refactoring dataprep to use dask and adding sample script

[tvt173]

hi @joamatab , this is a proof-of-concept using a dask backend for dataprep, which would enable parallel computation and visualization of task graphs. please see the example script i added which both generates a gds and the accompanying task-graph visualization, which details the computation flow.

i don't know yet how to customize the visualization to have a bit better information in the square nodes, but i think it's pretty neat that we can extract such a graph out of this computation, which is still defined concisely and naturally

[tvt173]

also note that we need to figure out a way to serialize region objects if we want to use multiprocessing or distributed backends for dask. right now it is just using the multithreaded backend

[tvt173]

and also, you need to of course install dask and also ipycytoscape for the demo to work

[joamatab]

looks great! thank you Troy!

[joamatab]

merged it to master, and merged the dataprep notebook examples