pm4py.util.compression.util module#
- pm4py.util.compression.util.project_univariate(log: EventLog | DataFrame, key: str = 'concept:name', df_glue: str = 'case:concept:name', df_sorting_criterion_key='time:timestamp') List[List[Any]] | None[source]#
Projects an event log to a univariate list of values For example, an event log of the form [[(‘concept:name’:A,’k1’:v1,’k2’:v2),(‘concept:name’:B,’k1’:v3,’k2’:v4),…],…] is converted to [[‘A’,’B’,…],…]
The method returns the compressed log
- Return type:
UCL- Parameters:
log – log to compress (either EventLog or Dataframe)
key – key to use for compression
df_glue – key to use for combining events into traces when the input is a dataframe.
df_sorting_criterion_key – key to use as a sorting criterion for traces (typically timestamps)
- pm4py.util.compression.util.compress_univariate(log: EventLog | DataFrame, key: str = 'concept:name', df_glue: str = 'case:concept:name', df_sorting_criterion_key='time:timestamp') Tuple[List[List[Any]], List[Any]] | None[source]#
Compresses an event log to a univariate list of integer lists For example, an event log of the form [[(‘concept:name’:A,’k1’:v1,’k2’:v2),(‘concept:name’:B,’k1’:v3,’k2’:v4),…],…] is converted to [[0,1,…],…] with corresponding lookup table [‘A’, ‘B’], i.e., if the ‘concept:name’ column is used for comperssion.
The method returns a tuple containing the compressed log and the lookup table
- Return type:
Tuple[UCL,ULT]- Parameters:
log – log to compress (either EventLog or Dataframe)
key – key to use for compression
df_glue – key to use for combining events into traces when the input is a dataframe.
df_sorting_criterion_key – key to use as a sorting criterion for traces (typically timestamps)
- pm4py.util.compression.util.compress_multivariate(log: EventLog | DataFrame, keys: List[str] = ['concept:name'], df_glue: str = 'case:concept:name', df_sorting_criterion_key: str = 'time:timestamp', uncompressed: List[str] = []) Tuple[List[List[Tuple[Any]]], List[List[Any]]][source]#
Compresses an event log to a list of lists containing tupes of integers. For example, an event log of the form [[(‘concept:name’:A,’k1’:v1,’k2’:v2),(‘concept:name’:B,’k1’:v3,’k2’:v4),…],…] is converted to [[(0,0),(1,1),…],…] with corresponding lookup table [‘A’, ‘B’], i.e., if the ‘concept:name’ and ‘k1’ columns are used for comperssion. The 2nd order criterion is used to sort the values that have the same trace attribute. The uncompressed arguments will be included, yet, not compressed (e.g., a boolean value needs not to be compressed)
The method returns a tuple containing the compressed log and the lookup table. The order of the data in the compressed log follows the ordering of the provided keys. First the compressed columns are stored, secondly the uncompressed columns
- Return type:
Tuple[MCL,MLT]- Parameters:
log – log to compress (either EventLog or Dataframe)
keys – keys to use for compression
df_glue – key to use for combining events into traces when the input is a dataframe.
df_sorting_criterion_key – key to use as a sorting criterion for traces (typically timestamps)
uncompressed – columns that need to be included in the compression yet need not to be compressed
- pm4py.util.compression.util.discover_dfg(log: List[List[Any]] | List[List[Tuple[Any]]], index: int = 0) DirectlyFollowsGraph[source]#
Discover a DFG object from a compressed event log (either univariate or multivariate) The DFG object represents a counter of integer pairs
- Return type:
Counter[Tuple[int, int]]- Parameters:
log – compressed event log (either uni or multivariate)
indes – index to use for dfg discovery in case of using an multivariate log
- pm4py.util.compression.util.discover_dfg_uvcl(log: Counter[Tuple[Any]]) DirectlyFollowsGraph[source]#
- pm4py.util.compression.util.get_start_activities(log: List[List[Any]] | List[List[Tuple[Any]]] | Counter[Tuple[Any]], index: int = 0) Counter[Any][source]#
- pm4py.util.compression.util.get_end_activities(log: List[List[Any]] | List[List[Tuple[Any]]] | Counter[Tuple[Any]], index: int = 0) Counter[Any][source]#
- pm4py.util.compression.util.get_alphabet(log: List[List[Any]] | List[List[Tuple[Any]]] | Counter[Tuple[Any]], index: int = 0)[source]#
