadata: {}
            data: {}
            target:
              entity_id: switch.pantry_light_switch
mode: restart

So, that makes you think “Man, this is a lot of ‘coding’ to get my head wrapped around.” What took me a little bit to realize is that everything above is configurable, and likely was configured, from the UI:

But, for getting help, it’s a lot better to post the raw YAML (since the UI is just filling out the YAML for you) since anyone can copy/paste into the right area. If you just rely on screenshots of the rules, then that simple automation I posted above would take four screenshots and would require being configured from scratch to emulate.

You can toggle between the UI and YAML in quite a number of places:

So, if you need help with something, you can use that option to flip to YAML, copy/paste the code in a forum post, and, if someone makes the changes for you, you’ll just have to copy/paste back. At any point, you can flip back to the UI config screen and are able to see what the YAML is doing from the UI side.

[SgtFlippy]

Oooh, that makes it a lot clearer, yes!
I thought every time I looked something “a little more advanced”, I’d stumble upon code like that and my head went: this is going over me
It’s something I’m able to read, but wasn’t ready to do everything like that.

@alex3m3ti8 That’s a very good explanation, makes me want to start, thanks!

PS: just ordered the Zigbee sticks. Will report back in a few weeks

[Didgeridrew]

If you need a primer on how to transcribe YAML automations that you find here or elsewhere using the UI Automation Editor, there is a (now slightly dated) tutorial available on the Resin Chem Tech Youtube channel.

[SgtFlippy]

Spent the day having a look around. Added both our phones for presence, added my Honeywell thermostat, cielo Breez, camera’s are added and the TV.

Tried my luck at adding the FGlair integration, but can’t get it to work at the moment.

Not looking at automations yet, by the end of the month the Zigbee antennas should arrive and when testing those, will give me more insight.

The thing I keep running up against, is when searching for integrations, there’s years of discussion, adding and adding and in the end, a fresh user has no idea where to start. I also see a lot of code or scripts being tossed around, without clear instructions where they even go. Simple configurations have instructions that become difficult because you have to enter data and it misses to tell you where to find it.

I know home assistant has a reputation to be a bit more advanced, but I feel these things are easily fixed and would attract a lot more people.
I’m not giving up, I really want to try out the antennas, but the first meeting is no smooth road for now While setting up the cameras I lost connection 5 times. After the setup it kept up fine.

[alex3m3ti8]

You aren’t wrong about your observations. Welcome to Home Assistant. Something to note, there’s a very clear difference between supported integrations and custom integrations.

The supported integrations (such as the one you installed for SmartThings or possibly your Honeywell Thermostat, as well as major platforms like Hue, ZigBee or Z-Wave) are going to be supported right out of the box and are subject to approval to merge from the Home Assistant developers in the github repository. These typically are monitored for things like how many folks are using it, what kinds of errors may be going on, did the vendor change the API etc, and get attention from a team of folks. These integrations also have to adhere to set of standards, like clear documentation, multi-language support, general supportability, and avoid doing things like scraping websites or other things that just don’t work well or might make Home Assistant a target for some company’s lawyers which no-body wants. Just look through this file and you’ll see the people who are marked as the person or persons who originally wrote and or support the integration. Several integrations have multiple people meaning more than one person helps to maintain that integration or approve changes. The full list of supported integrations are here.
Integrations - Home Assistant (home-assistant.io)

When you get to custom integrations (the ones you install through HACS like the FGlair Integration for example), then it’s a little more wild, wild west. In those cases, the developer might just be one person, a hobbiest like many of us, and may not have the time, ability or desire to write clear concise documentation and may assume users are advanced and know and understand things or take things for granted. I’ve seen it all the time. Thing is, most of the time, you can get a hint about an integration based on things like how many open issues are there, are they getting closed or addressed, when was the last PR merge, tells you a lot about an integration and whether or not it is still maintained or not. If it is maintained, I think you’ll be surprised how starting a question with “VERY new to HA here, but was trying to get this working, and I tried this and this and…” will illicit a friendly and helpful response. Not always, but most folks are pretty eager to help a newbie out. Just be clear in what you are doing, help them see what you tried and someone is likely going to try and help.

