Crawl for papers text

[AmitMY]

We are maintaining a large bibtex file here https://github.com/sign-language-processing/sign-language-processing.github.io/blob/master/src/references.bib

And I think that, for fine tuning language models (https://github.com/sign-language-processing/sign-gpt) it would be wise to get a collection of all cited books and pdfs. For that, we'll need to iterate all of the citations, and find a pdf, as well as crawl the citations of that citation, and if they are in the field of sign language (GPT classifier) then download them too, so we have the entire taxonomy. available.