nr-stream

nr-stream

This package provides utilities for writing functional-style code in Python. The package originally contained only the Stream class, hence the name, but since we've adopted the terminology for letting us streamline large chunks of our code.

API

Optional objects

Represents an optional value, i.e. one that either has a valid value or is None. The class is useful to chain modifications and have them execute based on whether a value is available or not.

Example

import os
from nr.stream import Optional

opt = Optional(os.getenv("SOMEVAR"))
value = opt.or_else_get(lambda: do_something_else())
value = opt.or_else_raise(lambda: Exception("SOMEVAR not set"))
opt = opt.map(lambda value: value + " another value")
len(opt.stream().count())  # 0 or 1

Refreshable objects

A Refreshable is a container for a value that can be updated and inform listeners. A chained operations on a refreshable will be replayed if the parent refreshable is updated. This is eager evaluation, not lazy evaluation and allows performant calls to .get() without going through a lazy chain of operations each time.

Unlike Optional or Stream, the Refreshable knows no "empty" state.

This class is often useful to pass configuration data around in your application. It allows making modifications to the configuration and have it automatically propagate throughout the application.

Example

from nr.stream import Refreshable

root = Refreshable[int | None](None)
child = root.map(lambda v: 42 if v is None else v)

print(root.get())  # None
print(child.get()) # 42
root.update(10)
print(root.get())  # 10
print(child.get()) # 10

Stream objects

The Stream class wraps an iterable and allows you to build a chain of modifiers on top of it. This often greatly simplifies consecutive operations on an iterable object and its items.

Example

from nr.stream import Stream

values = [3, 6, 4, 7, 1, 2, 5]
assert list(Stream(values).chunks(values, 3, fill=0).map(sum)) == [13, 10, 5]

Important: Stream objects always immediately convert the object passed to an iterator. This means that you cannot branch stream objects, as both forks will share the same initial iterator.

Supplier objects

The Supplier class allows you to lazily evaluate the retrieval of a value, as well as chain modifications on top of it and even trace the lineage of these modifications. It provides convenience methods such as .map(), .once(), .get_or_raise(). Unlike an Optional, a supplier will treat None as a valid value and instead separately track the state of "no value".

Trying to read a value from an empty supplier raises a Supplier.Empty exception. Note that suppliers always evaluate lazily, unlike Optional.

Example

from nr.stream import Supplier

sup = Supplier.of(42)
sup = sup.map(lambda value: print(value))
assert sup.get() == None  # prints: 42
assert sup.get() == None  # prints: 42

Supplier.void().get()  # raises Supplier.Empty