Python package to monitor Bluetooth Low Energy advertisements and store the data in a PostgreSQL database. The tool was developed to capture temperature and humidity measurements from Xiaomi Mijia devices with custom firmware but is configurable to work with other devices. The monitor captures all the transmitted service data and only needs to know about MAC addresses and service UUIDs.
Make sure you have PostgreSQL installed and running. Create a new user and database. Check out script/create_db.sql for an example.
Follow this guide to install the build dependencies for PyGObject.
Create a fresh Python 3.8+ environment, e.g. with
python -m venv ble
source ble/bin/activate
Install the requirements and the ble_monitor package with
pip install -r requirements.txt
pip install . # Use -e if you plan to make changes
The PGUSER, PGPASSWORD, PGHOST, PGPORT, PGDATABASE environment variables need to be set to access you newly created database. See script/set_env_vars.sh for an example.
Create a configuration file similar to config/ATC_V35a.yaml. If you are planning to capture advertisements from a Xiaomi Mijia Thermometer with custom firmware, you will likely only need to adjust the list of remote devices. Otherwise, you will at least want to change the service UUIDs and the data mask (leave blank if you do not care about masking your data). Run a bluetooth scan with the following commands to get some of the required information:
$ bluetoothctl
[bluetooth] menu scan
[bluetooth] clear # this should clear filters for BLE devices
[bluetooth] back
[bluetooth] scan on
Run the following line to set up the database
python3 scripts/run.py -c <path to config file> create
Start data collection with
python3 scripts/run.py -c <path to config file>
You can use the simple API and frontend in viewer to display your data. The frontend is geared towards temperature and humidity measurements from Xiaomi Mijia devices so you will have to adjust the data extraction if you work with other devices.
You need to have node and yarn installed. Navigate to viewer/api and install the dependencies for the API with yarn install and start the api with node api.js. Start the frontend with npx http-server viewer/frontend or your server of choice.
You may want to collect data on a remote device (like a Raspberry Pi, etc.). Make sure you allow remote connections to the database on this device by adding the following line to /etc/postgresql/<version>/main/pg_hba.conf (run show hba_file in psql if you cannot locate the correct file):
host all all 0.0.0.0/0 md5
Furthermore, adjust the respective line in /etc/postgresql/<version>/main/postgresql.conf to read
listen_addresses = '*'