Official reports on the 2026 Bundibugyo virus disease outbreak used to evaluate AI extraction

Archive of the 335 official reports used to evaluate AI-assisted extraction of case data in the study “AI-assisted extraction of case data from official outbreak reports for real-time Ebola transmission modelling: a development and evaluation study”. The reports concern the 2026 outbreak of Ebola disease caused by Bundibugyo virus in the Democratic Republic of the Congo, Uganda, Kenya and Nigeria, and were published between 15 May and 4 November 2026 by WHO (Disease Outbreak News; Regional Office for Africa news and weekly situation reports), ECDC, the UK Health Security Agency, national ministries of health and the DRC Institut National de Santé Publique (situation reports in French). 62 reports are in French and 273 in English. Contents of corpus.zip: archive/: each report in its original form (HTML or PDF); text/: the text extracted from each report; corpus_index.csv: one row per report, with its address, source, publication date, language, countries mentioned and SHA-256 fingerprint; FROZEN.json: the date the collection was frozen (4 November 2026) and the fingerprint of the whole collection (34dee558d9be7b06…). Reports remain the property of their original publishers and are reused under their respective terms; please cite the original source when using an individual report. The CC BY 4.0 licence applies to the index and metadata compiled by the authors. The code, the instructions given to the language model and the evaluation outputs are archived separately (doi:10.5281/zenodo.23941732).