TTS Ebook Preprocessing Tool

[danneauxs]

I've written a program that allows me to preprocess an ebook.
https://github.com/danneauxs/TTS_ebook_preprocessing/tree/main
It converts Roman numerals to normal numerals.
I fixes anacronyms - some should be spoken and others spelled out. ie CEO and NASA one is is C.E.O. and the other is said as a word nasa.
Pagination is automatically removed so you don't have to listen to page numers.
There's a data file with defined sections. One allows you to add word for spelling replacements burger -> bergur is done automatically. When I listen to a book I keep a note on my phone of words that TTS seems to have trouble pronoucing. Another section contains Heteronyms. The program will interactively display the text and allow you choose different spellings of the word to be replaced in the text for proper pronouciaiton. So, with read you are given the choice of reed or red.
The program also converts everything to lower case. That may not be needed with eb2ab but some of the other TTS had severe issues when authors TYPED IN ALL CAPS FOR EMPHESIS.
It does a bit more. This was orginally started when i was using Edge TTS, then Koroko. Some of the substitutions are irrelivent in eb2ab but they don't hurt anything.

Anyone wanting to give it a try let me know. Just be aware, AI wrote the code, I just kept bugging it to get it right until it worked.

[DarthDobber]

I would like to give it a test run.

[danneauxs]

I'm in the middle of doing an updatge. But I'll post when it's ready. I use linux so it's written for that. I'll also have to get to work and update the help file. I got distracted after posting the offer. I will let you know.

[danneauxs]

https://github.com/danneauxs/TTS_ebook_preprocessing/tree/main
give a try and leave me some feedback.

[ROBERT-MCDOWELL]

good intention to make things easier to convert to an audiobook!
it's always good to have a pre-check of ebook strcture before to use eb2ab,
even if eb2ab:

• convert roman numerals to numeral then to word (some tts engine cannot manage numerals
  if you created a git repo so we will add you link in our README, and why not find a way to add it on eb2ab as a plugin.

[danneauxs]

Let me work on it. It's very crude as I value function over form. I don't code so it's all AI which I just started to learn to use a few months ago. I was not aware eb2ab converted roman numerals. That's smart.
I'll make a git, get the code ready as best I can to run on other computers, and I have no idea how to make a plugin.
Some funcitons may be useless for eb2ab, but I still use koroko - it handles things like e.p.a better and some word pronounciation. But, I will share the code and if somone can make it better or it's usefull that's great.

[danneauxs]

https://github.com/danneauxs/TTS_ebook_preprocessing

This is my first git and really one of only a few programs I've written. Be kind.

[ROBERT-MCDOWELL]

there is always a start for everybody.... just do what you need and we'll see in the future what I can do....

[danneauxs]

New repository - revamp of old code.
https://github.com/danneauxs/Text_Prep_TTS

[ROBERT-MCDOWELL]

you can promote it on our discord if you wish. tell precisely for which language it's working.