Frigate Event Processing

Source code available on the frigate-export-ffmpeg repository

Now that the API has been flushed out to retrieve the proper events, a script is needed to download and begin the process of extracting the data from these events. This project was initially a study and practice of writing bash scripts, but has since been rewritten in Python for better performance and structure of execution.

The Process

Download the clips frigate_download.py

• executable bash file executed via cron
• when no date provided, use current date
• allow for input of custom date or date ranges
• place these files in a directory for later processing

Video Data process_files.py

• iterate over a default or specific directory
• use ffprobe to obtain length
• output JSON for DB insert
• update file names to <timestamp>.mp4
• move files to new directory for processing

Notes: by using the epoch timestamp, a unique identifier is created allowing for quick checks during file creating, processing, and DB insert.

Example Output:

{
	"length": 12.144,
	"src": "1719813362.mp4",
	"friendlyDate": "24-06-30--22-56-02",
	"epoch": 1719813362,
	"starred": false
}

Process the Videos ffmpeg_process.py

• iterated over a directory or JSON file
• crop, resize, mask blur video
• create png graphic of video for over

Whats Next?

• Create a NodeJS API MongoDB endpoint