How to use the composeml.LabelMaker function in composeml

def inactive_membership(transactions, window):
    transactions = transactions.sort_values('transaction_date')
    
    if len(transactions) == 1:
        cutoff_time, end_time = window
        elapsed_inactive = end_time - transactions['membership_expire_date'].iloc[0]
        return elapsed_inactive

    membership_expire_date = transactions['membership_expire_date'].iloc[0]
    next_transaction_date = transactions['transaction_date'].iloc[1]
    elapsed_inactive = next_transaction_date - membership_expire_date

    return elapsed_inactive


label_maker = cp.LabelMaker(
    target_entity='msno',
    time_index='lead_time',
    labeling_function=inactive_membership,
    window_size='100d',
)

# +
now = pd.Timestamp.now()

label_times = label_maker.search(
    transactions,
    minimum_data=0,
    num_examples_per_instance=2,
    gap=1,
    verbose=True,
)

from composeml import LabelMaker
from featuretools.demo import load_mock_customer

full_df = load_mock_customer(return_single_table=True)


def my_labeling_function(df_slice):
    label = df_slice["amount"].mean() > 80
    return label


lm = LabelMaker(
    target_entity="customer_id",
    time_index="transaction_time",
    labeling_function=my_labeling_function,
    window_size="2h",
)

lt = lm.search(
    full_df,
    minimum_data="1h",
    num_examples_per_instance=2,
    gap="2h",
)

lt

# flake8:noqa
import composeml as cp


def my_labeling_function(df_slice):
    """one slice of data inside of the prediction window for single instance of target entity"""
    label = df_slice["voltage"].mean()
    return label

# todo name
lm = cp.LabelMaker(target_entity="machine_id",
                   time_index="timestamp",
                   labeling_function=my_labeling_function,
                   window_size="7 days")

# describe the parameters to search for the labels
# returns a LabelTimes object, which is basically a pandas dataframe
lt = lm.search(dataframe=full_df,
               minimum_data="20 days", # minimum data before starting search
               num_examples_per_instance=10, # examples per unique instance of target entity
               gap="7 days") # time between examples

lt.summarize() # prints out distribution of labels

lt.describe() # prints out all the settings used to make the labels

# functions to modify label times and return a new copy of label times