Introduction
============

**nuts-flow** is a data processing pipeline that is largely based on Python's 
`itertools <https://docs.python.org/3/library/itertools.html>`_.

**nuts** are thin wrappers around itertool functions and
provide a ``>>`` operator to chain iterators in pipelines
to construct data flows. The result is an easier to read flow of data.

The following two examples show the same data flow. The first is
using Python's itertools and the second is using **nuts-flow**:

>>> from itertools import islice
>>> list(islice(filter(lambda x: x > 5, range(10)), 3))
[6, 7, 8]

>>> from nutsflow import Range, Filter, Take, Collect, _
>>> Range(10) >> Filter(_ > 5) >> Take(3) >> Collect()
[6, 7, 8]

Both data flows extract the first three integers 
in the interval [0, 8[ that are greater than five. However, 
the linear arrangment of processing steps with **nuts-flow** is
easier to read than the nested calls of itertool functions.

**nuts-flow** is the base library for 
`nuts-ml <https://github.com/maet3608/nuts-ml>`_, a
data pre-processing pipeline for deep learning.