Things to improve in the next version iteration

[gianluccacolangelo]

The ones I can think now are:

• Implementation of Issue #1
• A way to detect highlights that already had been converted to flashcards.
• Support for custom cards styles. Would be good if they can be directly imported from anki.
• Support for pictures and equations. This would have at least three parts:
  • Implementation of pictures and equations extractors in pdf_handler module.
  • Implementation of multi-modal models that can process images
  • Implementation of images and equations in context and FlashcardGenerator
  • Right now, we generate 1 flashcard per highlight. It would be good to introduce "modes" where the LLM can decide (or be obligated) to generate more than 1 flashcard per highlight.
  • To extract more metadata in pdf_handler, so we can add the title of the book as a tag in the card, for example.

[gianluccacolangelo]

Another one:

• The possibility to fine-tune the LLM in your existing cards. This even has a standard procedure

[Aguscarpio]

Another one:

• The possibility to fine-tune the LLM in your existing cards. This even has a standard procedure

what do you mean by "standard procedure" in this context?