Domoticz TEXT sensor triggering
-
@sushukka You can use netcat or socket connection from python script to send message directly to mysensors GW, omitting domoticz. Then you can link this script to your dummy switch in domoticz. It works.
netcat:
netcat <gw ip> <port>python:
#!/usr/bin/python3 import socket from time import sleep def gwsend(hostname, port, content): s = socket.socket(socket.AF_INET, socket.SOCK_STREAM) s.connect((hostname, port)) sleep(1) s.sendall(content) s.shutdown(socket.SHUT_WR) s.close() string = str.encode("some text") # Your TEXT message gwsend("127.0.0.1", 5003, b"2;1;1;1;47;" + string + b"\n")@monte I need to move to a next project so before conversion to Node-red and/or OpenHAB, how would you use netcat type of injection with MQTT? I mean now I have MySensors in W5100/MQTT mode and I want to send an update to my V_TEXT sensor via MQTT. Easiest way to do this with Domoticz would be by triggering some script. Sending an MQTT update via Python is probably easy (dunno, haven't done yet), but I have no idea how to form the message so that it would be in correct MySensors format...My node id is 31 and the text sensor child is 1:
#define MY_NODE_ID 31 #define LED_MODE_ID 1 MyMessage msgLedMode(LED_MODE_ID, V_TEXT);The payload is some number between -1 to 56. Really appreciate your help. :relieved:
-
@monte I need to move to a next project so before conversion to Node-red and/or OpenHAB, how would you use netcat type of injection with MQTT? I mean now I have MySensors in W5100/MQTT mode and I want to send an update to my V_TEXT sensor via MQTT. Easiest way to do this with Domoticz would be by triggering some script. Sending an MQTT update via Python is probably easy (dunno, haven't done yet), but I have no idea how to form the message so that it would be in correct MySensors format...My node id is 31 and the text sensor child is 1:
#define MY_NODE_ID 31 #define LED_MODE_ID 1 MyMessage msgLedMode(LED_MODE_ID, V_TEXT);The payload is some number between -1 to 56. Really appreciate your help. :relieved:
@sushukka I have limited knowledge of mqtt, I'm not using it in my setup, with ncat you can only talk directly to tcp/udp socket and mqtt is built above it. But there is mqtt library for python, and guides for using it. http://www.steves-internet-guide.com/into-mqtt-python-client/ try to follow those steps, I will try too :)
-
mqtt publish example in python
import paho.mqtt.publish as publish payload = 'Hey there' publish.single('mys-out/1/1/1/0/47', payload, hostname=<mqtthost>, port=<mqttport, retain=False)'mys-out/99/1/1/0/47' is the MQTT topic, and is decoded as follows:
99: Node ID
1: ChildId
1: command (Set in this case)
0: ACK (None in this case)
47: V_TYPE (V_TEXT in this case)Above pretty much follows the MySensors serial protocol, as found here, only difference is that sensor payload, and sensor id / msg type etc. is split into two parameters in MQTT (as in payload and topic), and doesn't come in as a single string.
-
I think, all this is not Domoticz problem, but mysensors problem.
Mysensors gateways has not got any universal input interface.
This is, why I am not using pure mysensors solution.
Imagine, that your gateway understood something like this:http://MS_GATEWAY/cmd?node=10&sensor=1&value=80
And translate it to mysensors network message, which sends value 80 to sensor number 1 of node number 10.
Then you can send into your node everything not only from Domoticz, but from simply Android application and etc.
With ESP8266 gateway you need only "ESP8266HTTPClient" library to parse this http call for example. -
I think, all this is not Domoticz problem, but mysensors problem.
Mysensors gateways has not got any universal input interface.
This is, why I am not using pure mysensors solution.
Imagine, that your gateway understood something like this:http://MS_GATEWAY/cmd?node=10&sensor=1&value=80
And translate it to mysensors network message, which sends value 80 to sensor number 1 of node number 10.
Then you can send into your node everything not only from Domoticz, but from simply Android application and etc.
