Public Suffix List parser implementation for Python 3.5+.

• Compliant with TEST DATA
• Supports IDN (unicode and punycoded).
• Supports Python3.5+
• Shipped with built-in PSL and an updater script.
• Written in Pure Python with no library dependencies.

publicsuffixlist can be installed via pip.

$ pip install publicsuffixlist

Basic Usage:

from publicsuffixlist import PublicSuffixList

psl = PublicSuffixList()
# Uses built-in PSL file

print(psl.publicsuffix("www.example.com"))   # "com"
# the longest public suffix part

print(psl.privatesuffix("www.example.com"))  # "example.com"
# the shortest domain assigned for a registrant

print(psl.privatesuffix("com")) # None
# Returns None if no private (non-public) part found

print(psl.publicsuffix("www.example.unknownnewtld")) # "unknownnewtld"
# New TLDs are valid public suffix by default

print(psl.publicsuffix("www.example.香港")) #"香港"
# Accepts unicode

print(psl.publicsuffix("www.example.xn--j6w193g")) # "xn--j6w193g"
# Accepts Punycode IDNs by default

print(psl.privatesuffix("WWW.EXAMPLE.COM")) # "example.com"
# Returns in lowercase by default

print(psl.privatesuffix("WWW.EXAMPLE.COM", keep_case=True) # "EXAMPLE.COM"
# kwarg `keep_case=True` to disable the case conversion

The latest PSL is packaged once a day. If you need to parse your own version, it can be passed as a file-like iterable object, or just a str:

with open("latest_psl.dat", "rb") as f:
    psl = PublicSuffixList(f)

The unittest and PSL updater can be invoked as module.

$ python -m publicsuffixlist.test
$ python -m publicsuffixlist.update

Additional convenient methods:

print(psl.is_private("example.com"))  # True
print(psl.is_public("example.com"))   # False
print(psl.privateparts("aaa.www.example.com")) # ("aaa", "www", "example.com")
print(psl.subdomain("aaa.www.example.com", depth=1)) # "www.example.com"

publicsuffixlist do NOT provide domain name and label validation. In the DNS protocol, most 8-bit characters are acceptable as labels of domain names. While ICANN-compliant registries do not accept domain names containing underscores (_), hostnames may include them. For example, DMARC records can contain underscores. Users must confirm that the input domain names are valid based on their specific context.

Partially encoded (Unicode-mixed) Punycode is not supported due to very slow Punycode encoding/decoding and unpredictable encoding results. If you are unsure whether an input is valid Punycode, you should use: unknowndomain.encode("idna").decode("ascii"). This method, converting to idna is idempotent.

If you need to accept arbitrary or malicious binary data, it can be passed as a tuple of bytes. Note that the returned bytes may include byte patterns that cannot be decoded or represented as a standard domain name. Example:

psl.privatesuffix((b"a.a", b"a.example\xff", b"com"))  # (b"a.example\xff", b"com")

• This module is licensed under Mozilla Public License 2.0.
• The Public Suffix List maintained by the Mozilla Foundation is licensed under the Mozilla Public License 2.0.
• The PSL testcase dataset is in the public domain (CC0).

This module and its packaging workflow are maintained in the author's repository located at https://github.com/ko-zu/psl.

A new package, which includes the latest PSL file, is automatically generated and uploaded to PyPI. The last part of the version number represents the release date. For example, 0.10.1.20230331 indicates a release date of March 31, 2023.

This package dropped support for Python 2.7 and Python 3.4 or prior versions at the version 1.0.0 release in June 2024. The last version that works on Python 2.x is 0.10.0.x.

• GitHub repository: (https://github.com/ko-zu/psl)
• PyPI: (https://pypi.org/project/publicsuffixlist/)