python-iso639

python-iso639 is a Python package for accessing ISO 639 language codes, names, and other associated information.

Current features:

• A representation of languages mapped across ISO 639-1, 639-2, and 639-3.
• Functionality to "guess" what a language is for a given unknown language code or name.

Installation

pip install python-iso639

Usage

python-iso639 revolves around a Language class. Instances of Language have attributes that you will find useful.

Note that while the package name registered on PyPI is python-iso639, the actual import name during runtime is iso639 (which means you should do import iso639 in your Python code).

Creating Language Instances

Create a Language instance by one of the classmethods.

from_part3, with an ISO 639-3 code

>>> import iso639
>>> lang1 = iso639.Language.from_part3('fra')
>>> type(lang1)
<class 'iso639.language.Language'>
>>> lang1
Language(part3='fra', part2b='fre', part2t='fra', part1='fr', scope='I', type='L', name='French', comment=None, other_names=None, macrolanguage=None, retire_reason=None, retire_change_to=None, retire_remedy=None, retire_date=None)

From Another ISO 639 Code Set or a Reference Name

>>> lang2 = iso639.Language.from_part2b('fre')  # ISO 639-2 (bibliographic)
>>> lang3 = iso639.Language.from_part2t('fra')  # ISO 639-2 (terminological)
>>> lang4 = iso639.Language.from_part1('fr')  # ISO 639-1
>>> lang5 = iso639.Language.from_name('French')  # ISO 639-3 reference language name

You Get None for Invalid Inputs

The user input is case-sensitive!

>>> None == iso639.Language.from_part3('Fra') == iso639.Language.from_name("unknown language")
True

Accessing Attributes

>>> lang1
Language(part3='fra', part2b='fre', part2t='fra', part1='fr', scope='I', type='L', name='French', comment=None, other_names=None, macrolanguage=None, retire_reason=None, retire_change_to=None, retire_remedy=None, retire_date=None)
>>> lang1.part3
'fra'
>>> lang1.name
'French'

Comparison

>>> lang1 == lang2 == lang3 == lang4 == lang5  # All are French
True
>>> lang6 = iso639.Language.from_part3('spa')  # Spanish
>>> lang1 == lang6  # French vs. Spanish
False
>>> 'French' == lang1.name == lang2.name == lang3.name == lang4.name == lang5.name
True
>>> lang6.name
'Spanish'

Guess a Language: Classmethod match

You don't know which code set or name your input is from? Use the match classmethod:

>>> lang1 = iso639.Language.match('fra')
>>> lang2 = iso639.Language.match('fre')
>>> lang3 = iso639.Language.match('fr')
>>> lang4 = iso639.Language.match('French')
>>> lang1 == lang2 == lang3 == lang4
True

The classmethod match is particularly useful for consistently accessing a specific attribute from unknown inputs, e.g., the ISO 639-3 code.

>>> 'fra' == lang1.part3 == lang2.part3 == lang3.part3 == lang4.part3
True

If there's no match, None is returned. You may need to catch a potential AttributeError:

>>> lang = iso639.Language.match('not gonna find a match')
>>> lang is None
True
>>> lang.part3
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'NoneType' object has no attribute 'part3'
>>> try:
...     code = lang.part3
... except AttributeError:
...     code = None
...     print("no match found!")
... 
no match found!

Macrolanguages and Alternative Names

>>> language = iso639.Language.match('yue')
>>> language.name
'Yue Chinese'  # also commonly known as Cantonese
>>> language.macrolanguage
'zho'  # Chinese
>>> language.other_names
[Name(print='Yue Chinese', inverted='Chinese, Yue')]
>>> for name in language.other_names:
...     print(f'{name.print} | {name.inverted}')
...
Yue Chinese | Chinese, Yue

Retired Language Codes:

>>> language = iso639.Language.match('bvs')
>>> language.part3
'bvs'
>>> language.name
'Belgian Sign Language'
>>> language.status
'R'  # (R)etired
>>> language.retire_reason
'S'  # (S)plit
>>> language.retire_change_to is None
True
>>> language.retire_remedy
'Split into Langue des signes de Belgique Francophone [sfb], and Vlaamse Gebarentaal [vgt]'
>>> language.retire_date
datetime.date(2007, 7, 18)

Into the Weeds

Attributes of a Language Instance

A Language instance has the following attributes:

[1] If the ISO 639-3 code is retired, then the type attribute is None, because its value is not clearly discernible from the SIL data source.

[2] A Name instance has the attributes print and inverted, for the print name and inverted name, respectively. If reference name, print name, and inverted name are all the same, then that particular (print name, inverted name) pair is excluded from the other_names attribute. For example, for Spanish (ISO 639-3: spa), one (print name, inverted name) pair is (Spanish, Spanish) from the SIL data source, but this pair is excluded from its list of other_names.

How Language.match Matches the Language

At a high level, Language.match assumes the input is more likely to be a language code rather than a language name. Beyond that, the precise order in matching is as follows:

• ISO 639-3 codes (among the active codes)
• ISO 639-2 (bibliographic) codes
• ISO 639-2 (terminological) codes
• ISO 639-1 codes
• ISO 639-3 codes (among the retired codes)
• ISO 639-3 reference language names
• ISO 639-3 alternative language names (the "print" ones)
• ISO 639-3 alternative language names (the "inverted" ones)

Only exact matching is done (there's no fuzzy string matching of any sort). As soon as a match is found, Language.match returns a Language instance. If there isn't a match, None is returned.

Language is a dataclass

The Language class is a dataclass. All functionality of dataclases applies to Language and its instances, e.g., dataclasses.asdict:

>>> import dataclasses, iso639
>>> language = iso639.Language.match('fra')
>>> dataclasses.asdict(language)
{'part3': 'fra', 'part2b': 'fre', 'part2t': 'fra', 'part1': 'fr', 'scope': 'I', 'type': 'L', 'status': 'A', 'name': 'French', 'comment': None, 'other_names': None, 'macrolanguage': None, 'retire_reason': None, 'retire_change_to': None, 'retire_remedy': None, 'retire_date': None}

Links

• Author: Jackson L. Lee
• Source code: https://github.com/jacksonllee/iso639

License and Data Source

The python-iso639 code is released under an Apache 2.0 license. Please see LICENSE.txt for details.

The data source that backs this package is the language code tables published by SIL. Note that SIL resources have their terms of use.

Why Another ISO 639 Package?

Both packages iso639 and iso-639 exist on PyPI. However, as of this writing (May 2022), they were last updated in 2016 and don't seem to be maintained anymore for updating the language codes. pycountry is a great package, but what if you want a more lightweight package with just the language codes only and not the other stuff? :-)

If you ever notice that the upstream ISO 639-3 tables from SIL have been updated and yet this package isn't using the latest data, please ping me by opening a GitHub issue.