Citation management

GrantKit uses pandoc-style [@key] citations backed by a references.bib BibTeX database. grantkit check verifies that every citation resolves.

How it flows¶

1. Response files use [@key] citation syntax.

2. references.bib holds the BibTeX database.

3. grantkit check extracts every [@key] and reports any that don’t resolve against references.bib (unresolved_citation), plus malformed citation syntax (citation_syntax) and citations used with no bib present (missing_references_bib).

4. grantkit build assembles the sections; your BibTeX travels with the project in git.

references.bib¶

Create a references.bib in the grant directory (a fresh grantkit init scaffolds an empty one):

@article{smith_example_2024,
  title = {An Example Research Article},
  author = {Smith, John and Doe, Jane},
  journal = {Journal of Examples},
  year = {2024},
  doi = {10.1000/example.doi}
}

@techreport{treasury_report_2024,
  title = {Government Policy Report},
  author = {{HM Treasury}},
  institution = {HM Treasury},
  year = {2024}
}

Institutional authors¶

Wrap organization names in double braces so they aren’t split into first/last names:

author = {{HM Treasury}}
author = {{PolicyEngine}}

Using citations¶

Recent research shows a significant impact [@smith_example_2024].

Multiple sources confirm this [@smith_example_2024; @doe_study_2023].

Don’t hardcode (Smith, 2024) — use the key so the linter can verify it and tools can render it consistently.

Checking citations¶

grantkit check

Relevant findings:

Programmatic use¶

from grantkit.references import BibTeXManager, CitationExtractor

manager = BibTeXManager("path/to/grant")
manager.load_bibliography()
keys = manager.get_all_keys()

used = CitationExtractor().extract_citations_from_text(text)
missing = {c.citation_key for c in used} - keys