Raspberry-Python-MeteoStation

This is my side project which helped me learn how to use Raspberry Pi. I used it to create some kind of meteo station with two sensors connected via I2C interface. With Python scripts Raspberry execute measurements and save them to the PostgreSQL database. All data are presented by Django application. This project was also a good opportunity to remaind how write Ansible playbooks for automatic deployment.

Stack

• Ubuntu
• Python
• Libs: SQLalchemy, smbus2, celery
• Django
• HTML
• Ansible
• Raspberry Pi with I2C interface

Requirements

• Raspberry Pi 4 model B with Ubuntu server 20.04
• GY-68 BMP180 sensor
• GY-30 H1750 sensor
• 1 Host with SSH access to Raspberry for Ansible (pref. Linux)

Installation

Prepare your Pie

Create user

First of all you need to create user that will be running application on Raspberry Pi. In my case it will be "meteo" (this user is already related to ansible inventory). You can create user with this command:

# sudo useradd -m <username>

This command will create system user with given username and home directory. Next, you have to add this user to two groups: 'sudo' and 'i2c'.

# sudo usermod -a -G sudo <username>
# sudo usermod -a -G i2c <username>

To be able to log in with this user, change the password with command:

# sudo passwd <username>

This user will be running our MeteoApp. Also we will use it as ansible user to deploy application.

Allow using sudo without password

Now it is very important to allow ansible execute some commands with sudo. If you use sudo, you have to provide user password for authentication. But we want deploy application full automaticly without waiting for password prompt. In this case we need to change one line in sudo group settings. Execute command:

# sudo visudo

And find line:

%sudo   ALL=(ALL:ALL) ALL:ALL

Change it to:

%sudo   ALL=(ALL:ALL) NOPASSWD:ALL

Now if you relog, you can use sudo without providing password.

Set environment variables

Last thing we have to do with Raspberry is setting environment variables for our installation. Application and scripts have following ENVs:

export DB_PASSWORD=
export DB_HOST=
export DB_USER=
export DB_NAME=
export DB_PORT=
export WEB_INTERFACE_SECRET=
export APP_PATH=
### Don't forget to fill this variables with values ###

I hope that everyone knows what fist 5 variable means :) . However WEB_INTERFACE_SECRET stores secret key for Django application and APP_PATH should be filled with path to directory for application's source code. Put this variables in environment script such as a:

/home/<username>/.profile

Connect your sensors

It is quite simple. Just take protoboard and connect wires as shown below.

Connect SDA bus to GPIO 2 and SCL bus to GPIO 3. Next connect GND and Vcc to Raspberry

Send your public ssh key

Now go to your system with Ansible installed. We need to create SSH keys for your user and send public key to Raspberry. To create pair of keys use command:

# ssh-keygen

Now to send public key to new user on Raspberry execute ssh-copy-id command:

# ssh-copy-id <raspberry_username>@<raspberry_ip>

After that If you try to log in with SSH to raspberry you won't be asked about password.

Customize your inventory

Firstly clone this repository and go to the 'ansible' directory.

There is some values that you have to change in inventory. All of them are listed below:

ansible_host=<raspberry_ip>
ansible_user=<raspberry_username>
ansible_ssh_private_key_file=<path to private key that we generated moment ago. (Path on ansible host ofc.)>
env_path=<path to script where you put environment variables>

Install ansible community modules

We need two modules (community.postgresql.postgresql_db and community.postgresql.postgresql_user) to manage PostgreSQL database. You can install them with following command:

ansible-galaxy collection install community.postgresql

Start installation

To start installation process you have to execute following command in 'ansible' directory:

ansible-playbook -i inventory.yaml main.yaml

This command will start installation and you have to wait until end.

Playbooks summary

system_init.yaml

This playbook will update apt packages and install python3 with all required libraries

app_configuration.yaml

This playbook will clone repository and will put source code in APP_PATH. At the end it will install all python requirements with pip.

required_software_apt.yaml

This playbook will install redis server for message broker. It will also install PostgreSQL where measurements data will be stored. There is additional libreries with 'i2c-tools' in case you need to debug sensors connection.

configure_postgresql.yaml

This playbook will create user and database in PostgreSQL. This user will be used to connect application with database through SQLalchemy.

setup_cron_task.yaml

This playbook will add cron task that run your appliction (WebApp with gunicorn, redis, celery worker) on every system reboot.

start_app.yaml

This playbook just restart your system. This is required to apply all new configuration and let cron run application.

Application overview

Web interface

All measurements are presented with Django application. There are three URI adresses:

/temperature (also '/')
/light
/pressure

Each page has three tables. First table shows current measurement. If you refresh website, Django will execute measurement and show current fresh data. Second table shows all measurements with date. Last table shows avarage values for given periods of time.

Celery

There is defined one task for celery in meteo_web_interface/tasks.py. If this task is executed, sensors are activated and measurement data are saved to database. This task is executed once per hour by default. It is configured in settings.py via crontab module.

MeteoLib

It is package with classes used to create this app. You can find there classes to controll sensors, or class which is used to connect with database.

TODO

• Create plot with database data and print it on websites