The m4a generate by the tools are lack of cover.

This table reading is already in place, but for the reader is not clear when it start and the iteration with them.

Desiderable: add a parameter to add verbosity reading of table

In many book is possible to find link to reference, but the reading of those numbers or phrase is not natural in a lecture section.

In all scripts exist a method used to extract chapter text, the analysis required here need to take this text as input and remove possible reference.
Most of times a reference Is defined with following syntax [1] where 1 is the reference.
Reference examples:

• [1]
• (Author, 1999)
• phrase.1

Desiderable: add an optional flag to remove the references from text.

Right now the bit_rate is the only solution I found to reduce the final file dimension.

This is a feasibility, then the desiderable - is to have as output

• pull request with a solution
  Or
• analysis of all compression path tried

As BACK_END_TTS Need to be move to m4b.py, maybe also other constante global Need to be moved to library.

This Activity Need to reduce duplicate code.

Need to create a blog something similar to explane how to use the tool and how contribute

Parsing and interpretate It in order to generate text

• #35
• #37
• use case diagram
• Activity diagram
• statechart diagram
• review all code to be consistent

1. The parser for each diagram shall stay under a specific folder plant_uml
2. Each parser shall return a graph represent the diagram read

Starting from PowerPoint diagram generate a description.

It can be done using phrase as:

The box "input" are linked to box "System A".
The box "System A" are linked with text "output"

The library used is pptx-python, the objects must be retrived are Shapes and Autoshapes (https://python-pptx.readthedocs.io/en/latest/api/shapes.html#shape-objects-autoshapes)

In order to reduce the bug injection and improve readability of code.

Desiderable: create a report of changes for each commit generated with a GitHub action

The functions in the scripts are not all documented.
To improve quality and readability of scripts, need to add at least header comments for each function.
Desiderable is to add the typing for each function and comments, header comments.

Sometime a book Is not wrote in our mother tongue, to solve It we can use api of DeepL to translate the documents in our preferiti language.

All the scripts right now are used only under Windows.
Can be useful to expand compatibility at least with MacOs.
Desiderable is to test it under ubuntu and other linux distro.

Update all scripts to be compliance with PyLint rule active

• fix code style (ba59592)
• refactor pptx2audio.py (df7e6c5)
• reduce cyclomatic complexity (39ccde8)
• reduce nesting function
• analyze all bottleneck performance
• remove duplicate code
• #30

when an epub are converted in m4b, the following steps are executed:

1. extractiong all data from epub in a temporary folder
2. extract guide from epub
3. for each file (not in guide)
   • extract text and save it (on the folder where the script is executed)
   • generate mp3 and save it (on the folder where the script is executed)
4. generate ffmetada starting from mp3 generate previously
5. merge all mp3 and ffmetada in unique M4B file

The goal of this activity is remove the temporary file.

Similar to other script, can be useful to have a reader for excel files.

Desiderable:

• add requirements
• read simple table and sheets
• say how many graph are present in the page (in the next step, we want read also graphs)

Create a script base on other in report, to read in input a class diagram (plantuml) and generate:

• plantuml class-diagram parser
• textual equivalent to diagram
• generate the final audio
• unit tests

Docs: https://plantuml.com/class-diagram

As last commit the GitHub action failed.

The tool support only italian language, add other language require to configure properly each backend provided to the tool.

To add new language, the developer Need to modify the global dictionary as following:

• LANGUAGE_DICT = {"it-IT":"it"}
• LANGUAGE_DICT_PYTTS = {"it-IT":"italian"} in m4b.py
• TITLE_KEYWORD = {"it-IT":"TITOLO", "en":"TITLE"}
• CHAPTER_KEYWORD= {"it-IT":"CAPITOLO", "en":"CHAPTER"} in docx2audio.py

Desiderable: looking for hard coded strings and generalize it with a global dictionary, as previous one.

Desiderable: Usage of Google Lens or equivalent service to analyze an image and retrieve information about it

• extract image from slides
• convert in JPEG format (to compresse image)
• sent to generative AI (as bing copilot or chatgpt) to retrived image context

Create a script base on other in report, to read in input a class diagram (plantuml) and generate:

• plantuml sequence diagram parser
• textual equivalent to diagram
• generate the final audio
• unit tests

Docs: https://plantuml.com/sequence-diagram

The tool shall permit to add custom pause track between two chapters, in this way the audio read are not "flat" and improve the listener focus.