Published February 20, 2022 | Version 1.0

Code for converting different formats to TEI XML and evaluating results

Authors/Creators

  • 1. University of Bologna

Description

The scripts contained in this publication allow to convert references written in  JSON, Jats and generic XML to TEI XML. For each reference only certain metadata are considered, on the basis of their publication type, the ones described in the paper https://doi.org/10.48550/arXiv.2202.08469. Also, a specific script allows to evaluate the tagging quality of the references by testing them against a Gold Standard, again in TEI XML, separately created.

The main folder contains two subfolders:

  1. conversion_to_tei containing 7 python files: Jats_to_TEI_CERMINEJats_to_TEI_scholarcy, json_to_TEIjson_to_TEI_anystyle and pure_xml_to_TEI are scripts that convert one input file to TEI XML. Jats_to_TEI_CERMINE accepts as input a file parsed by Cermine (Jats output), Jats_to_TEI_scholarcy accepts as input an output file of Scholarcy in Jats only, json_to_TEI accepts as input all the json files excepted by the ones parsed by Anystyle, which are converted by json_to_TEI_anystyle. Finally,  pure_xml_to_TEI accepts as input an XML files not based on specific rules sets. Differently, generate_new_xml and retrieve_jats_metadata contain ancillary functions called by the conversion scripts to carry out specific tasks.
  2. evaluation_scripts contains two script files, get_evaluation_metrics and meta_eval. The former takes as input two files in TEI XML, one that has to be tested and one that is almost correct (Gold Standard), against which the former has to be tested. The second script contains ancillary functions called by get_evaluation_metrics to carry out specific subtasks.

Files

references_extraction_CODE.zip

Files (29.8 kB)

Name Size Download all
md5:c58feeb576ac4dcbf6659646099c0705
29.8 kB Preview Download

Additional details

References

  • Santos, Erika Alves dos, Silvio Peroni, and Marcos Luiz Mucheroni. "An Analysis of Citing and Referencing Habits across All Scholarly Disciplines: Approaches and Trends in Bibliographic Metadata Errors." arXiv.org, February 17, 2022. https://doi.org/10.48550/arXiv.2202.08469