colss turns NumPy computation into math-like expressions

Hot take: this is a neat ergonomics layer, not a new category. It will appeal to people who write a lot of formula-heavy array code and want denser, more readable expressions, but it has to prove that the syntax gains are worth the extra abstraction.

• –The strongest angle is readability for complex expressions, especially when formulas get nested or repetitive.
• –The feature set is practical: comparisons, logical operators, ternary syntax, and common math functions cover the usual array-expression cases.
• –The main risk is crowded competition from existing expression evaluators and from plain NumPy/pandas idioms that are already familiar to most users.
• –The project will likely need benchmarks, clearer safety/docs around expression parsing, and stronger examples to convince users to adopt it.