genderBR v1.3.0

This is a new version of the genderBR package that includes a new function: get_gender_nn(), which uses a character-level neural network to predict gender from Brazilian first names. This model can generalise to names not present in the IBGE census dataset, so it can be used as a complement to the existing functionality in the package. The release also includes some improvements, tests, and documentation updates.

• get_gender_nn() is a new exported function that uses a character-level neural network to predict gender from Brazilian first names. Unlike get_gender(), this function can generalise to names not present in the IBGE census dataset.
• Added clear_nn_cache() to manage the in-memory model cache.
• Added download_gender_model(), an internal function that handles downloading and caching the neural network model weights and vocabulary from Hugging Face.
• Replaced iconv() with chartr() for stripping accents in name cleaning. The previous approach relied on iconv(name, to = "ASCII//TRANSLIT"), which is platform-dependent and returns NA on macOS for accented names (e.g., "joão"). The encoding argument in get_gender, get_gender_nn, and map_gender is now deprecated and will be removed in a future version.
• Improved test coverage for the new function and edge cases.
• Added torch to Imports; luz and httr2 to Suggests.

genderBR v1.2.0

This is a minor update of the genderBR package that includes several improvements and, more importantly, support for Brazilian 2022 Census data. Changes in this version include:

• Added support for the 2022 IBGE names API through a new year argument in get_gender, which allows users to access the most recent data for gender prediction in Brazil for both the national and state levels. Code changes do not break backward compatibility.
• Updated the default year in get_gender to 2022.
• Internal dataset nomes now provides probabilities for both 2010 and 2022, enabling offline predictions and threshold tuning for either year when internal = TRUE.
• Removed the magrittr dependency by switching to R's native pipe (requiring R 4.1.0 or higher).
• Replaced dplyr/tibble joins with a data.table backend to speed up internal merges and reduce dependencies.
• Updated and added new tests to ensure no errors when using 2022 data.
• Improved documentation and examples to reflect the new functionality and changes.
• Added a new section in the README discussing ethical considerations.

genderBR v1.1.2

Minor fixes and improvements

This patch version essentially makes the package more robust by:

• Introducing a few minor bug fixes
• Creating new internal input tests
• Creating unit tests