Pyspark: Use ffmpeg on the driver and workers

Question

I\'m trying to process a batch of audio files on my Spark cluster (in Databricks). I have an init script that installs ffmpeg (with apt) and pydub (with p