Implement fast NPY upload

[spopham]

Rewrite upload_npy_array and download_npy_array to do in-memory and have a npz compressed option

Reference: https://github.com/numpy/numpy/blob/master/numpy/lib/format.py

[anwarnunez]

👍

[anwarnunez]

Just had a quick look b/c it was on my mind. This should be pretty straightforward (see below). This should really only be allowed for uncompressed arrays and GZIP compressed arrays.

https://github.com/numpy/numpy/blob/master/numpy/lib/format.py#L326
https://github.com/numpy/numpy/blob/master/numpy/lib/format.py#L430

Using these those two functions, I think one just needs to dump that at the beginning of the StringIO (

cottoncandy/cottoncandy/interfaces.py

Line 631 in df64302

zipdata = StringIO()

and

cottoncandy/cottoncandy/interfaces.py

Line 624 in df64302

filestream = StringIO(array.data)

)