What problem does it solve? Actuarial models often need custom calculations that do not exist in the Gaspatchio framework, and naive implementations using Python loops or map_elements run 20-1000x slower. This Skill guides you to place each calculation correctly and build vectorized accessor methods that compose Polars expressions. ## Core Features & Use Cases - Performance Ladder Decision Framework: Routes every calculation to the right home — existing methods, inline operators, setup utilities, Rust kernels, or new accessors — before any code is written. - Accessor Templates: Provides complete column and frame accessor templates with registration, scalar/list column handling, and docstring conventions. - Anti-Pattern Catalog: Documents seven slow patterns (Python loops, map_elements, dict lookups per row) with their vectorized replacements. - Use Case: You need a Gompertz hazard rate function reusable across models. The Skill confirms no existing method via uv run gspio docs, then guides you to build a registered column accessor returning an ExpressionProxy. ## Quick Start Ask the AI to add a custom calculation such as a Macaulay duration or Gompertz hazard rate method to Gaspatchio as a reusable vectorized accessor.
