Repository restructured into a monorepo: the R package now lives under r/, alongside a Python counterpart under python/.
Prompt content (audience/verbosity/style/model instructions) is now sourced from a single canonical prompts/ directory at the repo root, shared with the Python package, and synced into inst/prompts/ via scripts/sync_prompts.py.
Internal prompt assembly (.assemble_sys_prompt()) now uses ellmer::interpolate_package() against a master template (inst/prompts/system_prompt_template.md) instead of hand-rolled paste0() string assembly. audience/verbosity/style instruction strings moved from one-file-per-setting into inst/prompts/config.yaml.
ellmer and yaml moved from Suggests to Imports (both are now used directly by package code, not just examples). ellmer requirement bumped to >= 0.4.1.
Prompt-type keys (e.g. lm, glm) renamed to language-neutral, semantic names (e.g. linear_model, generalized_linear_model) shared with the Python package, and prompt content generalized to remove R-specific framing.
explain() gained a language argument for requesting explanations in a specific language.
Distributed via r-universe and GitHub going forward, not CRAN.
Added
New “engine-specific” prompt notes for model types with multiple supported fitting engines across R and Python (currently linear and generalized linear models), describing the exact output format of the library that produced the summary.
Miscellaneous
Added CRAN badge to README. (Later removed as part of the switch to r-universe/GitHub distribution.)
statlingo 0.1.0
Added
Initial CRAN release (published under the name statlingua).
Changed
Major refactor, including new prompt layout in inst/prompts.
