sl10n

sl10n is a library that takes a unique approach to dealing with localization by using statically typed translation keys.

Features

• Type-safe: we use statically typed translation keys, so you don't have to worry about making a typo or using a wrong key.
• Explicit: you get the exact result that you expect.
• Versatile: you can use it in any project.
• Self-sufficient, no other tools required.
• Written purely on Python.
• Easy to use.
• Small.

Why?

Imagine you have a lang folder containing all our translation files and a parser that reads these files and stores them in a mapping.

{
    "my_key_1": "My text",
    "my_key_2": "Wrong text"
}

import parser  # your handwritten parser

locales = parser.parse('/lang')
locale = locales.get('en')

print(locale.get('my_key_1'))  # My text

You may probably think "Sounds pretty simple" and you'd be right. This approach is pretty common (e.g., Minecraft mods use it).

But it's really error-prone. You can easily make a typo or refer to a different key with similar name.

print(locale.get('my_key_I'))  # my_key_I
print(locale.get('my_key_2'))  # Wrong text

This becomes a real trouble once you change the schema of your translation files and even IDEs won't help you.

So?

sl10n fixes this by introducing locale containers.

In short, a locale container is a spec of your translation files that contains all possible keys. At parsing, all your localization gets collected into locale containers, so you can access them freely.

sl10n defines a base class for your locale container (SLocale) and a parsing system (SL10n).

from sl10n import SL10n, SLocale


class MyLocale(SLocale):  # your locale container, it MUST inherit from SLocale
    my_key_1: str
    my_key_2: str


l10n = SL10n(MyLocale, 'lang')  # creating a reference for system

l10n.init()  # execute parser

The key difference is that now you can access your translated strings as class attributes.

locale: MyLocale = l10n.locale('en')  # returns default one if this language wasn't found

print(locale.my_key_1)  # My text

That way, your IDE can suggest what keys you can use and also tell you if you made a typo.

print(locale.my_key_I)  # Unresolved attribute reference 'my_key_I for class 'MyLocale' 

You can still access your locale container dynamically if you want (e.g. when the key is known only at runtime).

key = f()  # returned 'my_key_1'

print(locale.get(key))  # My text

Other features?

• If your translation files don't follow the spec - SL10n will notify you and also try to fix it:

  • add all undefined keys
  • move all unexpected keys to the bottom of the file

• You can also create new translation files:

  from sl10n import SL10n, SLocale
  
  
  class MyLocale(SLocale):
      my_key_1: str
      my_key_2: str
  
  
  l10n = SL10n(MyLocale, 'lang')
  
  l10n.create_lang_file('de')  # copy the contents of default language file ('en') to a new file
  

• Define what filenames should be ignored:

  l10n = SL10n(MyLocale, 'lang', ignore_filenames=['config', 'tags'])
  

• Make a custom parsing implementation:

  from sl10n import SL10n
  from sl10n.pimpl import ParsingImpl
  import yaml
  
  
  class YAMLImpl(ParsingImpl):
      file_ext = 'yml'
  
      def __init__(self, module=yaml, *args, **kwargs):
          self.module = module
          self.args = args
          self.kwargs = kwargs
  
      def load(self, file):
          return self.module.load(file, self.loader)  # yaml.load and yaml.dump are not safe
  
      def dump(self, data, file):
          self.module.dump(data, file, self.dumper, *self.args, **self.kwargs)
  
  ...
  
  l10n = SL10n(MyLocale, 'lang', parsing_impl=YAMLImpl())
  

• Apply some modifiers to your file (todo: make a docs explaining modifiers):

  {
      "my_key_1": "My text",
      "my_key_2": "Wrong text",
      "$redump": true  // redumps the file anyway
  }
  

  {
      "my_key_1": "My text",
      "my_key_2": "<not finished>",
      "$exclude": true  // excludes the file from parsing
  }