PDFs are not always scraped correctly.

[emmano]

Describe the bug
PDFs are not always scraped correctly.

To Reproduce
Attempt to fetch content using the SemanticScholar retriever

Expected behavior
PDFs are scraped correctly

Additional context
The issue lies here:
https://github.com/assafelovic/gpt-researcher/blob/0034ad08d364b34306ab9c412afb61f93eccbf11/gpt_researcher/scraper/scraper.py#L84C1-L84C34

It assumes that link have to end with the .pdf extension to be parsed as pdf documents using PyMuPDFScraper. Not all pdf links returned by SemanticScholar end in the .pdf extension. Right now, those PDFs that do not end in the .pdf extension are being scraped as if they were regular web urls using the incorrect scrapper.