- Author: Tuomas Lähteenmäki
- License: MIT
- Version: 0.2.13
- Type: Library
- Status: stable, test, dev
- Canonical Repository (Codeberg): https://codeberg.org/lahtis/Self_Healing_Localization
- GitHub Mirror: https://github.com/lahtis/Self-Healing-Localization
- Documentation: https://codeberg.org/lahtis/Self_Healing_Localization/src/branch/main/docs
- Demonstration: https://youtu.be/t5veRZtt3dU?si=ADD4yS5C2VwNKx9f
Self‑Healing Localization Library (SHL) is a dependency‑free Python library that automates application localization and independently repairs missing translations. The library creates, updates, and synchronizes translation files throughout the application’s lifecycle without manual maintenance.
SHL follows a self‑healing principle: when the application requests a missing translation key, the library automatically generates it, fetches a machine translation if needed, and stores the result in the correct language file. This makes localization deterministic and reduces developer workload.
- Automatic creation of missing translations
- Automatic generation of new language files
- BCP‑47 language code support (e.g., fi‑FI, pt‑BR, zh‑TW)
- GLFM‑based language validation
- Machine translation support: Microsoft Translator, DeepL, Google Translate, MyMemory, LibreTranslate, LibreTranslate Community, Papago Translate, Yandex Translate and Localhost.
- Self‑healing localization pipeline
- Unified high‑level localization engine
- Offline and online support
- Zero‑dependency core
- Free and open‑source - SHL is fully free; paid providers may require separate subscriptions.
The following translation providers have been tested and are currently operational:
- MyMemory
- DeepL
- LibreTranslate
- LibreTranslate Community
The other supported providers are implemented but have not yet been fully tested.
Developer note: SHL is a zero-budget project. Provider support is implemented first, while full provider testing is performed as resources allow. MyMemory, DeepL, LibreTranslate, and LibreTranslate Community have currently been tested successfully. Some providers may require a temporary paid subscription or other access in order to complete testing.
SHL does not process user‑audited or user‑modified files. The library does not perform self‑healing corrections on user data files or configuration files; it operates strictly within the application’s own localization layer.
SHL’s architecture is based on a layered model where the localization engine routes translation requests through a router to different providers. The provider layer uses configuration that defines provider priorities, timeouts, environment variables, and content filtering.
pip install self-healing-localizationpip install -i https://test.pypi.org/simple/ self-healing-localization==0.2.12Create a .env file in your project root (optional):
MYMEMORY_EMAIL=your@email.com
MYMEMORY_API_KEY=your-api-key
LIBRETRANSLATE_API_KEY=your-api-key
LIBRETRANSLATE_COMMUNITY_API_KEY=your-api-key
DEEPL_API_KEY=your-api-key
GOOGLE_API_KEY=your-api-key
MICROSOFT_TRANSLATOR_KEY=your-api-key
NAVER_CLIENT_ID=your-api-key
NAVER_CLIENT_SECRET=your-api-key
YANDEX_API_KEY=your-api-key
LOCAL_TRANSLATOR_API_KEY=your-api-key
DETECTLANGUAGE_API_KEY=your-api-key
OPENROUTER_API_KEY=your-api-key
GEMINI_API_KEY=your-api-key
GROQ_API_KEY=your-api-key
CLAUDE_API_KEY=your-api-key
CHATGPT_API_KEY=your-api-keyEnvironment variables are optional, but required for providers that use API keys. If a provider has no API key, SHL will still work offline and fall back to local translation or self‑healing behavior
Create a config.conf in your project root:
[SETTINGS]
language = fin
base_lang = eng
m_translation_enabled = trueThe library generates
shl-config.jsonandshl-policy-config.jsonautomatically in the project root.
Initialize the engine and start retrieving text. Missing keys are added to your JSON files automatically.
from shl.engine import LocalizationEngine
shl.setup_logging("DEBUG")
# Initialize the engine (user language = Finnish, base = English)
engine = LocalizationEngine(base_lang="eng")
# If 'welcome_msg' is missing, it is created with the given default value
title = engine.ui_text("welcome_msg", "Welcome to the App!")
print(title) # "Tervetuloa sovellukseen!" (if translation exists)- base_lang="eng" → source code strings are English
- lang_code="fin" → user wants Finnish UI
So SHL does:
-
- Look for welcome_msg in fin.json
-
- If missing:
- Create the key in fin.json
- Use the default value "Welcome to the App!" as the English source
- Translate English → Finnish
-
- Return the Finnish result
Note: SHL uses ISO 639-3 (ISO 3-letter) language codes, such as eng for English, fin for Finnish, ita for Italian, and fra for French.
- base_lang = the language your source JSON files are written in
- lang_code = the language the user wants to see in the UI right now Everything else in SHL’s behavior flows from that.
This is the language of your canonical UI strings — the language your codebase “speaks”.
SHL uses base_lang to:
- know which JSON file is the authoritative source
- know what language missing keys should be stored in
- know what language to translate from when generating other languages
This is the language the user wants to see.
Examples:
- Finnish user → lang_code="fin"
- English user → lang_code="eng"
- Italian user → lang_code="ita"
SHL uses lang_code to:
- decide which JSON file to read from
- decide which JSON file to write new keys into
- decide which language to translate to
Machine translation is disabled by default. Enable it when you want missing texts to be translated automatically.
config = {"m_translation_enabled": True} # you can overwrite config in code
engine = LocalizationEngine(lang_code="fin", config=config)
text = engine.ui_text("new_key", "Hello World!")
# → "Hei maailma!" (automatically translated to Finnish)SHL handles localized AI prompt templates the same way as UI text.
prompt = engine.template("summarize_task", "Please summarize the following text:")If the template file for the current language does not exist, it is created automatically using the base language as the source.
Switch languages at runtime without restarting the application.
engine = LocalizationEngine(lang_code="eng", config={"m_translation_enabled": True})
# Switch to Finnish
engine.set_language("fin")
print(engine.ui_text("greeting", "Hello!")) # "Hei!" (Machine-translated)
# Switch to Swedish
engine.set_language("swe")
print(engine.ui_text("greeting", "Hello!")) # "Hej!" (Machine-translated)# Brazilian Portuguese and European Portuguese in separate files
engine = LocalizationEngine(lang_code="pt-BR") # → pt-br.json
engine = LocalizationEngine(lang_code="pt-PT") # → pt-pt.json
# Traditional and Simplified Chinese in separate files
engine = LocalizationEngine(lang_code="zh-TW") # → zh-tw.json
engine = LocalizationEngine(lang_code="zh-CN") # → zh-cn.jsonfrom shl.engine.translation import translate_text
# Automatically chooses the best provider
result = translate_text("Hello World", target_lang="fin")
print(result) # "Hei maailma"Check the latest documentation files.
Contributions are welcome.
This project aims to become a new standard for open‑source localization — simple, automatic, and self‑maintaining.
MIT License — free for personal and commercial use.
Localization should never be a burden.
With SHL, any project can become multilingual — automatically, reliably, and without manual maintenance.
No more missing translations. No more incomplete language packs. Localization that heals itself.
#localization • #i18n • #l10n • #self-healing • #translation • #multilingual #json • #python • #developer-tools • #automation • #templates • #cli #ai-assisted • #language-files • #internationalization • #localization-engine