With ESP8266 gateway you need only "ESP8266HTTPClient" library to parse this http call for example.@kimot then it would be a controller itself, not a gateway. You need gateway anyway (sick rhyme), or you would need to use only powerful enough hardware to implement web server to handle requests as you suggest. But it wouldn't be backward compatible with arduino gateways, thus it's not the way for now, maybe in the future we will come to it.
-
If the node requests the TEXT to domoticz, it gets the value, so domoticz can actually handle it, the problem is that it does not send the TEXT value if something changes the variable but only if the node requests it.
-
@kimot then it would be a controller itself, not a gateway. You need gateway anyway (sick rhyme), or you would need to use only powerful enough hardware to implement web server to handle requests as you suggest. But it wouldn't be backward compatible with arduino gateways, thus it's not the way for now, maybe in the future we will come to it.
@monte
It can be implemented into existing ethernet gateways, I think.
I experimenting with ESP8266 gateway only, not W5100.For inspiration look at code line 24 to 56 here:
https://diyprojects.io/driving-gpio-esp8266-web-server-domoticz-tcp-ip-wireless/
-
@monte
It can be implemented into existing ethernet gateways, I think.
I experimenting with ESP8266 gateway only, not W5100.For inspiration look at code line 24 to 56 here:
https://diyprojects.io/driving-gpio-esp8266-web-server-domoticz-tcp-ip-wireless/
@kimot I know, that you can make web server on esp8266, but now gateway code works on almost everything that is supported by arduino IDE and has enough memory. If you will make version of gw that only works on esp or raspberry pi it won't be right. For my taste we need affordable microcontroller based on ARM Cortex-M0 (M3, M4) with built-in Ethernet controller and hardware TCP/IP stack. For now there is one chip but it isn't nor affordable nor widely spread, and almost no one knows how to work with it (because no one has it). http://www.wiznet.io/product-item/w7500p/
-
mqtt publish example in python
import paho.mqtt.publish as publish payload = 'Hey there' publish.single('mys-out/1/1/1/0/47', payload, hostname=<mqtthost>, port=<mqttport, retain=False)'mys-out/99/1/1/0/47' is the MQTT topic, and is decoded as follows:
99: Node ID
1: ChildId
1: command (Set in this case)
0: ACK (None in this case)
47: V_TYPE (V_TEXT in this case)Above pretty much follows the MySensors serial protocol, as found here, only difference is that sensor payload, and sensor id / msg type etc. is split into two parameters in MQTT (as in payload and topic), and doesn't come in as a single string.
@tbowmo Thanks @tbowmo and everybody! This active community has really saved my day so many times. You rule guys and gals! :+1: :blush:
Here is a working script:import paho.mqtt.client as mqttClient import sys irCommand = sys.argv[1] print("Parameter is: " + irCommand) client = mqttClient.Client("Python") client.username_pw_set(username="xxx",password="xxx") client.connect("nnn.nnn.nnn.nnn", 1883) client.publish('domoticz/out/HomeNum_IR/0/0/1/0/47', irCommand, retain=False) # (ESP8266 "gateway" MQTT client -> node=0)And call it from Domoticz virtual selector switch:
script:///home/pi/domoticz/scripts/python/hnumber_led_mode.py 7Parameter is the desired mode for led animation. Made a similar on for IR remote commands.
Is there same ~25bit payload limit with MQTT too? I mean those NEC IR commands tend to be 32bit so probably cannot send them directly but need to be hardcoded in the sketch? Also planned to create a universal script for these needs where you could give also node number, child ids etc but then again...I chain myself even tighter to Domoticz shackles and again Domoticz should support this without scripting.
Couple of more questions:
- If you have several ESP8266 MQTT gateways, should one create a different topic for everyone or put them in the same MySensors queue? If so, will the MY_MQTT_CLIENT_ID make the separation?
- I installed nodered-contrib-mysensors and it seems very promising. Definitely going to play around with it. I probably still need some actual controller but Node-red seems to be important part whatever you're gonna do especially when using MQTT queues.
