Skip to content
  • Categories
  • Recent
  • Tags
  • Popular
  • Unsolved
Collapse
Discussion Forum to share and further the development of home control and automation, independent of platforms.
  1. Home
  2. Software
  3. Multi-System Reactor
  4. [SOLVED] Lights on all month UNTIL fourth Thursday of the month
Local expression in Rule does not evaluate as they used to do
CrilleC
Topic thumbnail image
Multi-System Reactor
Home Assistant 2025.11.2 and latest-25315
CrilleC
Topic thumbnail image
Multi-System Reactor
Reactor (Multi-System/Multi-Hub) Announcements
toggledbitsT
Build 21228 has been released. Docker images available from DockerHub as usual, and bare-metal packages here. Home Assistant up to version 2021.8.6 supported; the online version of the manual will now state the current supported versions; Fix an error in OWMWeatherController that could cause it to stop updating; Unify the approach to entity filtering on all hub interface classes (controllers); this works for device entities only; it may be extended to other entities later; Improve error detail in messages for EzloController during auth phase; Add isRuleSet() and isRuleEnabled() functions to expressions extensions; Implement set action for lock and passage capabilities (makes them more easily scriptable in some cases); Fix a place in the UI where 24-hour time was not being displayed.
Multi-System Reactor
Notice to Docker + ARM Users (RPi 3/4/5 and others)
toggledbitsT
This post does not apply to users of Intel/AMD-based systems. If you are using a Reactor image tagged latest-amd64 or stable-amd64, then this post does not apply to you. It also does not apply to bare-metal installs; it's for users of docker images on ARM-based systems only (principally Raspberry Pi hosts, but could be others). After January 15, 2026, I will no longer produce the aarch64-tagged docker image for Reactor. The ARM images will be arm64 for 64-bit operating systems, and armv7l for 32-bit operating systems. For those of you running a container from the aarch64 image today, this will be a relatively simple change: you just need to switch the image used for your docker container to a differently-tagged image. If you are using docker-compose, then this is a relatively simple matter of changing the image line in your docker-compose.yaml file and then stopping (docker-compose down) and restarting (docker-compose up -d) your Reactor daemon. But there's a catch... not all of you can safely just switch from the aarch64 image to the arm64 image. And, you can't just trust the output of uname -m, for example, because this exposes the CPU architecture, but not the word size of the OS running on that CPU. For Raspberry Pi systems, the transition to 64-bit operating systems was long (starting in 2016) and not always obvious — although there was a first "official" 64-bit OS for RPis in 2020, it did not become a default recommendation in the Raspberry Pi Imager until 2021, and then that was only the default for Pi 3/4 systems with >4GB RAM; it was 2022 before it was universally recommended for all 64-bit CPUs regardless of RAM size. Depending on when you first imaged your RPi system and what default you may have been offered/chosen, you could today easily have a 64-bit CPU Raspberry Pi running a 32-bit version of the operating system. Upgrades along the way would not change this; changing it to fully 64-bit requires a full reimage of the system. To establish if your OS is 64- or 32-bit, log in to your Pi and run: sudo dpkg-architecture -q DEB_HOST_ARCH. If the response is arm64 or aarch64, then you are running a 64-bit OS and you should use the arm64-tagged image. If it's anything else, you are running a 32-bit OS, and you should use the armv7l-tagged image. pi@rpi4-1:~ $ sudo dpkg-architecture -q DEB_HOST_ARCH armhf pi@rpi4-1:~ $ uname -m aarch64 pi@rpi4-1:~ $ In the example above, the uname command reports that the CPU is 64-bit architecture (aarch64), which is true for the host on which I ran these commands, but the DEB_HOST_ARCH value is armhf, indicating a 32-bit operating system. This system has to use