[FEA] Create a single RAPIDS library

[miguelusque]

Is your feature request related to a problem? Please describe.
Hi all,

I would like to share with you the following idea:

So just as we have the following imports:

import pandas as pd
import tensorflow as tf
import matplotlib as plt

why do not create a rapids library that could be imported as follows?

import rapids as rs

or

from rapids import cudf

I am aware that there are pros about having each library separate, but I think there are also pros about having a single library, such us

• easier for the data scientists to install: conda install rapids
• it creates a kind of identity around rapids (we do not think on cudf or cuml as independent identities, but as a whole).
• we might adopt 'rs' as a shortcut for rapids (currently, I have only found the following use of 'rs' import rhinoscriptsyntax as rs)

Please, share your thoughts about the above.

Thanks!

P.S.: There are also cons about having a huge library, but I think the pros exceed the cons.

[kkraus14]

This is out of scope for now, possibly to be revisited in the future.