Playing with Stackoverflow Question: 29762824

test.conf

key1:
   nested-key1 = val1
   nested-key2 = val2

key2:
   nested-key3 = val3

test.py

#!/usr/bin/env python


from pprint import pprint


data = {}
with open("test.conf", "r") as f:
    for line in f:
        line = line.strip()
        if line.endswith(":"):
            key = line.rstrip(":")
            data[key] = {}
        elif "=" in line:
            k, v = map(str.strip, line.split("=", 1))
            data[key][k] = v

pprint(data)

test2.py

#!/usr/bin/env python


from pprint import pprint


data = {}
lines = map(str.strip, open("test.conf", "r").readlines())
for line in lines:
    if line.endswith(":"):
        key = line.rstrip(":")
        data[key] = {}
    elif "=" in line:
        k, v = map(str.strip, line.split("=", 1))
        data[key][k] = v

pprint(data)

test3.py

#!/usr/bin/env python


from funcy import *  # noqa
from pprint import pprint


lines = open("test.conf", "r").readlines()
xs = map(str.strip, lines)  # strip surrounding whitespace
xs = filter(None, xs)  # strip blank lines

ys = partition_by(rpartial(str.endswith, ":"), xs)  # partition by : (denoting a key)

zs = chunks(2, ys)  # chunks of 2 [[key, values], [key, values]]

data = dict((x[0].rstrip(":"), dict(chunks(2, map(str.strip, mapcat(rpartial(str.split, "="), y))))) for x, y in zs)
pprint(data)

Trying to answer Stackoverflow Quesiton: 29762824

Transition from naive solution based on what you would normally do to a purely functional approach using the Funcy library.

• test.py -- traditions/naive approach
• test2.py -- using list comphrensions
  -- test3.py -- pure functional approach with list comprehensions and funcy

I think the last line in test3.py (data = ...) could be vastly improved. Also I think a lot of this could simply be done with itertools and list/dict comprehensions and a few functions.