the armv7l-tagged image. Other systems will have their own ways of determining the word size of the running OS. Since the majority of Reactor users running ARM systems are on Raspberry Pis, I am able to supply the above instructions, but if you happen to have a different ARM system, you'll need to do some web searching to figure out how to expose that information. Or, you can just try the arm64 image, and if it doesn't start up, try the armv7l image. Remember to always back up your system before making any changes. For everyone, please make this change as soon as possible, and if you have any trouble finding a working image, please (1) go back to the current aarch64 image; and (2) let me know in this thread along with as much detail about your host system as you can offer (including the output of the dpkg-architecture command mentioned above).
Multi-System Reactor
Requesting a proper ARM64/aarch64 Docker image (Pi 5 support)
M
Hi, I'm in the process of migrating from a Raspberry Pi 4 (ARMv7) to a Raspberry Pi 5 (ARMv8/aarch64), but I’ve run into an issue: there is no proper ARMv8/aarch64 image available. None of the existing images run on the Pi 5 - they all exit immediately with code 139 (segmentation fault), which typically indicates that the binaries inside the image are not compatible with the ARM64/aarch64 architecture used by the Pi 5. Would it be possible to publish a correct ARMv8/aarch64 (linux/arm64) image? Building one should be relatively straightforward using docker buildx with multi-arch support. For example, my own Node.js images are built this way: docker buildx build --push \ -t <localrepo>/<project>:<tag> \ --platform=linux/arm64,linux/amd64 \ --file ./apps/<project>/Dockerfile . This produces both the AMD64 and ARM64/v8 variants automatically. Also, as a side note, it may be best to avoid using Alpine as the base image for the ARM64 build, since musl-based builds often cause compatibility issues and unnecessary headaches. A glibc-based base image (e.g., Debian or Ubuntu) tends to work far more reliably on ARM64, especially for Node.js applications. @toggledbits - tagging you in case you missed this. Thanks, mgvra
Multi-System Reactor
Script action and custom timers
therealdbT
Sorry to write here without trying, but I’m flying today. Am I correct if i say that script action with alarm() makes it possible to execute a reaction in a given interval, lets say 15 seconds or 3.5 minutes? That sounds amazing, since I’ve used weird tricks, including a custom controller, just to do this.
Multi-System Reactor
Help resolve change in behaviour post update
CatmanV2C
Topic thumbnail image
Multi-System Reactor
Reactor w/HA 2025.11 error on set_datetime service call setting only time
CrilleC
@toggledbits Do you know if this is related to that PR or is it a change they made in 2025.11.1? [latest-25310]2025-11-11T13:16:24.319Z <HassController:INFO> HassController#hass perform x_hass_input_datetime.set_datetime on Entity#hass>input_datetime_vvb_dag with { "time": "10:45" } [latest-25310]2025-11-11T13:16:24.320Z <HassController:INFO> HassController#hass: sending payload for x_hass_input_datetime.set_datetime on Entity#hass>input_datetime_vvb_dag action: { "type": "call_service", "service_data": { "date": (null), "time": "10:45", "datetime": (null), "timestamp": (null) }, "domain": "input_datetime", "service": "set_datetime", "target": { "entity_id": "input_datetime.vvb_dag" } } [latest-25310]2025-11-11T13:16:24.321Z <HassController:ERR> HassController#hass request 1762866984320<2025-11-11 14:16:24> (call_service) failed: [Error] Not a parseable type for dictionary value @ data['date'] [-] [latest-25310]2025-11-11T13:16:24.321Z <HassController:WARN> HassController#hass action x_hass_input_datetime.set_datetime({ "time": "10:45" }) on Entity#hass>input_datetime_vvb_dag failed! [latest-25310]2025-11-11T13:16:24.321Z <HassController:INFO> Service call payload: {"type":"call_service","service_data":{"date":null,"time":"10:45","datetime":null,"timestamp":null},"domain":"input_datetime","service":"set_datetime","target":{"entity_id":"input_datetime.vvb_dag"},"id":1762866984320} [latest-25310]2025-11-11T13:16:24.322Z <HassController:INFO> Service data: {"fields":{"date":{"example":"\"2019-04-20\"","selector":{"text":{"multiline":false,"multiple":false}}},"time":{"example":"\"05:04:20\"","selector":{"time":{}}},"datetime":{"example":"\"2019-04-20 05:04:20\"","selector":{"text":{"multiline":false,"multiple":false}}},"timestamp":{"selector":{"number":{"min":0,"max":9223372036854776000,"mode":"box","step":1}}}},"target":{"entity":[{"domain":["input_datetime"]}]}} [latest-25310]2025-11-11T13:16:24.322Z <Engine:ERR> Engine#1 reaction rule-mgb8pfhs:S step 0 perform x_hass_input_datetime.set_datetime failed: [Error] Not a parseable type for dictionary value @ data['date'] [-] [latest-25310]2025-11-11T13:16:24.322Z <Engine:INFO> Engine#1 action args: { "time": "10:45" } [latest-25310]2025-11-11T13:16:24.322Z <Engine:INFO> Resuming reaction Sätt Schema VVB i Home Assistant<AKTIV> (rule-mgb8pfhs:S) from step 1 [latest-25310]2025-11-11T13:16:24.323Z <HassController:INFO> HassController#hass perform x_hass_input_datetime.set_datetime on Entity#hass>input_datetime_vvb_natt with { "time": "03:00", "timestamp": 0 } [latest-25310]2025-11-11T13:16:24.323Z <HassController:INFO> HassController#hass: sending payload for x_hass_input_datetime.set_datetime on Entity#hass>input_datetime_vvb_natt action: { "type": "call_service", "service_data": { "date": (null), "time": "03:00", "datetime": (null), "timestamp": 0 }, "domain": "input_datetime", "service": "set_datetime", "target": { "entity_id": "input_datetime.vvb_natt" } } [latest-25310]2025-11-11T13:16:24.324Z <HassController:ERR> HassController#hass request 1762866984323<2025-11-11 14:16:24> (call_service) failed: [Error] Not a parseable type for dictionary value @ data['date'] [-] [latest-25310]2025-11-11T13:16:24.324Z <HassController:WARN> HassController#hass action x_hass_input_datetime.set_datetime({ "time": "03:00", "timestamp": 0 }) on Entity#hass>input_datetime_vvb_natt failed! [latest-25310]2025-11-11T13:16:24.324Z <HassController:INFO> Service call payload: {"type":"call_service","service_data":{"date":null,"time":"03:00","datetime":null,"timestamp":0},"domain":"input_datetime","service":"set_datetime","target":{"entity_id":"input_datetime.vvb_natt"},"id":1762866984323} [latest-25310]2025-11-11T13:16:24.324Z <HassController:INFO> Service data: {"fields":{"date":{"example":"\"2019-04-20\"","selector":{"text":{"multiline":false,"multiple":false}}},"time":{"example":"\"05:04:20\"","selector":{"time":{}}},"datetime":{"example":"\"2019-04-20 05:04:20\"","selector":{"text":{"multiline":false,"multiple":false}}},"timestamp":{"selector":{"number":{"min":0,"max":9223372036854776000,"mode":"box","step":1}}}},"target":{"entity":[{"domain":["input_datetime"]}]}} [latest-25310]2025-11-11T13:16:24.324Z <Engine:ERR> Engine#1 reaction rule-mgb8pfhs:S step 1 perform x_hass_input_datetime.set_datetime failed: [Error] Not a parseable type for dictionary value @ data['date'] [-] [latest-25310]2025-11-11T13:16:24.324Z <Engine:INFO> Engine#1 action args: { "time": "03:00", "timestamp": 0 } [latest-25310]2025-11-11T13:16:24.325Z <Engine:INFO> Resuming reaction Sätt Schema VVB i Home Assistant<AKTIV> (rule-mgb8pfhs:S) from step 2 [latest-25310]2025-11-11T13:16:24.325Z <Engine:INFO> Sätt Schema VVB i Home Assistant<AKTIV> all actions completed.
Multi-System Reactor
Reactor Version 25310 : Office Light control via rule in reactor no longer working since last update.
P
Hello, I currently have an office light (connected via a Leviton Zwave Dimmer switch) controlled from a Gen5 Aeotech Zwave switch installed on my Synology 720+ NAS. I run HA(2025.11.10) in a virtual machine from my NAS and Reactor on the container manager of the same NAS. Prior to updating to 25304 the rule I had set to turn the light on to a specific dimming value worked correctly. Now the rule appears to follow the decision tree, however the reaction does not trigger setting the dimming or turning on the office light? Strangely I can still turn the light on and off as well as dim it directly from HASS..? I have tried using the ''try this action'' button in the rules reaction setting and it will not control the light and does not throw an error flagÉ Please help, P.S Reactor has been rock steady for me over the last few years and I'm a big fan of this solution.
