Python: How to pull a 'fake' modeled date from a dataframe header string and turn it into a new column

Question

I\'m working with a list of ~60k files that are modeled years that range from the year 1-59000. Each of those 60k files is in the following structure (numbers used are examples)