Update the reading of epub to use the semantic structure of the epub file and not only as chunks of text

[hlyl]

Problem:
Today you only read the text out of the epub file.
Solution:
we have updated the code to read and retain the structure of the epub file in beutiful soup so it can be used to change e.g the pitch of voice when reading the text.
Example:
We implemented a example of changing the pitch of voice when reading a Blockquote section of the test book.

We tested this with a text document that is also included in the code.

If you like this we will continue working in this direction.

[Bryksin]

I definitely like it! It looks way cleaner!
However, it feels like openai_tts_provider and piper were left unmodified to benefit from those changes?

[hlyl]

We spent ½ day on making the POC, and was focused on the implementation on Edge as this is what I use but would be willing to spend more time to expand the POC to be more clean, tested and look at the other TTS engines.
Should we continue to work in this direction ?

[Bryksin]

Oh yes! Please! We will definitely accept the PR