More intuitive approach to get_top_tokens()

[greebie]

Currently the window that produces a graph with top tokens runs get_top_tokens() in default, which is 20. Then it splices using OVERALL_OUTPUT_SIZE. This means that if STOP_WORDS has a number of items, it will only show 20 - len(STOP_WORDS) results, instead of the desired output.

POSSIBLE SOLUTION: change the get_top_tokens() instance to use OVERALL_OUTPUT_SIZE

[greebie]

The main component for this issue belongs in the au_notebook repository.