• This is Python script that uses the text file generated by 'Wikipedia Parallel Title Extractor - https://github.com/clab/wikipedia-parallel-titles' as an input.
• This script process the input text file (mentioned above) to generate a parallel corpus.
• Output of this script (parallel corpus) can be used to train transliteration model on MOSES.

• Moodser Hussain
• COMSATS University Islamabad, Lahore Campus
• Email: [email protected]

Special thanks to Dr. Rao Muhammad Adeel Nawab and Sir Muhammad Sharjeel for their continous support.

• Download the script file (splitter.py)
• Copy the input file (generated by wikipedia parallel title script) in same directory
• run the terminal/cmd command 'python splitter.py '
• Two output files will be generated for each language seperately.

• This Script is tested on English-Urdu parallel titles extracted from https://dumps.wikimedia.org/urwiki/20180801/ using https://github.com/clab/wikipedia-parallel-titles
• Python version 3.6 was used for testing this script.