Pubmed parser: a python parser for pubmed open-access XML subset and MEDLINE XML dataset XML dataset

How can biomedical XML become usable research data?

PubMed Parser converts biomedical article XML into Python data structures for text mining and bibliometric analysis. It supports the PubMed Central Open Access subset and MEDLINE, giving researchers a common starting point for extracting publication information without building a separate XML parser for each project.

What the study found

• The library exposes article metadata, references, paragraphs, and figure captions through separate parsing functions.
• MEDLINE and full-text article XML have different structures; the appropriate parser depends on the input collection.

How the study works

The software uses lxml to read XML and returns dictionaries or lists of dictionaries. Article identifiers and reference links allow extracted text and citation information to be joined in downstream analyses.

Scope and limitations

• Available fields depend on the source XML. Parsing does not correct incomplete metadata or establish the scientific validity of an article.
• Access and reuse terms for article content are separate from the software license.

Using this work

Start with the example XML files and parsing functions in the linked repository, then test a sample of your own corpus. Record the package version and input collection. The software is MIT licensed; cite the JOSS paper when using it in research.

Implementation and documentation

Cite this work

Titipat Achakulvisut, Daniel E Acuna, Konrad Kording (2020). Pubmed parser: a python parser for pubmed open-access XML subset and MEDLINE XML dataset XML dataset. Journal of Open Source Software. https://doi.org/10.21105/joss.01979

@article{achakulvisut2020pubmed,
  title = {Pubmed parser: a python parser for pubmed open-access XML subset and MEDLINE XML dataset XML dataset},
  author = {Achakulvisut, Titipat and Acuna, Daniel E and Kording, Konrad},
  year = {2020},
  publication_date = {2020-02-08},
  journal = {Journal of Open Source Software},
  volume = {5},
  number = {46},
  pages = {1979},
  doi = {10.21105/joss.01979},
  url = {https://joss.theoj.org/papers/10.21105/joss.01979}
}