DMG is an application that aims to give ease to orientalists, arabists and Islamic scholars in the German-speaking area by providing automatic transliteration.
The app is currently deployed on PythonAnywhere (https://transliteration.eu.pythonanywhere.com). Try it out now!
git clone https://github.com/theRealProHacker/dmg.git
pip install -r requirements.txt
py app.pyNow navigate to the localhost URL provided (http://localhost:5000)
- Generally letter replacement
- Hyphenation of particles wa-, fa-, sa-, bi-, li-, ka-, al-
- Sun letter assimilation الشَمس: aš-šams
- Detection of idafah and transliteration of ta marbutah as either "t" or "h"/"" accordingly:
- مَكتَبَةُ كَبيرَة: maktaba kabīra or maktabah kabīrah
- but مَكتَبَةُ الأُستاذِ: maktabat al-ustāḏ
- Consideration of hamzatul wasl
- انْكَسَرَ: inkasara
- الَّذينَ: allaḏīna
- اخرُج: uḫruǧ
- فَانتَقَلَ: fa-ntaqala
- هُم الكُتّاب: hum ul-kuttāb
- عَن الْكِتابُ: ʿan il-kitāb
- Inseration of missing alifs: هَذَا: hāḏā
- Removal of silent letters: قَلوا: qalū
- Nisba as ī: al-ʿarabī
- Many different (boolean) settings
- whether the text should be transliterated in pause or not
- whether ta marbutah should be transliterated as h or not
- whether diphthongs should be transliterated as ai/au or ay/aw
- whether geminated half vowels should be transliterated iyy/uww or īy/ūw
- whether iyy/uww at the end should always become ī/ū
- whether a hamza in the first position should be transliterated
- whether the personal suffixes -hu and -hi should be transliterated as they are pronounced
- Automatic vocalization
Please don't hesitate to raise an issue or to make a pull request. More information for contributing can be found here
The software is licensed under CC BY-NC except for the directories arab_tools and data which are licensed under GPL v3.
However, the license for data/ner-gold-standard is in that directory.
This project is part of my bachelor thesis that will be published soon.