Drop Numpy as a hard requirement
nicktimko opened this issue · 1 comments
nicktimko commented
Including Numpy as a necessary requirement is extremely heavy for a terminal printer. It seems like you just use it to convert a dataframe to an array and to do a floor function on a scalar(!). The latter can be easily replicated through math.floor in the stdlib, and for the former, you could attempt to import Numpy if it exists inside the dataframe function (if someone's giving you a Pandas dataframe you know there'll be Numpy).