FStringInterpolation 0.1.1

interpolation for python's `configparser` that does f-string substitution

FStringInterpolation - interpolation configparser that does f-string substitution

Extended Interpolation for the normal Python configparser

• Apply f-string interpolation on values.

Usage

>>> import configparser
>>> from fstringinterpolation import FStringInterpolation
>>> config = configparser.ConfigParser(interpolation=FStringInterpolation())
>>> config.read_string("""
[DEFAULT]
_pi = 3.141592653589793238462643383279502884197169399375105820974944592307816406286
[full]
pi = {_pi}
[short]
pi = {float(_pi):.2f}
[stringy]
pi = {_pi:.2s}
""")
>>> config["full"]["pi"]
'3.141592653589793238462643383279502884197169399375105820974944592307816406286'
>>> config["short"]["pi"]
'3.14'
>>> config["stringy"]["pi"]
'3.'

More examples

This will automatically create the proper title, depending on the values of a and b

Input

[DEFAULT]
a = 7
title = {a} is {"more" if float(a) > float(b) else "less"} than {b}
[A]
b = 12
[B]
b = 3
[C]
a = 3.14
b = 2.71

Output

section	title
A	"7 is less than 12"
B	"7 is more than 3"
C	"3.14 is more than 2.71"

Installing

TODO

Building

python3 -m build

Caveats

• no recursion! e.g. the following is invalid:

  [DEFAULT]
  pi = 3.141592654
  [int]
  pi = {float(pi)}
  

• the type of each value is always . if you need something else, you must explicitely convert the values:

  [DEFAULT]
  pi = 3.141592654
  [bad]
  halfpi = {pi/2}
  [good]
  halfpi = {float(pi)/2}
  

• care must be taken to not override python built-ins (that you want to use) with variables:

  [DEFAULT]
  a = 10
  b = 12
  [good]
  realsum = {sum(float(_) for _ in [a,b])}
  [bad]
  realsum = {sum(float(_) for _ in [a,b])}
  sum = 22
  [bad recursion]
  sum = {sum(float(_) for _ in [a,b])}
  

• currently, recursion is somewhat broken (unless the interpolated values are re-used as strings):

  [test]
  a = 3
  # b becomes '4'
  b = {int(a)+1}
  # GOOD: c becomes '34' ('3'+'4')
  c = {a+b}
  # BAD: d should become 7, but really fails with "int('{int(a)+1)}')"
  d = {int(a)+int(b)}