Implements a collection of multiple comparison correction methods (Bonferroni, Holm, Sidak, FDR Benjamini-Hochberg, FDR Benjamini-Yekutieli) as a standalone function. The module is self-contained, relying only on NumPy and SciPy, and can be extracted and reused in any project requiring p-value adjustment. It demonstrates careful handling of edge cases (e.g., negative p-values, NaN propagation) and follows the standard statistical definitions.

A collection of core utility functions used across the entire library: `remove_na` for listwise deletion of missing values from multiple arrays, `_check_dataframe` for input validation, `_postprocess_dataframe` for consistent output formatting (rounding, column naming), and helper functions for permutation p-values and dependency version checks. These functions are self-contained (only external dependencies: NumPy, pandas, SciPy) and can be reused in any data analysis pipeline. The `_postprocess_dataframe` pattern, which applies global formatting options via a config, is particularly elegant.

A minimal and clean module implementing a global options dictionary for the library. It provides a thread-safe way to set formatting preferences (e.g., rounding precision) that propagate to all output functions. This small module demonstrates a reusable pattern for managing user-configurable settings in a library, avoiding the need for per-function parameters. It can be extracted and adapted for any project needing a simple, global configuration system.