documenting package conventions

Author: divine7022

inst/STYLE_GUIDE.md

@@ -0,0 +1,43 @@
+# betydata Style Guide
+
+This guide documents the coding and data conventions for the `betydata` package.
+
+## R Code
+
+- **Pipe operator:** Use the native pipe `|>`, not `%>%`.
+- **Data frames:** Use `tibble::tibble()`, not `data.frame()`.
+- **Grouped operations:** Use `.by = c(col1, col2)` argument, not `group_by() |> ... |> ungroup()`.
+- **Early returns** over nested `if/else` blocks.
+- **Package-qualified calls** in package code: `dplyr::filter()`, not `filter()` after `library(dplyr)`.
+
+## Documentation
+
+- **Roxygen2** with markdown enabled (`Roxygen: list(markdown = TRUE)` in DESCRIPTION).
+- **Tags:** `@param`, `@return`, `@export`, `@author`, `@examples`.
+- **Title:** Short, one line, no trailing period.
+- **Description:** Second paragraph after a blank `#'` line.
+- Backtick-wrap code references in roxygen and vignettes.
+
+## Data Conventions
+
+- The primary table is `traitsview` -- a denormalized view combining traits and yields.
+- Source CSVs live in `data-raw/csv/` (version controlled, human-editable).
+- Built `.rda` files live in `data/` (generated by `data-raw/make-data.R`).
+- `checked = -1` records are excluded. `checked = NA` is converted to `checked = 0` during build.
+- Only public data (`access_level == 4`) is included.
+
+## Testing
+
+- Use `testthat` edition 3 (`Config/testthat/edition: 3` in DESCRIPTION).
+- Test files go in `tests/testthat/` with naming pattern `test-*.R`.
+
+## Version Control
+
+- Update `NEWS.md` for user visible changes.
+- Update `CITATION.cff` for non-trivial contributions (add yourself to `authors:`).
+- Keep commits focused -- one logical change per commit.
+
+## Data Corrections
+
+- To report errors: file a github issue using the "Data correction" template.
+- To submit fixes: edit the relevant CSV in `data-raw/csv/`, rebuild with `source("data-raw/make-data.R")`, and open a PR.