literature-review

Define a research question with PICO, then search at least three databases such as PubMed, bioRxiv, and arXiv using documented Boolean search strings. Aggregate results, deduplicate by DOI, screen titles and abstracts, and synthesize findings thematically following PRISMA guidelines.

Run the verify_citations.py script on your markdown document to extract every DOI, check that each resolves correctly, and pull metadata from CrossRef. It produces a verification report and properly formatted citations in your chosen style.

The generate_pdf.py script supports APA, Nature, Chicago, Vancouver, and IEEE citation styles via the --citation-style flag. It uses pandoc with xelatex to produce a formatted PDF with table of contents and numbered sections.

Broad web search uses parallel-cli, which requires installation and authentication. AI schematic generation requires an OPENROUTER_API_KEY. PubMed and arXiv searches via gget and direct APIs work without keys, while Semantic Scholar benefits from one.

PDF generation fails when pandoc or xelatex is not installed on the system. Run generate_pdf.py --check-deps to verify dependencies, then install pandoc and a TeX distribution such as MacTeX or texlive-xetex.

Yes, including preprints captures the latest unpublished findings, but they are not peer-reviewed so findings should be verified cautiously. Check whether each preprint has since been published via CrossRef and note the preprint version and date.