Pre-submission

Author: meirelesff

NEWS.md

@@ -1,6 +1,6 @@
 # genderBR 1.4.0
 
-This is a new version that introduces one important new feature: asymmetrical thresholds for gender classification. In the previous versions, users could set a single threshold to classify names as female or male -- and they could overlap if the threshold was too low. Now, users can set a different threshold for each sex, and there is a new check to prevent overlapping thresholds. The release also includes other improvements and bug fixes.
+This is a new version that introduces a new feature: asymmetrical thresholds for gender classification. In the previous versions, users could set a single threshold to classify names as female or male -- and they could overlap if the threshold was too low. Now, users can set a different threshold for each sex, and there is a new check to prevent overlapping thresholds. The release also includes other improvements and bug fixes.
 
 - The `threshold` argument of `get_gender()` and `get_gender_nn()` now accepts a vector with two values, allowing users to set a different threshold for each sex: the first value is used to classify females and the second, males (e.g., `threshold = c(0.9, 0.8)`). The two values can also be named, in any order (`c(Female = 0.9, Male = 0.8)` or `c(F = 0.9, M = 0.8)`). A single value keeps setting a symmetrical threshold, as before.
 - Thresholds that sum to less than 1 are now rejected with an informative error. These values make the female and male bands overlap, so a name could be classified as both. This also applies to a single `threshold` below 0.5, which was silently accepted before and gave inconsistent results between `get_gender()` and `get_gender_nn()`.
@@ -50,7 +50,7 @@ In this minor release, the genderBR package was improved in two ways. First, bug
 * Added a `NEWS.md` file to track changes to the package.
 * Added input checks to the `get_gender` function.
 * Reduce the time between requests to the IBGE's Census API.
-* Fixed a problem on vectorization in the internal `round_guess` funcion.
+* Fixed a problem on vectorization in the internal `round_guess` function.
 * Included an internal dataset with all Brazilian first names and their predicted gender extracted from the IBGE.
 * Update the `get_gender` function to work with internal data.
 

cran-comments.md

@@ -1,17 +1,34 @@
-This is a new version of the package that introduces a new method for predicting gender from Brazilian first names using a neural network model. In addition, this version includes some updates to the documentation, code improvements, and several new tests.
+This is a new version (1.4.0) that introduces asymmetrical thresholds for
+gender classification: users can now set a different threshold for each sex
+in `get_gender()` and `get_gender_nn()` (e.g., `threshold = c(0.9, 0.8)`),
+and threshold combinations that would make the female and male bands overlap
+are now rejected with an informative error. The release also adds a
+`nn_size` argument to split large inputs into batches, a `device` argument
+to run neural network inference on a GPU (`cuda` or `mps`), fixes a
+vocabulary mismatch in the neural network character encoder (`<UNK>` vs
+`<PAD>`), and moves `torch` from `Imports` to `Suggests` -- the neural
+network model is now built only when the NN backend is actually used, and a
+clear message tells users to install `torch` when it is missing. See
+NEWS.md for full details.
+
+Note on file downloads: `download_gender_model()` downloads pre-trained
+model weights from Hugging Face to `tools::R_user_dir("genderBR", "cache")`.
+This happens only on explicit user request, never at load or install time,
+and the corresponding examples are wrapped in `\dontrun{}` so R CMD check
+does not trigger it.
 
 ## Test environments
 
-* macOS Tahoe 26.3 (personal computer), R 4.5.3
-* Fedora 43 (personal computer), R 4.5.3
+* macOS Tahoe 26.5.1 (personal computer), R 4.5.3
+* Fedora 44 (personal computer), R 4.6.1
 * GitHub Actions - (ubuntu): release, devel, oldrel-1
 * GitHub Actions - (windows-latest): release
 * GitHub Actions - (macOS-latest): release
 * Winbuilder - (windows server 2022): release, devel
 
 ## R CMD check results
 
-0 errors | 0 warnings | 0 note
+0 errors | 0 warnings | 0 notes
 
 ## Reverse dependencies