- When using MQTT instead of Ethernet GW, I lose small but very important check: node pinging. I can see from Domoticz log warning messages when some of my ESP8266 Ethernet nodes are down, but when using MQTT there are no warnings, just no data received. I assume this is quite logical because the MQTT node's address in Domoticz is actually broker's (Mosquitto) address + right MQTT topic. However, when sniffing Domoticz MQTT traffic, it seems to send PING messages quite regularly (but don't know if this triggers anything if the ping success or fails...):
// MQTT topic: domoticz/out/HomeNum_IR/# PING qos : 0, retain : false, cmd : publish, dup : false, topic : domoticz/out/HomeNum_IR/0/0/3/0/18, messageId : , length : 40Update: Seems that the Domo has it's own requirements for MQTT queues and for preventing overlapping ESP8266 = GW = Node ID 0 have an option to have also Node ID in the MQTT queue definition. Still would be nice to be able to define the MQTT queue names freely in Domo...
Also my bad: MySensors payload size is 25 bytes, not bits... This seems to be because of NFR24L01 limitations. However, it would be nice to see an extended payload size for passing eg. longer text information with more advanced microcontrollers like ESP8266 in the future releases. Minor need though.
-
@kimot
So MQTT is not a common protocol? Or serial? What would the benefit be of a http request, over any of the other transport options, when the controller doesn't transmit when changed?In my opinion then http requests aren't that usable as a transport in a home automation setup. MQTT is much better for this, if you really want to go over tcp/ip.
-
And update here about V_TEXT and OpenHAB: Yes it works like it should. When I update the V_TEXT sensor in OpenHAB, it will send the update without any extra scripts or hacks to the MySensors node.
However, after spending several days and nights setting up OpenHAB I have to say that it is far from being ready. MySensors seems not to be qualified yet in the list of "downloadable ready" bindings in PaperUI. Had several errors with MQTT nodes. It is customizable yes, but requires way too much time to set up and maintain. It seems to be more like an architectural masterpiece than easy-to-use home controller.
You can define nearly everything in PaperUI or HABmin like me (a hell of mouse clicking and scrolling) or define everything in multiple configuration files. When you finally have all the nodes and sensors defined (bindings, things, items), you have to build your own gui to see them in usable format. You can either use HabPanel for dashboards or configuration files for sitemaps. Dashboards, which has the best outlook imo, seems to have a problem of working with different devices. I tried to avoid the manual sitemap building by installing Imperihome binding. Connection ok, but alas, you have to manually tag every damn item you have to show them in Imperihome. With all manual configurations you actually need to define quite much of information and they are nicely connected making bigger future changes complicated. Having now around 100 sensors, all this stuff really takes time. Moreover it seemed that the reliability at least with MySensors Ethernet Gateways (one main for NFR24L01 nodes and several ESP8266s) was not in par with Domoticz. Startup problems, rebooting, lost nodes etc...And this was the last drop --> changed back to Domoticz.
Have to say that after a couple of years OpenHAB could maybe be the number one open source home automation controller, but it requires lots of more work for getting there. Now it's too much of engineering and configuration porn vs. the real benefits it gives. Also I'm little bit afraid that the they are getting lost to this super elegancy by trying to be everything and pleasing everybody. PaperUI, HABmin, HomeBuilder, HabPanel, Basic UI, Classic UI, some third party UIs, configuration files... The concept is nice, but as with any product, you need to focus somewhere or otherwise you are just semicore in all areas. Domoticz has its problems, but at least it is reliable and allows putting the most of the available tinkering time to the actual node building than tuning the controller.
-
And update here about V_TEXT and OpenHAB: Yes it works like it should. When I update the V_TEXT sensor in OpenHAB, it will send the update without any extra scripts or hacks to the MySensors node.
