Youtube Python API repeating comments

I am trying to use Google's Youtube API to scrape comments from videos. Since videos can have many comments, I have decided to break this job into many smaller jobs

Ideally, the workflow would be:

1. scrape x number comments from a youtube video
2. keep track of the timestamp for the latest comment processed and save to a database (dynamodb)
3. when scraping is started again for the same video, only process comments new than the latest timestamp

However. It is not clear if there is a way to search a comment newer than a timestamp in order not to scrape comments already processed?

For reference, I'm currently trying to implement a solution similar to:

https://github.com/aws-solutions/discovering-hot-topics-using-machine-learning/blob/main/source/lambda/ingestion-youtube/util/video.py

But I am not sure how that solution goes about avoiding comments that have already been processed.