New PathReader features

[louizcerca]

Added the following features to PathReader, none of which impact current catalog configurations:

• Filtering the date of files based on a minimum date (min_date parameters on read_args) instead of only through a moving window (history_length and time_scale)
• Allows using a wildcard(*) on path arg to solve specific cases where only the last date is not enough to determine if a file should be loaded
• Applying a function passed as a string on read_args to paths to more advanced filtering
• Setting the file name as a column on loaded dataframes. Useful when the file have no versioning column in it
• Applying a function passed as a string on read_args during file loading, before the concat of all read dataframes. The main objective is to allow files to be treated individualy, grouped, filtered, have specific columns created to reduce memory usage while and after using the PathReader. Example: A daily dataset that will be only used in monthly aggregation can consume much less resources during the parallel I/O of multiple files if it is grouped right after loading and not after loading all the files. This is very important for low RAM environments