Linux service to collect and transfer Xiaomi Mi Flora plant sensor data via MQTT to your smart home system, with cluster support π±πΌπ₯π‘π³
A simple Linux python script to query arbitrary Mi Flora plant sensor devices and send the data to an MQTT broker, e.g., the famous Eclipse Mosquitto. After data made the hop to the MQTT broker it can be used by home automation software, like openHAB or Home Assistant.
The program can be executed in daemon mode to run continuously in the background, e.g., as a systemd service.
The Mi Flora sensor offers the following plant and soil readings:
Name | Description |
---|---|
temperature |
Air temperature, in [Β°C] (0.1Β°C resolution) |
light |
Sunlight intensity, in [lux] |
moisture |
Soil moisture, in [%] |
conductivity |
Soil fertility, in [Β΅S/cm] |
battery |
Sensor battery level, in [%] |
An MQTT broker is needed as the counterpart for this daemon. Even though an MQTT-less mode is provided, it is not recommended for normal smart home automation integration. MQTT is huge help in connecting different parts of your smart home and setting up of a broker is quick and easy.
On a modern Linux system just a few steps are needed to get the daemon working.
The following example shows the installation under Debian/Raspbian below the /opt
directory:
sudo apt install git python3 python3-pip bluetooth bluez
In case you intent to run the service on a Raspberry Pi OS (64-bit), you need to install this packages as well to prevent the bluepy Building wheel for bluepy (setup.py) ... error
error.
sudo apt install libglib2.0-dev
git clone https://github.com/ThomDietrich/miflora-mqtt-daemon.git /opt/miflora-mqtt-daemon
cd /opt/miflora-mqtt-daemon
sudo pip3 install -r requirements.txt
The daemon depends on gatttool
, an external tool provided by the package bluez
installed just now.
Make sure gatttool is available on your system by executing the command once:
gatttool --help
To match personal needs, all operation details can be configured using the file config.ini
.
The file needs to be created first:
cp /opt/miflora-mqtt-daemon/config.{ini.dist,ini}
vim /opt/miflora-mqtt-daemon/config.ini
Attention: You need to add at least one sensor to the configuration. Scan for available Mi Flora sensors in your proximity with the command:
$> sudo hcitool lescan
LE Scan ...
4B:47:E2:DE:CE:9A (unknown)
C4:7C:8D:62:72:49 Flower care
84:C0:EF:46:B2:8A (unknown)
10:0B:F1:43:59:16 (unknown)
C4:7C:8D:62:40:29 Flower care
By the way: Interfacing your Mi Flora sensor with this program is harmless. The device will not be modified and will still work with the official smartphone app.
Some configuration options can be set via environment variables, see config.ini
for details.
A first test run is as easy as:
python3 /opt/miflora-mqtt-daemon/miflora-mqtt-daemon.py
With a correct configuration the result should look similar to the the screencap above. Pay attention to communication errors due to distance related weak Bluetooth connections.
Using the command line argument --config
, a directory where to read the config.ini file from can be specified, e.g.
python3 /opt/miflora-mqtt-daemon/miflora-mqtt-daemon.py --config /opt/miflora-config
You most probably want to execute the program continuously in the background. This can be done either by using the internal daemon or cron.
Attention: Daemon mode must be enabled in the configuration file (default).
Systemd service - on systemd powered systems the recommended option
sudo cp /opt/miflora-mqtt-daemon/template.service /etc/systemd/system/miflora.service
sudo systemctl daemon-reload
sudo systemctl start miflora.service
sudo systemctl status miflora.service
sudo systemctl enable miflora.service
A Dockerfile in the repository can be used to build a docker container from the sources with a command such as:
docker build -t miflora-mqtt-daemon .
Running the container in interactive mode works like this:
docker run -it --name miflora-mqtt-daemon -v .:/config miflora-mqtt-daemon
To run the container in daemon mode use -d
flag:
docker run -d --name miflora-mqtt-daemon -v .:/config miflora-mqtt-daemon
The /config
volume can be used to provide a directory on the host which contains your config.ini
file (e.g. the .
in the above example could represent /opt/miflora-mqtt-daemon
).
You may need to tweak the network settings (e.g. --network host
) for Docker depending on how your system is set up.
It can be worth deleting any redundant images after building a new image:
docker image prune
In the "mqtt-json" reporting mode, data will be published to the MQTT broker topic "miflora/sensorname
" (e.g. miflora/petunia
).
An example:
{"light": 5424, "moisture": 30, "temperature": 21.4, "conductivity": 1020, "battery": 100}
This data can be subscribed to and processed by other applications. From this point forward your options are endless.
Enjoy!
The following shows an example of a textual configuration using the MQTT binding introduced with openHAB 2.4. The example also uses the new internal broker.
Bridge mqtt:systemBroker:MqttBroker "MQTT Broker" [ brokerid="embedded-mqtt-broker" ]
{
Thing topic FicusBenjamin "Ficus Benjamin"
{
Channels:
Type number : light "Light Intensity" [ stateTopic="miflora/FicusBenjamin", transformationPattern="JSONPATH:$.light" ]
Type number : battery "Battery Charge" [ stateTopic="miflora/FicusBenjamin", transformationPattern="JSONPATH:$.battery" ]
Type number : temperature "Temperature" [ stateTopic="miflora/FicusBenjamin", transformationPattern="JSONPATH:$.temperature" ]
Type number : conductivity "Soil Fertility" [ stateTopic="miflora/FicusBenjamin", transformationPattern="JSONPATH:$.conductivity" ]
Type number : moisture "Soil Moisture" [ stateTopic="miflora/FicusBenjamin", transformationPattern="JSONPATH:$.moisture" ]
}
}
Number:Illuminance Miflora_Ficus_Light "Light Intensity Ficus [%d lx]" <light> { channel="mqtt:topic:MqttBroker:FicusBenjamin:light" }
Number:Dimensionless Miflora_Ficus_Battery "Battery Charge Ficus [%d %%]" <battery> { channel="mqtt:topic:MqttBroker:FicusBenjamin:battery" }
Number:Temperature Miflora_Ficus_Temperature "Temperature Ficus [%.1f Β°C]" <temperature> { channel="mqtt:topic:MqttBroker:FicusBenjamin:temperature" }
Number Miflora_Ficus_Conductivity "Soil Fertility Ficus [%d Β΅S/cm]" <lawnmower> { channel="mqtt:topic:MqttBroker:FicusBenjamin:conductivity" }
Number:Dimensionless Miflora_Ficus_Moisture "Soil Moisture Ficus [%d %%]" <humidity> { channel="mqtt:topic:MqttBroker:FicusBenjamin:moisture" }
To integrate with ThingsBoard.io:
config.ini
set reporting_method = thingsboard-json
config.ini
assign unique sensor names for your plantsAccess token
as Credential type
and the chosen sensor name as tokenTo integrate with Wiren Board in your config.ini
set:
reporting_method = wirenboard-mqtt
hostname
with address of Wiren Board controller and optionally username
and password
Your sensors will automatically appear on Wiren Board as separate devices.
Xiaomi and Mi Flora are registered trademarks of BEIJING XIAOMI TECHNOLOGY CO., LTD.
This project is a community project not for commercial use. The authors will not be held responsible in the event of device failure or withered plants.
This project is in no way affiliated with, authorized, maintained, sponsored or endorsed by Xiaomi or any of its affiliates or subsidiaries.