Multi-System Reactor
[Solved] alarm() in global expression throws error in log.
CrilleC
Topic thumbnail image
Multi-System Reactor
[Solved] Define function issue in latest-25304
CrilleC
Topic thumbnail image
Multi-System Reactor
No Upgrade Notification for Build 25308?
CatmanV2C
FWIW I'm no longer getting a notification from MSR that there's an update. Just thought I'd mention it C
Multi-System Reactor
Strange behavior in MSR latest-25304 with disabled groups in Reaction
therealdbT
Topic thumbnail image
Multi-System Reactor
[Reactor] Variables not updating correctly in latest-25201-2aa18550
therealdbT
Topic thumbnail image
Multi-System Reactor
The reaction stopped working (Google Nest max playing a video)
F
Topic thumbnail image
Multi-System Reactor
Handling Dead Entities and Renamed Entities
PablaP
Hello all.. been a minute! I recently rebuilt my Z wave network and migrated to a new z wave stick. In order to prevent any downtime I kept my original z wave network up and ran a docker version of Z Wave JS UI with my new controller. This way I could add device by device without having any devices down. I finally moved all the devices over to my new stick today. The final step was to migrate everything from my Docker instance of Z Wave JS UI to the HA add-on of Z Wave JS UI. However during this migration some of the names didn't populate correctly which I later managed to import back into Z Wave JS UI. The issue was in Reactor it is stuck on the default names and the entities are not updating. I removed the controller from Reactor, restarted, hard refreshed, and added the controller back however the new entity names have not updated. Also it seems like the old entities from my previous instance of Z Wave JS UI are lingering and not being marked as dead (I believe a certain amount of time needs to lapse before they're marked as dead in Reactor). My goal is to basically purge all the entities for the 'ZWaveJS' controller in Reactor so it can pull all the updated entity names and only the entities that exist in Z Wave JS UI. I cannot find a quick way to do this, I know entities can be deleted one by one, but with over 100 entities this would take long I am guessing that if I added the controller with a new name in in the Reactor config it would pull the updated entities and names but I think that would break my rules since the entity IDs would change (I made sure to name all the entities the exact same as they were previously to prevent this issue).
Multi-System Reactor
Strange behavior for MQTT templates using payload and attributes
therealdbT
Topic thumbnail image
Multi-System Reactor
[MSR] reactor-mqtt-contrib package for additional MQTT templates
therealdbT
I'm slowly migrating all my stuff to MQTT under MSR, so I have a central place to integrate everything (and, in a not-so-distant future, to remove virtual devices from my Vera and leave it running zwave only). Anyway, here's my reactor-mqtt-contrib package: https://github.com/dbochicchio/reactor-mqtt-contrib Simply download yaml files (everything or just the ones you need) and you're good to go. I have mapped my most useful devices, but I'll add others soon. Feel free to ask for specific templates, since I've worked a lot in the last weeks to understand and operate them. The templates are supporting both init and query, so you have always up-to-date devices at startup, and the ability to poll them. Online status is supported as well, so you can get disconnected devices with a simple expression. Many-many thanks to @toggledbits for its dedication, support, and patience with me and my requests
Multi-System Reactor
HA 2025.9.4 Supported Yet?
CatmanV2C
Tangentially did I miss 2025.9.4 getting blessed in MSR? I've been holding off Cheers C
Multi-System Reactor
Rule Set UI bug - RESOLVED
3
Topic thumbnail image
Multi-System Reactor