For example, I looked up FGLair and I found this community post
Home Assistant support for Fujitsu General Airconditioners [FGLair] - Share your Projects! - Home Assistant Community (home-assistant.io)
which lead me to this github repo

pyfujitsu_for_homeassistant/README.md at master · Mmodarre/pyfujitsu_for_homeassistant (github.com)

So this is definitely and advanced thing. First off, this particular repo hasn’t been updated for 5 years. That’s probably a red flag #1. Second, if you look at the forks you can see that there is one fork which was last updated 10 months ago. Third, if you look at the issues, you’ll see the latest issue has a solution which points you to bigmoby/fglair_for_homeassistant: FGLair heat pump controller integration (github.com)

This one was last updated last month, so is likely gonna be at least somewhat working. With custom integrations this happens all the time, the OP made a thing, got bored with it, hasn’t updated it for 5 years, and someone else forked it, took it over and is doing their own thing with it. Unfortunately, you are right, this isn’t beginner stuff and does take a bit of digging to figure out and requires some knowledge about github to get a hunch about a lot of these things.

The difference between the supported and the custom is that there’s simply not enough time to allow every custom integration to become supported. A lot of them do, but they typically have to go through code review and approval, meet the threshold of standards etc before it can get merged in, and a lot of folks just don’t wanna spend that time to do understandably. So lots of folks leave it as custom and maintain it for themselves. But custom integrations tend to get updated very quickly because they don’t require the amount of additional code review or approval to get published. Lots of custom integrations I’ve seen a bug filed and the fix is available within hours.

What I’ve found from my multi-year journey to Home Assistant is now before I make a purchase, I check Home Assistant Integrations and look for a standard local (non-cloud dependent) integration and that really does sway my purchase decision. I also look at the stats and see how many folks have that integration installed, that typically gives me an additional sense of security that I’m not going on my own into something. I get how this isn’t possible in retrospect and what you have in your house for HVAC isn’t something you can simply change overnight.

Sorry for the long ass post, but I hope as a new person it helps to clarify some of the challenges you’ve observed.

[nickrout]

If you are looking for integrations, look at the integrations page Integrations - Home Assistant

As you’ll see there is no fglair integration in home assistant. There may be a custom integration. If you want to ask about a custom integration, ask in the GitHub repo for that custom integration.

[SgtFlippy]

It comforts me It’s not just me that sees this.
With Smartthings, Groovy or now Edge, the thing was when you found an integration it was a one click install and it either worked or it didn’t. Spending hours looking at code, finding some other code, trying again, read further, find some code edits, … Makes the task a bit more tedious.

It also makes it more difficult to simply say: oh well, it doesn’t work Now I keep searching and thinking, this has to work! Even though at the moment I have the same devices integrated in a rather similar way as I have with Smartthings. So can’t really be mad about it. The FGlair thing would simply be great because I have 3 of them and only 1 Cielo to control 1 of them. Not having to buy 2 more and simply integrate the other 2 would be awesome.

I should maybe have waited for the sticks to start all this. Adding the Zigbee devices and making them work, I assume will be easier and would have been a tad more satisfying. Now I’m stuck with messing with the custom integrations untill they arrive

PS: as in that FGlair github, it states it needs the python library. That’s probably the reason it’s not working. Where do I put that?
I have the Samba connection for file exploring, I have added components and added fglair to the config. So I know how to edit those and return to original.

OMG I found it, got all 3 A/C’s on my dashboard

[FriedCheese (Ryan)]

Another thing to look out for…Helpers.

My read is that most of the Helpers that can be setup started as template sensors that had very common use cases. So, when you go looking for “how do I combine temp sensors to get an average” or “how do I see the rate of change of a humidity sensor” you’re likely to find a million posts with setting up template sensors. Both of these can now be done through Helpers that are simple to setup from the UI.

Part of that will be to be cognizant of the date that a post was made (I know I’ve ran across some from years ago where creating an entity group was only available through by updating configuration.yaml). The other part is just checking (like was alluded to above) to see if there’s already a baked in capability to do what you want.