Scholarly publications are the most formal research artifacts, yet the full pipeline from citation discovery to PDF acquisition is surprisingly fragile. DOIs resolve to paywalled pages, preprint servers reorganize, and personal reference libraries live in proprietary databases that do not interoperate cleanly.
This section catalogs tools for building durable, version-controlled collections of scholarly references and their full-text content.
Citation Discovery and Curation – Finding all works that cite, are cited by, or are related to a given set of publications. citations-collector automates this across CrossRef, OpenCitations, DataCite, and OpenAlex.
PDF Acquisition – Obtaining full-text PDFs through legal open-access channels (Unpaywall, publisher OA repositories, preprint servers) and archiving them with provenance metadata in git-annex.
Reference Management – Synchronizing curated collections with reference managers like Zotero for collaborative bibliography management, BetterBibTeX export, and integration with writing workflows.
A reference collection is a living dataset. New citations appear as papers accumulate downstream citations. PDFs get updated when authors post corrections. Metadata improves as aggregators reconcile records.
By storing references and PDFs in a DataLad dataset, every change is tracked, reproducible, and attributable – the same principles that apply to code and data apply to the scholarly record itself.