[SOLVED] Lights on all month UNTIL fourth Thursday of the month

Scheduled Pinned Locked Moved Multi-System Reactor
8 Posts 3 Posters 599 Views 3 Watching
  • Oldest to Newest
  • Newest to Oldest
  • Most Votes
Reply
  • Reply as topic
Log in to reply
This topic has been deleted. Only users with topic management privileges can see it.
  • G Offline
    G Offline
    gwp1
    wrote on last edited by gwp1
    #1

    I've a scene for Arming for Xmas Season that works great, starting on Black Friday (Friday after fourth Thursday/Thanksgiving) and it uses latch to achieve the goal.

    Now I want to have DIFFERENT lighting for November 1 - Thanksgiving night, fourth Thu of November. So far I have:

    Screen Shot 2021-11-01 at 9.11.30 PM.png

    I know that this will only end up with lights coming on ON the fourth Thu, not prior to it. Am I looking at Output Control or Constraints to make my end date/time Thanksgiving night?

    Solved here: https://smarthome.community/post/10256

    *Hubitat C-7 2.4.3.149
    *Proxmox VE v8, Beelink MiniPC 12GBs, SSD

    *HASS 2025.11.1
    w/ ZST10-700 fw 7.18.3

    *Prod MSR in docker/portainer
    MSR: latest-25310-dc2bb580
    MQTTController: 25139
    ZWave Controller: 25139

    toggledbitsT 1 Reply Last reply
    0
    • LibraSunL Offline
      LibraSunL Offline
      LibraSun
      wrote on last edited by LibraSun
      #2

      Heck, does this mean MSR needs a "Movable Feasts Calendar" or "National Holidays" component in addition to Date/Time, so you can simply pick [BEFORE] [THANKSGIVING] [OFFSET: 0 DAYS]? I recall Calendar Creator 2.0 on my Apple //c had something similar. It figured out things like Easter, Lent (and by extension, Mardi Gras) and so forth. Even Google Calendar can't do all that, nearly 35 years later!

      1 Reply Last reply
      1
      • LibraSunL Offline
        LibraSunL Offline
        LibraSun
        wrote on last edited by
        #3

        And geez, thanks for sending me down a huge rabbit hole. I'm back! I contemplated the reasonableness (you be the judge) of using an API of some kind to simply fetch the always-changing annual date of Thanksgiving. Turns out there are many such APIs, but they cost money, so I focused on the tried-and-true Google Calendars API since Google hosts a robust calendar of U.S. holidays which MSR can query.

        Let's take a little stroll down the path I traveled lo these past couple of hours (again, thanks!):

        STEP 1. Create an API app in Google developer account https://console.developers.google.com

        STEP 2. From the "Credentials" tab, create an API key. You get something like this "AIzaSyBcOT_DpEQysiwFmmmZXupKpnrOdJYAhhM"

        STEP 3. Then you can pull events from Google's holiday calendar using this URL -
        https://www.googleapis.com/calendar/v3/calendars/en.usa%23holiday%40group.v.calendar.google.com/events?key=yourAPIKey

        STEP 4. To retrieve only events called "Thanksgiving" -- there are 3 of them... last year, current year and next year -- one must insert a 'q=' modifier, so the URL we will use in MSR looks like -
        https://www.googleapis.com/calendar/v3/calendars/en.usa%23holiday%40group.v.calendar.google.com/events?q=Thanksgiving&key=yourAPIKey

        STEP 5. Create a new Rule in MSR which (I guess every January 1st) retrieves the next occurrence of Thanksgiving Day and stores that value into a blank expression. The Rule would have (at minimum) the following parts:
        CONDITION = Date/Time set to execute on January 1st of each year
        SET REACTION = An [HTTP] request that performs a [GET] on the URL from Step 4 and captures the response to expression [Events].
        EXPRESSIONS =
        'Events' := (blank)
        'Items' := Events.items
        'TGend' := T = first Occur in Items with (substr(Occur.end.date,0,4) == dateparts(time()).year), T.end.date

        The magic is happening in the 'TGend' expression, which returns a date as string, such as "2021-11-26". You could then, in theory, have your Thanksgiving Lights rule active between November 1 and {{ TGend }}, with some judicious use of an added expression and/or conversion of that end date into epoch time (e.g. (number) 1637884800000) for mathematical purposes. Your call.

        BED TIME NOW!

        G 1 Reply Last reply
        0
        • LibraSunL LibraSun

          And geez, thanks for sending me down a huge rabbit hole. I'm back! I contemplated the reasonableness (you be the judge) of using an API of some kind to simply fetch the always-changing annual date of Thanksgiving. Turns out there are many such APIs, but they cost money, so I focused on the tried-and-true Google Calendars API since Google hosts a robust calendar of U.S. holidays which MSR can query.

          Let's take a little stroll down the path I traveled lo these past couple of hours (again, thanks!):

          STEP 1. Create an API app in Google developer account https://console.developers.google.com

          STEP 2. From the "Credentials" tab, create an API key. You get something like this "AIzaSyBcOT_DpEQysiwFmmmZXupKpnrOdJYAhhM"

          STEP 3. Then you can pull events from Google's holiday calendar using this URL -
          https://www.googleapis.com/calendar/v3/calendars/en.usa%23holiday%40group.v.calendar.google.com/events?key=yourAPIKey

          STEP 4. To retrieve only events called "Thanksgiving" -- there are 3 of them... last year, current year and next year -- one must insert a 'q=' modifier, so the URL we will use in MSR looks like -
          https://www.googleapis.com/calendar/v3/calendars/en.usa%23holiday%40group.v.calendar.google.com/events?q=Thanksgiving&key=yourAPIKey

