Open chungjryan opened 3 weeks ago
To my knowledge, this is isn't a thing.
Likely, we can just add something like this to our nlp library:
def replace_symbols_with_words(text):
replacements = {
'%': 'percent',
'&': 'and',
'$': 'dollar',
'@': 'at',
# etc.
}
for symbol, word in replacements.items():
text = text.replace(symbol, word)
return text
This is easier than using a switch case, and although it's not the prettiest it will get the job done.
@chungjryan unless you are against me doing this, I will implement it and adjust test cases and process_text accordingly.
@ianhajra naw ur good
I think we should figure out how to make it such that output is only words and no symbols are allowed. Lets look into the Speech_recognition documentation and hope theres a toggle there because otherwise we're gonna have one hell of a switch case statement.
Discovered issue in #30, in tests/nlp_tests/test_stt_use_cases buggy output is commented