Coverage for jstark / mealkit / order_periods.py: 100%

1"""OrderPeriods feature""" 

2 

3import functools 

4import operator 

5 

6from jstark.features.feature import DerivedFeature 

7 

8from pyspark.sql import Column 

9import pyspark.sql.functions as f 

10 

11from .order_count import OrderCount 

12from jstark.feature_period import FeaturePeriod 

13 

14 

15class OrderPeriods(DerivedFeature): 

16 """OrderPeriods feature""" 

17 

18 def column_expression(self) -> Column: 

19 exprs = [] 

20 for period in range(self.feature_period.end, self.feature_period.start + 1): 

21 exprs.append( 

22 f.when( 

23 OrderCount( 

24 as_at=self.as_at, 

25 feature_period=FeaturePeriod( 

26 self.feature_period.period_unit_of_measure, period, period 

27 ), 

28 first_day_of_week=self._first_day_of_week, 

29 ).column 

30 > 0, 

31 1, 

32 ).otherwise(0) 

33 ) 

34 return functools.reduce(operator.add, exprs) 

35 

36 def default_value(self) -> Column: 

37 return f.lit(None) 

38 

39 @property 

40 def description_subject(self) -> str: 

41 return ( 

42 f"Number of {self.feature_period.period_unit_of_measure.name.lower()}s" 

43 + " in which at least one order was placed" 

44 ) 

45 

46 @property 

47 def commentary(self) -> str: 

48 return ( 

49 f"The number of {self.feature_period.period_unit_of_measure.name.lower()}s " 

50 + "in which at least one basket was purchased. The value will be in the " 

51 + f"range 0 to {self.feature_period.start - self.feature_period.end + 1} " 

52 + f"because {self.feature_period.start - self.feature_period.end + 1} is " 

53 + f"the number of {self.feature_period.period_unit_of_measure.name.lower()}" 

54 + f"s between {self.start_date.strftime('%Y-%m-%d')} and" 

55 + f" {self.end_date.strftime('%Y-%m-%d')}. When grouped by Customer and" 

56 + " Product this feature is a useful indicator of the frequency of" 

57 + " which a Customer purchases a Product." 

58 ) 

59 

60 @property 

61 def feature_name(self) -> str: 

62 return ( 

63 f"Order{self.feature_period.period_unit_of_measure.name.title()}s" 

64 + f"_{self.feature_period.mnemonic}" 

65 )