          STEP 5. Create a new Rule in MSR which (I guess every January 1st) retrieves the next occurrence of Thanksgiving Day and stores that value into a blank expression. The Rule would have (at minimum) the following parts:
          CONDITION = Date/Time set to execute on January 1st of each year
          SET REACTION = An [HTTP] request that performs a [GET] on the URL from Step 4 and captures the response to expression [Events].
          EXPRESSIONS =
          'Events' := (blank)
          'Items' := Events.items
          'TGend' := T = first Occur in Items with (substr(Occur.end.date,0,4) == dateparts(time()).year), T.end.date

          The magic is happening in the 'TGend' expression, which returns a date as string, such as "2021-11-26". You could then, in theory, have your Thanksgiving Lights rule active between November 1 and {{ TGend }}, with some judicious use of an added expression and/or conversion of that end date into epoch time (e.g. (number) 1637884800000) for mathematical purposes. Your call.

          BED TIME NOW!

          G Offline
          G Offline
          gwp1
          wrote on last edited by
          #4

          @librasun Part of me tried, TRIED, to sympathize - but I just couldn't do it. I was poking at various iterations of this rule based off of reading and re-reading the manual and staring endlessly at the xmas rule (which is the opposite, it STARTS on the fourth Friday in November) during an incident bridge at work that did not require me or my team's attention any longer.

          *Hubitat C-7 2.4.3.149
          *Proxmox VE v8, Beelink MiniPC 12GBs, SSD

          *HASS 2025.11.1
          w/ ZST10-700 fw 7.18.3

          *Prod MSR in docker/portainer
          MSR: latest-25310-dc2bb580
          MQTTController: 25139
          ZWave Controller: 25139

          1 Reply Last reply
          1
          • G gwp1

            I've a scene for Arming for Xmas Season that works great, starting on Black Friday (Friday after fourth Thursday/Thanksgiving) and it uses latch to achieve the goal.

            Now I want to have DIFFERENT lighting for November 1 - Thanksgiving night, fourth Thu of November. So far I have:

            Screen Shot 2021-11-01 at 9.11.30 PM.png

            I know that this will only end up with lights coming on ON the fourth Thu, not prior to it. Am I looking at Output Control or Constraints to make my end date/time Thanksgiving night?

            Solved here: https://smarthome.community/post/10256

            toggledbitsT Offline
            toggledbitsT Offline
            toggledbits
            wrote on last edited by toggledbits
            #5

            @gwp1 If you want it to end Thanksgiving night, that means you don't want the lights on starting with the 4th Friday (so on from the Nov-1 to the end of the day on the 4th Thursday).

            55e85987-a1d8-4b85-8265-3ea323b01ebb-image.png

            Theory of operation:

            The first date/time condition bounds the operating time of the entire thing. It can only be active in the month of November.

            For the remainder, I'll work from the inside of the groups out...

            The weekday condition obviously selects the 4th Friday of the month. It is set for latching (options), so when that weekday is encountered, it sticks. Because it's latching, it needs another condition partnered with it to tell it when to reset, and that is the purpose of the accompanying date/time condition in the group: it will be true throughout the end of the month, starting with the earliest possible day that a fourth Thursday could be. So when the end of the month rolls off and the date goes to the first of the month, that condition goes false and causes the interior group to go false, and the group going false causes the latch to be released on its child condition. In all, that means this interior group is true from the fourth Friday through the end of the month, and false otherwise.

            Now, since we want the rule reset between that fourth Friday and the end of the month, we add a wrapper group with a NOT to invert the state of the interior group. Very important point here: you cannot take a shortcut and make the interior group NOT AND, because of the latched condition. You change the behavior of the group/latch relationship if you do that -- the latch relies on its parent group to go false to reset it, so you need to leave the logic state of that group unaltered to get correct latch and reset behavior for this solution. The wrapper group allows you to change (NOT) the inner group's state as needed for the rule without changing the interrmost group's native logic state.

            Key takeaway: the parent group of a latched condition must be a pure AND group (cannot be NOT AND, OR, XOR, NUL).

            Author of Multi-system Reactor and Reactor, DelayLight, Switchboard, and about a dozen other plugins that run on Vera and openLuup.

            G 1 Reply Last reply
            1
            • LibraSunL Offline
              LibraSunL Offline
              LibraSun
              wrote on last edited by
              #6

              Just to reinforce the patently obvious truism: If you ever see me and Patrick posting in the same thread, go with his sage advice!

