AHU-Brasil Catalogue: a searchable database of the Arquivo Histórico Ultramarino summaries on colonial Brazil

Read full text
Authors: Saulo Rogério Pacheco Rocha
Cited by: 0

Abstract

Catálogo AHU-Brasil: banco de dados e motor de busca dos resumos do Arquivo Histórico Ultramarino sobre o Brasil colonial Version of 28 October 2026 Saulo Rogério Pacheco Rocha (UFSC) · ORCID 0000-0003-3715-6706 Versão em português: LEIAME.md. This deposit is the new version of record 10.5281/zenodo.20451215 ("Semantic Search Engine andVernacularity Predictor for the Southern Macro-Region of Brazil AHU Catalog"). The first versioncovered only the 7,051 entries of the AHU-Sul collection; this one brings together the 167,344documents of the Catálogo AHU-Brasil, the catalogue of the Arquivo Histórico Ultramarino (AHU,Lisbon) concerning colonial Brazil, described in the preprint Rocha (2026),doi:10.1590/SciELOPreprints.16461. Contents File What it is catalogo-ahu-brasil-codigo-e-dados.zip Snapshot of the repository github.com/saulorrp/catalogo-ahu-cone-sul (branch main, commit 7b573a1), without the git history auditoria.zip The files of the human audit of the SRSP score (Section 3.1 of the preprint): auditool.py, calculista.py, auditoria_class.json (human scores) and catalogo_ahu.json, the same as in the first version LEIAME.md This file, in Portuguese README.md This file Inside the zip: corpus/colecoes/: the two collections (ahu-sul, 7,051; ahu-brasil, 160,293), in Parquet, withthe provenance of each document. corpus/camadas/: the analysis layers (typology, DeepSeek v3 inferences, people andbiographies of the AHU-Sul collection, semantic vectors), each with its own camada.json. app_dados/: the files the application reads (documents, BM25 lexical index, vectors frommultilingual-e5-small and from the Model2Vec static model), with manifest.json and hashes. ahu/, build/, app.py: the code of the application and of the data build (Python, Streamlit). coleta/: the scripts for collection, verification transcription and quality control. docs/METODO_COLETA.md: the step-by-step account of the corpus expansion, with the figures(in Portuguese). docs/prompt_deepseek_v3.txt: the prompt used for the AHU-Sul inferences. tests/: the automated tests. Licences Code: MIT. Derived data (layers and application artefacts): CC BY 4.0. Catalogue summaries and call numbers: they belong to the Arquivo Histórico Ultramarino (Lisbon)and to the Projeto Resgate "Barão do Rio Branco" (Biblioteca Nacional / Ministério da Cultura,Brazil), and remain credited as such. How to cite Rocha, S. R. P. AHU-Brasil Catalogue: a searchable database of the Arquivo Histórico Ultramarino summaries on colonial Brazil [Catálogo AHU-Brasil: banco de dados e motor de busca dos resumos do Arquivo Histórico Ultramarino sobre o Brasil colonial] (version of 28 October 2026). Zenodo.doi:10.5281/zenodo.18772666 (concept DOI, which always resolves to the latest version). Rocha, S. R. P. Desenvolvimento de um Classificador do Catálogo do Arquivo HistóricoUltramarino. SciELO Preprints, 2026. doi:10.1590/SciELOPreprints.16461.

Fields of Study & Subjects

Digital Humanities and Scholarship Historical Linguistics and Language Studies Natural Language Processing Techniques

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
1 January 2026
Language
EN
Document type
Dataset