tld

• Drop Python 2.7, 3.5 and 3.6 support. Minimum required version now is Python 3.7.

• Make sure to fail silently on bad URL patterns.
• Tested against Python 3.11.
• Tested against Python 3.10.
• Updated bundled tld names.

• Move Registry class from tld.registry to tld.base.
• Reformat code using black.
• Log information on updated resources of the update_tld_names.

Release dedicated to defenders of Armenia and Artsakh (Nagorno Karabakh)
and all the victims of Turkish and Azerbaijani aggression.

• Fixed lower-cased parsed_url attributes (SplitResult) when getting tld as object (as_object=True).

2021-01-02

• Tested against Python 3.9.

• Separate parsers for (a) public and private and (b) public only domains. This fixes a bug. If you want an old behaviour:

  The following code would raise exception in past.

  
    from tld import get_tld
  
    get_tld(
        'http://silly.cc.ua',
        search_private=False
    )
  

  Now it would return ua.

  
    get_tld(
        'http://silly.cc.ua',
        search_private=False
    )
  

  If you want old behavior, do as follows:

  
    from tld.utils import MozillaTLDSourceParser
  
    get_tld(
        'http://silly.cc.ua',
        search_private=False,
        parser_class=MozillaTLDSourceParser
    )
  

  Same goes for get_fld, process_url, parse_tld and is_tld functions.

• Add Mozilla license to dist.
• Fix MyPy issues.

Note

In commemoration of Armenian Genocide.

• Correctly handling domain names ending with dot(s).

• Use Public Suffix list instead of deprecated Mozilla's MXR.

• Minor speed-ups, reduce memory usage.

2020-02-05

• Python 2.7 and 3.5 fixes.

• Adding test TLDs list to the package.

• Minor fixes in setup.py and manifest. Include missing resources for python 2.7 dist.

Note

There have been no code changes since 0.11.2. The only change is that support for Python 2.7 and 3.5 has been added.

• Added support for Python 2.7.

• Targeted releases for all supported Python versions.

2019-12-12

• Targeted releases for all supported Python versions.

2019-12-12

• Changed order of the releases (Python 3.6 and up come first, then Python 3.5).

• Make all distributions except Python 3.5 universal.

2019-12-12

• Added missing resources to the Python 3.5 release.

2019-12-12

• Bring back Python 3.5 support.

• Minor speed ups.
• More on adding typing.

Note

Since introduction of parser classes, usage of NAMES_SOURCE_URL and NAMES_LOCAL_PATH of the tld.conf module is deprecated. Also, tld_names_local_path and tld_names_source_url arguments are deprecated as well. If you want to customise things, implement your own parser (inherit from BaseTLDSourceParser).

• Drop support for Python versions prior to 3.6.

• Clean-up dependencies.

• Introduce parsers.

• Drop tld_names_source_url and tld_names_local_path introduced in the previous release.

• Minor speed-ups (including tests).

Note

This is the last release which supports Python 2.

• Make it possible to provide a custom path to the TLD names file.

• Make it possible to free up some resources occupied due to loading custom tld names by calling the reset_tld_names function with tld_names_local_path parameter.

• Fix for occasional issue when some domains are not correctly recognised.

Note

This release is dedicated to my newborn daughter. Happy birthday, my dear Ani.

• Handling urls that are only a TLD.

• Accepts already splitted URLs.

• Tested against Python 3.8.

• Fix for update-tld-names returns a non-zero exit code on success (introduced with optimisations in 0.9.4).
• Minor tests improvements.

2019-09-11

• Tests improvements.

• Optimisations in setup.py, tests and console scripts.
• Skip testing the update-tld-names functionality if no internet is available.

• Added is_tld function.
• Docs updated.
• Upgraded test suite.

• Fix an issue causing certain punycode TLDs to be deemed invalid.
• Tested against Python 3.7.
• Added tests for commands.
• Dropped Python 2.6 support.
• TLD source updated to the latest version.

• Correctly handling nested TLDs.

    This release contains backward incompatible changes. You should update
    your code.

    The ``active_only`` option has been removed from ``get_tld``, ``get_fld``
    and ``parse_url`` functions. Update your code accordingly.

• Removed active_only option from get_tld, get_fld and parse_url functions.
• Correctly handling exceptions (!) in the original TLD list.
• Fixes in documentation.
• Added parse_tld function.

    This release contains backward incompatible changes. You should update
    your code.

    Old ``get_tld`` functionality remain is moved to ``get_fld`` (first-level
    domain definition). The ``as_object`` argument (False by default) has been
    deprecated for ``get_fld``.

        res = get_tld("http://www.google.co.uk", as_object=True)

**Old behaviour**

        In: res.domain
        Out: 'google'

        In: res.extension
        Out: 'co.uk'

        In: res.subdomain
        Out: 'www'

        In: res.suffix
        Out: 'co.uk'

        In: res.tld
        Out: 'google.co.uk'

**New behaviour**

        In: res.fld
        Out: 'google.co.uk'

        In: res.tld
        Out: 'co.uk'

        In: res.domain
        Out: 'google'

        In: res.subdomain
        Out: 'www'

When used without ``as_object`` it returns ``co.uk``.

**Recap**

If you have been happily using old version of ``get_tld`` function without
``as_object`` argument set to ``True``, you might want to replace ``get_tld``
import with ``get_fld`` import:

        # Old
        from tld import get_tld
        get_tld('http://google.co.uk')

        # New
        from tld import get_fld
        get_fld('http://google.co.uk')

• Move to a Trie to match TLDs. This brings a speed up of 15-20%.
• It's now possible to search in public, private or all suffixes (old behaviour). Use search_public and search_private arguments accordingly. By default (to support old behavior), both are set to True.
• Correct TLD definitions.
• Domains like *****.xn--fiqs8s are now recognized as well.
• Due to usage of urlsplit instead of urlparse, the initial list of TLDs is assembled quicker (a speed-up of 15-20%).
• Docs/ directory is included in source distribution tarball.
• More tests.

• The fix_protocol argument respects protocol relative URLs.
• Change year in the license.
• Improved docstrings.
• TLD source updated to the latest version.

• Added base path override for local .dat file.
• python setup.py test can used to execute the tests.

• Fix relative import in non-package for update-tls-names script. #15
• get_tld got a new argument fix_protocol, which fixes the missing protocol, having prepended "https" if missing or incorrect.

• Tested against Python 3.5, 3.6 and PyPy.
• pep8 fixes.
• removed deprecated tld.update module. Use update-tld-names command instead.

2016-01-23

• Minor fixes.

• Minor fixes.
• Updated tld names file to the latest version.

• Exposed TLD initialization as get_tld_names.

• Support for wheel packages.
• Fixed failure on some unicode domains.
• TLD source updated to the latest version.
• Documentation updated.

2014-09-28

• Minor fixes.

• Force lower case of the URL for correct search.

• Making it possible to obtain object instead of just extracting the TLD by setting the as_object argument of get_tld function to True.

• Softened dependencies and lowered the six package version requirement to 1.4.0.
• Documentation improvements.

2013-12-05

• Speed up search

• Fix for URLs with a port not handled correctly.
• Adding licenses.

• Minor fixes.
• Credits added.

2013-09-13

• Python 3 support added.

• Tiny code improvements.
• Tests added.