coarsen

coarsen(values, strategy='first_term')

General label extraction / coarsening.

Parameters

values : list or array Raw column values. Elements may be strings, lists of strings, None, or NaN floats. strategy : str or callable "first_term" — split on comma, take first token, lowercase. "raw" — convert to str as-is. "prefix_N" (e.g. "prefix_3") — first N characters. A callable f(str) -> str — applied after unwrapping.

Returns

labels : (n,) str array