However, after spending several days and nights setting up OpenHAB I have to say that it is far from being ready. MySensors seems not to be qualified yet in the list of "downloadable ready" bindings in PaperUI. Had several errors with MQTT nodes. It is customizable yes, but requires way too much time to set up and maintain. It seems to be more like an architectural masterpiece than easy-to-use home controller.
You can define nearly everything in PaperUI or HABmin like me (a hell of mouse clicking and scrolling) or define everything in multiple configuration files. When you finally have all the nodes and sensors defined (bindings, things, items), you have to build your own gui to see them in usable format. You can either use HabPanel for dashboards or configuration files for sitemaps. Dashboards, which has the best outlook imo, seems to have a problem of working with different devices. I tried to avoid the manual sitemap building by installing Imperihome binding. Connection ok, but alas, you have to manually tag every damn item you have to show them in Imperihome. With all manual configurations you actually need to define quite much of information and they are nicely connected making bigger future changes complicated. Having now around 100 sensors, all this stuff really takes time. Moreover it seemed that the reliability at least with MySensors Ethernet Gateways (one main for NFR24L01 nodes and several ESP8266s) was not in par with Domoticz. Startup problems, rebooting, lost nodes etc...And this was the last drop --> changed back to Domoticz.
Have to say that after a couple of years OpenHAB could maybe be the number one open source home automation controller, but it requires lots of more work for getting there. Now it's too much of engineering and configuration porn vs. the real benefits it gives. Also I'm little bit afraid that the they are getting lost to this super elegancy by trying to be everything and pleasing everybody. PaperUI, HABmin, HomeBuilder, HabPanel, Basic UI, Classic UI, some third party UIs, configuration files... The concept is nice, but as with any product, you need to focus somewhere or otherwise you are just semicore in all areas. Domoticz has its problems, but at least it is reliable and allows putting the most of the available tinkering time to the actual node building than tuning the controller.
@sushukka why don't you use configs? It's not too hard to write them and you don't need to find options in incomplete gui. But I agree that their approach doesn't suit well for tinkering and building system step by step adding few devices every couple of days, or changing the layout. I guess it has more use for complete systems that doesn't require adding new things more than once in few months.
I have to say that I like Domoticz' way of setting things up more, but the fact that it's almost not developing and amount of lag and unpolished corners here and there...It doesn't feel robust enough to suggest it to someone else than myself or people like ones on this forum. -
@sushukka why don't you use configs? It's not too hard to write them and you don't need to find options in incomplete gui. But I agree that their approach doesn't suit well for tinkering and building system step by step adding few devices every couple of days, or changing the layout. I guess it has more use for complete systems that doesn't require adding new things more than once in few months.
I have to say that I like Domoticz' way of setting things up more, but the fact that it's almost not developing and amount of lag and unpolished corners here and there...It doesn't feel robust enough to suggest it to someone else than myself or people like ones on this forum.@monte True that. Thing is that people like to use the guis instead of configuration files. That's probably the reason why OpenHAB "fastly" evolves in that direction.
For all the not-so-technical home automation builders I would myself suggest Domoticz as the choise for open source controller. OpenHAB maybe for the ones who buys ready-made stuff from Ikea, zWave etc and just want to control the stuff from one place. But still I'm pretty sure that many of those are not really interested or have required knowledge to start tinkering with OpenHAB. For MySensors users, as you said, it creates lots of additional work for little extra over Domoticz. Domo has this stupid V_TEXT sensor limitation and some people like that the default graphing stuff is limited. However, OpenHAB has nothing without setting up persistence services and their separate databases and definitions...and you can have the praised Influx-Grafana integration also for Domo. My personal opinion is that the normal Domo gui is still pretty useful plus you can have Imperihome, Dashticz type of connections or build your own with Rest API. With OpenHAB you don't have anything usable without manually building it first. I'm myself using Imperihome which basically gives little to complain about gui usability and visual aesthetics. With Domo you install MyDomoAtHome which has all the mappings done so that you can just start playing around with your Imperihome gui. With OpenHAB...
