This project implements text normalization for Farsi(Persian) language.
it contains types below :
- normalizing numbers
- normalizing dates
- normalizing times
- normalizing currency
- normalizing measurement (physical measurement)
- normalizing phone number and ID number
- normalizing punctuation
- normalizing miscellaneous abbreviations
for text-to-speech and speech-to-text (TTSv1(default) , TTSv2, STT).
python main.py [input file address] [output file address] [version] [type1, type2, ....]normalize all types for text-to-speech version 1
python main.py inp.txt out.txt normalize time and date for speech-to-text
python main.py inp.txt out.txt TTSv2 -t -d- by declaring a type the normalizer Limited to the declared type !
- The difference between TTS version 1 and TTS version 2 is in the way the punctuations are normalized