JarbasAI
JarbasAI
https://github.com/MycroftAI/lingua-franca/pull/2 handles this @penrods i do not think we should handle typos, that will quickly get out of hand, i wouldn't spend any effort on that, if someone needs it...
great feedback @filips123 , that does make sense!
related to #95 , this adds a replace_token to keep track of consumed words in the remainder, however the number normalization step converts all words to numbers added a TODO...
can this please use scientific notation? counting zeros to validate long/short scale is very error prone... with scientific notation we just look at the exponent and can immediately judge if...
add unittests and this should be good, i'm not sure about the behaviour for all_digits=False, i bet there is some language where we will run into an issue like catalan...
anchor date should fallback to datetime.now(), but it should be timezone aware from user config, wsnt there a util for that local_now() or something?
since the point of normalize was making intent parsing etc easier, this just makes it harder to detect numbers or percentages, eg, a voc file containing "percent" and "%" will...
then normalize the symbol into a word
in a computer a decimal number is always represented with a . regardless of language. I don't think we will get any stt transcription ever where this isn't the case...
i think this is a great idea