Skip to main content

A CST-based config editor for configparser

Project description


This is a module for making automated edits to an existing configparser-compatible ini file. It operates like a CST, parsing into a tree of nodes which can then be edited and written back, preserving whitespace and comments.

Quick Start

(Tested as test_example_from_readme in the test suite)

Let's say you have the following in setup.cfg:

# the package name
name = imperfect
# slurp the readme
long_description = file:

packages = imperfect

and you'd like to make an edit setting long_description_content_type in [metadata] but don't care where it goes. Default is at the end.

import difflib
import imperfect
import io
with open("setup.cfg") as f:
    data =

conf: imperfect.ConfigFile = imperfect.parse_string(data)
conf.set_value("metadata", "long_description_content_type", "text/markdown")


What if you want to have control over the odering, and want it right after long_description? Now with diffing and more internals...

import difflib
import imperfect
import io
with open("setup.cfg") as f:
    data =

conf: imperfect.ConfigFile = imperfect.parse_string(data)
metadata_section = conf["metadata"]

# Ignoring some whitespace for now, this looks like
# long_description_content_type =  text/markdown\n
# [                   entry                      ]
# [            key            ][eq][    value    ]

value = imperfect.ValueLine(
new_entry = imperfect.ConfigEntry(
    value = [value],
    whitespace_before_equals=" ",
    whitespace_before_value="  ",
for i, entry in enumerate(metadata_section.entries):
    if entry.key == "long_description":
        metadata_section.entries.insert(i+1, new_entry)

print(moreorless.unified_diff(data, conf.text, "config.cfg", end=""))
with open("setup.cfg", "w") as f:
# or
with open("setup.cfg", "w") as f:

A note on whitespace

Following the convention used by configobj, whitespace generally is accumulated and stored on the node that follows it. This does reasonably well for adding entries, but can have unexpected consequences when removing them. For example,

# this belongs to k1
k1 = foo
# this belongs to k2
k2 = foo
# k3 = foo (actually belongs to the following section)


An insertion to the end of section1 would go between k2 and the k3 comment. Removing section2 would also remove the commented-out k3.

I'm open to ideas that improve this.

A note on formats

The goal is to be as compatible as possible with RawConfigParser, which includes keeping some odd behaviors that are bugs that have been around for a decade and probably can't be changed now.

  1. Section names are very lenient. [[x]]yy is a legal section line, and the resulting section name is [x. The yy here is always allowed (we keep it in the tree though), even with inline_comments off.
  2. \r (carriage return) is considered a whitespace, but not a line terminator. This is a difference in behavior between str.splitlines(True) and list(io) -- configparser uses the latter.
  3. \t counts as single whitespace.

Supported parse options

Option                 Default  Imperfect supports
allow_no_value         False    only False
delimiters             =,:      only =,:
comment_prefixes       #,;      only #,;
empty_lines_in_values  True     True (False is very close to working)


We use hypothesis to generate plausible ini files and for all the ones that RawConfigParser can accept, we test that we accept, have the same keys/values, and can roundtrip it.

This currently happens in text mode only.

If you would like to test support on your file, try python -m imperfect.verify <filename>

Why not...

  • configobj has a completely different method for line continuations
  • I'm not aware of others with the goal of preserving whitespace


Imperfect is copyright Tim Hatch, and licensed under the MIT license. I am providing code in this repository to you under an open source license. This is my personal repository; the license you receive to my code is from me and not from my employer. See the LICENSE file for details.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for imperfect, version 0.3.0
Filename, size File type Python version Upload date Hashes
Filename, size imperfect-0.3.0-py3-none-any.whl (13.4 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size imperfect-0.3.0.tar.gz (16.0 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page