Source code for pm4py.algo.discovery.minimum_self_distance.variants.pandas

from enum import Enum
from pm4py.util import exec_utils, constants, xes_constants, pandas_utils




class Parameters(Enum):
    ACTIVITY_KEY = constants.PARAMETER_CONSTANT_ACTIVITY_KEY
    CASE_ID_KEY = constants.PARAMETER_CONSTANT_CASEID_KEY



CONCAT_ACT_CASE = "@@concat_act_case"
INT_CASE_ACT_SIZE = "@@int_case_act_size"
DIFF_INDEX = "@@diff_index"




def apply(df, parameters=None):
    """
    This algorithm computes the minimum self-distance for each activity observed in an event log.
    The self distance of a in <a> is infinity, of a in <a,a> is 0, in <a,b,a> is 1, etc.
    The minimum self distance is the minimal observed self distance value in the event log.
    The activity key needs to be specified in the parameters input object (if None, default value 'concept:name' is used).


    Parameters
    ----------
    df
        Pandas dataframe
    parameters
        parameters object;

    Returns
    -------
        dict mapping an activity to its self-distance, if it exists, otherwise it is not part of the dict.
    """
    if parameters is None:
        parameters = {}

    activity_key = exec_utils.get_param_value(
        Parameters.ACTIVITY_KEY, parameters, xes_constants.DEFAULT_NAME_KEY
    )
    case_id_key = exec_utils.get_param_value(
        Parameters.CASE_ID_KEY, parameters, constants.CASE_CONCEPT_NAME
    )

    df = df.copy()
    df = df[list({activity_key, case_id_key})]
    df = pandas_utils.insert_ev_in_tr_index(
        df,
        case_id=case_id_key,
        column_name=constants.DEFAULT_INDEX_IN_TRACE_KEY,
    )
    df[CONCAT_ACT_CASE] = df[case_id_key] + df[activity_key]
    df[INT_CASE_ACT_SIZE] = df.groupby(CONCAT_ACT_CASE).cumcount()
    df_last = df.groupby(CONCAT_ACT_CASE).last().reset_index()
    df_last = df_last[df_last[INT_CASE_ACT_SIZE] > 0]
    df = df[df[CONCAT_ACT_CASE].isin(df_last[CONCAT_ACT_CASE])]
    df = df.merge(df, on=[activity_key, case_id_key], suffixes=("", "_2"))
    df[DIFF_INDEX] = (
        df[constants.DEFAULT_INDEX_IN_TRACE_KEY + "_2"]
        - df[constants.DEFAULT_INDEX_IN_TRACE_KEY]
        - 1
    )
    df = df[df[DIFF_INDEX] >= 0]
    ret = df.groupby(activity_key)[DIFF_INDEX].agg("min").to_dict()
    return ret