Source code for pm4py.objects.ocel.util.event_prefix_suffix_per_obj

from pm4py.objects.ocel.obj import OCEL
from typing import Optional, Dict, Any
from enum import Enum
from pm4py.util import exec_utils, pandas_utils




class Parameters(Enum):
    PREFIX_OR_SUFFIX = "prefix_or_suffix"
    INDEX_ATTRIBUTE = "index_attribute"
    LEFT_SUFFIX = "left_suffix"
    RIGHT_SUFFIX = "right_suffix"





def apply(ocel: OCEL, parameters: Optional[Dict[Any, Any]] = None):
    """
    Gets for each event and each one of the related objects,
    the prefix or the suffix.
    E.g. if there is an object "o1" with lifecycle "e1", "e2", "e3",
    then:

    The prefix of "e1" for "o1" is empty.
    The prefix of "e2" for "o1" is ["e1"].
    The prefix of "e3" for "o1" is ["e1", "e2"].

    The suffix of "e1" for "o1" is ["e2", "e3"].
    The suffix of "e2" for "o1" is ["o3"].
    The suffix of "e3" for "o1" is empty.

    Parameters
    --------------
    ocel
        Object-centric event log
    parameters
        Parameters of the algorithm, including:
            - Parameters.PREFIX_OR_SUFFIX => establishes if the prefix or the suffix of the event per object is
                                                needed. Possible values:
                                                    - "prefix"
                                                    - "suffix"
            - Parameters.INDEX_ATTRIBUTE => the index attribute which is artificially inserted in the relations df.
            - Parameters.LEFT_SUFFIX => the suffix for the left-part of the merge.
            - Parameters.RIGHT_SUFFIX => the suffix for the right-part of the merge

    Returns
    --------------
    ret
        A dictionary where the first key is the event, the second key is the object, and the value is a list
        of events which are the object-specific prefix or suffix of the event.
    """
    if parameters is None:
        parameters = {}

    prefix_or_suffix = exec_utils.get_param_value(
        Parameters.PREFIX_OR_SUFFIX, parameters, "prefix"
    )
    index_attribute = exec_utils.get_param_value(
        Parameters.INDEX_ATTRIBUTE, parameters, "@@index"
    )
    left_suffix = exec_utils.get_param_value(
        Parameters.LEFT_SUFFIX, parameters, "_LEFT"
    )
    right_suffix = exec_utils.get_param_value(
        Parameters.RIGHT_SUFFIX, parameters, "_RIGHT"
    )

    relations = ocel.relations.copy()
    relations = pandas_utils.insert_index(
        relations, index_attribute, reset_index=False, copy_dataframe=False
    )
    relations_merged = relations.merge(
        relations,
        left_on=ocel.object_id_column,
        right_on=ocel.object_id_column,
        suffixes=(left_suffix, right_suffix),
    )
    if prefix_or_suffix == "prefix":
        relations_merged = relations_merged[
            relations_merged[index_attribute + left_suffix]
            > relations_merged[index_attribute + right_suffix]
        ]
    else:
        relations_merged = relations_merged[
            relations_merged[index_attribute + left_suffix]
            < relations_merged[index_attribute + right_suffix]
        ]
    relations_merged = relations_merged[
        [
            ocel.event_id_column + left_suffix,
            ocel.event_id_column + right_suffix,
            ocel.object_id_column,
        ]
    ]
    relations_merged = relations_merged[
        relations_merged[ocel.event_id_column + left_suffix]
        != relations_merged[ocel.event_id_column + right_suffix]
    ]
    relations_merged = relations_merged.to_dict("records")

    ret = {}
    for el in relations_merged:
        e1 = el[ocel.event_id_column + left_suffix]
        e2 = el[ocel.event_id_column + right_suffix]
        obj = el[ocel.object_id_column]
        if e1 not in ret:
            ret[e1] = {}
        if obj not in ret[e1]:
            ret[e1][obj] = []
        if e2 not in ret[e1][obj]:
            ret[e1][obj].append(e2)

    return ret