search_capes_text() accepts several terms and several fields (#5, hugoavmedeiros/capesR). A row is kept when any of the terms occurs in any of the fields, e.g. search_capes_text(data, c("varicela", "catapora"), c("titulo", "resumo")); the new match = "all" argument requires every term instead. Previously a vector in term raised an error.
read_capes_data() text filters (titulo, resumo) also accept a vector of terms, matching rows that contain any of them.
When nothing matches, search_capes_text() returns an empty data.frame with the same columns as the input (previously a data.frame with no columns).
Added unit tests (testthat).
capesR 0.2.0
Data updated from the CAPES open data portal: years 2023 and 2024 added, and 2019, 2021 and 2022 replaced by the revised files CAPES published in 2025 (institution names in those files now use (CAMPUS) instead of ( CAMPUS )). The package now covers 1987 through 2024.
André Leite is now the package maintainer (previously Hugo Vasconcelos Medeiros).
Data hosting moved off OSF. The yearly Parquet files are now listed in the new capes_years dataset (year, file, url, bytes, md5), which replaces years_osf. download_capes_data() reads the URL from that table, so the host can change without touching the code.
download_capes_data() gains a base_url argument (default getOption("capesR.base_url")) to download from a mirror, downloads to a temporary .part file first, checks the file size, and warns about requested years that are not available.
read_capes_data() text search now works on the actual column names of the data (titulo, resumo); it previously only recognised title, so filters = list(titulo = "...") silently matched nothing.
read_capes_data() no longer loads the whole dataset into memory just to read the column names.
