Skip to main content

Library for working with Valve's VDF text format

Project description

Latest version released on PyPi MIT License Test coverage Build status of master branch
SonarCloud Rating SonarCloud Rating SonarCloud Rating

Pure python module for (de)serialization to and from VDF that works just like json.

Tested and works on py2.7, py3.3+, pypy and pypy3.

VDF is Valve’s KeyValue text file format

Supported versions: kv1
Unsupported: kv2 and kv3


You can grab the latest release from or via pip

pip install vdf

Install the current dev version from github

pip install git+

Problems & solutions

  • There are known files that contain duplicate keys. This is supported the format and makes mapping to dict impossible. For this case the module provides vdf.VDFDict that can be used as mapper instead of dict. See the example section for details.

  • By default de-serialization will return a dict, which doesn’t preserve nor guarantee key order on Python versions prior to 3.6, due to hash randomization. If key order is important on old Pythons, I suggest using collections.OrderedDict, or vdf.VDFDict.

Example usage

For text representation

import vdf

# parsing vdf from file or string
d = vdf.load(open('file.txt'))
d = vdf.loads(vdf_text)
d = vdf.parse(open('file.txt'))
d = vdf.parse(vdf_text)

# dumping dict as vdf to string
vdf_text = vdf.dumps(d)
indented_vdf = vdf.dumps(d, pretty=True)

# dumping dict as vdf to file
vdf.dump(d, open('file2.txt','w'), pretty=True)

For binary representation

d = vdf.binary_loads(vdf_bytes)
b = vdf.binary_dumps(d)

# alternative format - VBKV

d = vdf.binary_loads(vdf_bytes, alt_format=True)
b = vdf.binary_dumps(d, alt_format=True)

# VBKV with header and CRC checking

d = vdf.vbkv_loads(vbkv_bytes)
b = vdf.vbkv_dumps(d)

Using an alternative mapper

d = vdf.loads(vdf_string, mapper=collections.OrderedDict)
d = vdf.loads(vdf_string, mapper=vdf.VDFDict)

VDFDict works much like the regular dict, except it handles duplicates and remembers insert order. Additionally, keys can only be of type str. The most important difference is that when trying to assigning a key that already exist it will create a duplicate instead of reassign the value to the existing key.

>>> d = vdf.VDFDict()
>>> d['key'] = 111
>>> d['key'] = 222
>>> d
VDFDict([('key', 111), ('key', 222)])
>>> d.items()
[('key', 111), ('key', 222)]
>>> d['key']
>>> d[(0, 'key')]  # get the first duplicate
>>> d[(1, 'key')]  # get the second duplicate
>>> d.get_all_for('key')
[111, 222]

>>> d[(1, 'key')] = 123  # reassign specific duplicate
>>> d.get_all_for('key')
[111, 123]

>>> d['key'] = 333
>>> d.get_all_for('key')
[111, 123, 333]
>>> del d[(1, 'key')]
>>> d.get_all_for('key')
[111, 333]
>>> d[(1, 'key')]

>>> print vdf.dumps(d)
"key" "111"
"key" "333"

>>> d.has_duplicates()
>>> d.remove_all_for('key')
>>> len(d)
>>> d.has_duplicates()

Project details

Download files

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

Source Distribution

vdf-3.4.tar.gz (11.1 kB view hashes)

Uploaded source

Built Distribution

vdf-3.4-py2.py3-none-any.whl (10.4 kB view hashes)

Uploaded py2 py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page