Adds a ghostscript timeout to `scripts/extract_pdf_with_pdftotext.sh`

The script extract_pdf_with_pdftotext.sh contains a call to ghostscript when it encounters a PDF that normal processing can't handle. For some pathological cases, ghostscript never returns useful output, and as a result the pipeline workers will keep processing the file and the pipeline may spawn multiple copies of this script.

This release adds a timeout command to the call to ghostscript, so that if the PDF fails to process within 30 seconds, the workers will receive a SIGINT back from the script.

Also includes:

• bump spacy==2.2.4
• bump pip==24.0 and setuptools==57.0 in .github/workflow