              1 Reply Last reply
              1
              • toggledbitsT toggledbits

                @gwp1 If you want it to end Thanksgiving night, that means you don't want the lights on starting with the 4th Friday (so on from the Nov-1 to the end of the day on the 4th Thursday).

                55e85987-a1d8-4b85-8265-3ea323b01ebb-image.png

                Theory of operation:

                The first date/time condition bounds the operating time of the entire thing. It can only be active in the month of November.

                For the remainder, I'll work from the inside of the groups out...

                The weekday condition obviously selects the 4th Friday of the month. It is set for latching (options), so when that weekday is encountered, it sticks. Because it's latching, it needs another condition partnered with it to tell it when to reset, and that is the purpose of the accompanying date/time condition in the group: it will be true throughout the end of the month, starting with the earliest possible day that a fourth Thursday could be. So when the end of the month rolls off and the date goes to the first of the month, that condition goes false and causes the interior group to go false, and the group going false causes the latch to be released on its child condition. In all, that means this interior group is true from the fourth Friday through the end of the month, and false otherwise.

                Now, since we want the rule reset between that fourth Friday and the end of the month, we add a wrapper group with a NOT to invert the state of the interior group. Very important point here: you cannot take a shortcut and make the interior group NOT AND, because of the latched condition. You change the behavior of the group/latch relationship if you do that -- the latch relies on its parent group to go false to reset it, so you need to leave the logic state of that group unaltered to get correct latch and reset behavior for this solution. The wrapper group allows you to change (NOT) the inner group's state as needed for the rule without changing the interrmost group's native logic state.

                Key takeaway: the parent group of a latched condition must be a pure AND group (cannot be NOT AND, OR, XOR, NUL).

                G Offline
                G Offline
                gwp1
                wrote on last edited by gwp1
                #7

                @toggledbits Ok, unpacking this in my head:

                Topmost trigger, no brainer: eligible during November. My CDO wants to use the 29th as the end but I get it.

                The inverting wrapper - I'm kicking myself because I completely forgot about NOT even though I've used it several times in various rulesets. Ten demerits for me.

                Here's where I'm slightly behind the curve - detect and latch. Talking aloud, weekday is Friday. Ok, good. The trigger grouped with it essentially says "eligible during this sub-range of that big ol' trigger range way up top".

                Got it.

                Am I implementing the latch because there is a availability of multiple Friday's in that 22-EOM after range and latching the extra layer of weekday fourth vs any Friday is what caps the grouped sub-range from going beyond that fourth Friday?

                Sidenote: when have you ever known me to take a shortcut - if anything, I go the waaaaay long way around lol

                @LibraSun - no worries, I did give your method a shot and was successfully up to the point where my mental KB fails me: adding the {{ TGend }} expression into the ruleset.

                The last thing I want to review in my rulesets is to ensure everything tied to Thanksgiving is done/gone at midnight Friday so at sunset Friday the Xmas ruleset kicks in without any race states. I've been a master at creating those these past couple weeks....

                Post-script: I did add one additional trigger at the top: sunset +10 to sunrise -15.

                *Hubitat C-7 2.4.3.149
                *Proxmox VE v8, Beelink MiniPC 12GBs, SSD

                *HASS 2025.11.1
                w/ ZST10-700 fw 7.18.3

                *Prod MSR in docker/portainer
                MSR: latest-25310-dc2bb580
                MQTTController: 25139
                ZWave Controller: 25139

                toggledbitsT 1 Reply Last reply
                1
                • G gwp1

                  @toggledbits Ok, unpacking this in my head:

                  Topmost trigger, no brainer: eligible during November. My CDO wants to use the 29th as the end but I get it.

                  The inverting wrapper - I'm kicking myself because I completely forgot about NOT even though I've used it several times in various rulesets. Ten demerits for me.

                  Here's where I'm slightly behind the curve - detect and latch. Talking aloud, weekday is Friday. Ok, good. The trigger grouped with it essentially says "eligible during this sub-range of that big ol' trigger range way up top".

                  Got it.

                  Am I implementing the latch because there is a availability of multiple Friday's in that 22-EOM after range and latching the extra layer of weekday fourth vs any Friday is what caps the grouped sub-range from going beyond that fourth Friday?

