feat: added progress bar for paragraphs
feat: consume wikipedia summary from keyword
feat: consume directly from pdb
feat: consume directly from webpage, with several fallback methods and
    ability to change what matters to something else than "p"
feat: consume directly from user inpu
style: renamed a lot of function to make it more clear
fix: instead of using csv library, use pandas to output as csv
feat: output as json
feat: output directly the pandas dataframe
feat: all functions include docstrings
feat: all code is PEP compliant
fix: set an environnement variable for the tokenizer otherwise
    you get a warning written at the top of the output file
fix: prints a message when autocard is loading as it can be long
    on slow devicecs
refactor: self.qg is called differently, it is now wrapped so as to be
    part of a dataframe
feat: txt containing wikipedia style reference like [5] is sanitized
fix: when outputting as csv, all included commas are escaped
feat: consume_web uses several ways to find the title of the webpage
fix: renamed self.clear to self.clear_qa for clarity
feat: added self.string_output() to allow the user to get directly the
    output instead of only being able to print it
feat: self.print and self.pprint allow to print and pretty print the
    output
new: added a commented section that allows to pause the script
    using ctrl+c then resume it later. Very useful when running on very
    large text but could potentially cause issues when uploading it to
    PyPi so I commented it

mentions downloading punkt for first time users

_sanitize_text was missing a docstring

fix: