This is probably not possible due to copyright issues, but assuming I've purchased a large number of Kindle books over the years, is there a way I can feed them into an AI such that I can have it provide code assistance as I work? I know I can do this with GPT and PDFs, but most of these texts are not in PDF format. I know CoPilot and JetBrains <i>claim</i> to have AIs that have all of this info, but for reasons we all know, I need to keep a "closed corpus".
Most ebook formats are some sort of container + metadata + html +css. So you need only the files, extract the html and feed the text in it to the statistic model.