Using PYCHARM remote develop the code onto a PI

This was difficult to setup for remote Raspberry PI dev. The general instructions for remote Python development using Pycharm, a Raspberry PI and SSH are here.

After securing an SSH connection, we then used the pycharm interpreter virtual env to manage the packages the project needed to run. This involved installing the following dependencies. I assume this is because the virtual environment isolates it's dependencies from the system dependencies to avoid collisions.

{
  "Pillow": "9.5.0",
  "RPI.GPIO": "0.7.1",
  "numpy": "1.25.2",
  "inky": "1.5.0"
}

To get numpy to work we had to manually install the following dependency onto the pi sudo apt-get install libatlas-base-dev.

To get PIL to work you need to install Pillow as a python package dependeny. Pillow did not work until we ran the sudo apt install libjpeg-dev zlib1g-dev on the pi.

To get the pin connections to work from python we needed to install the RPi.GPIO package with pip.

Microsoft Graph Connect

App registration was done via the azure portal. To be able to read calendar info you need to assign the Microsoft Graph - Calendars.Read API permissions to your app. You then need to set your azure application ID to the APP_ID python environment variable.

Running Python script on a schedule on Pi

To set the job to run on a regular cadence we set up the following cron script on the PI unix OS. As this script uses some setup from env variables etc, we setup a .sh script on the server to run through those pre-requisites.

Below are the steps to run the script every 20 mins.

// crontab -e
*/20 * * * * [path-to-sh] 1>/dev/null 2>[path-for-cron-err-log]

To enable cron debug logging we had to comment out the cron line in the rsyslog.conf file.

open /etc/rsyslog.conf
service rsyslog restart
service cron restart

We also had to ensure our .sh script had no access permissions.

chmod +x [path-to-sh]