Refactor harvest validation for on prem deployment

[aaron-collier]

The current harvest validation is based on the prior AWS deployment and retrieves data from S3, the task needs to be refactored to use the local data harvest location and archive data to validate the harvest.

• Compare the harvest output dataframe to the most recent archive dataframe
• If they are identical, no new data was harvested and the remainder of the ETL pipeline should be skipped
• If there are changes, update the harvested output data.csv with the diff from the new harvest