https://github.com/openhab/openhab2-addons/tree/master/addons/io/org.openhab.io.imperihome
The tagging part is not a small work to do...and this pretty much underlines the problem with OpenHAB. You can tune everything, but basically spend huge amount of time on doing that...or even getting things to work on basic level. The OpenHAB community seems to be active, but it has to, as there seems to be lots of problems too.But not bashing OpenHAB more: I want to like it, but it's just making liking it too hard for now. However, it has Eclipse foundation backing it up and has started evolving to the right direction out of the pure configuration file based setup/tuning. After year or two I believe it's much more user-friendly and more widely used. If not and Domo won't be developed anymore there is always the middleware/MQTT/NodeRed/cloud options to go...
-
@monte True that. Thing is that people like to use the guis instead of configuration files. That's probably the reason why OpenHAB "fastly" evolves in that direction.
For all the not-so-technical home automation builders I would myself suggest Domoticz as the choise for open source controller. OpenHAB maybe for the ones who buys ready-made stuff from Ikea, zWave etc and just want to control the stuff from one place. But still I'm pretty sure that many of those are not really interested or have required knowledge to start tinkering with OpenHAB. For MySensors users, as you said, it creates lots of additional work for little extra over Domoticz. Domo has this stupid V_TEXT sensor limitation and some people like that the default graphing stuff is limited. However, OpenHAB has nothing without setting up persistence services and their separate databases and definitions...and you can have the praised Influx-Grafana integration also for Domo. My personal opinion is that the normal Domo gui is still pretty useful plus you can have Imperihome, Dashticz type of connections or build your own with Rest API. With OpenHAB you don't have anything usable without manually building it first. I'm myself using Imperihome which basically gives little to complain about gui usability and visual aesthetics. With Domo you install MyDomoAtHome which has all the mappings done so that you can just start playing around with your Imperihome gui. With OpenHAB...
https://github.com/openhab/openhab2-addons/tree/master/addons/io/org.openhab.io.imperihome
The tagging part is not a small work to do...and this pretty much underlines the problem with OpenHAB. You can tune everything, but basically spend huge amount of time on doing that...or even getting things to work on basic level. The OpenHAB community seems to be active, but it has to, as there seems to be lots of problems too.But not bashing OpenHAB more: I want to like it, but it's just making liking it too hard for now. However, it has Eclipse foundation backing it up and has started evolving to the right direction out of the pure configuration file based setup/tuning. After year or two I believe it's much more user-friendly and more widely used. If not and Domo won't be developed anymore there is always the middleware/MQTT/NodeRed/cloud options to go...
@sushukka I'm using Domoticz myself at home, but have chosen OpenHAB as a platform for my friend's home automation project I'm working on. We are using some KNX modules and for now only openhab has native support for it (among platforms that are kept developing). Another big advantage of openhab is number of different bindings for different hardware and community that is willing to write the new ones. On the other hand developers of Domoticz either ignore your suggestions or tell you to implement them yourself. Can't see a bright future with such an attitude.
-
@sushukka I'm using Domoticz myself at home, but have chosen OpenHAB as a platform for my friend's home automation project I'm working on. We are using some KNX modules and for now only openhab has native support for it (among platforms that are kept developing). Another big advantage of openhab is number of different bindings for different hardware and community that is willing to write the new ones. On the other hand developers of Domoticz either ignore your suggestions or tell you to implement them yourself. Can't see a bright future with such an attitude.
@monte I fully agree with this. I have tried to get some explanation to this V_TEXT problem through multiple channels and no one is really answering anything. Either the developers have dumped the whole project or are just nasty dwarfs in a cave. GitHub shows some activity but the project seems to lay very heavily on "gizmocuz" shoulders who probably has a nasty attitude but has also answered to this V_TEXT question 7/2017 not easily implemented -> Best to help with development...so it seems that they are lacking hands...but in other hand, if the attitude is bad there aren't probably many willing to come to help.
So...same here, I'll continue using Domoticz, but actively seeking a better alternative for it and trying to build as little as possible Domoticz specific stuff.
