This function is meant to be triggered by an mp3 file uploaded to the podcast S3 bucket after episode info has been added to a DynamoDB table.
It reads podcast episode info from DynamoDB, figures out the mp3 duration and file size, and builds a new podcast rss feed.
Be sure that the rss file update doesn't trigger this function in an infinite loop
Limit the trigger to a prefix that has only the mp3 files, or to *.mp3 files.
Python modules are installed into a "package" folder on a linux machine (I used Cloud9) according to instructions here: AWS Lambda deployment package in Python