Load the $20$newsgroups sample dataset into Python from the scikit$-$learn li$-$

brary. Using the initial list of document data $($Hint: Make sure to set sub$-$

set$=$'all' and shuffle$=$False in order to retrieve the full dataset without ran$-$

domized reordering$),$ develop a function to tokenize each document into a list

of constituent words $($terms$).$ Limit text processing to removal of punctuation

and special characters, splitting the text using whitespace as a delimiter.