Source code for featuretools.primitives.standard.transform.cumulative.cum_min

from woodwork.column_schema import ColumnSchema

from featuretools.primitives.base import TransformPrimitive


[docs]class CumMin(TransformPrimitive): """Calculates the cumulative minimum. Description: Given a list of values, return the cumulative min (or running min). There is no set window, so the min at each point is calculated over all prior values. `NaN` values will return `NaN`, but in the window of a cumulative caluclation, they're ignored. Examples: >>> cum_min = CumMin() >>> cum_min([1, 2, -3, 4, None, 5]).tolist() [1.0, 1.0, -3.0, -3.0, nan, -3.0] """ name = "cum_min" input_types = [ColumnSchema(semantic_tags={"numeric"})] return_type = ColumnSchema(semantic_tags={"numeric"}) uses_full_dataframe = True description_template = "the cumulative minimum of {}" def get_function(self): def cum_min(values): return values.cummin() return cum_min