                  Sidenote: when have you ever known me to take a shortcut - if anything, I go the waaaaay long way around lol

                  @LibraSun - no worries, I did give your method a shot and was successfully up to the point where my mental KB fails me: adding the {{ TGend }} expression into the ruleset.

                  The last thing I want to review in my rulesets is to ensure everything tied to Thanksgiving is done/gone at midnight Friday so at sunset Friday the Xmas ruleset kicks in without any race states. I've been a master at creating those these past couple weeks....

                  Post-script: I did add one additional trigger at the top: sunset +10 to sunrise -15.

                  toggledbitsT Offline
                  toggledbitsT Offline
                  toggledbits
                  wrote on last edited by
                  #8

                  @gwp1 said in Lights on UNTIL fourth Thursday of the month:

                  Am I implementing the latch because there is a availability of multiple Friday's in that 22-EOM after range and latching the extra layer of weekday fourth vs any Friday is what caps the grouped sub-range from going beyond that fourth Friday?

                  No, you're using latch because 6 of 7 days of the week aren't Fridays, and as many as 8 days, including another Friday, or as few as 2 with no Friday, may follow the fourth Friday in November of any given year. So once we encounter the fourth Friday, we want to continue to hold that condition state on Saturday, Sunday, Monday, and every other day that may follow through the end of the month. The latch takes care of that easily.

                  I know @LibraSun is into expressions, so another possible solution is to use expressions to gin up the timestamp for midnight on the 22nd day of November (the earliest day of the month a fourth Friday could be) of the current year, and then get its day of the week. Knowing that Friday is weekday index 5, you can increment the date to figure out when the fourth Friday occurs. For example, if 11/22 is already weekday 5, you're on target and no offset is needed (that date is the fourth Friday). But if 11/22 is weekday 1 (Monday), you need to add 4 (days, 5-1) to the timestamp to get it to a Friday. I'll leave what to do if 11/22 is a Saturday (weekday index 6) as an exercise for the reader. 🙂 Once you've incremented the date, just compare it to the current time to see if you're on or before it.

                  Author of Multi-system Reactor and Reactor, DelayLight, Switchboard, and about a dozen other plugins that run on Vera and openLuup.

                  1 Reply Last reply
                  2
                  • toggledbitsT toggledbits locked this topic on
                  Reply
                  • Reply as topic
                  Log in to reply
                  • Oldest to Newest
                  • Newest to Oldest
                  • Most Votes


                  Recent Topics

                  • Local expression in Rule does not evaluate as they used to do
                    CrilleC
                    Crille
                    0
                    1
                    15

                  • Home Assistant 2025.11.2 and latest-25315
                    G
                    gwp1
                    0
                    6
                    58

                  • Reactor (Multi-System/Multi-Hub) Announcements
                    toggledbitsT
                    toggledbits
                    5
                    129
                    73.5k

                  • Notice to Docker + ARM Users (RPi 3/4/5 and others)
                    toggledbitsT
                    toggledbits
                    1
                    1
                    43

                  • Requesting a proper ARM64/aarch64 Docker image (Pi 5 support)
                    M
                    mgvra
                    1
                    3
                    109

                  • Script action and custom timers
                    toggledbitsT
                    toggledbits
                    0
                    4
                    127

                  • Help resolve change in behaviour post update
                    CatmanV2C
                    CatmanV2
                    0
                    12
                    342

                  • There is an alternative to homebridge-mqttthing
                    akbooerA
                    akbooer
                    1
                    2
                    106

                  • Reactor w/HA 2025.11 error on set_datetime service call setting only time
                    CrilleC
                    Crille
                    0
                    6
                    140

                  • Reactor Version 25310 : Office Light control via rule in reactor no longer working since last update.
                    toggledbitsT
                    toggledbits
                    0
                    17
                    407

                  • Shelly Wall Display XL
                    akbooerA
                    akbooer
                    2
                    9
                    773

                  • [Solved] alarm() in global expression throws error in log.
                    toggledbitsT
                    toggledbits
                    0
                    26
                    767
                  Powered by NodeBB | Contributors
                  Hosted freely by 10RUPTiV - Solutions Technologiques | Contact us
                  • Login

                  • Don't have an account? Register

                  • Login or register to search.
                  • First post
                    Last post
                  0
                  • Categories
                  • Recent
                  • Tags
                  • Popular
                  • Unsolved